The Man Who Sells Excitement
Marvin Caesar, the president of Aphex speaks about how the Aural Exciter came into being, and for the first time reveals news of some forthcoming interesting new products...
Marvin Caesar, the president of Aphex talks to Paul White about his exciting career.
Aphex is best known for inventing the Aural Exciter. How did that device come to be invented?
The original concept was developed by my partner, Curt Knoppel, over ten years ago and was really the result of an accident. Curt was building a stereo valve amplifier kit, and having completed it, he found that while one channel worked perfectly, the other had been wired wrongly and sounded rather odd. For no apparent reason he mixed the two sounds together and was surprised to find that the result sounded better than the original. We didn't know at that time what we had technologically, but we knew that it worked. Moreover, at the time, we had no idea at the time how to develop the system or how to modify it.
We were very undercapitalised to start with and the most we could do was scrape together enough money to build a few units and until we could obtain patent protection, we decided only to hire them out so that we could control what happened to them. We sealed up the circuits and didn't tell anyone what was inside.
Our first big break came with the Wings over America tour done by Showco. Peter Asher who had previously been on the James Taylor tour with the same system and the same piano couldn't believe that the piano now sounded so much better. On enquiring how it was done, the engineer pointed to this little box, the Aural Exciter. He asked if the effect could be put onto tape and when we said yes, it was rushed down to The Sound Factory in Hollywood where they were almost finished with a Linda Ronstadt album. They processed the remaining half of the album with the Aural Exciter and the first half sounded dull by comparison, so in the end they remastered the whole thing using our box and Aphex got a big credit on the sleeve. Then things moved from there, first throughout the country and then internationally.
Has the design changed much from the early models?
Originally, we didn't know what we'd got, and the first models used valves. Later we isolated the various components that made up the effect and found various harmonic distortion components, some phase distortion and some amplitude distortion in terms of frequency response. As we came to understand the circuit more, we developed the Aphex Two, on which all the significant parameters are adjustable. The aspects of the effect that could be achieved by EQ, we took out of the design, and concentrated on those parameters that couldn't be duplicated by conventional means. The Aural Exciter effect is much more of a textural effect than an EQ change, and this later became variable when we added the Timbre control which allows you to balance the proportion of even and odd harmonic enhancement. For things that are meant to be warm and soft sounding, you would use even harmonics but aggressive instruments, sound effects or percussive sounds become even more aggressive when odd harmonics are added. We were learning more and more about the device and the technology behind it and we wanted to get into other markets, so we looked for ways to simplify the device in terms of its circuitry and its controls.
This led to the development of the 13700 VCA circuit which we use in the Type B model and has become tremendously successful. We were quite open about how the thing worked because I was getting quite sensitive about all this black magic image and I wanted to dispel the suspicion that the Aphex effect was all hype.
Of course we have competitors in the areas where we haven't got patent protection and one area of concern is that any bad ones will give Aural Excitement a bad name, but there's very little we can do about it. One or two European so-called exciters turned out to be pretty nasty. I only hope that people will realise that the Aphex name represents a certain standard of quality and know that in buying ours they are buying the best. Of course there is nothing more compelling than price and so we had to do what we could in that area not only by improving the product but also by making our own ICs to reduce manufacturing costs. This has led to the development of the Type C which will be introduced in about a month's time. You're the first to hear about this in the UK.
It's 10dB quieter than the Type B, it's smoother in operation, and the setting of the drive control is much less critical so you don't have to worry so much about the levels. The drive range is around 20dB instead of the previous 10dB or so. We're hoping to sell a stereo unit for around £250 so we will be able to equal the competition on price as well as offering a new generation of product. At this price, the unit should be attractive to the home recording fraternity as well as to the professional studio owner.
Do you envisage this model being quiet enough to be compatible with digital recording techniques?
Yes, though you must remember that although the Exciter is inherently quiet, it will tend to emphasise any noise already present in the programme material.
So far, most of your work seems to have been done in the area of psychoacoustics. Do you have any plans for product diversification other than the Compellor?
