With news of Neil Young pushing to get studio quality audio out to the consumer, it occurred to me that it would be good to get some perspective on how we go about comparing audio quality.  One of the most problematic issues surrounding audio quality and file resolution these days has been that, more and more, we hear about people not being able to tell the difference between compressed audio and full-resolution audio, and by extension some people are claiming that efforts like Neil Young's are out of touch and irrelevant.  I couldn't disagree more.

In my opinion, one of the biggest pitfalls in this arena is the dreaded A-B test.  In our seemingly infinite desire to manipulate the response of our fellow humans, we tend to stage A-B tests in two ways:

1) Almost-Barely: A-B because you can barely tell, and thus proving that the two items are close enough to be interchangable.

2) Absolutely-Boldy: A-B because you want to show how different two things are (and typically the superiority of one of them, which the test-designer is likely selling).

In my opinion, both types of A-B tests won't tells us what we want to know, but it's the first type, the Almost-Barely tests, that I take issue with here.  (And I'm also inadvertently and respectfully drawing into question what, Ethan Winer, (who I adore) has said about "scientific" A-B testing in Tape-Op #88, p.66.)

We've all done it at some point.  A.  B.  A.  B.  Flip.  Flip.  Flip, flip, flip.   This is B?  Ok, A.  B.  Ok, is that B?  Ok, do A?  Gosh, they're close!

The Almost-Barely tests seem so objective.  Here's A, and here's B.  See, soooo close!  The conclusion: because most people can't tell the difference, there's not really a difference.  And then you stack up the stats on these results, and suddenly we're doing science.

The problem is that these tests assume that because two things are close enough in a quick test that the difference will also be indistinguishable over long stretches of time. However, this assumption totally misses how it is that we tend to actually experience things in our very real lives.

For example, we hear people talk about how one can't make out the difference between a hi-res MP3 and a 24bit WAV file (assumedly a difference similar to the one Neil Young feels is worth fighting for).  Admittedly a hi-res MP3 and a 24bit WAV are relatively close enough in resolution that many people will not be able to pick them out in an A-B test.

But, we don't live with music like that.  If your'e anything like me, you listen to a lot of music in a lot of styles and - over the course of, say, a month - perhaps you've absorbed well over a hundred listening hours across many different albums on a few different playback systems.

How can flip flip flip replicate what it is to live with that much music for that long?  How can a drop of water emulate what it is to swim for hours?

If you want to do a real test of the differences, give people a music collection that's all MP3s for a month, then give them that same collection as 24bit WAVs for a month, and then ask which one's which, and I bet you will start to get some correct answers. 

Why?  Part of the answer is that, if given enough time, subtle differences will reveal themselves to us.  Subconsciously at first, and eventually consciously, we become aware of new details, subtleties, nuances.  We humans need time to truly come to perceive things in full detail.  But details, once revealed, become important features in the big picture.

For example, when I started living with my partner I introduced her to what I call "good coffee."  At first she kind of shrugged it off as my snobbery at work, and she couldn't really taste the difference.  But then, after months of drinking the good stuff, she found herself to be a bit of a coffee snob, too.  She could taste the difference because she had, simply, spent time with the good stuff.  The coffee revealed itself to her, slowly and subtly.  Her palate developed.  And the thing about good coffee is that it holds more detail, nuance and, therefore, interest. 

But it takes a while to become aware of that depth and complexity.  Had she done a flip-flip-flip A-B and made her choice to only drink the cheaper stuff because, "you know, they're basically the same," she'd have missed an opportunity to develop her palate.

I think the same thing can be said for the resolution of music, and it breaks my sonic heart to think of the A-B tests out there designed to convince someone that because they can't tell the difference today they won't tell the difference in a month or a year.  A-B tests may be designed to show that subtle differences don't matter, but what they really do is shut down the possibility that those subtle differences could be the key to someone's aesthetic awakening.

If you can't tell the difference when you flip flip flip between two subtly different sounds, please know that I'm here cheerleading for you to slow down and be a real human being. Live with the better quality for a while, and see if over time you too start to hear things you never heard before.  There is nothing like real life to truly test the quality of something.

So, for this reason, I herald Neil Young (as I always have, come to think of it) for sticking to his beliefs and making efforts to bring affordable and easily accessible hi-resolution audio to the market.  It's time we left the MP3 behind and opened back up the possibility of people developing their ears and loving great sound once more.  Who knows where such high standards could lead us.

Thank you, Neil. 

The aesthetic revolution will be beautiful! 

Allen Farmelo
www.farmelorecording.com

-----------------
4-5-12
I'd like to note that this thought piece was not intended as a direct rebuttal of Justin Colletti's piece on Neil Young, but his article - along with the articles I came across at Paste, Pitchfork, Rolling Stone and Neil Young's many advocacy interviews - did inspire me to write.

Unfortunately, a quick Google search will reveal countless MP3 vs. WAV A-B tests.

It's also worth noting that Neil Young seems to be proposing 24bit, 192k files on new-generation pocket-sized playback devices.  Give anyone 192k files, a decent D-A converter, a robust headphone amp and some great ear buds, and then put them back on an MP3/iPhone rig and I think we might not need a month to tell the difference.  This also supports Mr. Young's claims that an MP3 only contains 5% of the information found on a professional analog recording.  Neil Young seems to be proposing a far greater leap than the short gap between 44.1k WAVs to AAC files.

"My goal is to try to rescue the artform that I've been practicing for the past fifty years." - Neil Young

4-11-12: I just came across this statement from Neil Young's website.  I hadn't seen it before, but thought I'd share it here.  Digging the part about "senses opening up."

"The spirituality and soul of music is truly found when the sound engulfs you and that is just what 2012 will bring.  It is a physical thing, a relief that you feel when you finally hear music the way artists and producers did when they created it in the studio.  The sound engulfs you and your senses open up allowing you to truly feel the deep emotion in the music of some of our finest artists….This is what recording companies were born to give you and in 2012 they will deliver."   - Neil Young, May 2011

 

Tape Op is a free magazine exclusively devoted to
the art of record making.

 
Sun, Oct 26, 2014 - 12:05AM
Add your two cents to the discussion below:
:
:
:
:
:
 
Wed, Apr 4, 2012 - 6:01PM
Calvin Locklear said about this:

Really liked this article and it made me also think of how available and affordable storage is these days. Why do we need to compress music down to 5MB files?

 
Wed, Apr 4, 2012 - 6:02PM
Dean k said about this:

Ever had an eye test?

"Is A a little bit better then B? No, about the same? Flip Flip. Good? Done."

Had to say it but, in principle I agree with you. But I also wear glasses and have to live with those A/B snap decisions. And as an engineer we have to do it all the time.

 
Wed, Apr 4, 2012 - 6:02PM
Don said about this:

Perhaps you mean "Palate"...... unless of course you plan on shipping coffee on a pallet..

Is this nit-picky? No. Spell check let it go because pallet is a legitimate word, albeit not the one that should be used.

 
Wed, Apr 4, 2012 - 6:07PM
Justin Colletti said about this:

Although A/B listening tests seem to show that even trained listeners are unable to tell 256kbps AAC files from full-resolution WAVs, I think Allen makes a fair point:

It's possible that listeners could show an *unconscious* preference for High Resolution Audio.

I believe that we can test for that, too instead of leaving the question to voodoo.

As far as Neil Young's comments that "?We?re in the 21st century and we have the worst sound that we?ve ever had. It?s worse than a 78 [rpm record]"....

...Well, I think the A/B tests prove that those comments are grossly overstated and generally pretty silly.

And this is coming from someone who's madly in love with Neil Young - and audio fidelity!

 
Wed, Apr 4, 2012 - 6:08PM
Andy said about this:

Great piece.
Hope Young sees it.
The "live w/MP3s for a month" and then same songs w/WAVs point is important. And the quality coffee analogy works.

 
Wed, Apr 4, 2012 - 6:11PM
Bjorn said about this:

Even if we agree that the delivery format is a problem, it's not the biggest problem. Today's recordings have no dynamic range, and therefore don't even need 16-bits, much less 20 or 24. Why are we trying to create a better delivery format when the recordings themselves (and not the delivery format) is the problem? I don't know. Producing better masters will fix the problem, but fixing the format will not.

 
Wed, Apr 4, 2012 - 6:15PM
shabbyrobe said about this:

"[It] breaks my sonic heart to think of the A-B tests out there designed to convince someone that because they can't tell the difference today they won't tell the difference in a month or a year."

This is the reasoning I used when I was initially digitising my music collection. Space was at a premium, so the decision to go whatever the equivalent of --preset-extreme was in 2000 wasn't without its drawbacks when my main hard drive was 40gb and half full.

I rightly assumed that although I couldn't hear the difference between the CD and a 128kbps MP3 at the time, I would be mightily unhappy if, over the lifetime of the library, I did come to be aware of the sonic shortcomings of such an aggressive encode (I bin 128s on sight these days unless I have absolutely no other way of getting the material or the recording has no hi-hat in it).

Unfortunately, once again after a very large number of listening hours logged, I'm having to replace some of my major favourites with FLAC. It's a time-sink, but it's worth it.

The coffee analogy is perfect.

 
Wed, Apr 4, 2012 - 6:19PM
Allen Farmelo said about this:

Don, thanks for the spell-check! My French needs some work...

 
Wed, Apr 4, 2012 - 7:10PM
Jtron said about this:

Nice article. I'd consider swapping your month-long test to B/A though. I feel like it's easier for me to notice when the good stuff is replaced with lesser quality.

 
Wed, Apr 4, 2012 - 7:16PM
Anu Kirk said about this:

Sound quality is only one piece of the picture, though.

Another is the convenience trade-off. That's how we ended up with cassettes. And vinyl.

Big files mean fewer files on devices. Many listeners like access to lots of tracks.

The biggest obstacle for listeners isn't sound quality of the reproducing format, it's things like:
- They're not paying attention, the music is background
- Their headphones/speakers/gear is bad
- Their listening environment is bad
- They're listening to music that isn't well-recorded to begin with.

I tell people the same things all the time. If you have 256 kbps or 320 kbps MP3, you're probably fine. But you should buy decent headphones, get a GOOD recording, sit down, close your eyes, and just LISTEN to it.

 
Wed, Apr 4, 2012 - 7:27PM
chris porro said about this:

something i read some time ago was about most of the information in audio washing over you. there are all kinds of very fast transients, details just above the masking threshold. it takes a while to process this stuff. i also think the A/B test (which i do find useful) is kinda like what pepsi did in the 80's with their "sip test". you got a sip of coke and pepsi. pepsi was sweeter and for a sip that tastes better to a lot of people. but after lots of sips it starts to taste too sweet. i think the artifacts introduced in popular 128 kbps mp3s get old after a while.

 
Wed, Apr 4, 2012 - 9:57PM
Cameron said about this:

Great insight. I completely agree with Allen & have expressed the same insight in the past (with a wine analogy). For the sake of argument, if you A & B an mp3 of Miles Davis' Bitches Brew to a CD or to vinyl (on a good system with good monitors) the difference is very noticeable. For some music, the difference is minuscule if not placebo. What is really important to remember is that higher quality is better, even if it isn't sonically apparent. The higher resolution music can always be downgraded to a lesser format, but not the other way around.

 
Thu, Apr 5, 2012 - 2:20AM
Stephane said about this:

Please remember what the scientist try to achieve with the A/B testing : removing the "a priori" bias. "Hearing is Believing" we say. In the case of Audio, it is much the opposite : "Believing is Hearing". When differences become subtle (or inaudible), psychology plays a big role in our hearing perception. That's the effect scientists try to minimize with A/B testing. Other than that, A/B tests have no value on their own, it all depends on the testers abilities. Give deaf people to A/B uncompress 24bit audio and 64kbit mp3, and the conclusion will be that there is no difference at all. My biggest complain is that the audience used for A/B test are often average listeners... we need top-of-the-cream instead.

 
Thu, Apr 5, 2012 - 3:27AM
Russ Hughes said about this:

I think you are missing a vital part to the argument, market conditions.

It has been proved time and again that if you offer the mass market something that is cheaper v higher quality then most will chose based on cost.

Do I agree with it? Of course not, but what Apple showed us is that when it comes to people buying music, then quality is way down the list in decision making.

Compressed audio sounds bad? No sh*t! But that's not the challenge we have, tell the average person they can have better quality and they just go 'Huh?'

 
Thu, Apr 5, 2012 - 4:16AM
Thomas said about this:

We are so use to hearing crappy quality music that we have forgotten high fidelity. the AB test is irrelevant unless you grew up listening to high fidelity. Kids have know idea what that is. IPODs killed all of that. At least an old guy with balls is trying to bring it back.

 
Thu, Apr 5, 2012 - 5:43AM
Benjamin Lowengard said about this:

I like the coffee analogy . I'm a student of Golden Ears which lends itself to training to hear various frequency ranges which has been wonderful for my listening ability to distinguish quality issues- but annoying for my music creation ability because - for instance - my effects chain always affects frequencies. Dynamic issues are also in play- but these are not new memes that have appeared in the digital age- I remember arguments over whether speakers or needles were the most important piece of your stereo. We all want to record quality material and have people listen back and NOT think about the delivery - but about the content.

 
Thu, Apr 5, 2012 - 6:51AM
Allen Farmelo said about this:

Russ - I think you're right about iTunes and the past decade or so, but then I ask myself how technology managed to increase fidelity for nearly a century (1890 - 1990), or even why we all suddenly agreed to pay more for CDs in 1989. I think the market is more complex than "cheaper wins." Look at HD-TV sales, for example. People will pay for better quality if it's marketed to them properly, I think.

Stephan - I totally agree about the A-B tests using "average listeners." Concluding there's no difference between audio formats based on average listener response is like saying that there is no earthquake because the geiger counter got jammed. Or, worse, it's corporations concluding that they can get away with selling smaller files at full price because no one notices. That, in essence, is pollution.

 
Thu, Apr 5, 2012 - 7:18AM
Bruce Bartlett said about this:

Great article, thanks!

I used to think that FM radio sound was pretty good. But then I heard one of my recordings played on the air. Yuck. So much distortion and compression. No surprise; the broadcast/receive signal path is much longer than the studio signal path.

My recording sounded so bad on FM because I was used to hearing very clean sound in the studio. Comparing the studio sound vs. the radio sound was a very slow A-B test, but the difference was obvious.

 
Thu, Apr 5, 2012 - 10:13AM
Tom Dennehy said about this:

Mass acceptance of high resolution lossless digital audio (in which I am a believer) faces two challenges. Let's break down the name.

The target market for high resolution, an older demographic, is not crossing over for two reason. One, they tend to buy less music, relying more on a sizeable personal collection. Second, they seek fidelity and a sizeable percentage believe absolutely that anything analog is better then everything digital.

The target market for lossless digital, a younger demographic, isn't crossing over for two reasons. One, compressed digital is all a substantial part of the population has ever known and they think it sounds just fine. Second, they seek convenience and lossless digital is no more convenient to carry around or stream through the cloud than compressed digital.

Neither is insurmountable but the combination is daunting.

 
Thu, Apr 5, 2012 - 10:52AM
John said about this:

Where is all this high resolution content going to come from? I'll tell you where: it's going to be the already available CD master, only upsampled. It's already happening.

Maybe the FTC should be developing marketing guidelines on the subject, because there's no way the consumer is going to tell the difference without a lot of effort.

Also, assuming you see a benefit to high sample rates (a topic for another debate), 192kHz is pretty ridiculous. I don't know anyone in mixing or mastering who routinely works at that sample rate.

 
Thu, Apr 5, 2012 - 11:13AM
Stephane said about this:

Another nicely written article, on the same topic. Allen will probably disagree with Monty's viewpoint, but it's a very interesting reading anyway.

http://people.xiph.org/~xiphmont/demo/neil-young.html

(the author is a contributor to the FLAC audio format and the inventor of the Ogg Vorbis codec).

