Like it or not, Pro Tools is the closest thing to a standard in the world of DAWs; most every commercial studio needs to run Pro Tools, and if you’re working at a certain level, you need Pro Tools|HD. So, about 7 years ago, I upgraded my Digi 002 (Tape Op #33) and Apogee Rosetta (#40) system running Pro Tools LE to a pre-owned PT|HD2 PCI system. Personally, I never had any problems with the Digi 002. I made a lot of records with it. It worked great, never ran out of CPU power, and for the way I worked, it had plenty of tracks available. (The Digi 002 system is still used on a daily basis in our B-room with no complaints.) But clients demanded PT|HD, especially when they were bringing in sessions with track counts in the triple digits. The HD system worked out great, and it was nice not to have clients who were frustrated with the LE system not meeting their needs. Over the years, I bought two more pre-owned HD cards and a Magma expansion chassis (#67) to house them in. The system worked — and still works — fine. But, I can see the writing on the wall. My legacy system uses PCI, not PCIe cards, so I knew at some point, it would be another candidate for the electronics recycler — four PCI cards originally valued at over $20,000 retail rendered worthless by the inevitability of computers relentlessly marching forward. I also realized that we rarely came close to using the TDM processing power of the cards; most of our clients tried to avoid TDM plug-ins because they didn’t transport as well and didn’t mix well with RTAS plug-ins. I often had less experienced engineers call me into a session to “solve the problem” of a TDM plug-in inserted before an RTAS plug-in, at which point I’d have to explain that you can’t actually do that. As computers began to get more powerful, I started to wonder why I really needed all those on-card DSP chips for the TDM plug-ins? Most DAWs besides Pro Tools had been running native for years now. Was TDM really the reason Pro Tools was “Pro”? My guess is that it wasn’t.
So, when Avid announced Pro Tools|HD Native, based on a single PCIe Core card, I was eager to try it out. It promised many of the benefits of PT|HD, including low-latency monitoring and unabridged track-counts — minus TDM plug-ins — but at a fraction of the cost of an HD TDM system. When I looked at all the plug-ins I owned, there were very few that were TDM-only that I really used. Conversely, some of my favorite plug-ins were part of the Universal Audio UAD family, and my aging PCI-equipped Mac could only run UAD-1 cards. Even though I had two of these cards, many of the newer plug-ins wouldn’t run on my cards, and some of my more power-hungry favorites, like the Moog Filter and Neve 33609, really bogged them down. So, as part of my upgrade, I also installed a new UAD-2 Quad PCIe card ( Tape Op #67, #76). My gamble was that my two new PCIe cards that cost about $5000 total would replace the six PCI cards that originally cost over $20,000. Another factor in my decision was looking towards (nobody looks forward to them) future Pro Tools upgrades. Every PT upgrade I’ve ever done has gone fairly smoothly — except for the plug-ins. Every major upgrade to PT has meant losing a few plug-ins forever and spending days getting the rest authorized, upgraded, and blessed by the appropriate authorities. When I got my first Pro Tools system, third-party plug-ins were essential because the stock Digidesign ones were terrible, but that’s not the case anymore. The Avid EQ is really quite good, as are all their effects and instrument plug-ins. Moreover, when I rely mostly on the Avid plug-ins, my sessions transport better. But, sometimes I really want that EMT plate (Altiverb, TL Space, UAD), Pultec (Bomb Factory, UAD), or Fairchild 670 (Bomb Factory, UAD). What’s the common thread to that last sentence? Right, UAD. How much easier will future upgrades to Pro Tools be if the majority of my plug-ins are from two brands, Avid and UAD? Lots easier. I’ve also found both companies to be very responsive and quick when I needed support. As I went down my list of plug-ins, there were really only a few non-Avid and non-UAD ones that I cared about or clients asked for: Eventide ( Tape Op #55), SoundToys EchoBoy (#62), Auto-Tune, Altiverb (#28), Drumagog (#84), Wave Arts Restoration Suite (#60, #63), and Sonnox (#32, #41, #68). If future upgrades only involved a limited number of plug-in bundles, they would go much smoother. (Note to Eventide: Get on the UAD platform and license your code to them! Your sales would go way up.) So, I bought a new 8-core Mac Pro, and I was ready to put my theory to the test.
