Ken Bogdanowicz

Ken Bogdanowicz is the man behind Soundtoys, the boutique software company known for their feature laden and very analog flavored effects plug-ins. Ken's history with designing effects goes back to the pre-plug-in hardware days when he was a key player at Eventide. We talked about the evolution of digital audio effects and how he looks to past success stories when designing his plug-ins.

I know you started out doing hardware long before there were any plug-ins. What is your background?

My first job was a summer job working at a military contractor in New Jersey, which exposed me to exactly what I didn't want to do — working with a bunch of people bored out of their skulls designing God-awful defense systems. I graduated college with an engineering degree and went through the magazines — Guitar Player, Keyboard — and made a list and sent a resume out to all of those guys and ended up at Eventide. This was about 1984- 1985. By this time they had already designed the first Harmonizer and the first digital delay. My first job that had to do with audio there was to try to do an emulation of an AMS non-linear reverb algorithm from the RMX 16... that was kind of cool.

A lot of people really dig the old AMS stuff.

I just bought an old RMX 16 on eBay and one of the 1580s too, which are pretty cool.

Do you buy the old gear to try to emulate various pieces in your plug-ins?

We're not really making a conscious effort to emulate anything, but we do want to have some of the flavor of the older gear and try to introduce kind of "analogness" to things. As I started to work on EchoBoy we ran into this whole world of old delay effects. You know, everybody has their own personal favorites. So I just made it a mission to try to get one of everything. I'm not quite there yet.

Ha, that's an expensive mission! You must have quite an arsenal building up.

I would talk to so many people and you'd hear things about what makes certain things sound great. On mic pres and EQs some people think it's the tranny [transformer] that gives it the sound. Sometimes it's hard to exactly hear this, so I wanted to have these in my hands to check it out. I'll hook it up to an analyzer and run sweeps and distortion and spectrum analysis, and then I start looking at schematics. For example, with the [Neve] 1073 I went back to my engineering roots and worked out all of the equations for the EQ parts and generated these giant, God-awfully ugly equations. Then I generated the theoretical way it is supposed to sound, and then I did a lot of listening. Some of these old schematics are really badly written. It's kind of like a mystery for me, trying to suss out what's going on with this stuff.

How long did you work at Eventide?

Eight years. I think at the time the [H-]969 Harmonizer was the last in the line and they decided they wanted to redo it and make it stereo. I worked for Richard Factor, with Bob Belcher and Dave Derr [Empirical Labs] [ Tape Op #33 ]. We were really right out of school and new at things and excited to get a crack at designing this new product. Eventually Richard did let us work on it. It was basically supposed to be a stereo pitch shifter and we just went overboard. Every once in a while they would ask us how it was going and why it was taking so long and we just made things up just to keep things going. Bob, who pretty much designed the entire hardware top to bottom, would sometimes sleep in the officeafter spending all night working on circuits, and then wake up and start working on software. My part of it was to do all of the DSP stuff — the algorithms. We actually didn't have any hardware working, and I had gotten an old IBM PC XT with maybe a 10 or 20 megabyte hard drive, and whacked together an A to D converter card so I could sample some audio to test out some algorithms. I would test things out and try it out on my computer. It would have to churn overnight — I'd go home have dinner and come back and there would be a few seconds of audio. Compared to what we have now it was just unbelievable.

It sounds like they gave you total free rein.

They gave us a lot of free rein, but I think they were really nervous.

At what point did you leave Eventide?

Well, at one point I was starting to get the bug to do something on my own. While I was at Eventide, Richard agreed to let me do a ROM add-on for the Harmonizer that I sold on the side through an ad in the back of Mix Magazine. After a while I was ready for a change. I left Eventide, did a bit of consulting and eventually convinced Digidesign to let us to do a plug-in — that was PurePitch, which we still develop today.

So eventually you started Soundtoys. You started out programming for the Pro Tools TDM platform and now you've released native versions of the Soundtoys suite. How different are the two environments — TDM versus native — when it comes to developing your plug-ins?

The TDM world is DSP chips that are all coded in assembly language, pretty much watching every little bit almost — it's very low level and close to the machine. The rest of the native world — RTAS, AU — are way easier. Everything is floating point, which means you have infinite dynamic range — where 24 bit fixed point is where TDM runs, and you have to be very careful even on something simple like an EQ or you'll have crappy clipping or horrible distortion — which is why you probably see a flood of all these native plug-ins, because it really is a lot easier to do. The environment is easier and you don't need specialized hardware. Designing a TDM plug-in is sort of like building a ship in a bottle, because you have these TDM cards and you have a bank of six or nine DSP chips on them, and when you're developing your shoving code onto them and there are really no great tools to figure out what's going on in these chips. Working on the host computer, there are these great tools from Microsoft and Apple that make it so much easier.

Is that the result of the way the Digidesign hardware evolved?

It's really because Digi hardware is kind of equivalent with what you would see in a dedicated hardware effects processors. It's called "embedded" — it's more specialized and there for pure performance.

What are the nuts and bolts of designing a plug-in that is trying to have some of the sound of a vintage hardware piece?

Instead of trying to do exact emulations we try to do one where you can dial in a bunch of different colors. A lot of our stuff follows a pattern where we start out with of our stuff follows a pattern where we start out with and deeper. With EchoBoy we were trying to come up with something that had a good range of tone colors, but to fill out what we thought was missing, which was a good vocal echo — something that was warm, but not too dark. Our first reference for that was when we got an [Ampex] ATR-102 tape machine and an old Elvis-style tape echo. We wanted to add darker colors, so we looked at some of the pedals and Echoplexes, [Boss] DM-2, [Lexicon] PCM 42, [Roland] RE-501. I was talking to someone who said that what he liked about EchoBoy is that it sort of has a built in de-esser, so when you hit something really hard it pulls back on some of the highs and adds in some warmth. It blends in really nice and subtle. What I was trying to create was something where you could get things in more musical terms — things like shuffle and dealing with the groove without having to go in and tweak fifteen different delay times. With FilterFreak the first thing that we started out with is trying to get some tone color, so we picked up a Minimoog as our reference, then we got a [Mutronics] Mutator — that British box — and then a Sherman FilterBank and also a [Musitronics] Mu-Tron. It was a little harder because the filter as an effects processor is something that has not been as developed as echo, reverb, EQ or compression. The Crystallizer is the one crazy thing that we did that was sort of a nod to the [Eventide] H3000. There was that preset in the H3000 called "Crystal Echoes", which was based on this reverse pitch shift algorithm in one of the early revs of the H3000. That was one of those patches that you would just load up and play one guitar chord in and end up with something completely different. With Crystallizer we just thought that we had to do something like that.

What do you do when it comes predicting what's coming down the pike, both on the technical and the artistic sides of what you do?

I just came back from a conference where a lot of the people who went were coming from the serious, academic music composition world and they're looking for more ways to control timbre and sound. They're not just approaching it to push a button and lock to 120 BPM, but they're looking for more freedom and control. In a way, a lot of the stuff that these early computer music people were doing twenty years ago that seemed wacky is considered pretty normal today. I think when it comes to using computers and software for music, it's sometimes good to look at that world. A lot of what was developed years ago is used in pop music today.