[sdiy] Speaking of the Elektor Vocoder (and the Korg Vocoder)

[Eric Brombaugh]

Ben Lincoln wrote:
> On Wed, February 20, 2008 9:44 am, Eric Brombaugh said:
> 
>> As an aside, I'm interested that analog processing is 'sort of
>> instantaneous', while digital processing is apparently not. Where does
>> this belief come from?
> 
> It's a fairly common statement (I remember it being in the study guide for
> the Computers merit badge when I was briefly a Boy Scout about 15-20 years
> ago). I imagine that a completely parallel clockless digital system would
> behave more or less the same as an analogue system in that respect, but in
> anything with an architecture based on CPUs and registers (or more
> generally, serial data of any kind) there will necessarily be latency
> because the work is being performed as a series of discrete steps.

I know about latency. :)

My point was that if sample rates are high enough (and hardware is 
sufficiently complex), digital processing need not be less instantaneous 
than the analog equivalent. The time constants of the system being 
implemented should dominate the speed of processing, while the latency 
of the digital processing should be negligible by contrast.

In this particular case (vocoders), the filter banks and envelope 
detectors all have certain time constants (as Richard Wentk pointed out 
earlier). This will be true for both analog and digital designs.

Some manufacturers will design complex DSP systems with hardware 
constraints that lead to significantly increased latency over the best 
case. That long latency isn't inherent to digital implementations in 
general though - if you're willing to spend enough money you can get 
good digital processing done quickly.

</soapbox>

Eric