 
Thu, Apr 5, 2012 - 12:12PM
Allen Farmelo said about this:

@John - I think you're right about CD masters, so there's a catch-22 in some cases, but for records tracked and mixed to tape, and newer records recorded at higher sample rates, it's the way forward.

As for 192 being ridiculous, I'm with Mr. Young in feeling that we're rounding the corner with both storage capacity and download speeds to support it. As for the sound - that's a hot topic I'd rather not get into here, but I will say that I believe the higher the SR the better. (I actually am looking into writing something about that, but I'm still too unclear about the math and theory of it all...I'm groping at understanding it, though).

 
Thu, Apr 5, 2012 - 12:45PM
Allen Farmelo said about this:

@Stephane - fascinating article. I am trying to have lunch with someone I know who has a different (and frustratingly unpublished) take on how lower SR generates artifacts, too. I hope to gain insight if I can keep up with the math/physics. Will share if I do.

 
Thu, Apr 5, 2012 - 3:24PM
Stephane said about this:

@Allen - Yes, the article is one of the best you'll find on the internet, reflecting the engineer's point of view on the topic. Both of your opinions are complementary, I think.

BTW, I can modestly help you with the math and physics, if needed ;-)

 
Thu, Apr 5, 2012 - 4:52PM
James said about this:

Yes! This is why I still prefer CDs and records to digital downloads and why I can't bring myself to pay for mp3s.

 
Fri, Apr 6, 2012 - 10:47AM
Ron Magness said about this:

Does anyone remember mp3 Pro? It looked to me like it was going to be "the future". But its pitfall was identical to so many others in technologically based audio It required both the sender and receiver (of an email in those days) to own the hardware. But it did attack the biggest problem with mp3s ... the Bass.

Try this little test. Record one of your favourite tracks with a good strong low end on WAV or AIFF. Then roll off ALL the bass. Measure both files. You'll be shocked.

 
Sat, Apr 7, 2012 - 11:15AM
Sam Rogers said about this:

Helpful article. I'm trying to learn to listen, and I'll try the idea of giving each format time. I should have known that was a good thing to do being one that really learns things slowly. I was doing A-B tests and not really hearing much difference.

I liked your references to our being human, and real life, very much.

With respect and gratitude to you and Mr. Young.
Sam

 
Sat, Apr 7, 2012 - 9:44PM
Arghhh... said about this:

We record 24 bit to prevent clipping, the dynamic range of a live musical performance is unlikely to exceed 96db (16 bit) and almost all music is mastered to under 20db.

This article makes no arguments in favour of extended dynamic range and just posits that people may hear a difference. Not of the same master they won't, not unless they're in a properly treated listening room in which case the noise floor _may be_ lower.

There is a (just about) valid argument for 20 bit distribution formats but since we don't listen to music in anechoic chambers it's not very convincing.

 
Sun, Apr 8, 2012 - 9:42AM
Allen Farmelo said about this:

@Sam - thanks for the kinds words, and just keep listening like a human, not a human under testing conditions.

@Arghhh - there certainly are many interacting variables to consider with digital formats: dynamic range, compression codecs, bit depth, sample rate, dithering, truncation methods, intermodulation, heterodyning, embedded errors and more. my aim here, however, was to encourage real world listening over flip flip A-B'ing. but, your comment about dynamic range and bit-depth is very relevant to Neil Young's efforts, and seems to be very much a part of the general discussion around his spiel.

 
Tue, Apr 10, 2012 - 1:26AM
eliot said about this:

Thanks for this article, Allen. I too was intrigued by Neil's announcement. I think that his approach to marketing hi-resolution audio is a good thing, but not just because of the subjective perception of audio fidelity. Perhaps more importantly, it asserts that there could be digital audio that actually has value (in comparison with mp3s, which in their free form are ubiquitous enough, even through legal channels, to be effectively valueless).

But if you (or anyone else reading this) wants to go through the whole AB testing thing again, I've found that certain source material is much worse affected by mp3 or aac encoding than much of the material that is used for AB and ABX testing. In my experience, almost everyone (even non-musicians and untrained listeners) can immediately hear the difference between high-quality, full-bandwidth orchestral recordings (such as Debussy's La Mer) or Steve Roach's ambient music and *any* current form of compressed audio, as the compression leaves rhythmic artifacts that are audible well below 15khz. I've found that brief listening tests have encouraged students of mine to ditch their mp3s and pursue FLAC and uncompressed audio (and that wasn't even an intended effect). So AB might have a redeeming value, after all...

 
Tue, Apr 10, 2012 - 8:58AM
Allen Farmelo said about this:

@Eliot - this is a great insight. Again, we're seeing dynamic range play a role in fidelity, but not in a linear "more dynamics is better" way, but as a feature in the program material that can affect which digital format presents it best.

And I'm not at all surprised to hear that ambient music would fall into this category. The typical lack of a rhythm section and the amount of space in the music leaves from for the artifacts of data compression to be heard.

On the tracking side, I hear engineers often say, "It's rock, so 44.1k is fine," or "Well, this is a sparse recording, so let's run at 96k." Where there will be emptiness in the music, especially where reverb tails will be evident, higher sampling rates do audibly do a better job.

And so, many classical buffs still demand vinyl while many blown out pop music fans are content with MP3s, or can't hear the difference.

Agreed about the perceived value of music going back up with the resolution - absolutely.

 
Tue, Apr 10, 2012 - 11:44AM
Jp said about this:

Below is a forward I wrote recently about our new download format we're offering to clients. I think this is very relevant to this thread and those interested in high-end dynamic audio. Check out www.fdrhd.com

(PS Don - There could be some spelling errors - please try to concentrate on the comments about audio!!!)

"We've always thought about the mastering process as orientated around the RMS of the program material and not the peak. This concept is something that has been lost in my opinion, after the introduction of digital the focus has shifted to the peak hence the quest for loudness continues unabated without the previous physical restrictions of older formats such as vinyl to consider.

We have always seen the restriction of the peak as an output format requirement and not wholly as part as the mastering process as with the restrictions of the physicality of cutting to vinyl and its differences to Audio CD. This means we as with may other engineers have always made a full dynamic range masters but have had to limit them because of the client?s requirements.

All this has eventually led us to the launching of the download store selling these full dynamic range master versions in high definition at the same grade as the mastering process.

This facilitates those Artist and Labels to sell a differing product, one that accesses a different market. Not the hyper-compressed i-tune?s, Audio CD market but the more discerning audiophile and musician. Both of have access in general to play higher resolution files and better quality playback systems thus will benefit greatly from this increase in quality. To the point of the original post, they will hear the differnce when sustanined listening.

In our ethos we are not trying to say Artists shouldn?t release music that?s hyper-compressed. We?re just giving them the opportunity to sell another product to a different consumer and maybe in the process make all consumers aware of what they?re listening to.

This is a format we want other mastering engineers and studio to get involved with. We want to bring as many albums in a full dynamic range high definition format to the public as we can.

Let people hear the music as we do ? no restrictions ? no comprise.
This is what we should all be interested in, hearing music at it best!"

Jp Braddock - Mastering Engineer @ www.formationaudio.co.uk

Store Website: www.fdrhd.com

 
Tue, Apr 10, 2012 - 2:24PM
Ryan Biggs said about this:

I'm sure Neil Young has a killer hi-fi set, a fabulous turntable, and a meticulously cared for collection of vinyl records. I grew up listening to vinyl on my parent's stereo, and even though it was good quality and they tried to keep records in sleeves, etc, I remember and awful lot of pops and hisses. I also remember that vinyl records were very difficult to play in the car or while jogging.

People prefer digital music because it is infinitely more convenient. My iPhone is the size of a single cassette, sans box, and yet it provides me with instant access to dozens of albums worth of music. Through Pandora and Spotify, I have access to thousands more. This level of convenience and control is what we traded sound quality for (not price, as some are suggesting).

We've done the same thing with photography. Companies like Polaroid and Kodak figured they were safe, because people would ultimately value the quality of film, and digital isn't really any cheaper (bought printer ink recently?). But the convenience of seeing our photos instantly, cropping, manipulating, removing red-eye, and only printing the ones we want - that convenience trumps quality.

But that doesn't stop people from buying better quality digital cameras as they become available. Digital audio needs to come up with it's equivalent of a digital SLR camera. But just like an SLR won't help your photos if your printer sucks, better digital audio won't impress anybody who is listening through crappy computer speakers and headphones. Maybe Neil should be rallying against earbuds!!!

 
Wed, Apr 11, 2012 - 8:18AM
Allen Farmelo said about this:

@JP - nice. I think we're just generally about to round the corner and leave the MP3 behind us. Efforts like yours present important alternatives.

@Ryan Biggs - The cool thing about technology right now is that we can store hi-res files on our phones and other tiny devices. It's only going to get better as storage shrinks and becomes more affordable. The recent headway made in atomic-level data storage points to a vastness previously unheard of.

http://www.nytimes.com/2012/01/13/science/smaller-magnetic-materials-push-boundaries-of-nanotechnology.html

As for earbuds, I've got a $90 pair of Grados that are lovely to listen on, and I'm seeing more and more people willing to spend $$ on their buds and phones. It's time for that $$ to be spent on fidelity, rather than celeb models, however.

But, agreed: every piece of the chain matters. I'm very curious about portable D-A converter upgrades, myself.

 
Thu, Apr 12, 2012 - 12:31PM
Scottie Walker said about this:

While I agree that the time is near to upgrade past lossy compression into, if not full-res audio, at least very close to it (storage size/streaming speed is still a concern), it's not the biggest reason new music sounds crappy.

We here all know about the loudness wars, but the general public could still care less (although that whole Metallica-Rock Band thing was nice to see). Over-limiting is still the most pervasive, unnecessary, and most easily avoidable blight on sonic quality.

What Neil Young is doing is great, but I'd rather see he and other household names throw their weight behind greater dynamics before less data compression.

 
Thu, Apr 12, 2012 - 3:26PM
Allen Farmelo said about this:

@Scottie - Dynamics play a big role in sound, no doubt. But I don't see it as an either/or situation. We need both better res and the option of making more dynamic recordings.

A blanket "squished is bad" approach to the topic of dynamics is tricky because in some cases a squished dynamic range is either (a) presented in the music itself, (b) a creative choice, or both.

One of my favorite bands, Stereolab, has relatively narrow dynamics on their records, but that's not due to squishing the master; it's because they play with linear, trance-inducing non-dynamics. Those records sound great, despite minimal dynamics.

But, in terms of the music that's just squished to be at war with it's peers, I agree 100%. That shit has done major damage.

I'm guessing that for Neil Young dynamics are a given. Check out how his records rate at the Dynamic Range Database

http://www.dr.loudness-war.info/index.php?search_artist=Neil+Young&search_album=

 
Fri, Apr 13, 2012 - 5:34PM
Mixerman said about this:

Thank you Allen for this blog post.

I've explained this concept to Ethan many times on the Internet.

When evaluating sound, you must take into account emotional impact, and you can?t evaluate emotional impact by flip, flip, flipping back and forth between two audio files (which is exactly what they become, nothing but audio files). Our brains aren?t geared to evaluate sound (or any other external information for that matter) without an emotional component. This is especially so with music, which is a medium of expression. When you flip back and forth, the music is irrelevant, hence the term ?audio files.? When you?re recording music, the sound and the performance are inextricably attached. They go hand and hand. You can?t evaluate sound as sound alone. It?s how the sound works in the context of the music that offers us some useful basis of comparison.

It?s also nearly impossible to make comparisons when you don?t have a stake in the product. I can hear insanely subtle details due to my twenty-plus years of making records, but if you hand me a song I don?t know or don?t like, those differences won?t matter to me, and therefore I will struggle to hear them. I have no stake and therefore I?m not being affected emotionally. Sonic differences in that context are irrelevant. We can?t discount the brain in all of this.

As producers and engineers we automatically have a stake in the product we make. We?re usually so intimately familiar with the work that subtle changes in how a track makes us feel are noticed instantly. This explains why professional record-makers cringe when we hear MP3s of our own work. We hear and feel exactly how the product is affected. We have the basis of comparison between our artistic intentions and the reality of an MP3. Ethan might be right about our lack of auditory memory in the abstract, but if you attach a song to that audio, our auditory memory becomes outstanding. If it didn?t we wouldn?t be able to sing back a melody we just heard.

A consumer who emotionally invests herself in a full resolution version of a song, given the same setting will recognize the poor quality of an MP3 version instantly. For the untrained ear (and even for the trained, really), a large percentage of the difference between the high-resolution track and the MP3 is felt rather than heard. Once there's an understanding (and memory) of the artistic intent, the MP3 then pales in comparison.

Back and forth A/Bs are fine for quick decisions within the context of the mix. There will be times when we evaluate audio purely as audio. The further along we get into the process, the more the music and emotion come into play, and that?s when the critical decisions are made.

I believe Michael Wagner once said that as record-makers ?we manipulate emotions.? I would just add, sound is merely the canvas from which we achieve that goal.

Enjoy,

Mixerman

 
Sat, Apr 14, 2012 - 10:59AM
Ethan Winer said about this:

Allen,

I love you too man. <thumbs up>

You are both correct and incorrect in your assessment of A/B testing. On the correct side, it's absolutely true that we can learn to recognize certain types of degradation and artifacts. Once you hear the swishy swirly sound of low bit-rate lossy compression, it's much easier to hear it in small doses at higher bit-rates. Same for subtle types of distortion and other artifacts. I agree that testing a bunch of uneducated listeners, then concluding that 96 kbps MP3 sounds as good as uncompressed PCM, is invalid.

But you're not an uneducated listener! And neither are many of the people who read Tape Op. After all your years of producing music, you have already learned to recognize the sound of a flaky patch connection, lossy MP3 compression, slight clipping, air handler room noise, and so forth. So if someone with *your* experience can't tell a difference in a side by side comparison, then that difference really is too small to matter or non-existant.

Another thing that many overlook is the ABX test. This lets you compare things over as long a time as you want, including weeks or even months. The original ABX testers were hardware switches, but now a number of free software versions are available. There's no "pressure" to perform in front of someone else who's testing you, and you can go back and forth as many times as you want to be sure you're not missing anything. You can also use any music you want, including tunes you're intimately familiar with.

Finally, the short-term nature of our hearing for audio quality is unrelated to memorizing a tune. And the issue of audio quality comparisons has nothing to do with the emotional aspect of music. I've explained this to Mixerman many times on the Internet. :->)

--Ethan

 
Sat, Apr 14, 2012 - 12:41PM
dwoz said about this:

we can all agree to disagree about what is important.

I have a little aphorism I spout now and again...that most people listen to music, while audiophiles listen to sound systems.

What this means, of course, is that there is simply not a single objective standard for "quality." And a particular person will apply different objective standards of quality in different situations.

i.e. "context matters."

In illustration of Mixerman's point, I reminisce about a recording situation I was in once, where an instrumental part I had contributed was not, shall we say, exactly ripping the socks and panties off the listeners. I listened to the candidate mix they were discussing and though it sounded quite "ok" and well-constructed and "good..." they were right. It was NIGHT AND DAY difference to the scratch mix I had checked my part with. Turns out the instrumental part I had contributed had been flown into the mix a mere handful of millis late-to-clock.

Just a handful of milliseconds. It utterly torpedoed the emotional impact of the part, though of course you'd be hard-pressed to argue that such a minor timing difference was discernable in an A/B test. (it isn't).

Additionally, the mixer had made a small EQ change to the part, which served to move it out of the sonic space of a second part, but which eliminated the "alike-ness" of that frequency spectra with a third, rhythmic part, which disassociated the two parts and made them less cohesive. A/B of the two versions simply did not reveal this, though I, the person most invested in the part, most certainly missed it, like I'm miss my car that was stolen out of the supermarket parking lot when I came out with my bag of milk and eggs.

