Muff's Modules & More

[Analog>Plugins]

I read this on Gearslutz:

[os]

I guess that would work, but it seems to me that you'd have some lag in the CVs, and you'd have to low-pass filter them a bit to remove the AC component. The poster alludes to that ("guess the higher the pitch the better response time"). Given that one of the main attractions of Silent Way is the sample-accurate sync with your audio, losing that to some "response time" seems undesirable.

But if anyone tries it I'd love to know how well it works. I could add the extra signal generation to Silent Way Voice Controller.

Can you link to the original post?

[ndkent]

Well there are ways to extract a DC signal from a standard audio signal's amplitude, you might already have one, it's an envelope follower. Even if you don't have a dedicated module, one can be made with a slope or EG generator module with CV parameters.

The concept is you create a signal that doesn't confuse your envelope follower. You want a steady signal out not a wavering out from a steady tone. The wave's amplitude will generate a DC value. change the amplitude any you get a different DC

One flaw is the speed of your envelope follower is the highest frequency you can achieve, so for instance if you have an LFO from amplitude change going and increase the encoded speed at some point you will just get the maximum peak as a steady voltage... because you are following the envelope, not the frequency. Also this is a positive only affair. If you want a negative DC value you'd have to mix in a DC offset and compensate. (say 0v out is now -2.5v and +5v out is now only 2.5v)


The person suggesting the ease should clearly be challenged to to show something that works if it's so easy. It clearly has something to do with encoding a 3rd value into 2 channels then extracting the encoded value with custom built hardware. But the "8-output audio interface could output just seven CV's" claim seems just plain wrong. You would need a pair for each extracted CV (4 cvs from 8 channels). If you needed a signal as a reference you could just as easily generate that signal inside your custom decoder circuit while you were at it. But this sort of thing could not involve a single reference audio range voltage that was hared by 7 different channels. I think I can follow the basic concept, that is to build a custom vst to encode a DC value into 2 AC audio signals and then build some hardware to decode it back to DC. Definitely possible but not exactly trivial or they would have done it as proof of concept.

As to someone spending "a day or 2" writing a free VST and maybe skimping on the GUI to save time... hmmm I'd think there would be more free plugins for MOTU devices then. I know people were doing basic DC plugins with MAX (before Cycling 74 signed a pact with Ableton and dumped VST support)... anyway the reality is it's easy to do something in DIY software with DC out a MOTU interface but far from trivial to get a nice reliable feature set and it should be much the same only from scratch in making some kind of encoder that does the same things.

Anyway I bring that up since while writing this I was wondering what progress has come along the lines of bypassing coupling caps from cheap existing audio cards. I mean if one is building new hardware and software to do this it might just be easier to roll a clone of volta/silent way and then unblock the DC on some other model card?

Anyway this all sounds like taking several steps back to save buying a MOTU card and a commercial plug. Considering people often find the "proper" (= commerically sold) solution difficult to deal with, taking some steps back to reinvent the software and build hardware doesn't sound easy at all.

[os]

Thinking about this some more, it strikes me that you don't even need an envelope follower.

Assuming you have balanced outputs, if you had 2 diodes per channel to rectify the signal, and output the CVs by amplitude modulating a square wave at half the sample frequency (e.g. 22.05kHz for 44.1kHz operation), then you could just use the rectifier output directly. If you needed bipolar CVs, a simple DC shift would do it.

Then you have CV output at 22.05kHz, sample accurate, with no envelope follower lag.

Am I talking sense? Anyone got a few diodes handy? I can quickly knock up a plug-in that modulates CVs (e.g. from Silent Way) as described, if anyone would like to give it a go.

[frijitz]

[meridic]

You are talking sense.

Square wave will not give you any better DC then a sine wave in the end, you will need to filter the output either way, but there will be no lag since the frequency of the carrier is twice its output frequency. You will also want to follow the rectifier with a buffer since it would otherwise present a high impedance to anything it is feeding.

I used to control my ARP Axxe this way many years ago.

[os]

Awesome. I think this is well worth a try. I'm now regretting getting rid of a box full of random components I had in my attic - I would've had all the necessary diodes, caps and op-amps in there to try this.

I repeat my offer to write the software pronto if someone knocks up the circuit.

[jenamu6]

Yes please.....1 want one.

To anyone building this, if the price is right.....Here's your first order.

[1nput0utput]

[os]