Synthesizer/Drum Machine resources help

[koisk]

I got a essay/project to do that's on Drum Machines. I've actually been able to find very little information on how they work, with the exception of Wikipedia (but it's all uncited). Does anyone know of some good places I can look?

[adam]

I am a massive wealth of information. I've taken courses, read books, etc, etc. Can you be more specific about what you're looking for? There's been a few different synthesis methods over the years; subtractive, FM, and physical modelling being the most important. Are you looking to concentrate on one? Present a historical overview? What's the plan?

[Renze]

The bible of analog synthesizers is "The A-Z of analogue synthesizers" by Peter Forrest. I think that's a good starting point for doing research for your essay.

[adam]

But the analogue in that title excludes a lot of valuable stuff...depends on what you're looking for...

[koisk]

I bascially had the freedom to make up the topic, so long as it related to the overall subject "Science of music". So mine is on "Drum Machines". I'm covering how they work and thier impact on contemporary music.

What I need material on is how drum machines work. Whatever is involved with making the handclap or snare drum is what I need to talk about. Obviously it could be craaazy involved, but this isn't a thesis so it will only be as in depth as it needs to be.

Adam I'd appreciate you sending any materials/links my way on the subject.

[adam]

Well, there are some interesting angles you could pursue.

See, you could emphasize the difference in synthesis techniques. Early drum machines worked on subtractive synthesis, meaning that they had a sample in memory (or generated an original sound through oscillators), and then you shaped the original, "full" sound by taking away from it - hence the term. So early drum machines, like the Roland 909, worked on these principles. The hi-hat and cymbals were samples; the rest were generated by oscillators that were shaped by the filters and envelope generators into their best imitation of real drums.

The thing is, these drum machines were commercial failures. They sounded nothing like real drums. And they tanked. They were sold off cheap, and ended up in the hands of dance producers, where it was found that the characteristics could be shaped in pretty intereseting ways once you abandoned the objective of realism.

In the meantime, other methods of synthesis were developed. FM synthesis (frequency modulation) is made by having one oscillator (something that produces a wave) have its frequency (its pitch) changed by another oscillator. This produces an entirely different set of characteristic sounds.

With the constant increase in computer processing ability and storage capacity, we've seen a couple of other significant changes. Drum samplers can now hold massive, high quality samples. Whereas before they were limited by extremely small amounts of RAM, we now have drum machines and computers that can hold great amounts of high quality samples. Drum sampler programs on computers are at the forefront of this - Native Instruments' Battery 3 comes with 12 gigs of samples. These can be layered so that you can get different samples playing depending on how hard you hit the key; or there are drum kits in that program that have 4 different mics on a single drum, that can be blended as you'd like, which is more what you'd find in an actual recording studio. They've also added a "humanize" control, to add variations to timing and playing style to make things sound more human.

On the non-sampling front, though, due to exponential increases in processing power, we now have physical modelling as a synthesis technique. Rather than take a recording of a drum, or take sounds that sound vaguely like drums, physical modelling uses complex algorithms to calculate the actual behaviour a real drum. This allows unprecedented controls...in, for example, Roland's V-Drums, you can alter individual characteristics of drums, like the diameter of your kick drum. It then calculates the physics of such a drum, and outputs what the algorithm says such a drum would produce.

The interesting thing is where drums have come in terms of how those drum machines shaped it. Those Roland 909s, which produced terribly non-realistic sounds, are now desired, as their sound became popular, and then, iconic. Much more realistic drum sounds can now easily be produced, but for some purposes, people want those old, supposedly awful sounding drums.

On a similar note, though increases in computer power have allowed much higher quality samples, effects are often used to downgrade the quality of these samples to make them sound more like the old machines. Two factors typically decide sample quality: word length and sample frequency (CDs have a 16 bit word length, and sample at 44100 times a second). Though computers are typically capable of processing sounds at far greater quality, effects are often used on one or both of those factors to make them sound rougher, rawer and crunchier: a sound that at one time was an undesirable side effect of sampling has now become desirable.

This is similar to the entire recording industry. Accidental sonic characteristics of older equipment have become sought after. Recording on tape produces an unavoidable compression; we now have lots of effects available to simulate that sound. Neumann microphones don't have a level frequency response; that is, they emphasize certain frequencies and produce distortion. But those characteristics are considered excellent, and people pay a fortune for their microphones. Those are but two examples; the entire recording industry replays this story over and over.

For a modern drum machine, check out MachineDrum. That is considered one of the more desirable standalone drum machines currently on the market, and is interesting because it maintains some of the interface elements of the early drum machines, and also uses (I think) all of the above synthesis methods.

Also, look up Native Instruments' Battery for the software side, and the classic Roland machines for the old-school machines.

[patrick]

well that was super informative, and there is your essay para-phrased... nice work adam! where'd you learn all that?

[adam]

took a three year program in audio engineering. included some synthesis courses. work on my own stuff. read gear reviews when i'm bored at work.

[Aaron Contreras]

Nice post.

[adam]

Originally Posted by Aaron Contreras

Nice post.

thanks


fuggin ten characters...