What does this mean? That A/B testing CAN simply have zero relevance in the context of the production of master recordings, and in fact may OFTEN have zero relevance in that context.

 
Sat, Apr 14, 2012 - 12:48PM
Ethan Winer said about this:

> What does this mean? That A/B testing CAN simply have zero relevance in the context of the production of master recordings, and in fact may OFTEN have zero relevance in that context.

Agree totally dwoz. As you know, all I ever address is raw fidelity - faithfulness to the source - as relates to transparency (or not) of audio gear. So while you're correct that A/B comparisons are not so useful for assessing the impact of a mix, they are *absolutely* valid for comparing converters and preamps etc. If you can't tell whether you're hearing a budget sound card or a top of the line converter in an A/B test, then IMO you can't tell at all.

Note that I don't claim that all converters are indistinguishable, as people often attribute to me. All I'm saying is that you need to do a proper blind test to know for sure if you really can tell a difference.

--Ethan

 
Sat, Apr 14, 2012 - 2:15PM
dwoz said about this:

respectfully disagree, Ethan. I think you're talking past the point.

let's use another example: a convertor that does something like throw everything above 4k 135 degrees out of phase. As you yourself say, that will be completely indistinguishable in an A/B test listening to 'raw' program content. But use that convertor in a mix CONTEXT where not all the mix goes through it, and you'll suddenly find that the differences are VERY discernible with some program content, less so on other program content.

Again, you're simply reiterating your premise, which I personally think is flawed. Music production systems and music auditioning systems are SYSTEMS that have a chain of components. You can't ignore the context around a particular component and just discuss the component alone. It's ALWAYS used in-situ.

And due to the often profound tradeoffs made in design and manufacturing, between ideal performance and ideal cost, any gear that you care to mention, cheap or esoteric, will NEVER behave the same in situ that it does stand-alone.

Let's also not blur the line between definitive truth and relative truth. If something is indistinguishable in an A/B test to a particular person, there's two absolutely equal possibilities: first, that there really is no difference, and second, that the test has simply not been designed in a way that would illustrate the difference.

Again, that means CONTEXT. THIS convertor on THIS material is indistinguishable from THAT convertor on THIS material. Well and good. However, this in no way speaks to whether the result can be extrapolated to THAT OTHER material.

 
Sat, Apr 14, 2012 - 3:11PM
DJMC said about this:

I think some of this debate is talking at cross-purposes.

Studio professionals and people making work intended for posterity should obviously use the highest-quality formats available to them, for a whole lot of pretty obvious reasons. Consumers will tolerate compromises of many and various sorts. Some people would buy 1/4" reel-to-reel duplicates of the masters if they were available, but the vast majority of people will tolerate some compromise for things like convenience and cost.

Even "coffee snobs" rarely fresh-roast their own beans. Those who grind usually do so only to the degree that they have easy access to relatively inexpensive grinders and whole beans. They're not in their hotel room crushing up coffee beans at 6am.

From that POV, the goal should be to improve the quality of convenient music formats, or to improve the convenience of quality. Lecturing consumers on why their hearing should be more refined is not a winning strategy.

RE: the validity of A/B testing, some of the OP is misguided. For starters, rapid A/B testing as described is not even close to a scientifically fair test of subtle differences, but for completely different reasons than those listed. By a wide margin, the big problem with that kin of AB testing is simple self-delusion: confirmation bias, placebo effect, whatever you call it... experts are often the easiest to fool, simply because we bring such a truckload of experiences and biases to the test. This has been proven time and again in all kinds of different fields of subjective expertise. Years of in-depth forensic examination of a thing build not only finely-attuned perceptiveness and sensitivity, but also well-entrenched preferences and prejudices that can be tricked, if the test is not a double-blind one.

If you can discern even a tiny difference in one technology vs. another, then you should use the better one in the studio, if it is available to you. If it's not, then you should use the best that you can lay hands on. Better to make music than to let perfection become the enemy of the good.

But if you or someone you know is just listening to music, then listen to whatever is convenient and accessible, favoring the better where practical. MP3 might be sub-optimal sonically, but it's got more people listening to a wider variety of music than ever before in the history of the world, and that good so vastly outweighs any evil, that only a myopic audio professional would count it as something to lament.

Finally, and perhaps most to the point, I think it is extremely debatable to suggest that the average overall audio quality experienced by everyday consumers is worse now than it was 20 or 30 years ago. I am sure that in, say, 1982, Neil Young was listening to records on sound-systems that vastly exceed most 2012 iPod speakers and home-theaters-in-a-box, but was the same true of typical middle-class households? I remember an awful lot of pretty atrocious stereos from those days...

 
Sun, Apr 15, 2012 - 12:26PM
Ethan Winer said about this:

DJMC, those are all good comments. Another thing that's often overlooked with regard to the quality of gear used by pros (not consumers) is that degradation is cumulative. Even if three budget devices are each audibly transparent individually, after sending the music through all three in a row the end result may no longer be transparent. So for that reason I'd prefer a device with 0.001 percent distortion over one with 0.01 percent, even though 0.01 percent is probably not audible by itself.

--Ethan

 
Sun, Apr 15, 2012 - 12:37PM
Ethan Winer said about this:

@dwoz:

> "a convertor that does something like throw everything above 4k 135 degrees out of phase. As you yourself say, that will be completely indistinguishable in an A/B test listening to 'raw' program content. But use that convertor in a mix CONTEXT where not all the mix goes through it, and you'll suddenly find that the differences are VERY discernible with some program content"

That's a very contrived example that would never happen in practice. Phase shift in audio gear occurs at the frequency extremes, due to coupling capacitors at the low end and parasitic capacitance or anti-aliasing filters at the high end. Even then, most of the phase shift is outside the audible range. Further, even if the design of a converter or preamp did add such a large amount of phase shift in the middle of the band, it would be the same on all channels and thus not a problem when the tracks are mixed.

> "THIS convertor on THIS material is indistinguishable from THAT convertor on THIS material. Well and good. However, this in no way speaks to whether the result can be extrapolated to THAT OTHER material."

This also seems flawed, but I remain open to hearing an example that proves your point. Often when people object to my examples showing the audibility of artifacts at various levels, they say it doesn't match their experience. So I *always* request they post their own examples using any musical source, and any added artifacts they believe best makes their case. If someone actually posted files showing that some type of artifacts are audible when 90 dB below the music, I'd change my opinion in a heartbeat. But they never do.

If you can post files showing that one pass through some model converter is transparent on one piece of music, but is audibly degrading on another, I'd love to hear it. And - being as respectful as possible - if you can't, perhaps you'd consider changing your opinion?

--Ethan

 
Sun, Apr 15, 2012 - 8:49PM
Allen Farmelo said about this:

Fascinating stuff, everyone. Ethan, glad you jumped in.

I'm still staunchly anti-AB. I believe AB-ing itself causes one to disengage from the emotional aesthetic experience in which subtle differences in sound quality can have an impact on how one feels music. More to the point, I think AB-ing looks in the conscious mind for results that, for most people, are rendered in the subconscious.

And for me, I'll notice little weird things like my shoulders being slightly more relaxed when listening to analog sources over time, or an general higher fatigue rate when listening to lower-res files. I really do think much of this stuff sits on the fence between our conscious minds and our subconscious, rocking back and forth in and out of our awareness, affecting us in ways we're barely aware of, but affecting us nonetheless.

 
Mon, Apr 16, 2012 - 6:47AM
Allen Farmelo said about this:

Quoting Ethan: "If someone actually posted files showing that some type of artifacts are audible when 90 dB below the music, I'd change my opinion in a heartbeat. But they never do."

Isn't there more to this than the audible presence of artifacts? Aren't there qualities such as center and stereo imaging, depth of field, and a general sense of space that can be changed by the presence of artifacts, even if we can't hear the artifacts per se?

In other words, is it the residue of a null test that we're actually looking for, or the affects of that residue on the overall sonic image?

 
Mon, Apr 16, 2012 - 9:39AM
Mixerman said about this:

Allen,

According to Ethan there's no such thing as "depth of field." that's just more of that "magical thinking" that Ethan refers to in his article.

Mixerman

 
Mon, Apr 16, 2012 - 1:58PM
Ethan Winer said about this:

>> AF: Isn't there more to this than the audible presence of artifacts? Aren't there qualities such as center and stereo imaging, depth of field, and a general sense of space that can be changed by the presence of artifacts, even if we can't hear the artifacts per se?

Yes, imaging and space are important qualities. (Mixerman's typical stuffing of words I never said into my mouth notwithstanding.) But those are more related to acoustics (imaging and phantom center), and artificial spaciousness (added reverb and echo). I don't see how a converter, preamp, or summing amp could affect depth or width. Again, I'm glad to be proven wrong if someone would post an example. Say, a single track or mix as is, then again after passing through some device.

Another way you can add "width" to a mono source is by adding stereo hiss, as happens when mixing to analog tape. I've tried that, and found the apparent width of a mono source increased even with the hiss mixed in fairly softly.

> In other words, is it the residue of a null test that we're actually looking for, or the affects of that residue on the overall sonic image?

I don't see how soft artifacts could affect these non-tangible qualities. It's sort of like the belief that we can be influenced by subliminal spoken messages under music, when the speaking is too soft to hear. If it's too soft to hear, then it's too soft to hear! This in particular has been proven repeatedly, so it's not just an opinion.

That's why controlled testing is so important. Hearing and perception are variable, and the only way I know to separate real from imagined when the changes are subtle is with a blind test. If you know another way to conclusively determine these things, in a way that makes sense logically and all can agree with, I'd love to hear it. I can also point out that blind testing is the gold standard in all branches of science, especially for things that relate to perception. So it makes no sense to me that for some reason audio is exempt.

--Ethan

 
Mon, Apr 16, 2012 - 2:02PM
Ethan Winer said about this:

Allen,

If you think it would accomplish anything useful, I'd be glad to drive out to your place some weekend to discuss and test some of these things in person. Not even test, we could just compare stuff and talk about what we hear. The weather is finally nice, and it'd be great to meet you in person.

--Ethan

 
Mon, Apr 16, 2012 - 6:40PM
Allen Farmelo said about this:

Ethan wrote: "I don't see how a converter, preamp, or summing amp could affect depth or width. Again, I'm glad to be proven wrong if someone would post an example. Say, a single track or mix as is, then again after passing through some device."

Ethan, here you go: a full-on straight-up scientific A-B of my own creation of a mix summed digitally and one summed analog. Plus, a long explanation of why analog summing is a key ingredient in my work.

http://www.farmelorecording.com/in-the-press/analog-vs-digital-summing/

And for some odd reason this is like the only really apples-to-apples summing test I know of, with the fewest changes to the variables.

Give a read and a listen and then get back to me and let me know if you still think depth and such can't change from analog summing.

But, don't take it from the quick AB. Spend time with the whole track and see how it hits you.

BTW, the difference is more pronounced at higher resolution (but those are 320kps MP3s).

 
Mon, Apr 16, 2012 - 6:48PM
Allen Farmelo said about this:

Ethan, I'd love to have you down for some science (and anything else) sometime. I always learn so much from you. Maybe we make this some kind of public discussion here in the city...?

I will say, though, that I feel my career has been a giant series of long-term AB tests that will be hard to undo in a quick visit - and I say that humbly, recognizing my proclivity for romantic notions in the face of science.

But, I will still stand by the idea that there is no test alive that doesn't change the subject being tested - and this notion has been well documented and integrated into scientific method. I don't feel that music is the one exception to this; I believe that almost anything being tested will be changed by the testing itself to a degree, and I believe that AB testing in particular is a really obvious condition in which to find oneself, and can thus have a great impact on the very perception one is trying to test neutrally.

I've very interested in how my own falling in and out of awareness around what I perceive affects my overall perception - such a terribly complex thing to consider when you start to look at how the subconscious plays into this while discussion.

 
Tue, Apr 17, 2012 - 6:36AM
Allen Farmelo said about this:

More to the point on the science side of things, I think I fall into an actual romantic camp here which rejected empiricism around 1900. This is from wiki on Qualitative Testing:

"In the early 1900s, some researchers rejected positivism, the theoretical idea that there is an objective world about which we can gather data and ?verify? this data through empiricism. These researchers embraced a qualitative research paradigm, attempting to make qualitative research as ?rigorous? as quantitative research and creating myriad methods for qualitative research. In the 70s and 80s, the increasing ubiquity of computers aided in qualitative analyses, several journals with a qualitative focus emerged, and postpositivism gained recognition in the academy. In the late 1980s, questions of identity emerged, including issues of race, class, and gender, leading to research and writing becoming more reflexive. Throughout the 1990s, the concept of a passive observer/researcher was rejected, and qualitative research became more participatory and activist-oriented."

I think my beef with AB'ing lies here at the heart of how we frame up scientific research I'm definitely postpostivist.

And Ethan, I appreciate your positivism, but after soaking in close listening for the past (ahem)...well, let's say many years, I feel an AB test just frames up a very unrealistic situation. Of course, we do some AB'ing on the daily in the studio, but Neil Young is talking about albums people are going to buy and live with for life, likely without anything to compare them to.

And just poking around Wiki, I'm seeing that what I was recommending in my original post was something closer to "Usability Testing," the most famous example being Henry Dreyfus building ocean-liner suites in a warehouse and having people move in with their luggage for weeks on end to see how they do over time. Dreyfus designed everything from the table-top telephone to the John Deer tractor, so he's an interesting cat to consider.

Let's put people in a warehouse with Neil Young's new device with 192k files and a regular iPod with MP3s (same music) for a week and see what they end up using.

 
Tue, Apr 17, 2012 - 11:21AM
Ethan Winer said about this:

I downloaded the files and had a quick listen, and heard little difference on small speakers. I went back and forth a few times - one time the analog mix sounded wider, but another time the digital mix sounded wider. I'll listen again later at full volume through my large "good" speakers and report back.

In the mean time, is it possible for you to post or email me two short (ten seconds each) clips at full quality? I'd like to put them in a DAW to null them, which will reveal more clearly what's different. Due to content being removed during encoding, MP3 files don't null as reliably as Wave files. I'm sure you calibrated your converters closely since the files *sound* like the parts are at the same volume. But still, a null removes perception from the equation making it easier for me to know for sure what's different.

I'd definitely be down for a "public" discussion. Do you mean you'd invite a few other people to participate? Regardless, it's probably easier to discuss stuff in person than via forum posts.

--Ethan

 
Tue, Apr 17, 2012 - 1:15PM
Mixerman said about this:

Allen,

Sending files to Ethan for critical listening can be somewhat problematic given his monitoring chain.

Ethan on GS: "The speakers in my studio are large old-school JBL 4430s, with a 15-inch woofer and bi-radial horn tweeter. They're bi-amped at 1,200 Hz using a Rane crossover, which feeds a pair of Crown PowerBase amps totaling just over 1 KW. I don't use a "monitor controller" per se. Audio goes in and out of my computer through a Mackie 1202 mixer, and there's also a Rane "DJ" type mixer in the path where I select the source of Mackie, cassette deck, or Sony DAT player. It wouldn't be too difficult to patch around any or all of these pieces."

I'm sorry, but no one is going to notice, what you and I would consider to be rather large differences through that monitoring chain. Not even you or I.

Mixerman

 
Tue, Apr 17, 2012 - 1:53PM
Ethan Winer said about this:

^^^ Are insults really all you have to offer the discussion?

 
Tue, Apr 17, 2012 - 3:54PM
ColinG said about this:

I take the author's point about A/B testing, but any meaningful comparison has to be blinded; tell listeners one is better and they'll hear a difference even if both are identical. If you want to prove hi-res formats are better, send test subjects away and make them live with each for a month, but, crucially, don't tell them which is which. That is the minimum you need to say there's a genuine benefit to hi-res audio.
Having said that, I can't see any possible benefit in going as high as 192khz when human ears generally can't hear above 20 kHz.

