It had been a long wait, and it had very nearly ended in failure, but at last he agreed to come, and he was talking. Our man selected to conduct the debriefing was Paul White.

There were two of them. They were sitting in a sleazy Indian restaurant in down-town Chicago, seemingly at ease, relaxing between courses. One of them was speaking, slowly but intensely, while his eyes from time to time flicked around the room, taking in the appearance of any new customer who entered. This was none other than Keith Barr, the man behind Alesis. The rain poured down outside and the street began to look like the set for Blade Runner.

Keith knew his stuff. The man who took up advanced organic chemistry as a hobby and nuclear physics as an alternative to the Times crossword contemplated the contents of the dish before him, before speaking.

'My dad was first a teacher and then a physicist. When I was a child, he would teach me about what he called God's laws, by which he meant the way that the universe worked. Then I started to pull old TVs apart and build things. I guess a lot of kids get into electronics like that. Later on, in the early 70s, I went into business by starting MXR with Terry Sherward. That company is probably best remembered by musicians for the pedal effects it built.

'We saw a market for digital reverb and so I designed the MXR 01 which was a pretty simple hardware implementation, but unfortunately, I didn't really understand the software at that time, so it wasn't the best design ever. That business ultimately failed, so I went my own way to start Alesis and the others went on to become ART.'

The Obsession

'Would it be true to say that when you started Alesis, you had an obsession for developing a digital reverb?'

'Obsession is the right word, I would work all day long thinking about digital reverb, and I'd go to bed at night with a pad of paper thinking about it. I have a tendency not to look at or listen to competitors' products, because that can be really negative from a design point of view. The influence that other products can have on you can be extremely detrimental, and can stifle any original thought on the subject.

'I modelled the sound of my reverb from the finished sounds on record, but I wouldn't check to see which reverb they were using or how it worked. I suppose that shows my ignorance on one hand as every manufacturer should know something about his competition. From time to time I'd listen to comments from users about competitors' products and from their description I could deduce what they were doing, but I did live in my own little vacuum! For the past few years too, I've lived in my own world, thinking about the reverb process and how it can be simplified in terms of hardware and yet retain quality.

"I modelled the sound of my reverb from the finished sounds on record, but I wouldn't check to see which reverb they were using or how it worked."

'But what's your musical background? Presumably you must have some musical involvement to be able to judge the artistic merit of your product?'

Well, we all admire music. I'm 36 years old, so I grew up listening to the Beatles and playing guitar. I'm not a performing musician, I don't have that sort of courage. I drive the people at the plant crazy by playing the same records over and over again in the lab. I get locked onto one album and play it non stop for weeks. They get sick of it, but I don't. I guess they think I'm a moron sometimes.'

'Alesis has made its reputation on very few products so far; there was the XT, the XTC and more recently the Midiverb...'

'That's about it so far. But these products are not like any other kind of effect. You can't just take basic building blocks and put them together. They are synergistic systems where the hardware and the software have to be designed to complement each other, and this all has to be done in house. It has to be this way with any budget priced reverb that's going to work properly: the circuitry, the software and the logic paths all have to lock into each other, so you can't just snap your fingers and design one like you can with, say, a compressor or an equaliser. A good design might take the best part of a year, whereas you could do a reasonable compressor in a couple of days.'

'The Midiverb was your first truly revolutionary product. How did you establish the design criteria for that before you actually started on it?'

'We started off by saying that it had to be inexpensive. There was a market out there that couldn't afford the $1000 foolishness that was going on at the time, and my understanding of the algorithms was getting to a point, through my experimentation, where it was possible to produce the Midiverb using inexpensive hardware but supporting good sounding, high quality algorithms. Initially it was just supposed to be inexpensive but from the comments we've had back from the field, it seems as though we have done a good job.'

'Did you model the sound treatments on things you'd heard on record or did you just go ahead and produce something that you liked?'

'I'm not a crazed 'Rock 'n' Roller'. In fact I'm quite conservative in many respects but I really got hooked by the sounds on Billy Idol albums. To my mind, it's wonderful music: beautifully produced and impeccably performed. As to the reverb sounds, I'm guessing they used a Lexicon 224XL. I really enjoyed the sense of bass energy rumbling off into the distance and I built that into several of the algorithms. The gated sounds I modelled on Phil Collins. He's a master. I guess it's an AMS and that's been a very popular sound. I listened to a lot of his work before designing the sounds and I really appreciated the smoothness and the lack of ring.

'I developed the reverse algorithm myself and it looks as though I've taken a different direction to everyone else in that field.'

Programmability

The rain beat continually on the windows, almost drowning the ticking of the clock on the wall above their heads. The heavy clouds outside made it appear dark inside. A sullen waiter came to remove their plates, and both men instinctively paused in their conversation.

'The Midiverb is so much cheaper than the competition, you must have cut corners or rethought the basic principles.'

