[sdiy] Blackmer VCA cell

[Brian Willoughby]

How do you get fast Attack with PWM smoothing?

Put another way, why is cost such an issue that discrete PWM smoothing circuits would be chosen? Each PWM channel must have it's own smoothing circuit, and it's own PWM generator.

In contrast, a single DAC can be shared among multiple CD outputs with an analog multiplexer configured as Sample and Hold. This is a circuit design that's been used in analog synths for decades.

Is it really cheaper to dedicate independent PWM+smoothing circuits when a single DAC and mux/hold is faster? ... not to mention it's easier to expand to more channels with the DAC+mux circuit.

Brian


On Oct 23, 2021, at 14:57, Mike Bryant wrote:
>> ... of the EG itself (at least without varistors
> 
> Which takes us back to the PWM thread :-)  
> PWMing each of the ADSR parameters is quite easy as all of the PWM frequency gets well smoothed away.
> 
> 
> 
> -----Original Message-----
> From: Brian Willoughby
> Sent: 23 October 2021 22:50
> 
> The simplest envelope generator is exponential, based on pots and switching. But these are more difficult to voltage-control the parameters of the EG itself (at least without varistors). Great for cheap systems where the pots on the front panel are the only way to control the EG.
> 
> A linear EG should be possible in analog using op-amp integrators and current sources, which are more easily voltage-controlled. These are probably more expensive, but I'm not familiar with the costs.
> 
> A linear EG into a exponential VCA should sound very natural. Varying between linear and exponential would be difficult in the analog domain, though.
> 
> Brian
> 
> 
> On Oct 23, 2021, at 14:41, Mike Bryant wrote:
>>> Would it be possible to redesign the CV envelope generator curve to work with an exponential control VCA?
>> 
>> Yes it is.  I think Neil was being a little bit sweeping as quite a few synths used to have log VCAs with linear envelope generators.  But the accuracy of the log VCA doesn't need to be that good so an expensive solution would be overkill.  In my digital designs I always allow the envelope to be varied anywhere between pure linear and pure log anyway (easy in digits) as they can give quite differing nuances to the sound.  
>> 
>> 
>> -----Original Message-----
>> From: Brian Willoughby
>> Sent: 23 October 2021 22:26
>> 
>> On Oct 23, 2021, at 13:53, Neil Johnson wrote:
>>> Their control function is exponential, linear-in-dB which is great for audio fading so they are not suitable for the traditional envelope generator that needs a linear control VCA.
>> 
>> Would it be possible to redesign the CV envelope generator curve to work with an exponential control VCA?
>> 
>> I haven't thought this through, but I can imagine different ways to generate an envelope - both linear and logarithmic. What I don't know is whether you'd actually need anti-log CV to get the right results out of an exponential control VCA for envelope purposes.
>> 
>> Brian