 
Tue, Apr 17, 2012 - 4:00PM
Mixerman said about this:

Ethan,

Explaining your monitoring chain isn't an insult. It's meant to illuminate what you consider a critical listening chain.

Once again, you can't critically listen to audio unless you have accurate monitoring. Your chain alone leaves much to be desired. You're using a Soundblaster converter, which all on it's own is going to hide a great many things that would be obvious on even mediocre professional converters. While monitors are largely a matter of taste, I would be surprised if those JBL 4430s were capable of emitting audible sound above 16k (and I'm being kind).

It's perfectly reasonable for me to illuminate your monitoring chain (a chain that you put forth, and that is in the public record), when you are insisting you can't hear a difference. My point is, no one can. That's not a chain appropriate for judging.

If you're uncomfortable with, or embarrassed by your chain, then perhaps you should consider upgrading it rather than accusing me of hurling insults.

Enjoy,

Mixerman

 
Tue, Apr 17, 2012 - 4:55PM
Allen Farmelo said about this:

Ok, hold on a second. I feel like a flame war started in another prominent forum has been dragged over here. I'm not any referee or anything, but maybe we can get away from old battles and keep the discussion more open and exploratory!

Ethan, I would agree that the quality of your monitoring chain is capable of masking certain subtleties. We've all been through the revelatory experience of hearing hidden things revealed on excellent playback systems.

I'll look for those files - been a while.

However, I will still say that a null test won't reveal "space." I think for me the conversation becomes more interesting and productive when we talk about the perceptual effects of the differences, or about what the differences are doing to the parts of the audio that are the same. Like hiss making things sound wider - now that's fascinating. Or dithering masking artifacts. Or Walter Sear's 30k shelf eq totally opening the sonic image (even though we can't hear up there (eq curve probably reaches a bit below 30k, but the point is still interesting).

@ColinG - I know what you mean about 192k, but I think much of what higher sampling rates are capable of doing is moving different distortions caused by intermodulations and some other more recently discovered problems out of our perceptual range (above 20khz).

I feel I have to bite my tongue here because I don't know enough, but a very intelligent and frustratingly non-publishing tech here in NYC is onto something really interesting in terms of how sampling rate and freq response interact, and it's not intermodulation distortion. It's a whole different thing that **might** help reveal how higher sampling rates reduce distortion in the high end below 20khz by moving weird behaviors way out of hearing range. I hope to learn more and write about it (assuming I grasp it and it seems solid - been meaning to do this for years, and it seems the time is now).

In short, just because we can't hear it doesn't' mean it isn't affecting what we can hear, or the equipment's behavior, etc. Tape bias anyone?

 
Tue, Apr 17, 2012 - 8:13PM
DJMC said about this:

@Ethan Winer

RE: the layering-up of sub-optimal sound quality, that's actually what I meant to start with, but got side-tracked.

A side-by-side of an mp3 stereo render might be hard to distinguish from the CD, 24 or 48 or 120 tracks of mp3s rendered to on stereo file is likely to sound noticeably "wrong" even to causal listeners, compared with a mixdown of "good" quality tracks.

Having read a lot of your supremely intelligent and thought-provoking writings on the topic of sound quality, I suspect that you and I might disagree on some finer points regarding the definition of "highest-quality medium available", but I'm fairly certain that we would agree that whatever fills that bucket is what the recordist should use, for exactly the reasons you stated.

Most listeners are going to experience compromised sound-quality, however we define it. There is not a whole lot we can do about that on the supply-side.

If I had one wish that could be granted on the consumer-end, it would not be for the elimination of mp3, nor even for realtraps and good speaker-placement in every living room, but for a pushbutton compression+"loudness" circuit to be included on every consumer stereo, so that the purveyors of sound (some of whom get apoplectic about mp3) would stop feeling compelled to compete with bad recordings the source.

Just make a 2-band, -14dB, look-ahead limiter chip standard equipment on everything, and everyone could get back to the art and science of making the best-sounding records, by whatever metric we choose to use.

What bothers me much more than compromised sound on the consumer end is compromised sound on the supplier side. Perfectly-good records are frequently wrecked before the consumer even gets a chance to select a high-quality playback system.

 
Wed, Apr 18, 2012 - 11:43AM
Ethan Winer said about this:

@coming: "If you want to prove hi-res formats are better, send test subjects away and make them live with each for a month, but, crucially, don't tell them which is which."

Yes, that totally avoids accusations that short-term blind testing is somehow flawed.

--Ethan

 
Wed, Apr 18, 2012 - 11:44AM
Ethan Winer said about this:

^^^ Sorry, I missed that my spell checker changed ColinG to "coming." :->)

 
Wed, Apr 18, 2012 - 11:58AM
Ethan Winer said about this:

@Mixerman:

> Explaining your monitoring chain isn't an insult.

I'm certain that's how you intended it.

> you can't critically listen to audio unless you have accurate monitoring. Your chain alone leaves much to be desired. You're using a Soundblaster converter

First, I do not use a SoundBlaster sound card for monitoring. I've told you this a dozen times, and yet you still repeat that. Allen is correct, you are bringing friction into this discussion in part by continuing to stuff words into my mouth. Until my last computer died six months ago I used an M-Audio Delta 66 as my main playback. My new PC doesn't have the old style PCI slots, so I bought a Focusrite Scarlett 8i6 USB sound card. I do have a SoundBlaster card, but I use it (rarely) only to edit SoundFonts.

> I would be surprised if those JBL 4430s were capable of emitting audible sound above 16k (and I'm being kind).

Right, they're 3 dB down at 16 KHz. Which is fine since at 63 years of age I can't hear much above 14 KHz. :->) But these speakers are extremely clean, with very low distortion, and far better dispersion than most nearfields.

Your problem is you dismiss perfectly competent pro grade equipment as sub-par with no evidence. Yes, gear with high distortion can indeed mask detail. However, many people overlook that the distortion of even very good loudspeakers is ten times higher than that of most electronic gear. If you have evidence that my Mackie and Rane equipment masks detail more than, say, going through multiple stages in an SSL console, now's the time to show it.

But I know you have only beliefs, no evidence. Heck, I'm still waiting for dwoz to post an example to back up his claim that a converter can add 135 degrees of phase shift at 4 KHz, or that a converter can sound totally transparent on one mix and obviously colored on another.

--Ethan

 
Wed, Apr 18, 2012 - 12:05PM
Ethan Winer said about this:

@Allen:

> I would agree that the quality of your monitoring chain is capable of masking certain subtleties.

Of course.

> I will still say that a null test won't reveal "space."

A null test will *absolutely* reveal all differences between two files. Have you ever used nulling to remove vocals from a complete mix? Most lead vocals have reverb, so after the files are nulled the main vocal is gone, but a reverb "ghost" remains and is clearly audible. That's the difference between the two channels.

> Walter Sear's 30k shelf eq totally opening the sonic image (even though we can't hear up there (eq curve probably reaches a bit below 30k, but the point is still interesting).

If you can hear a difference after applying EQ at 30 KHz, then the EQ is absolutely affecting frequencies below 20 KHz, not just "a bit" below 30 KHz.

> non-publishing tech here in NYC is onto something really interesting in terms of how sampling rate and freq response interact ... I hope to learn more and write about it (assuming I grasp it and it seems solid

I'll be glad to give an opinion if you send me any information.

> Tape bias anyone?

Good point! Though that's different from the notion that we can be affected by ultrasonic content even if we can't hear it directly. Tape bias is an electronic solution to a magnetic problem - hysteresis.

--Ethan

 
Wed, Apr 18, 2012 - 1:40PM
Mixerman said about this:

Fair enough Allen. I've responded to Ethan's post on the Womb in regards to his monitoring chain. We'll continue this portion of the conversation there.

For anyone interested, just click on the "mixerman" at the time stamp that begins this post.

Enjoy,

Mixerman

 
Thu, Apr 19, 2012 - 8:26AM
Allen Farmelo said about this:

Ethan writes: "A null test will *absolutely* reveal all differences between two files."

Of course, in terms of information. But it won't reveal the difference those differences make when one is listening.

My questions:
(a) how do those differences impact the overall sonics?
(b) how does that impact on sonics affects human perception of immeasurable qualities like the perception of spacial dimensions in music?

Nulling is like removing a tiny drop of blue in a gallon of paint and saying "See, hardly any difference," while ignoring that that tiny drop can change the hue and depth of the whole gallon of paint. Any interior designer will tell you these tiny differences mean a lot when living with a color for decades, and that checking out two paint chips won't tell you eff-all about what it'll feel like to wake up five years from now and pour coffee in that room.

Ethan writes: "If you can hear a difference after applying EQ at 30 KHz, then the EQ is absolutely affecting frequencies below 20 KHz, not just "a bit" below 30 KHz."

Walter Sear (RIP) and I talked about this, and his contention was that inaudible frequencies have harmonic relationships to audible ones, and will interact, and therefore impact the audible range. It's also one of the reasons he was staunchly committed to analog tape, because the highend didn't crumble into noise like in lower sample-rate digital. Analog maintains it's waveform structure way up into the high high kHz's and therefore real-world relationships to the audible frequency range.

Now, with that stated, if we start to talk about 192k, we're talking about waveform representation up into the inaudible frequencies, and if you're like me and think Walter was onto something, then perhaps your mind is more open to further investigation.

The non-publishing tech I mention is none other than Sear Sound's in-house tech. I'm buying him lunch soon to finally unpack (if I can) the concepts he and Walter were discussing and, apparently, measuring. Very interesting stuff, not intermodulation artifacts, not harmonic content, but something else entirely.

I don't mean to be cryptic; I just don't get it yet. But have a hunch they're onto something important. Walter was a long-time participant in the discussions around audio (via AES and his general outspokenness), and my feeling is he was - like many people who spend decades dedicated to something - really onto something.

Back to the real world: I think Walter was onto an explanation of why I feel fatigued after 12-hours of digital work at lower sampling rates, feel better after a day at 96k, and really just fine on tape.

I need to buy that lunch, ASAP!

Also, I hope readers will find in my (admittedly somewhat uninformed) investigations here a broader message: Let's give Neil Young's ideas a chance!

"It's so hard to make love pay
When you're on the losing end,
And I feel that way again."
- N.Y.

 
Thu, Apr 19, 2012 - 12:58PM
Tomás Mulcahy said about this:

Hi Alan,

I like your paint chip analogy, it makes the point very clear. But some of your other points need addressing:
1. I've studied both electronics and ethnomusicology. Qualitative research was essential to the second, but I've never come across it in the first. Maybe I'm out of date. Perhaps it is used in the field of psychoacoustics? I don't know.
2. You've stated a common misconception about sampling. As per the Nyquist-Shanon-Koletnikov theorem, frequencies near the sampling frequency are represented with 100% accuracy. In audio there are sometimes issues with the anti-aliasing filter, particularly in poorly designed systems. But for a long time now ADCs have been oversampling. This means the audio is measured at rates higher than even 192kHz, then converted on chip to what the recorder expects. You'll find a very good proof and explanation of all this in John Watkinson's book "Basic Digital Audio". I think he was being facetious when he called it "Basic"! :)

 
Thu, Apr 19, 2012 - 2:32PM
Jason Spatola said about this:

So...

Someone needs to (blind) test Allen's two-month hypothesis.

I'm sure Sarafin and Winer could buddy up and finance such a study, yes?

 
Thu, Apr 19, 2012 - 2:48PM
Ethan Winer said about this:

Wow, I wrote so much I had to split this into two posts!

@Allen,

> But it [nulling] won't reveal the difference those differences make when one is listening.

Yes and No. As you may know, often people will report hearing a difference even when no difference exists at all. So a null test is perfect for that, to prove that the perceived difference is in fact entirely imagined. This is more important than you may think, as shown by the continued sale of replacement AC power cords to the pro audio market.

Even if there is a residual, the level of that residual gives us a solid clue as to its audibility. For example, if the residual after nulling two Wave files leaves only broadband noise that's 80 dB below the music, I think it's safe to conclude both files sound the same even if someone believes they sound different.

Toward that end, I got a chance to listen to your two example files loudly through good speakers. I don't hear any difference. And again, a few times I thought the analog sounded ever so slightly wider, and a few times the digital seemed wider. I wasn't watching which was playing at the time as I listened. When I nulled the two files, the residual was that swirly bubbly sound of lossy compression. So I really do need the original Waves to assess the level of the differences. Again, if one really is "wider" than the other, the difference will absolutely be revealed by nulling.

> (a) how do those differences impact the overall sonics?

Again, it depends entirely on their magnitude.

> (b) how does that impact on sonics affects human perception of immeasurable qualities like the perception of spacial dimensions in music?

Same answer. Further, width and depth are not immeasurable. There's no direct metric for width such as THD or frequency response. But we know that what creates the perception of width is differences between the left and right channels. A mix that has identical content left and right sounds like a single narrow source emanating from between the two speakers. At least it does in a room devoid of early reflections. So again, nulling shows what is different between two mixes. With the two original Wave files I could null A against B, and also null left against right for both files and see how those two residuals compare.

 
Thu, Apr 19, 2012 - 2:48PM
Ethan Winer said about this:

> Nulling is like removing a tiny drop of blue in a gallon of paint and saying "See, hardly any difference," while ignoring that that tiny drop can change the hue and depth of the whole gallon of paint.

I don't think that's a valid analogy because that's not how nulling works. Nulling doesn't remove stuff - it *reveals* stuff - the difference. So using your paint analogy, you'd null one mixed gallon against the other, and the result would reveal the size of the "tiny drop" that was added to only one of the gallons.

> Walter Sear (RIP) and I talked about this, and his contention was that inaudible frequencies have harmonic relationships to audible ones, and will interact, and therefore impact the audible range.

There's some truth to this, but probably not in the way Walter believed. Frequencies don't interact except in the presence of nonlinearity. So unless the equalizer was adding at least 1 percent IM distortion, any newly generated content would be at least 40 dB below the music. Now, IM distortion absolutely occurs inside our ears, and at levels much larger than 1 percent when the SPL is fairly loud. But that too is different than hearing increased clarity after boosting a knob labeled "30 KHz." I address all of this in my book, in excruciating detail. :->)

> It's also one of the reasons he was staunchly committed to analog tape, because the highend didn't crumble into noise like in lower sample-rate digital. Analog maintains it's waveform structure way up into the high high kHz's and therefore real-world relationships to the audible frequency range.

This is very easy to disprove by recording a few test tones and looking at the recorded output on an oscilloscope or high-res RTA. If you don't believe in using test tones, you can just as easily record music and see which medium's output waveform more closely resembles the original. I promise you digital wins by a large margin every time. Try it, you'll see.

> Very interesting stuff, not intermodulation artifacts, not harmonic content, but something else entirely.

I look forward to hearing more.

--Ethan

 
Fri, Apr 20, 2012 - 9:42AM
Allen Farmelo said about this:

Tomás Mulcahy wrote:

"1. I've studied both electronics and ethnomusicology. Qualitative research was essential to the second, but I've never come across it in the first. Maybe I'm out of date. Perhaps it is used in the field of psychoacoustics? I don't know."

I guess I just think that we're really talking about studying the phenomenological moment of interaction between an electronic system and a human system, and that quantitative analysis isn't going to show us much of what's happening on the human side. AB testing is grossly inadequate for measuring such a system. I'd guess some modern brain activity analysis might do the trick. I'd love to see what lights up in my neural web with the different files.

" 2.You've stated a common misconception about sampling. As per the Nyquist-Shanon-Koletnikov theorem, frequencies near the sampling frequency are represented with 100% accuracy."

This is true, but what Walter and his tech were getting at has to do with a very different phenomenon that I will make as informed an account of as I can ASAP. Again, apologies for crypticness. But, general rule seems to be that the higher we can push the sampling rate, the further out of our perception the artifacts they're finding will be. Again, not intermodulation distortion - more on this soon, I hope.

