@svanya The external seems to behave in the same way in 64-bit (widows) vanilla as in extended.
http://msp.ucsd.edu/pdrp/latest/files/dodge-song/lib/ if you dump the executable somewhere that Pd is looking......
The middle inlet seems to be hot, with a range from 0 to 2.
The left inlet is only hot with the middle inlet between 0 and 1, and floats are not accepted.
So the help patch needs a [sig~] to work as "expected??"...?
The help patch for [wrap~] might shine a light on your "expectations"....
Is it a simple combination of [tabread4~] and [wrap~] with convenient arguments?
@Sebastien Very sorry then...... to have taken you on a wild goose chase.
Yes, for 3 channels you could use the R G and B pins.
The colour bit depth of VGA will be good for a smooth waveform output.
You would need to boost the voltage enough for your synth controls...... simple op-amps will do.
The only problem will probably be the "black burst" in the analog output.
The "horizontal blanking interval" here,.......... http://www.eecg.toronto.edu/~tm4/rgbout.html
A small capacitor would smooth that (and help with noise), as it occurs at 50-75Hz usually for VGA.
If you need more channels you would need to "window"....... use the HV sync and a clock to mask the data...... capturing individual pixels or blocks of pixels, and then a larger capacitor for smoothing.
Splitting the signal to a monitor you could have a very nice colourful display of your output as well!
I am very rusty on the electronics.
The Pd patch would be easy (I think).
@Sebastien I have seen discussion of the dac >> VCA.
I think most have gone the midi >> CV route, with proven success.
I don't know whether the systems they have built could translate NRPN for 14-bit resolution, or if they were stuck with 0-127 (8-bit signed).
I am asking myself if I read your OP correctly.
I read VCA (voltage controlled amplifier) instead of VGA.
Did you mean the analog video VGA output from your computer?
VGA sends data in discrete steps, so PWM-ish data..... RGB + H+V sync (0-0.7Volts).
You might be able to use capacitors (pF probably) to smooth the RGB channels and then op-amps to get the desired control voltages.
If you could "window" the data first you could also get a large number of channels.
Every screen pixel could be set with [pix_set] giving 3x (screen resolution) channels.
An interesting project?
@Sebastien Sending square wave pulses into an analog system would most likely generate some nasty high frequency harmonics, if not simply noise.... which you might be able to screen or smooth out, or you might not. Hard to tell without trying....... but?........
P.S. motorised potentiometers with digital control are another option...... but you would have "steps".
For PWM and motorised pots you might need to go with a RPI / arduino.
Which makes me think..... there might be some cheap "robotics" usb stuff that you could talk to through comport.
Motorised pots keeps the pulses away from the analog circuit with any luck. I don't mean motorising the pots on the synth. I mean using a motorised pot to send a voltage to the synth.
@Sebastien I think you will have to get your soldering iron out.
As you say, audio dacs will probably not do 0Hz and will not supply enough voltage.
PWM will be bad news (noise on analog buss).
A usb data acquisition box would be expensive-ish and not report through the os as an audio device.... so useless.
Even a very expensive audio dac that can go to DC would be channel restricted.
You would still need an op-amp
There are chips that can do frequency to voltage conversion..... that would take your existing dac output.
You would need to scale and trim to the output voltage to match the VCA voltages of your synth, or maybe just use the synth voltage rail if the chip can cope with that voltage.
And you would need to build a patch sending appropriate frequencies and scaling..... that is.... do the opposite conversion in your patch...... voltage to frequency.
Otherwise....... midi to CV.....