We're making a new limiter called the Dominator and as it happens you're the first person in this country to hear about those as well. The Dominator is a 3-band limiter with a peak processor. The Compellor on the other hand is an average level gain riding device. The Dominator is three band using subtractive filters; if you put a square wave in to the system, you'll get a square wave out. Even though the signal is going through a crossover and being re-summed, there is no effect on the audio. The value of a 3-band limiter can be seen when you look at the action of a conventional circuit. If you have a lot of high end, it tends to be modulated by the low end and typically a kick drum comes and punches holes in the hi-hat. The Dominator doesn't suffer from these problems, so you end up with a more dense signal with a better peak to average ratio. Also, to maintain maximum loudness, there's an automatic threshold control which decides whether it should limit or clip. If the width of a transient is very short, the ears are incapable of detecting clipping and so we can squeeze an extra dB or two out of the system by clipping the very narrow transients. Another feature is a transient enhancer circuit. With a conventional circuit, the leading edge of a transient is slowed down and so you lose your high end and it all sounds dull. Our basic circuit inevitably suffers from this drawback too but we build the transient back up again using a specially designed psychoacoustic enhancer circuit about which I won't say too much at this time.
These designs are based around our own VCA chip which is very fast and very accurate and can run at up to 300MHz depending upon the application. RCA called us up when they were using some of these chips as video switchers and the guy said that the 3MHz squarewaves were starting to lose their corners by the time they had passed through the circuit. It turned out to be the op-amps and the capacitance around the chip that was causing the problems, not a limitation of our VCA. When you're looking at the Dominator circuit, which has to recognise transients and then decide on whether to clip them or limit them, you'll appreciate that we need this sort of speed. Likewise the enhancement part of the circuit has to be very fast to build up the leading edges of the transients.
Have you looked at producing anything for the digital audio market yet? The concensus of opinion seems to be that digital audio lacks the warmth of analogue, possibly because we're used to hearing tape compression and have come to expect it. It would seem to be right up your street to produce a box which could process digital recordings to actually put some of this warmth back.
We could spend days talking about the problems with digital recordings. I think one of the problems of sampling at 44.1 kHz is that it isn't fast enough to faithfully reproduce transients; you've got a high end but it's not accurate.
Isn't this a good case for using something like an Exciter to re-synthesise this missing high end detail? Although the harmonics added are not going to be accurate, the brain will interpret them as being natural due to the way in way in which the effect is perceived.
Yes, the harmonics from an Exciter sound pleasant and natural but if you have filters ringing at around 14kHz in a digital system, this adds harmonics which are totally unrelated to the programme material and so are going to sound pretty unpleasant.
Could you use filters to roll off the top end at say 12 KHz before these effects take place and then re-synthesise the missing harmonics?
Well this is done on occasions. When the hi-hats sound like rice crispies you can gently filter out the top end and then put it through the Aural Exciter and tweak up the top end a bit. This generates a much more natural sounding high end. Digital is fine for the low frequency end, it's certainly fast enough for that.
Do you have any new ventures in the pipeline that you are prepared to talk about at the moment?
We're producing a system for decoding surround sound material which will offer better seperation than any previous system and we see this as being attractive to the hi-fi video market as well as the more obvious audio areas. I don't anticipate it being attractive to very many serious audiophiles though, as they tend to be very conservative and stick to basic stereo systems.
We are also working on a very low noise mic amp which actually utilises the matched transistor array in our VCA chip. This is in response to those people who wish to record directly onto a digital recorder, such as the Sony F1, where the requirement is for low distortion, good headroom and very low noise. Recordings of classical music done in this way could be made to a very high standard with the minimum of equipment.
The company obviously started from very small beginnings. How many people work for you now?
We've grown over the years so that we now employ around 40 people. There's still a long way to go with the Exciter principle and I can see it being installed directly into mixers and discos under licence from us; this is already happening in some countries. In other areas it's still a matter of educating people because there are still a lot of people who do not appreciate what an Aural Exciter can do and in this respect, your magazine is doing a great deal to improve the situation.
Interview by Paul White
Previous article in this issue:
Next article in this issue:
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!