 
Fri, Apr 20, 2012 - 9:47AM
Allen Farmelo said about this:

Jason Spatola wrote:
"Someone needs to (blind) test Allen's two-month hypothesis. I'm sure Sarafin and Winer could buddy up and finance such a study, yes?"

EXACTLY! See, I think the real issue as I said earlier is that you'd need a major grant to study this in any kind of meaningful way. Good luck getting that grant when someone else is looking at autism or cancer or something that really matters.

And still, AB'ing is the MP3 of research - cheap, easy to do on the net, and filled with crappy artifacts :)

 
Fri, Apr 20, 2012 - 9:57AM
Allen Farmelo said about this:

Ethan writes:
"I got a chance to listen to your two example files loudly through good speakers. I don't hear any difference."

You can't hear something I can hear. Ok. Interesting. Perhaps a researcher bias is at play in your insistence that there is no difference. Admittedly, it's a small difference, but it's a difference. I'll find the files, but even just looking at the wave forms was like looking at two different songs.

Ethan writes:
"Again, if one really is "wider" than the other, the difference will absolutely be revealed by nulling."

Totally disagree. The results of a null test will isolate what's different, but it won't tell us how those differences play out in the perception of width.

Ethan writes:
"Further, width and depth are not immeasurable. There's no direct metric for width such as THD or frequency response. But we know that what creates the perception of width is differences between the left and right channels."

Again, I disagree. Mono mixes can be more or less deep than each other, AND they can sound wider. Differences between L&R can certainly create a sense of width, but that's not the only thing. Try an Antelope clock on a Digi192 converter and the sound gets wider - don't ask me why, but it sells antelopes (which sound great, BTW). Add mono tape hiss, as you said, the the images gets wider - why is that? There's more to the story than L&R differences.

 
Fri, Apr 20, 2012 - 10:10AM
Allen Farmelo said about this:

Ethan writes:
"I don't think that's a valid analogy because that's not how nulling works. Nulling doesn't remove stuff - it *reveals* stuff - the difference. So using your paint analogy, you'd null one mixed gallon against the other, and the result would reveal the size of the "tiny drop" that was added to only one of the gallons."

That's exactly what I meant. Sorry I wasn't more clear about there being two gallons. The hue changes between two gallons due to the drop being in one. Agreed. Now, maybe go back and reconsider my analogy. I think it's valid.

Ethan writes:
"Frequencies don't interact except in the presence of nonlinearity. So unless the equalizer was adding at least 1 percent IM distortion, any newly generated content would be at least 40 dB below the music. Now, IM distortion absolutely occurs inside our ears, and at levels much larger than 1 percent when the SPL is fairly loud. But that too is different than hearing increased clarity after boosting a knob labeled "30 KHz." I address all of this in my book, in excruciating detail. :->)"

Sounds painful! :-) Ethan, stay open to what Walter was onto, and learn about his custom console and that high end shelf before judging what's going on there. Remember, he was "in the lab" every day doing real world work and had amazingly discerning ears (the man could very quickly tell you which of his 251s was on any source - eerie!).

Ethan writes:
"This is very easy to disprove by recording a few test tones and looking at the recorded output on an oscilloscope or high-res RTA. If you don't believe in using test tones, you can just as easily record music and see which medium's output waveform more closely resembles the original. I promise you digital wins by a large margin every time. Try it, you'll see."

I can't say I know enough here, but I can say that what is "most accurate" and what is going to reproduce through a tiny pair of drivers a result that a human will perceive in a way that delivers the end user's intentions - well, those are different things.

But, more to the point, what does your oscilloscope say about the structural integrity of waveforms of the very high end content of 44.1khz digital. What does 192khz look like in a 44.1khz recording? Tell me that, and then we're talking about the same thing.

Ethan writes:
"> Very interesting stuff, not intermodulation artifacts, not harmonic content, but something else entirely.
I look forward to hearing more."

Me too - could be a red herring, but my hunch is we'll learn something interesting (or at least ask some good questions). I promise to report as soon as I can.

I'm off to sum through an analog console to an analog tape machine! If the placebo effect is in effect, I'm cool with that for now :) I wish someone could hook up bran scanners to my head today and see what's happening!

 
Fri, Apr 20, 2012 - 1:19PM
Ethan Winer said about this:

@Allen:

> You can't hear something I can hear. Ok. Interesting. Perhaps a researcher bias is at play in your insistence that there is no difference.

I didn't insist there's no difference, I only said I don't hear a difference. I may be biased, and I may be deaf. This is another reason to examine the original Wave files, to better evaluate the difference objectively.

> The results of a null test will isolate what's different, but it won't tell us how those differences play out in the perception of width.

Yes, null tests don't reveal how stuff is perceived. Nobody is arguing that. But a null test will show what's different, and the magnitude of the difference. That's very useful!

> Differences between L&R can certainly create a sense of width, but that's not the only thing. Try an Antelope clock on a Digi192 converter and the sound gets wider

I know I'm now treading on dangerous turf, but I'll ask anyway: Have you ever tested this blind, with someone else switching clocks while you listen? And have you compared that since treating your room? This is why I offered to come by for a visit. I'd like very much to switch those clocks while you listen. If we do that ten times in a row we'll know for sure if the sound really does seem wider.

As for Walter Sear, I'd be happy to join you when you meet your friend if you'll have me. I'll even sign an NDA if needed. :->)

> what does your oscilloscope say about the structural integrity of waveforms of the very high end content of 44.1khz digital.

Structural integrity is the same as fidelity, which means little change to the waveform and sound. Yes, this is a perfect application for an oscilloscope. If you compare any analog tape recorder to any modern digital system, the digital version will have more "structural integrity." This is easy to prove, as I said. Now, some people enjoy the colored sound of analog tape, and that's fine. But to claim that analog tape is more true to the source than digital is simply wrong.

> What does 192khz look like in a 44.1khz recording?

A 44.1 KHz sample rate can't capture content at 192 KHz, though neither can analog tape. I don't understand what this has to do with what we're discussing.

--Ethan

 
Fri, Apr 20, 2012 - 1:20PM
Ethan Winer said about this:

@Jason: I've tried to arrange a get-together with Eric, but he's too hostile to have anything useful come of such a meeting.

--Ethan

 
Fri, Apr 20, 2012 - 3:28PM
Jason Spatola said about this:

Two honest questions for anyone involved, not meant to disparage anybody:

1. Is it true that Ethan Winer has conducted two studies with botched results due to flawed methodology?

2. Is it true that the Mixerman Show featured a segment about how FTPing a file changed the spectral qualities of a mix?

If either of these are true, it's not the end of the world. People make mistakes.

Still, regarding an Eric/Ethan collaboration in a two-month study, I'd suggest their only involvement be financial.

Allen wrote:
"Good luck getting that grant when someone else is looking at autism or cancer or something that really matters."

...Inarguable point, indeed! :)

 
Fri, Apr 20, 2012 - 9:13PM
Allen Farmelo said about this:

Ethan: "...a null test will show what's different, and the magnitude of the difference. That's very useful!"

I don't agree that the magnitude is the issue. I think all kinds of things go on in a stereo mix that are individually unperceived but contribute to depth, interest, degradation of the sound.

And let's agree to disagree.

Ethan: "Have you ever tested this blind, with someone else switching clocks while you listen?"

Yes, and I can pin which clock is which every time (before and after room treatments). It's got to do with width and depth and openness in the highend. But I don't care that I can pick it out or not. More important is what it's like to live with one or the other for 12 hours a day for years. That's what matters to me. I want to stack up those little differences and soak in the cumulative effects of clocks, converters, sample rates, balanced power, good speaker stands, integrating tape - lots of little differences that form what I consider a high-functioning audio system.

Again, let's agree to disagree. You dig AB'ing as a form of science, and I find it an artificial experience that make me tense and not able to live with an aesthetic experience over time.

Ethan writes:
"...to claim that analog tape is more true to the source than digital is simply wrong."

I don't know, is it? I mean, what's the sampling rate of analog? Like a gazillion kHz? I just think the questions here should stay open for a while longer, and that the benefits of higher sampling rates should be considered.

Ethan writes:
"A 44.1 KHz sample rate can't capture content at 192 KHz, though neither can analog tape. I don't understand what this has to do with what we're discussing."

These folks cut a 122khz tone to vinyl from the bias freq on the analog tape! Wow! (http://www.positive-feedback.com/Issue2/mastering.htm) So, not quite 192khz, but if we need 192khz to get clear on 96k without aliasing in a digital system, then we're kind of talking about 192khz representing frequencies below those that are showing up on analog tapes.

Again, I just don't see that the book is closed on this issue, and I (obviously) don't want it closed by AB-testing and null tests. I want Neil Young to push the envelope, help develop the technology, and bring us into a new era of hi-fi digital sound. We need this push because we've been pushed to settle for shit for too long.

If we're going to be fooled, let's have more than we need, not less that we deserve.

And with that I have to say I'm truly out of my depth once we get into the more technical side of things - though I remain skeptical and curious. The intersection of aesthetics and technology is an infinitely complex arena that's going to need more than AB-tests to reveal what's really going on here.

And, Ethan, I totally agree that selling people fancy power cables via placebo effect is lame. So is selling MP3s with the same techniques.

 
Fri, Apr 20, 2012 - 10:42PM
Jason Spatola said about this:

Allen wrote:
"I want Neil Young to push the envelope"

Just keep in mind that the last time he tried that, we got "Trans".

(Ba dum tish.)

 
Sat, Apr 21, 2012 - 7:46AM
Allen Farmelo said about this:

I love Trans, and that he made it, and the ideas behind it.

 
Sat, Apr 21, 2012 - 11:42AM
Ethan Winer said about this:

@Jason: Yes, I did record some Wave files through a sound card with the wrong settings in the sound card's control panel. The methodology was fine, I just screwed up. So sue me. You never recorded something wrong? That doesn't make my core points incorrect, or me incompetent, or any of the other off-topic things Mixerman would love for people to believe. But that's all the Ethan haters have, so two years later they still use that as an argument.

--Ethan

 
Sat, Apr 21, 2012 - 11:52AM
Ethan Winer said about this:

Jason, I didn't mean my tone above toward you to sound harsh. It was meant for those who keep bringing up my goof as if that somehow invalidates everything else I've done.

Allen, we can agree to disagree if you prefer, but this stuff is not unknowable, even though some people don't understand that. Magnitude of artifacts is *everything*. And after years of searching, those who believe we can hear or be influenced by ultrasonic content have never shown proof that stands up to scientific scrutiny.

I'd still love to visit you to hear those clocks, and all the other stuff we discussed.

--Ethan

 
Sat, Apr 21, 2012 - 2:47PM
Allen Farmelo said about this:

Cool, Ethan.

I'm stepping down off the debate platform now, and want to invite any readers who feel they'd like to chime in to do so. I've certainly made my points, and would love to hear other opinions and ideas.

AF

 
Sat, Apr 21, 2012 - 4:22PM
Jason Spatola said about this:

I'm stepping down before I get punched.

I'm optimistic that further research will in fact happen and this won't be a topic of debate (regardless of the outcome) much longer.

 
Sun, Apr 22, 2012 - 8:34AM
Allen Farmelo said about this:

Or, Jason, that we'll all just move on to higher res audio like NY is suggesting and, for that reason, won't be debating much longer. Technology usually causes the big changes, not testing!

 
Mon, Apr 23, 2012 - 12:59AM
Jason Spatola said about this:

Allen,

I wholeheartedly agree that the high res master should be delivered to the consumer losslessly. The listener should be able to hear the album exactly as intended.

I'm a bit dubious on the need for Neil Young to be the "THX certification" of this particular movement (as the business model has existed for years anyway - and the technical formats for two decades), but if he finally brings something to the table I'm all for it.

"Technology usually causes the big changes, not testing!"

Come on now - you think Tape Op readers are going to be satisfied to leave it at that? :)

Really though, I'm a bit perplexed by that statement - testing births technology.

Testing is ubiquitous throughout all phases of bringing nearly any new technology to the marketplace - especially tech with progressive paradigmatic implications.

I realize you probably know this (considering that you value the input of Sear Sound's tech and are interested in his research), but that comment sounded kind of post-9/11 Bill O'Reilly to me and I couldn't let it go.

Cheers!

 
Mon, Apr 23, 2012 - 7:36AM
Allen Farmelo said about this:

"Testing is ubiquitous throughout all phases of bringing nearly any new technology to the marketplace - especially tech with progressive paradigmatic implications."

But testing is just as likely to halt the development of a technology. Case in point: the Swiffer. It was shelved because people were given a quick mop/swiffer comparison and the results were "no." Then the developer forced Proctor and Gamble to let people live with prototypes, and those test results were "hell yes!"

"I'm a bit dubious on the need for Neil Young to be the "THX certification" of this particular movement..."

He's a rare kind of expert with 50 years of recordmaking as his artform. And his passion to set new standards in motion comes from a very genuine place: true frustration with the limitations of technology to satisfy his vision. Standardizing is part of putting any new format into the market.

 
Mon, Apr 23, 2012 - 9:52AM
Jason Spatola said about this:

"But testing is just as likely to halt the development of a technology."

We're going to have to agree to disagree on the "just as likely" part.

Now - again - trying to flee before getting punched...

:)

 
Tue, Apr 24, 2012 - 12:43AM
Tomás Mulcahy said about this:

Hi Allen,
Your example of the mop is a test of consumer opinion, I'm pretty sure that's not the kind of testing Jason is referring to. When I think of testing, I always think of the British military term "proving ground" i.e. a test is to prove something works. If it doesn't, further development is needed. They're pretty tough about that in the military, and those high standards filter through to audio electronics design.

 
Tue, Apr 24, 2012 - 5:28AM
Allen Farmelo said about this:

Tomas, you're pointing right at something very important regarding testing. Thank you!

I think when we talk about audio file resolution mattering or not, we get into both a technological discussion and a consumer opinion discussion. My issue with AB'ing as a method is that I strongly believe that AB'ing creates consumer opinion as much as it tests anything significantly. "Flip flip flip - see, no difference," and the consumer is convinced. That's barely scientific, and it's not getting me the info I want.

I'm calling for a thorough and ongoing investigation of the technology. I want to really know how sampling rates, but depths, codecs, playback systems, clocks, converters, intermodulation, heterodyning, aliasing - all of it - affects digital audio. This investigation wont end, and there's lots to learn.

Then, I want to know more substantially how human perception systems (ears, brains, neural webs) process what's in those files over longer periods of time, precisely because it is over longer stretches that we live with recorded music.

If all we did was listen to small snippets, or one song per week, maybe an AB test would be adequate. But we immerse ourselves in recorded music for vast stretches of our lifespans.

The mop analogy was meant to show how a quick AB test can breeze right past a strong consumer opinion. Clearly the Swiffer is a hit, but Proctor and Gamble shelved it due to quick AB tests.

So, I just want us to live with Neil Young's suggested audio formats (or the equivalent) before concluding in a quick AB test that no one will care about these differences.

 
Tue, Apr 24, 2012 - 4:33PM
Tomás Mulcahy said about this:

Thanks Alan I see your point. So I think the correct path is to design the tests properly. It needs clearly defined goals, control groups and statistical analysis. That's non trivial, but would still involve AB- just not a quick one, not by a long shot!

Have you read "This is your Brain on Music"? I think it would answer your questions on perception etc. I'm no expert- I need to reread that book, and follow up on the references.

Ben Goldacre's "Bad Science" blog is good on stats.

