I want to take the attack of a sample like the scraping of the violin bow or a Guitar pic, and put that onto the front of a different sound. The problem with SC is it’s all MouseX to play around with sounds. There is no way to select a section of a sound in an editor and use all the interesting things in the library to affect a sample. Maybe a little editor could be made to do interesting transformations with SC.
I think in order to do this attack thing I have to use a different app. But I wanted to ask at least what others might do. I know cross synthesis is done, but a true application of one attack onto another would be nicer. I think CDP offers this
You can try to detect the splitpoint between attack ans sustained sound algorithmically. But if it’s only about a few soundfiles, I’d probably just use a DAW or sound editor. Who says you need to do everything in SC? Pick the right tool for job!
I have audacity but I don’t see anything. Im not trying to do Envelop following, but take the inharmonic part of the attack and put onto something else. AudioSculpt is supposed to do this, Maybe some apps in Terminal? Not sure. Good luck getting anything from Ircam.
What I meant was: separate it manually in an audio editor. Once you have the two parts, it’s easy to recombine them. (Digital synthesizer have been doing this since the late 1980s.)
If you want to do this programmatically in SC, you can try to combine onset detection with frequency analysis. For example, you can use “spectral flatness” to estimate the current harmonicity of the signal.
But as I said, if it’s only about a few soundfiles, you are probably quicker when doing the separation by hand.
I think this might be what you’re looking for: Learn FluCoMa
You mean like cross synthesis? Or cross fading?
I agree, FluCoMa is cool, that is def on the to do list.
Might try blending samples with the stk modeling library too, I don’t know why I started this thread
tbh, sorry
I’d first try with simple crossfading. After all, that’s what many classic digital synths used to do; they would simply play the attack sound and then quickly fade to a wavetable oscillator or looping sustain sample. If you’re not convinced by the result, try something more fancy.
I’d second that – try this first.
Stravinsky’s got, somewhere, I forget which piece, a sf staccato trumpet note for the attack, and a p long note in the flute, same note, for the sustaining part. It’s a cool sound! (If it weren’t, I wouldn’t have taken note of the orchestration detail.) It’s literally a crossfade between one instrument’s attack and another’s sustain… really works, try it before spending a lot of time on something complicated.
hjh
1 Like
What would be the best way to isolate the inharmonic attack portion of a sound?
Is there something in the FFT Ugens? Or Plugins? Something that removes all harmonic content is really what I want. The LPC stuff sounds like it does, but for vocals
What would be the best way to isolate the inharmonic attack portion of a sound?
Again, please have a look at Learn FluCoMa. That’s exactly what HPSS (harmonic-percussive source separation) is about.
Oh, cool, thank you very much. I battle with adhd, so much reading with sc, I’m trying tho. Def going with FluCoMa for percussion work.
+1 on FluCoMa. HPSS is great, would recommend checking it out for this use.