Hi !
Here is a small demo video of my Puredata “gamAray” patch for Lich (and computer), which is a guitar amp simulation with chorus, delay and 2 slow LFOs at fixed speed and already attenuated.
Delay and chorus can be used for the “L” input at the same time as the “R” input is used for the Guitar amp!
The guitar amp and delay have a stereo spatialization effect!
Here is the demo video:
Patch: https://www.rebeltech.org/patch-library/patch/gamAray
Have fun !
Sounds pretty good!
I think that relying on naive hard clipping is a problem in feedback patches, it would create noise from aliasing artifacts. You could try using soft clipping if something like this cubic clipper works. It should sound better, but still not ideal.
The best approach would be to use an antialiased clipper/saturator. Math behind it is actually fairly simple, but I’m not sure if it can be expressed easily in PD.
Hi !
Thank you !
This may be a misconception, but I seem to have understood that aliasing is only a problem for very high frequencies, and as the distortion goes directly into a bandpass filter to emulate a guitar cabinet, the high frequencies are filtered anyway, so aliasing should not be heard. (?)
You’re mostly correct, but you’re thinking about output, while you should be considering all the intermediate data that is processed in your patch instead. So if you hardclip, there would be inevitably added some amount of harmonics that go above 1/2 Nyquist frequency (above 24kHz) as soon as you clip the signal. They would get reflected back from that frequency, i.e. if you have a signal with harmonics at 5K, 10K, 15K, 20K, 25K, 30K, etc, 25K would become 23K, 30K would become 18K, and so on. In worst case the reflected signal reaches 0Hz and gets reflected from it again, and so on.
So all those aliased partials are usually not harmonically related to your audio and they end up near the same frequencies. This sounds like a buzzing noise. You can’t just filter it away completely with BPF, because they would be interwoven with your audio contents. It may become not very prominent depending how much of your output signal is left, but it typically increases noise floor or gets unexpected side effects later.
That’s awesome, I love it! The guitar sounds great.
Do you use an instrument interface module, or how do you get the guitar signal into the modular?
Thank you for sharing!
Great patch ,great demo, great video!!
Gonna test it as soon as I get home 
@antisvin : Thank you for all this precious information !!!
For the moment I’m worried about other things, but I’m already looking forward to getting my hands dirty !
@mars : Thanks ! In this demo, the guitar goes directly into one of the VEILS2020 VCAs with ofset, then directly into the Lich!
But as I don’t have a real output module yet, I haven’t really verified that all the levels are correct, so i guess there may be a saturation of the DAC with all the effects activated if all the input signals are in hot modular level, I will check that when I have the output module.
Thank you Befaco & maaark ! Hope you had fun!
In the new firmware you can open the two audio inputs over usb-audio. Could check there if there is enough headroom perhaps?
@dreamer, I think that actually won’t work in practice. Data is in float format in OWL internally during processing and can exceed ±1.0 at that time, but for USB audio it gets converted to 16 bit integers. This would cause clipping if the signal isn’t normalized before that.
And another issue is that currently OWL is usable as USB audio interface to record its inputs, not patch output. This is supposed to be changed with the next release and we might support switching between both configurations eventually.