Also, are you familiar with the work of JJ Johnston, co-inventor of mp3? The problem with mp3 is that the designer's intent was almost often not implemented. His blog:
audioskeptic.blogspot.com

 
Tue, Apr 24, 2012 - 5:18PM
Mixerman said about this:

@Jason: "Yes, I did record some Wave files through a sound card with the wrong settings in the sound card's control panel. The methodology was fine, I just screwed up. So sue me. You never recorded something wrong? That doesn't make my core points incorrect, or me incompetent, or any of the other off-topic things Mixerman would love for people to believe. But that's all the Ethan haters have, so two years later they still use that as an argument. "

Actually, whether you "screwed up" the test or not is in question. In one post, you say you "screwed up" but then discuss two other instances where you completely mislead your subjects, in one case, never even telling them. In the other post, you admit outright that you did indeed deliver tampered files in the hopes to prove expectation bias. That ALONE shows how little you know about testing in the field, because any competent record maker is going make sure the files don't null completely before listening for differences.

Anyone who wants to see how "innocent" Ethan is when it comes to "scientific" (if only) testing, well click on the links and read what Ethan admits to in his OWN words.

http://www.gearslutz.com/board/attachments/studio-building-acoustics/288684d1335220883-hat-off-jules-ethan-admits-.jpg

http://www.gearslutz.com/board/attachments/studio-building-acoustics/288685d1335221432-hat-off-jules-more-impropriety.jpg

I would ask that Allen not consider this flaming from arguments past. Ethan is the one that brought up his test in Tape Op Magazine, which should make it fair game for rebuttal.

Enjoy!

Mixerman

 
Tue, Apr 24, 2012 - 5:23PM
Mixerman said about this:

Ethan said to Jason: "Yes, I did record some Wave files through a sound card with the wrong settings in the sound card's control panel. The methodology was fine, I just screwed up. So sue me. You never recorded something wrong? That doesn't make my core points incorrect, or me incompetent, or any of the other off-topic things Mixerman would love for people to believe. But that's all the Ethan haters have, so two years later they still use that as an argument. "

Actually, whether you "screwed up" the test or not is in question. In one post, you say you "screwed up" but then discuss two other instances where you completely mislead your subjects, in one case, never even telling them. In the other post, you admit outright that you did indeed deliver tampered files in the hopes to prove expectation bias. That ALONE shows how little you know about testing in the field, because any competent record maker is going make sure the files don't null completely before listening for differences.

Anyone who wants to see how "innocent" Ethan is when it comes to "scientific" (if only) testing, well click on the links and read what Ethan admits to in his OWN words.

http://www.gearslutz.com/board/attachments/studio-building-acoustics/288684d1335220883-hat-off-jules-ethan-admits-.jpg

http://www.gearslutz.com/board/attachments/studio-building-acoustics/288685d1335221432-hat-off-jules-more-impropriety.jpg

I would ask that Allen not consider this flaming from arguments past. Ethan is the one that brought up his test in Tape Op Magazine, which should make it fair game for rebuttal.

Enjoy!

Mixerman

P.S. Sorry for the dbl post. My first one makes it seem like Jason was who I was quoting.

 
Tue, Apr 24, 2012 - 9:16PM
Allen Farmelo said about this:

@Tomas, thank you for those links. I'm going to read and learn. I did read "This is your Brain on Music," and also need to look back at that one. The topic of perception is so complex! That's why these simple AB tests don't cut it for me.

@mixerman - I do consider this flaming from arguments past. Let's not go back there again, ok? You and Ethan might consider couples therapy! :)

 
Wed, Apr 25, 2012 - 2:32PM
Ethan Winer said about this:

LOL at "couples therapy" Allen.

MM as written *hundreds of thousands of words* in dozens of forum threads trying to defame me. Amazing! But why?

--Ethan

 
Thu, Apr 26, 2012 - 3:59PM
Steve Weinstein said about this:

Small correction to this claim made by Tomas:

"As per the Nyquist-Shanon-Koletnikov theorem, frequencies near the sampling frequency are represented with 100% accuracy."

Actually, the theorem says that frequencies up to one-half the sampling frequency are represented with complete accuracy.

 
Thu, Apr 26, 2012 - 6:49PM
Allen Farmelo said about this:

Steve, thank you for this correction - and it's a very important one when considering sampling rates.

For example, 192khz becomes accurate at 96khz, which lies beneath the frequency that most tape machines generate their bias frequency. And it's been shown that some machines can cut frequencies as high as 122khz onto vinyl.

If nothing else, this allows us to understand better where analog and digital formats overlap in terms of frequency representation. It's much closer than we maybe realize, and perhaps there are still some things to learn about these ultrasonic frequencies (Ethan, this is not a claim that we can hear them!)

 
Mon, Apr 30, 2012 - 8:31AM
Allen Farmelo said about this:

I've done a lot of thinking about all of this, and I'm going to throw down my final conclusion at this point:

AB testing is the MP3 of perceptual science. It is filled with distortions and artifacts as a trade-off for convenience, especially when using the internet.

What I think needs to happen to make an AB test at all compelling is to use brain-scanning technologies like MRI to study how the perceptual centers of the brain are working when people listen. And, in particular, we need to see how people's brains are lighting up or not during AB-ing itself.

My hypothesis: AB-ing turns off the all-important right brain that compliments focused left-brain activity. When AB-ing we miss the forest because we're so focused on the trees.

This hypothesis is based on my own experiences as a listener, and this hypothesis is supported by Neil Young's intuitions as well, that sound will wash over you and give you that experience.

Until someone does this level of study, I urge everyone to question the validity of AB-ing as some kind of conclusive evidence of anything. Like so many tests, the effects of the testing itself often turn out to be as much worth study as the thing the test aims to study was.

Don't drink the Kool-AB-Aid!

 
Tue, May 1, 2012 - 1:11PM
Tomás Mulcahy said about this:

Damn, I meant to say "cutoff frequency" not "sampling frequency" (which was leading to my point about the filter ringing problem). Thanks Steve!

Allen, I think there may indeed be something in your hypothesis, but also I think a lot of problems are due to using the tool incorrectly. Statistical analysis is often omitted, because it requires a high level of expertise not generally found in internet forums...

There are some interesting studies of people with no physical connection between left and right hemispheres, where left and right hand actually fight with each other, which seem to show that there is an issue of dominance. However, the notion that the brain is differentiated in this way is currently being discredited, due in part to things revealed by MRI technology.

As always, science answers some questions while creating ever more unknowns.

 
Tue, May 1, 2012 - 9:49PM
Allen Farmelo said about this:

Tomas, you're totally right about the complexity of the human brain and that the roles of the two hemispheres are being revealed as way more complicated than originally proposed.

But, it's been well documented through MRI that areas in the right side of the brain are responsible for making a bigger picture of sensory information that gets processed in areas on the left side. In other words, the "right brain" helps us to integrate the smaller pieces of info into a bigger picture.

My hypothesis (that extensive AB trials causes the right brain's integration to cease) would need to be tested, of course.

However, on an experiential level, how many of us have spent too long focusing on the particulars of a mix only to loose the big picture and need a day away to get "fresh ears" and hear the whole? Or how many of us have to step outside the room, or read a book, or have a conversation, to get our brains to listen more holistically to a mix?

AB testing, in my own experiences actually shots down my ability to really hear more holistically. I recently took a blind AB test to check this out (two identical recordings except for different converters) and I nailed it on the first few tries. Then after a bunch more tries, I got caught up in details of the sound and couldn't get the whole impression. Then I kind of forced myself to relax and just hear the whole again (a mixing skill that's built up, I guess) and voila, back to nailing the test.

I'm really curious about what went on in my brain activity to account for these shifts in my ability to perceive. And I'm fascinated by the paradox of my intentionally focusing harder causing me to hear less accurately.

Anyways, Tomas, you're brought to light the need for me to not casually taut brain theories, since I really am not an expert on that.

 
Thu, May 3, 2012 - 9:49PM
Justin Colletti said about this:

Wow guys. Great discussion here. I was amazed to see how civil and detailed this whole thread was. (A couple of needlessly aggressive remarks from MM notwithstanding.) On the whole, the lot of you have somehow given "debating on the internet" a good name!

I don't have much to add here, except that I greatly identify with Ethan's statement that:

"Blind testing is the gold standard in all branches of science, especially for things that relate to perception. So it makes no sense to me that for some reason audio is exempt."

As much of a romantic as I can be, I just don't think it's safe to ignore how easily our sensual perceptions can be misled or manipulated.

I also agree with Colin G when he writes:

"I take the author's point about A/B testing, but any meaningful comparison has to be blinded; tell listeners one is better and they'll hear a difference even if both are identical. If you want to prove hi-res formats are better, send test subjects away and make them live with each for a month, but, crucially, don't tell them which is which. That is the minimum you need to say there's a genuine benefit to hi-res audio."

Until we do that, I'm content to sit out on any further speculation.

I only have one thing to add, and it's a tiny, silly rebuttal to Allen's analogy about product testing:

"Testing is just as likely to halt the development of a technology. Case in point: the Swiffer. It was shelved because people were given a quick mop/swiffer comparison and the results were "no." Then the developer forced Proctor and Gamble to let people live with prototypes, and those test results were "hell yes!"

My humble experience has been that after the dazzle of the Swiffer's marketing materials and opt-in incentives wear off, most users eventually come to realize the fact that the damn thing doesn't work nearly as well as a decent dustpan and broom. (Not to mention that it's far less cost-effective.)

I for one, wish that Proctor and Gamble had heeded their research on that one, instead of pushing through a huge and brilliantly effective marketing campaign for a product that's essentially inferior to a 5,000 year old piece of technology.

Of course, your own preferences may vary: While it's still true that the mop version of the Swiffer doesn't clean nearly as well (or economically) as a decent sponge mop, it sure is tidy and convenient in a way.

I guess when it comes to every type of product, we all bring our own deep-seated preferences and priorities to the table.

 
Fri, May 4, 2012 - 1:59PM
Allen Farmelo said about this:

I wasn't intending a debate about the efficacy of The Swiffer, and one's personal floor-cleaning preferences don't really impact the fact revealed by the Swiffer case that short and long-term comparison tests produce different results, which was my point.

I've really made my point about AB testing, and would only ask that those of you who are convinced that AB'ing is the "gold standard" take a closer look at what's happening with the explosion of MRI and other brain-scanning technologies when it comes to understanding how human perception works.

One simple general truth from MRI testing: different perceptual tasks use different processing centers in the brain, and (this is key) even when the same exact thing is being perceived.

Participating in an AB comparison is just one kind of perceptual task, and is not the end-all of how the brain perceives sound. And, AB testing it does a great deal to set up the conditions of listening, which in turn sets certain brain processing centers in motion.

Just some of the messages that the AB test sends to the same brain doing the perceiving: try, compare, focus, you're being tested, etc...

These messages prime the brain to perform a certain type of task - that is, a particular WAY of listening is set in motion.

So, if one of the points of AB testing is to show that you "just don't think it's safe to ignore how easily our sensual perceptions can be misled or manipulated," perhaps it's time that the AB testers of the world started to come to terms with how AB testing itself can mislead and manipulate perception.

 
Fri, May 4, 2012 - 2:12PM
Ethan Winer said about this:

Allen, all comparative testing might not be perfect, but that doesn't mean that all such tests are useless either.

--Ethan

 
Fri, May 4, 2012 - 5:14PM
Tomás Mulcahy said about this:

Allen, firstly thanks for facilitating a genuine debate! :)

I'm certain your hypothesis is worth testing. It would be great to see what JJ Johnston or Dan Levitin would think of this idea. I suggest you email one of them.

 
Fri, May 4, 2012 - 6:31PM
Allen Farmelo said about this:

Tomas, yes, good idea. I'll ask them to give a look at my idea.

Ethan, of course, not useless, but not conclusive either.

Lots of ways to listen, lots of ways for the brain to process and filter what it receives from our senses. Lots of ways to influence how the brain does that work.

 
Sat, May 5, 2012 - 1:42PM
Ethan Winer said about this:

Well, some listening tests are indeed conclusive. If you compare a mix of a typical pop tune with an alternate version that has 6 dB of boost at 1 KHz with a Q of 1, everyone will easily notice. So then it's a matter of how small a change can be heard reliably. As I mentioned in my email to you earlier today, this stuff has been researched extensively for many years. So it is not unknowable or even unknown.

--Ethan

 
Sat, May 5, 2012 - 9:00PM
Allen Farmelo said about this:

Ethan, I'm going to stop arguing back on our contention, as I think we've both made our points abundantly clear, and thus our differences in perspective, preferred methodologies, and opinion.

I do like this notion that you raise about "how small a change can be heard." It reminds me of the progressive anthropologist of mind Gregory Bateson's definition of information as "a difference which makes a difference."

I think any good debate like ours (and I include all of us here, and any who join in next) should be raising more and more questions. So, using Bateson's definition of information....

Are there differences that make a difference for some people and not others?

Given this question, where are the thresholds of perception for different people?

To what degree can one improve their perception such that smaller and smaller differences can start to make a difference (developing one's ear, or eye, we'd say), or are there fixed thresholds?

Could one person's difference that makes a difference be another person's placebo effect?

Can we sit on a threshold of perception and bounce back and forth between perceiving and not perceiving?

If so, how does AB testing affect that bouncing back and forth?

What does the brain doing an AB test look like compared to a brain doing, say, an hour, day, year, of relaxed listening?

How many listening types might there be, and how can we discover them, and can they be willfully entered into or not?

What are the most effective brain states for perception of small differences that lie near the thresholds of perception?

Are there sensory inputs that we can't perceive while AB testing but that we could perceive when listening in a different way?

I'd love to hear more questions arise out of this discussion....

 
Sun, May 6, 2012 - 11:17AM
Justin Colletti said about this:

Definitely a great debate Allen! Thanks for helping to facilitate it.

I also enjoyed your own unrelated AB test, which you shared here:

http://www.farmelorecording.com/in-the-press/analog-vs-digital-summing/

In this example, I also think I hear subtle differences between the two versions, and I believe that I too prefer the analog-summed version.

I believe it enough that I'd happily submit to a blind ABX test on it. In this case, I'm pretty sure I could tell the two apart.(And I'd bet that you could too!)

At this point, I've had enough experience with blind ABX tests to feel that they don't just exist to make guys like us look dumb. In fact, ABX tests often prove that some listeners *can* hear subtle differences that the majority might not. I've been happy enough with my results on enough of these tests to have lost any insecurity about taking them, and to see it as a valuable lesson if I get stumped.

(Ethan, I'd be happy to play guinea pig for you on this or anything else. I'd still love to try a blind ABX of a 24/192 WAV file vs a 256kbps AAC myself. And I promise not to mess up on purpose! :)

With that said Allen, your two most compelling points still stand: I agree that people can learn to listen better, and to hear differences that they may not have heard at first. I also agree that the conclusions we come to in short-term tests don't always hold up over time.

I'm looking forward to seeing more studies in this field too! Sure, they may seem frivolous at first glance, but when you consider the massive cost of bringing a new technology to market, the expense is trivial.

I just hope we can all agree that whenever the science seems to be missing something, we should commit to doing more and better science... not to throwing the whole thing out the window!

-Justin

 
Sun, May 6, 2012 - 2:57PM
Ethan Winer said about this:

Both posts above are right on the mark. But I have to say that while we should continue asking questions, a lot of stuff being discussed here has already been settled. Some people may not understand that it's settled, but a lot of it is. One issue is JND, or Just Noticeable Difference. Using my EQ example, it seems we all now agree that at some point a change is too small to matter. This is a huge step forward. Now we just need to get Allen to agree to a method for determining the JND. :->) Once we have that, all the rest - clocks, dither, summing, etc - falls into place.

One problem with comparing analog versus digital summing is ensuring that nothing else has changed other than the summing method. Pan laws are a problem because identical settings can yield real level differences. This can be accounted for and corrected, but it's a lot of extra work.