My day one experience went smoothly. Number of hours spent installing the HD Native card, Pro Tools 9, the UAD-2 card, and the UAD plug-ins: less than one. Calls to tech support: none. Visits to websites: a single 1-minute visit to uaudio.com to authorize plug-ins. In one hour, I had 85% of the new system up and running in my B-room, with an Avid HD Omni interface connected to the HD Native card. The rest of my plug-ins transferred over pretty easily — a few upgrades and downloads needed here and there.
Once everything was more or less loaded and ready to run, I pulled up a mix I had done for the Portland, OR band Archeology that was pretty dense, using 87 of the 192 available voices and a fairly high plug-in count. It ran flawlessly without a glitch. I opened the system usage meter and could see we had horsepower to spare. CPU usage was under 5%! Disc usage was closer to 40%, but that’s to be expected with so many tracks. When I increased the buffer size, it went down to the low 30s.
Next, I started inserting some additional plug-ins to see how the system reacted to them. On our HD4 TDM system, whenever you inserted a plug-in during playback, you’d get an annoying pause in the audio lasting at least a second or so — pretty distracting to the mix process. With HD Native, plug-ins become active almost instantaneously; even the UAD plug-ins can be inserted on-the-fly with barely an audible glitch. Speaking of the UAD plug-ins, these run smoother in Pro Tools than they ever have. Even the extremely CPU-intensive ones that require 60-70% of a UAD-1 card only utilize 2-3% of the UAD-2 Quad.
The next big test was to move the computer into the A-room and see how well it connected with our existing PT|HD interfaces, a first-generation 192 I/O and an Apogee Rosetta 800 with an X-HD card. I wish I could say this went smoothly, but it definitely did not. To make a long story short, the system was glitchy and buggy, and the interfaces seemed to randomly disconnect. We spent a frustrating few weeks having to swap back and forth between the old TDM rig and the new HD Native rig as sessions were going down. The folks at Avid were really helpful and helped me troubleshoot the system. (Thanks to Kyle Splittgerber and Bobby Lombardi.) We initially suspected the cabling, but finally, the support crew at Apogee confirmed that the X-HD card in the Rosetta needs a firmware upgrade, that unfortunately, would not be available in time for this review. As we were in the middle of a week-long mixing session, we swapped out the Rosetta for the HD Omni, and the system was totally stable. Since then, we’ve upgraded our I/O to an Apogee Symphony (more on this in a future issue), and the system has been totally stable.
Over the span of a few months, all of our regular clients mentioned that the HD Native system is much faster than our old TDM system. Also, not a single client ever mentioned that latency was an issue for tracking, overdubbing, or mixing on the new HD Native rig, with the typical headphone monitor mix being created in the internal Pro Tools mixer (more on this below). Most of the sessions we’ve hosted so far have been at 44.1 or 48 kHz, but the San Francisco band Tartufi had a very dense 96 kHz session running with no problems. I oftentimes took a peek at the system usage when clients were in, and it was always reassuringly low. Later, I reopened the Tartufi session, which utilized 89 of the available 96 voices at 96 kHz, and I set the buffer to the lowest available setting of 64 samples; upon playback, the CPU usage was less than 15%, and disc usage was around 40%. I added plug-ins to see how the system would handle them, and even power hungry plug-ins seemed to have minimal effect. I’m sure if I tried hard enough, I could create a huge, 192 kHz session with lots of plug-ins that the system would choke on, but I run a studio making records on real-world budgets, and we don’t really encounter sessions like this. I’m guessing for most recordists, HD Native coupled with a fast computer will more than meet their needs.
