Fork me on GitHub
Slack Archive

This page is not created by, affiliated with, or supported by Slack Technologies, Inc.

2023-11-05

Channels

#overtone
<
2023-11-05
>
plexus10:11:23

Just realized something that has been bothered me for ages. When using samples in Overtone, half of the time they would have a loud audible click at the end. I would try to work around it with an extra envelope, but it wasn't great. Turns out it's actually Overtone being too clever. The built-in sample synths do a bunch of extra stuff to cater for playing loops, adding an asr env and a phasor. If I play the sample buffer directly with play-buf then I don't have the clicking.

plexus10:11:47

instead of using play-buf, they use a phasor to go through the individual samples it seems

plexus10:11:39

to reproduce 

;; loud click at the end (at least for me, on linux)
((freesound-sample 26878))
;; no click
(demo (play-buf 2 (:id (freesound-sample 26878))))

diego.videco23:11:19

I think it may be an old supercollider idiom, prior to PlayBuf

diego.videco23:11:18

Would be definitely nice to fix that.

pfeodrippe21:11:10

Not that I would have time to work on it at the moment, but I was thinking if we could have audio diffs analogous to browser screenshot diffs, but where the “screenshot” is the 2d representation of the audio (in frequency or time or FFT time windows). Example, you build some sound, play it and save the output; then you make some changes and compare the second output with the first one visually. For automated testing of these diffs, we would have to make it robust enough to minimize false positives (can we control the randomness seed in SC? Hopefully we can), but it could help with some regressions. Let me know what you think o/

plexus08:11:40

Interesting idea... my main thought is that the effort it would take to build this testing harnass would be better spent on overall development. But I do think it's doable, take both a waveform and a spectogram, and do an image diff over them.

🙏 1