I do agree that some people can hear things others can't, and that some people have learned to pick out artifacts that others won't notice. But we still have the dilemma of how to test this in a way all can agree is correct and conclusive. BTW, the Artifact Audibility section of my AES Audio Myths video addresses this stuff directly.

--Ethan

 
Sun, May 6, 2012 - 7:56PM
Justin Colletti said about this:

A lot of great points there, Ethan. I'll admit that I do lean quite a bit toward your objectivist line of thinking.

With that said, I think Allen still makes a great point about analog summing:

Although you're right to say that it may be pan laws or a variety of things other than the actual "summing" which create the difference he hears, that doesn't negate him hearing or preferring those differences.

It could be that Allen or I enjoy analog summing because of how it re-interprets our pan choices, or because it adds pleasing and familiar THD, transformer saturation, or even plain-old noise across multiple channels instead of just two. Regardless of the cause, "if it sounds good, it is good."

With that said, I'm all for blind listening. I don't care too much *why* I prefer one sound over another (I'll leave that to you to figure out.) All I want to know is that I'm not misleading myself in the first place! Blind listening is definitely great for that.

 
Sun, May 6, 2012 - 8:18PM
Allen Farmelo said about this:

There's usually a heck of a lot more than panning differences in analog summing situations. Take my console alone, and each channel strip is going to color the hell out of the signal, and then the whole center section is going to do that again. The introduction of THD, especially via transformers and op-amps in my board, is a big part of why analog summing continues to be the preference of so many people.

To make some perfect AB testing situation (which my own comparison test has done as best as possible - see above) doesn't even come close to replicating what is a typical summing method, which usually involves a console filled with circuits that will radically alter the tone of the signals hitting it.

And here's the clincher: you can't get individualized channel treatments like that if you're summing in digital (unless you insert hardware channel strips into the DAW, which adds 2 conversion points, plus the summing output conversion). I mean, there's no way to get those individual circuits into action unless you send out individual hardware channels.

From here the number of variables just goes up and up and up and the fact is that, for about 20 different reasons, summing through a console is a radically different thing than summing in the box.

And, even the DAW summing algorithms will be different from DAW to DAW, and the clocks and converters used will change the sound of the DAW. And, even the most stable, clear, clean-circuit summing boxes (such as the Dangerous D-Box used in my testing) will still impart some level of THD and other coloration.

So, apples & apples? I'm not sure apples even exist.

 
Sun, May 6, 2012 - 8:25PM
Allen Farmelo said about this:

As for anyone checking out my AB of analog and digital summing, please consider these factors in doing any AB listening there:

- you're listening to MP3s made with some codec that pro tools devised. higher res files will reveal more differences.

- what are you listening on? do you have a high quality monitoring path to hear these files on? the better the monitoring path, the more you'll hear the sound of the monitoring path.

- most people are streaming those files, so who knows what streaming does to the sound.

And these are some of the reasons that AB tests done on the internet are limited means of comparing differences.

For those who care:
http://www.farmelorecording.com/in-the-press/analog-vs-digital-summing/

But I'd even steer you away from my own AB test.

 
Sun, May 6, 2012 - 8:38PM
Allen Farmelo said about this:

Ethan writes: "One issue is JND, or Just Noticeable Difference. Using my EQ example, it seems we all now agree that at some point a change is too small to matter."

To matter? What does that mean? To whom? In what context? Being used in conjunction with what? I mean, if we agree that some differences are small enough that only some people can consciously make them out, they'd still "matter" to those who could.

And, there is no evidence at all that because someone can't make something out consciously during AB testing that that difference doesn't "matter" in some way that's yet to be discovered or understood.

Ethan continues: "Now we just need to get Allen to agree to a method for determining the JND. :->) Once we have that, all the rest - clocks, dither, summing, etc - falls into place."

I've recommended many possible avenues of investigation above. And, NO, none of that stuff "falls into place." And, in fact, that technology is on a constant rate of change itself.

And, not to get into mixerman territory here, but Ethan we'd really need to know which clocks and which summing mechanisms, and which converters you're even talking about. You might want to spend some time getting up to date on the cutting edge of this technology and get some people with developed ears involved.

One of my new rules of thumb: it's hard to get expert subjects listening to great gear because both tend to be really busy.

And with that, I have to say again that anyone wishing to use an AB test to refute the claims of people like me and Neil Young at least not use the internet as the means for conducting the test. The variables that exist in the end users' playback system alone are enough to make it obvious that such a test is not going to hold much water.

 
Sun, May 6, 2012 - 8:43PM
Allen Farmelo said about this:

TYPO: "the better the monitoring path, the more you'll hear the sound of the monitoring path." should read "the worse the monitoring path..."

 
Sun, May 6, 2012 - 10:07PM
Justin Colletti said about this:

"As for anyone checking out my AB of analog and digital summing, please consider these factors in doing any AB listening there"

Do you really think anyone other than us three nerds made it this far? :D

For the record Allen, I think that blind ABX testing would probably *prove* your claims about running your mixes out onto the console and through some transformers -- not refute it.

Anyway, great hearing all your thoughts, guys! For real. I think I'm gonna bow out now, but it's been both subjectively *and* objectively awesome.

 
Mon, May 7, 2012 - 7:07AM
Allen Farmelo said about this:

Thanks for popping in, Justin.

From the liner notes to the 1954 RCA Victor LP "Hearing is Believing."

"Many people are confused about what High Fidelity means...and a lot of the confusion comes from the technical language sound engineers use when they try to explain it. So in this record we are not trying to explain the technique that made High Fidelity possible - we are simply giving you a chance to hear the tremendous improvement it makes in recorded sound. First the music is played as in ordinary recordings. Then you hear the same composition in RCA Victor "New Orthophonic" High Fidelity...the most faithful sound on records. As you compare the two, the principal points that distinguish High Fidelity sound from the old are clearly brought out."

they go on...

"You can probably tell the difference between the old sound and the "New Orthophonic" High Fidelity recordings when you play this record on your present instrument. But for a really spine-tingling experience in sound [sounds like Neil Young's description, right?] you should hear it on one of the RCA Victor's "New Orthophonic" High Fidelity Victrolas."

And there you have it - the "probably" of old-school playback systems revealing differences, the marching forward of technology, the AB test as the way around the mumbo-jumbo, and a new, better playback system meant to improve everyday listening.

Neil Young was 9-years-old, my parents had just married, the first color TV was released (by RCA), and Toscanini (who appears on this record) retires due to memory loss.

It's a hell of a thing to consider the history of humankind striving to capture and replay one of its most beautiful creations in the best way possible.

The MP3 was a huge setback, and the bleeding-edge of new digital technologies (from clocks, to converters, to sub-atomic info storage devices, to hand-held-hi-fi systems) will culminate soon in new audio formats and playback system that will get us back on track.

The aesthetic revolution will be beautiful!

 
Mon, May 7, 2012 - 2:22PM
Ethan Winer said about this:

Allen wrote: "if we agree that some differences are small enough that only some people can consciously make them out, they'd still 'matter' to those who could. And, there is no evidence at all that because someone can't make something out consciously during AB testing that that difference doesn't 'matter' in some way that's yet to be discovered or understood."

Do you really believe that *no* change to an audio signal is too small to matter in any situation?

Also, about your summing box adding distortion, that's outside of what I discuss. Yes, of course, a summing box that adds intentional distortion will change the sound audibly if the distortion is great enough. But that's very different from the claim I often see that clean analog summing sounds different than clean digital summing. If both are properly implemented, any change to the sound should be too small to notice.

Finally Allen, it would be great if you replace your MP3 fragments with shorter excerpts that are not compressed. Or did we discuss that already? Not only will it more clearly reveal whatever differences there may be, it will also allow using nulling / subtraction to identify the differences and their magnitude.

--Ethan

 
Mon, May 7, 2012 - 7:33PM
Allen Farmelo said about this:

Ethan: "Do you really believe that *no* change to an audio signal is too small to matter in any situation?"

Of course not. And I also know placebo effects very well, first hand - who in a recording studio hasn't by accident from time to time tricked themselves? We laugh it off and move on.

What I'm talking about are differences that lie in the area at the borders of perception, where some can hear them and some can't, where they might even matter even if someone can't consciously make them out. As far as I can tell no one has done that test.

I feel like a broken record at this point, Ethan, so let's drop it here and let some other voices chime in (or not).

I don't have time to replace my MP3s on that test, and I'd have to dig up some old drive somewhere. Many people can make it out on the MP3s no problem - it's a spacial thing, mostly.

 
Tue, May 8, 2012 - 12:10PM
Ethan Winer said about this:

Allen said: "What I'm talking about are differences that lie in the area at the borders of perception, where some can hear them and some can't, where they might even matter even if someone can't consciously make them out. As far as I can tell no one has done that test."

Allen, *I* have done that test. Many of them, many times. That's a big part of my AES Audio Myths video, and there's even more in my book. The real question for me is why people believe that material too soft to hear can still have an affect.

--Ethan

 
Tue, May 8, 2012 - 8:33PM
Allen Farmelo said about this:

Ethan, you make it hard to let it go, man.

Let's work by analogy for a sec:

I used to race bicycles, and was a climber, and we'd obsesses over lightness of bikes and bodies.

So, like, the weight of say my brake levers could be reduced. Could I tell? No.

But so could my seat post. Could I tell? No.

And my seat? No, can't tell.

And the pedals? I don't think so. No.

And the tires? Nope, can't tell.

And the fork? The stem? The handlebars? Nope, can't tell.

Shoes? No.

How about the chain with holes in it? Now that's dumb. Really?

What about the crank, and what if you throw your water bottle in the gutter when it gets steep?

No. None of those things matter.

What about a ceramic axle set?

No, can't feel it.

Ok, so we'll just put all the old stuff back on the bike, ok?

Hell no! The cumulative effects are over 2 lbs!

That's what I mean about stuff having an impact even if we can't make it out individually. Cumulative effects.

You can go ahead and isolate them for an AB test, but when I'm making records there are at least four small differences that improve sound across the whole digital system:

- balanced power
- excellent clocks
- excellent converters
- higher sampling rates

Apparently all red herrings according to your AB testing, but I prefer to go climbing on a serious racing bike, myself.

 
Tue, May 8, 2012 - 9:32PM
Allen Farmelo said about this:

Justin Colletti nails it in his new piece:

"Whenever I?m able to hear subtle difference between nearly identical sounds in AB tests, it?s because I stop thinking and listening on a conscious level. It?s much better to treat the sound like a 3D poster..."

And here in lies the blind-spot of blind AB testing: no one is looking at the brain-states of the people doing the listening!

This points to my hypothesis: people MAY be failing or passing AB tests due to certain WAYS of listening employed.

Until we know that, we don't know much about what AB testing is revealing.

 
Tue, May 8, 2012 - 9:35PM
Allen Farmelo said about this:

Colletti's new piece is here:

http://trustmeimascientist.com/2012/05/05/results-from-our-audio-poll-neil-young-and-high-definition-sound/#.T6ltGTV8YAw.facebook

And I'm feeling like this forum is becoming a bit too much like some others where arguing and personalities take over the vibe. Going to bail until some new ideas fall on the screen here.

Thanks everyone!

 
Wed, May 9, 2012 - 12:47PM
Ethan Winer said about this:

Allen, I agree fully about cumulative effects. The difference between your bike example and audio is that the weight difference with the bike is easily measured! And that 2 pound difference is significant. A racing bike is 15 pounds or less, so 2 pounds is more than a 1 dB change. :->)

All of the stuff you listed either makes no measurable change at the output of the gear, or the change is so small as to not matter. Now, you can argue that it does matter, and you can hear that it matters. But then we're back to "prove it's not just flawed perception" which then gets back to blind tests.

Yet again, I'll consider any reasonable method for you or others to prove they can hear this stuff. It doesn't have to be a traditional blind test. But it has to be compelling. If you think I'm being unreasonable insisting people prove they can really hear what they claim, I'd love to know why!

--Ethan

 
Wed, May 9, 2012 - 6:27PM
Allen Farmelo said about this:

"Please accept from me this unpretentious bouquet of very early-blooming parentheses: (((())))."

- J. D. Salinger

 
Thu, May 17, 2012 - 2:31PM
Mixerman said about this:

Ethan wrote: "Allen, *I* have done that test. Many of them, many times. That's a big part of my AES Audio Myths video, and there's even more in my book. The real question for me is why people believe that material too soft to hear can still have an affect."

Ethan, you're nearly deaf (sue me, I'll prove it in court), and you have no idea what a critical listening chain is. When you finally did deliver a file that wasn't purposely Winered by you, nearly everyone could hear the difference even though you said it was too small to be perceptible. You were wrong. Still are.

Enjoy,

Mixerman

 
Fri, May 18, 2012 - 3:11PM
Ethan Winer said about this:

Ah geez, Eric is back with more insults (only). I guess you haven't found my rebuttal yet. :->)

 
Mon, May 21, 2012 - 11:19AM
Allen Farmelo said about this:

Given the long (-winded) history of Ethan and Mixerman having at each other in other forums, I am going to request (softly and kindly) that if either of you have something NEW to say, then jump in. But please stop dragging out the same-ole same-ole....thanks.

 
Mon, May 21, 2012 - 12:12PM
Ethan Winer said about this:

Agreed 110 percent Allen. As you have seen, I'm not the problem. I never chase people around audio forums to insult and demean them. It's *always* others, like Eric, who do that to me. As for something new, hopefully this will put a stop to MM's harassment for a while:

[url=http://www.ethanwiner.com/mixerman.htm]Mixerman Exposed[/url]

--Ethan

 
Mon, May 21, 2012 - 12:13PM
Ethan Winer said about this:

Sorry, wrong forum format. :->)

Mixerman Exposed:

http://www.ethanwiner.com/mixerman.htm

 
Mon, May 21, 2012 - 5:35PM
Allen Farmelo said about this:

Not exactly complying with my request for something NEW, Ethan.

Let's return to something more elegant than a cyber-fued, please.

 
Wed, May 23, 2012 - 4:17PM
Ethan Winer said about this:

Okay Allen, you asked for something new to add to the discussion. This hasn't been mentioned here yet:

The Dishonesty of Sighted Listening Tests:

http://seanolive.blogspot.com/2009/04/dishonesty-of-sighted-audio-product.html

Sean Olive is a serious audio-science guy, and his arguments are well thought out and expressed clearly IMO. See the About Me section partway down the right side of the page for his credentials and background.

--Ethan

 
Wed, May 23, 2012 - 10:31PM
Allen Farmelo said about this:

Olive's article is helpful if you're the type who's worried that the $2500 you just spent on 7' speaker cables was a jive deal.

otherwise it's just another audiophile saying that blind A-B tests are as good as we can do.

Again, I still heartily disagree....probably more now than I did when I wrote the original post, though my tolerance for the topic is waning faster than Pluto's second moon.

 
Thu, May 24, 2012 - 10:11AM
Ethan Winer said about this:

Sean Olive is hardly "just another audiophile." He is a serious researcher-dude who knows a huge amount about psychoacoustics, and all the related issues.

It's obvious that nothing will change your opinion, no matter how compelling, and no matter who explains it. As I told you by email, this stuff has been researched extensively for many years. So believing that serious researchers / experts / scientists have somehow missed a key failing of blind testing falls under "extraordinary claims demand extraordinary proof."

Anyway, you did ask for something new to the discussion!

--Ethan

 
Thu, May 24, 2012 - 1:23PM
Allen Farmelo said about this:

NEWS FLASH: advertising and branding sway human perception! Call the press!

 
Mon, Jun 18, 2012 - 5:05PM
Luke said about this:

Let's say you witness a car crash. It makes a strong impression. Let's say you see a car crash in a movie in the theater. Still could be a strong impression based on circumstances, but probably less so, no? What about on TV? Even less so, maybe??

Now, let's say your friend comes up to you and says someone you know crashed their car. Now, let's say your friend calls you up and says someone you know crashed their car. Now, let's say your friend texts you and says someone you know crashed their car. In my experience, much less difference between those three.