'Programmability had to go. It's an expensive option to put on a digital reverb, and it can be confusing to the user. I am coming more and more to the conclusion that programmability is an almost useless feature on a reverb unit as long as you have enough choices. It's very satisfying to design the software for a programmable machine, but it doesn't necessarily benefit the user. Synths, on the other hand, must be programmable because of the number of variations that are possible. Some people have applied that logic to the digital reverb market but in reality there's only a limited number of pleasing sounds for modern music production. It seems rather pointless to me to buy a digital reverb that allows you to work out a modest selection of good sounds from all the possible bad sounds available. All those bad sounds are a waste of software and make matters harder for the user.'

'The reason that we have so many reverb sounds is that rock and roll has several different applications. Snare drums, kick drums, vocals and instruments all require different treatments if you are going to go to the wall with your production. Classical music is certainly different, but some of the rock and roll settings would probably apply very nicely to classical music too.'

"I am coming more and more to the conclusion that programmability is an almost useless feature on a reverb unit as long as you have enough choices."

'So how can a customer tell a good digital reverb from a bad one?'

'Well... I think that a bad reverb is probably one that rings. Ringing is the most common symptom of a bad reverb and manifests itself as a note or several discernable notes in the decay of the reverb. It's particularly evident on percussive sounds. A short, bright snare drum from a drum machine shows it up best. It might sound like several steel chimes being struck at once, and this results from poor algorithm design. The algorithms vary a great deal from manufacturer to manufacturer because there really aren't any printed papers on the subject to guide designers down the proper path. Designers tend to be locked onto ideas developed by Schroeder 20 years ago which don't really sound all that good as they stand.

'Density is another aspect which is important; the reverb has to be dense enough to fill in all the little spaces in the sound, without leaving any gaps. Insufficient diffusion can give a grainy or coarse sound but on the other hand, a lot of the ringing evident on some reverbs comes about because the designer has tried to overdiffuse his reverb. Again this is evident on percussive or transient sounds, but there's an application for low diffusion programs too; for the treatment of vocals for example. So there are two evils to listen for: ringing and graininess!

'Another problem is called ducking, and this can be heard best if, once again, you put a snare drum into the unit. Just a fraction of a second after the reverb starts, there will be little spaces, pieces missing from the reverb sound, and then it will continue to decay normally. It's the result of poorly chosen Schroeder type algorithms.'

Basic Requirements

'What do you think are the basic requirements for a digital reverb unit that will satisfy all the normal rock and roll requirements?'

'I think all the sounds should be diffuse, even though some vocal treatments might benefit marginally from a less diffuse effect. In terms of decay time, I don't think that you need to go shorter than 400mS or longer than six seconds and it's not difficult to design in a long decay time if you want to. It's not like a digital delay. Long reverbs are created by using a recirculation process and it's easy to build in a 1000 second decay if you want to, just by tweaking the software.

Apart from gated and reverse effects, I can't see that we need a lot more on the Midifex. I've developed this program called Bloom, where the reverb rushes up, rather than starting immediately. But gated and reversed sounds don't share many, or even any, of the parameters that you might ascribe to real reverb, so a programmable machine is encumbered in having to handle the different names for the different parameters.

"... quite frankly, I'm not really in favour of the MIDI control of general purpose effects."

'The Midifex offers effects based on delay, reverb and EQ and it's really designed for people that have small recording set-ups and want to be able to get into more advanced production techniques, without having to have extra channels on the desk, or patch cords everywhere. It does stereo synthesis, multi-tapped delay and a few reverb things, some of which are utility effects and some of which are very specialised things. There's nothing really revolutionary in it; it's just presented in a convenient and affordable package.'

'Do you intend to continue producing budget products like the Midiverb and Midifex or do you have other ideas?'

'We're working towards products of that nature, and I'm interested in the MIDI control of harmonisers, though quite frankly, I'm not really in favour of the MIDI control of general purpose effects. You have enough trouble keeping the MIDI control of pitch together but it's quite satisfactory for switching programs using patch information.

'We do have a new product which acts as a simple remote control for MIDI effects, which is inexpensive and useful if you have several MIDI effects which you can assign to different channels. This way you can control a whole rack of processors from a single point.

'Also new is an OEM reverb module which we will be marketing to console manufacturers and so on, to build into their products. It's a 16-bit device with a custom built chip using our RISC architecture, and it offers a selection of preset reverb effects which can be varied according to the manufacturers' requirements.'

'I have some... ideas... that do not fit into the budget market and I'll pursue those too. I'm not going to be limited to making products that only fit into one price category.'

'Nevertheless, I do see the home recording market as one of the most important aspects of modern music both from the business and musical points of view. Nowadays you don't need to go into a $200 per hour studio to produce unique music.'

'That's all he would tell us... but there's more going on at Alesis than meets the eye. Turning his collar up against the rain, Keith Barr strode out into the night and was gone.