Regarding the sound quality of HD Native, I was recently chatting with engineer John Paterno (Tape Op #54), who was curious about our impressions as he had considered HD Native but decided to stick with his HD TDM system. He felt like the audio in PT Native was lacking in some areas compared to his HD TDM system — especially when mixing in the box; he preferred the sound of the fixed-point math over that of the floating-point math in PT Native systems. To be honest, I hadn’t paid much attention to this debate, and as I mentioned earlier, I have made a lot of records with LE and a 002 that I was really happy with. But, I was curious about this, so I called Bobby Lombardi at Avid to try and get a better understanding of this, as I had assumed that HD TDM and HD Native were functionally identical except for the plug-ins they could run. It turns out I was mistaken. HD TDM systems use a 48-bit fixed-point mix engine. HD Native as well as Pro Tools 9 on other interfaces use a 64-bit floating-point mix engine. They are definitely different and will sound different. How different will depend on a lot of factors: how many tracks, how many plug-ins, etc. Moreover, TDM plug-ins dither audio to 24 bits as it moves through one plug-in to another or to the TDM mix engine. PT 9 and HD Native dithers down to a 32-bit floating-point word between plug-ins and mix engine. Some HD Native and PT 9 users have said that they feel 64/32-bit floating-point sounds better than HD TDM, while others like John prefer the TDM mix engine. Does HD?Native sound as good as, better, or worse than HD TDM? I can’t answer that question for you. It almost certainly will sound different, especially with a higher track count and lots of plug-ins when you’re mixing in the box, which I?don’t do. If this is how you work, you should probably listen carefully to both systems. It’s also important to note that older native systems using PT LE 8 utilize a 32-bit floating-point mix engine, a significant difference from the 64-bit mixer in PT 9 and HD Native.
One area where HD Native and HD TDM are identical is in the driver layer that handles the I/O between the software, the audio files on the drive, and the Pro Tools cards. Both systems use the same controller hardware and firmware with a driver latency of 7-9 samples; whereas using a USB or FireWire interface, you might see latency figures as high as several hundred or even thousands of samples. In this way, HD Native and HD TDM are pretty much identical. Where they do differ significantly is again in the mix engine. Importantly, the TDM mixer runs on the TDM subsystem, where there are no additional latencies, while the HD Native mixer runs on the host CPU, resulting in additional latency to get the samples to and from the software mixer via the Hardware Buffer. Therefore, if you are relying on the Pro Tools mixer to create a monitor mix, your total end-to-end, analog-to-analog latency will be higher with HD Native than with HD TDM. But as I mentioned already, none of our clients have complained about the latency of HD Native — for us it’s a non-issue. But if you find yourself running a session with extremely high track and plug-in counts, and therefore, you need to increase the size of your H/W Buffer to offset too-high CPU usage, HD Native gives you a special low-latency monitoring feature that bypasses the PT mixer. In the Output tab of the I/O Setup window, you can specify a stereo pair of outputs to operate as your Low Latency Monitoring path. Assigning an aux send or channel to this output pair bypasses both the H/W Buffer and the software mixer, once again dropping latency down to the industry-leading figures that PT|HD TDM users have enjoyed for years. There are some idiosyncrasies associated with use of the LLM feature — for example, when feeding an aux send to the LLM path, the channel faders affect the levels, not the aux send faders as you would normally expect — but for the most part, it’s a seamless way to create a usable headphone mix even when an overly-taxing session requires bumping up the buffer size.
So, is Pro Tools|HD Native the right system for you? Although some HD TDM users wouldn’t consider HD Native an upgrade, as it lacks dedicated TDM processing, Studios with older HD systems should definitely consider purchasing a new HD Native rig. It’s a solid, cost-effective option if you can forgo the use of TDM plug-ins and if you don’t do any tracking with TDM plug-ins inserted — and if 64 tracks of simultaneous recording is enough for you (as opposed to TDM’s maximum of 160 tracks). Another option is to configure a Pro Tools 9 system with third-party interfaces and even add Complete Production Toolkit 2 to increase track counts, but you’d be giving up the aforementioned special LLM mode, plus HD’s input monitoring, TrackPunch, automatic delay compensation on hardware inserts, Sync HD support, and a few other features. And potentially, you’d also be dealing with much higher latencies all-around, especially with USB and FireWire interfaces. Also, if you’re at all like me, an owner of a studio that caters to freelance engineers, I bet the last thing you want to do is deal with third-party drivers and I/O settings — and the inevitable questions you’d get from engineers who just want a Pro Tools system that works like the ones they’re used to. For me, HD Native is a big step up from our older HD TDM system and was a much more affordable purchase. I think for a lot of music recordists and working studios, Pro Tools|HD Native will more than meet their needs and expectations — for years to come. (HD Native Core card $3495 street; external I/O hardware sold separately; bundles and exchanges available; www.avid.com) –JB
by Chachi Jones
Sometimes the best way to make a new song is to just mess around, dive-in headlong, and improvise until something you like starts to take shape. In other words, jam. Jamming is a fundamental...