I equate the quality of the recording to the first hypothetical and the quality of the song to the second one.

 
Fri, Jun 29, 2012 - 3:20AM
Tomás Mulcahy said about this:

It bothers me that one side is arguing hypotheticals and using poetic analogies, while the other side has a bunch of disparate test data that superficially has commonality, but really is not unified at all because there's no test where several of the issues in contention were tested together. Most worrying is that neither side appears to have more than a passing awareness of cognitive science- which I think is the key here. Can we stop these tit for tat arguments and go look for real data, or at least papers documenting actual research? Because I think both "sides" can learn a lot from each other!

 
Sat, Jul 21, 2012 - 7:24PM
Allen Farmelo said about this:

Thomas, I know what you mean. However, I think you mean perceptual studies, rather than cognitive (how we sense things, rather than how we think), and I also feel whole-heartedly that - unless there's some benefit to humankind - such research won't, and shouldn't, get done.

As for my use of "poetry": language is language, and the construction of my sentences is generally spot-on grammatically and does use quite a bit of metaphor and simile, which may seem "poetic," but by no means does the use of metaphor and simile somehow magically undo the logical structure of my ideas, nor is the use of such devices not to be found in scientific writing. To the contrary: the best scientific writers uses metaphor and simile to add clarity to the expression of their ideas.

 
Sun, Jul 22, 2012 - 1:43PM
Tomás Mulcahy said about this:

Alan,

No, I meant cognition. The whole process of decision making.

I was not objecting to good writing or poetry. I do feel it's problematic when people rely on thought experiments (at best) or assumptions (at worst) rather than actual tests or experiments. Then try to paper over the gaps with flowery language. That is my point. It is quite amazing how much "the brain" fools us all the time when making decisions. This stuff is discussed a bit in "This is your brain on music".

 
Mon, Aug 20, 2012 - 11:24AM
Allen Farmelo said about this:

Tomas - the hegemony of "scientific language" truthfully cracks me up. I mean, it's like if I use any form of passionate expression, then I'm not being scientific. Focus on what's being said or not - on the ideas - and not the word choice. It's what I have to do when coming up against "scientific language" myself; I filter out the seeming lack of passion (while noting that to have written at all indicates a motive, perhaps quite passionate) and look to the nugget of logic being expressed.

I can't be held responsible for the kinds of tests I'm calling for having not been done, but calling for them does put the current A-B paradigm into question. It's 100 years old, this critique of mine, and it's called The School of Qualitative Research.

"paper over the gaps with flowery language" - nice one, that. My language is surely not any worse than people trying the just-got-off-my-horse-swagger of scientific objectivity's own style of sentence construction and vocabulary.

The presence of passion does not indicate a lack of logic. Don't fall for that old game.

Again, let's just engage the ideas and not sling scientific approaches around like some kind of truth-guaranteeing-silver-bullet.

All linguistic concerns aside: Has anyone ever checked to see if A-B tests themselves are affecting how people listen/taste/see/etc...? If you can answer that, then this discussion can move forward without falling into critiques of each others word choices.

 
Sun, Sep 9, 2012 - 8:09AM
Tomás Mulcahy said about this:

"The presence of passion does not indicate a lack of logic". Alan, again, I never said that, and that is not the objection I am making. The language in itself is not the problem. It is the lack of evidence that is the problem.

My issue is that all we have here in this blog are at best thought experiments. We don't have actual empirical evidence of a lot of the things that people are taking a stand on. It is all assumption, nothing is concrete.

At the very least, we could apply some thought to designing tests for the ideas/ arguments/ objections being put forward. Demonstrate that they work or not. But arguing about them does not make the point anything more than imagined. Evidence makes it real. Or not.

It's actually very difficult to design good tests, not least because of the necessity to apply statistical analysis. AB is only a very general category, it is a starting point on which to build rigor. It is not in itself rigorous.

Your question about "testing AB tests" shows that perhaps you haven't read how tests are actually done? Bias and placebo are just two things that researchers try very hard to eliminate in a test, for example by using control groups. It's not a good test if it does not have built in procedures for validating the methodology and controlling for/ reducing variables.

Ben Goldacre's "Bad Science" is very good on how tests are done, and critiquing them. Actually I'm just reading a book called "The Runner's Body" which has some very good critiques of various research. Either of these books would be good pointers on test design. If you're into running the second one especially is great! :)

 
Mon, Sep 10, 2012 - 12:06PM
Ethan Winer said about this:

Very good post Tomás!

--Ethan

 
Tue, Sep 11, 2012 - 11:50AM
Tomás Mulcahy said about this:

Thanks Ethan. I failed to address the issue of qualitative research. Alan I agree that that kind of approach is relevant here, but it does not preclude quantitative research.

Basically what I think is needed is for us to design some comprehensive tests, as a first step. Creating dichotomies such as "emotion" v "knowledge", "positivism" v "postpositivism", "pro" v "scientist" etc. etc. is divisive and counter productive. We need to use several approaches to give us different "angles" on the issues under discussion here. Otherwise, we'll just go around in circles.

What we need is evidence. How we get it doesn't really matter as long as it is achieved with genuine rigor and self reflection.

 
Mon, Oct 1, 2012 - 7:55PM
Allen Farmelo said about this:

Tomas wrote: "At the very least, we could apply some thought to designing tests for the ideas/ arguments/ objections being put forward. Demonstrate that they work or not. But arguing about them does not make the point anything more than imagined."

I believe I've been very clear in stating in various places that what I'm presenting is a hypothesis. Given the level of commitment to, and knowledge of, the scientific method, I don't need to tell you guys that a hypothesis is "imagined" until tested. Hypothesizing is the imagination part of the scientific method - the insight, the dreaming up an idea, the impulse from which new investigation emerges. That's the point from which all of what I'm saying, and did say, exists.

So, again, let me state my hypothesis: A-B testing of sensory input itself (the testing) affects how the subject perceives the input.

How to test this hypothesis? This is an open question, and you're right: tests need to be designed. I have strongly suggested that the rapidly developing technologies that allow us to watch the brain's activity would be a way to compare different modes of listening. I have suggested that in such a test one compares how the brain is operating during an A-B test and during a more "passive and immersive" mode of listening.

 
Mon, Oct 1, 2012 - 7:56PM
Allen Farmelo said about this:

...continued from above due to space.....

One of my greatest opponents in this discussion, Justin Colletti, has said as much, and I quote: "Whenever I?m able to hear subtle difference between nearly identical sounds in AB tests, it?s because I stop thinking and listening on a conscious level. It?s much better to treat the sound like a 3D poster or Zen koan, allowing your mind float off, and letting each sound wash over you for 15, 30, 45 seconds or more." (http://trustmeimascientist.com/2012/05/05/results-from-our-audio-poll-neil-young-and-high-definition-sound/)

What Justin is pointing to in this quote is exactly what I want to have tested: the possibility that there are what I call "modes of listening" and that we can intentionally engage them.

Again, this is a HYPOTHESIS, and as far as I know it has not been tested.

And, Tomas, I am not trying to say that it's an either/or situation: A-B'ing has its very important place, but to say it's the be-all-and-end-all of empirical measurement and then apply it to every listening situation is - as you've noted by suggesting that we need multiple approaches - not good enough.

And this brings me back full-circle to my original post: We just don't listen to music in our everyday lives in and A-B'ing mode - and in fact there is no B, just an A (a bit of the Zen Justin mentions creeping into my language here, one hand clapping and all that - I apologize if it's too floral, not intentional...perhaps inherent to the topic).

So, when I saw people using the A-B mode of listening to prove something about a product that is intended for a very different mode of listening, I had to draw their conclusions and methods into question.

I am a strong strong believer in the power of intentional passive and immersive listening as a way to engage sensory experiences that more active and focused listening wont provide.

And I just want someone to accept that it's POSSIBLE that A-B testing itself can be shutting down another mode of listening that allows us to hear the differences Neil Young is putting on the market.

Colletti's quote - filled with poetic language - points the way.

Tomas, the evidence you seek is likely not going to happen because the time/money/energy will - I hope - be spent on more important research. I can only hope a cure for some disease is being studied instead of this.

I'll buy a Pono system ASAP. No one will get hurt.

 
Thu, Nov 8, 2012 - 5:33AM
Tomás Mulcahy said about this:

Well Alan, I think the test you described in your initial post would be fairly straightforward to do, and worthwhile. Now that you've clarified your hypothesis we might have a sound basis for a paper that could be submitted to AES. I'm willing to help with making the test happen. None of us are medical researchers so no humans will be harmed in this test! :)

However, I still think you should look at some papers where AB testing is done, especially the good ones which apply statistical analysis and controls to account for the variables. It's not clear whether or not you have. It could only strengthen your point. Take a look at the AES library. There are definitely a few good ones in there. Some terrible stuff as well!

Finally, there is quite a lot of good info out there about how our hearing actually works. You might be surprised at how much is known- definitely nowhere near everything, but certainly illuminating. Take a look at the various US universities who make course content available. A lot of what is contentious in this thread would be debunked. I have a bunch of links from a colleague, I can post those if you like.

 
Wed, Nov 14, 2012 - 10:41AM
Ethan Winer said about this:

Yes Tomás, please post some links. I'd like to learn more about what's already known. Mostly I go by my own tests that I've done alone and with friends.

 
Tue, Dec 18, 2012 - 6:34AM
Tomás Mulcahy said about this:

Hi all,

Here are some links.

http://www.smashingmagazine.com http://tech.ebu.ch/loudness/
http://en.wikipedia.org ? search on peak meter, DB, SPL, digital audio
http://www.sengpielaudio.com/Calculations03.htm
http://www.independentrecording.net/irn/resources/freqchart/main_display.htm
http://www.thewelltemperedcomputer.com/KB/1_intro.htm
http://www.sfu.ca/sonic-studio/handbook/Decibel.html
http://www.sfu.ca/sonic-studio/handbook/Intro2.html
http://www.aes.org/
http://www3.ebu.ch/cms/en/sites/ebu/home.html
http://www.w3schools.com/
http://www.physics.gla.ac.uk/~johannes/SciFun/1X%20Physics%20lecture%202.pdf
http://www.sengpielaudio.com/calculator-wavelength.htm
http://www.epa.gov/radiation/understand/ionize_nonionize.html
http://www.intmath.com/trigonometric-graphs/music.php
http://hyperphysics.phy-astr.gsu.edu/hbase/sound/loud.html
http://www.acousticecology.org/wind/winddocs/noise/Thorne%20thesis%20intrusive%20noise%20low%20amplitude%20sound.pdf
http://www.independentrecording.net/irn/resources/freqchart/main_display.htm
http://audiophysic.de/aufstellung/index_e.html

Happy Christmas! :)

 
Tue, Feb 26, 2013 - 2:48AM
Tomás Mulcahy said about this:

One more- this is good on desacribing different kinds of tests above and beyond blind AB, and does a great job of introducing statistical analysis:
http://www.tonmeister.ca/main/textbook/intro_to_sound_recordingch6.html

 
Sat, Mar 16, 2013 - 1:23PM
bern said about this:

Just a reminder - Neil and I are the same age. Our hearing is no longer what it used to be, and I have not spent the years on the road with a loud band, I listen at less than 85dB in the studio. So, can we really have something valuable to say about what a teenager hears?

 
Wed, Apr 17, 2013 - 3:50AM
D7100 said about this:

This is my first time go to see at here and i am
in fact pleassant to read everthing at single place.

 
Wed, Jun 19, 2013 - 3:23PM
bern21 said about this:

All well and good, but my wife, an intelligent and educated person, doesn;'t care about fidelity, lyrics, harmony, melody, solos, etc. If it makes her dance, it's good. If not, who cares? I'm an engineer who goes through convulsions to make the sound in the studio as perfect as possible and highlight the fidelity of my mixes/releases. If they make her dance, they are great. If not, who cares? I hate dance music. She hates not-dance music. So, where are we in the quest for fidelity???

 
Tue, Nov 5, 2013 - 5:21PM
la loi duflot said about this:

je m?intéresse à ce sujet car nous avons aussi le même problème en France avec la loi duflot qui s'occupe des logements neufs et des investissements immobiliers.

 
Thu, Apr 10, 2014 - 5:50PM
Ignacio Etchegaray said about this:

This is great! Someone knows what he is doing and I know this will even improve music quality and it's demand!

How can people who has never heard hi-res recording be able to tell the difference without having time to internally process and feel it for a while is a great point. We are not computers.

For me high quality sounding music is not meant to be listened only in the studio by the personeel in any way, and I think any artist would like to deliver their work in the best possible medium available.

I think that brick was missing, and even today when the digital technology is at a great state the best distributable format to buy in stores is the same as it was in the 80's.

The end of the chain wasn't only the mastering for a storage medium, but the storage medium itself.
And we have been aiming to the CD and radios for years, and now mp3s.
I think we should be happy about what Neil Young has started.
We put all our efforts in getting the best sounds for our music. It's time for us as a community to work hard to improve the formats too.

The market has been just fine selling cds and even more fine selling digital compressed files, in some point the quality went off the road and lost the war against quantity with loosy encoding.
We know it can sound better!

There are lots of audio guys who just go: why care if this is ending in an mp3 or 44.1/16.
This is gonna change that, and let people start hearing great sounding records in devices that can translate them as they were meant to sound.

 
Mon, Aug 18, 2014 - 5:07AM
electronic cigarette said about this:

e cigarette oxyelite pro reviews dangers http://topelectroniccigarettesreviews.com

 More Entries 
Larry Crane · July 5, 2011
What? Huh? That's right, now you can "use classic studio gear online." This guy Fredrik created some way to remotely set compressors and equalizers and then pass your audio through them. Pretty genius and probably the wave of the future for certain...
Larry Crane · Jan. 19, 2009
Here's some links and stuff from a thread I started at the Tape Op Message Board forum: David Byrne interviews Thom Yorke about the Radiohead "In Rainbows" download in Wired Magazine. Pandora Radio is an internet radio site that allows the...
Scott Evans · June 14, 2011
  Lots of laptops use an accelerometer to detect that the computer is falling. When that happens, the computer parks the hard drive heads to prevent them from bouncing on the HD platter. Apple calls this the Sudden Motion Sensor. On...
Dave Middleton · May 17, 2013
With 3 number one albums in 2012, Grammy winning mix engineer Ruadhri Cushnan is at the top of this game. Shortly after Mumford and Sons' album "Babel" was awarded 'Album of the Year' at the Grammys, Ruahdri made time in his busy schedule to chat...
Larry Crane · Sept. 5, 2010
I'll be attending and pretending to be a panelist at The 2nd Annual Recording Summit in Nashville, held Nov 12-14 at Welcome to 1979 Studios. Check the link above to become part of this crazy cool event!
Larry Crane · Aug. 14, 2008
Quantegy , despite the oddity of still being an active link, posted this announcement recently: "Plans are being formulated for the revival of Quantegy 499 Gold Studio Mastering and GP9 Platinum Studio Mastering audio products, as well as the Black...
Allen Farmelo · April 6, 2014
As we exit the era of digital audio data compression and begin the ascent into higher sample rates for consumer consumption (via multiple platforms, though most popularly Neil Young's PONO system), we are finding ourselves as a community of...
Larry Crane · March 23, 2010
For those that couldn't make it, we'd like to report that the Tape Op Party held during SXSW at the Beauty Bar on March 19th, 2010 was a freaking blast. We had a wonderful set by the band Quasi, DJ sets by Ian McLagan, Rebecca Gates, Richard Barone...
Larry Crane · July 2, 2013
A year ago I read this quote from self-proclaimed expert music biz blogger, Bob Lefsetz:  "We live in a lo-fi era. You can focus on sound quality, but most people can't hear it. There's a chance hi-fi is coming back, but do you really have...