Recent Posts

1

Help and Questions / Re: Pitch sample without affecting length/speed?

« Last post by Saga Musix on Yesterday at 18:35:23 »

And to make this solution even better, you can also apply the Pitch Shift operation to sample selections starting from the next OpenMPT update. I completely forgot that I only implemented that option for time stretch, not pitch shift.

Other pitch shifters like Rubberband do something similar and re-sychronize at transients.

2

Help and Questions / Re: Pitch sample without affecting length/speed?

« Last post by peterpiper0815 on Yesterday at 16:35:50 »

To avoid destroying the original Attack of (for example) a kick you can copy the sample cut away the trasient, pitchshit (I noticed that a high quality and low FFT setting works good for kicks) and copy/paste it back to the end of the original sample. Now cut away the original release part of the sample. Yes, its not a one click solution but it can give good (at least interesting) results.

3

Help and Questions / Re: Pitch sample without affecting length/speed?

« Last post by Saga Musix on Yesterday at 08:15:57 »

Please note that it is simply mathematically impossible to build a perfect pitch shifter or time stretcher. There are a few commercial ones that work surprisingly well despite these physical limitations, but it will never be perfect. Autotune may for example work well with vocals but I have doubts it would get the job done for kicks and stabs. You can give Rubberband a try (which is free, but not compatible with OpenMPT's license so we don't use it), but I can tell you that both "classic" ways of implementing pitch shifting and time stretching (namely resynthesizing the frequency spectrum with fast fourier transformation, and temporal overlapping of small snippets of audio, both used by the pitch shifter and time stretcher in OpenMPT respectively) do not handle drums very well. It's just in the nature of time stretching to smear transients.

What may be more helpful in your specific case is editing the samples by hand, e.g. to apply a fadeout to make a kick shorter, or delete small bits of it, or copy&paste bits to make it sound longer. I personally would never use a time stretcher for that.

4

Help and Questions / Re: Pitch sample without affecting length/speed?

« Last post by Rakib on Yesterday at 00:07:52 »

Buy autotune and load the samples there? The pitch engine in openmpt is not the best but is usable for many applications.

5

Help and Questions / Re: Pitch sample without affecting length/speed?

« Last post by Grid on Yesterday at 00:04:09 »

Doing this will usually make it sound echoed/reverbed and rather strange for me. I'm using the .IT format for most of my work and attempting to pitch stuff with that never sounds like the 'original' but pitched, it sounds like an off recording of the original sample.

6

Help and Questions / Re: Pitch sample without affecting length/speed?

« Last post by Rakib on January 24, 2021, 23:58:09 »

In the sample view,  pitch shift/ time stretching.

7

Help and Questions / Pitch sample without affecting length/speed?

« Last post by Grid on January 24, 2021, 23:34:24 »

Hey! I've tried searching on the forums on and off before and on the internet and I haven't been able to find an answer that helps with my workflow or that tells me a process with good results.
I'm trying to find if there's a way I can pitch up certain samples without altering their speed too badly, e.g. certain kicks or stabs, which in theory should be easier to work for pitching up & down without affecting them too much as opposed to straight up long loops & samples. I haven't been able to find anything that works other than bouncing the samples to another program like a DAW then bouncing them back in, but I'd like to do this purely in OpenMPT. Is there a way around this? Thanks!

8

Development Corner / Re: High-precision pitch offset command

« Last post by Saga Musix on January 24, 2021, 12:02:03 »

Thinking one step further... This command should also work with plugins. This can of course be achieved by translating the cents to MIDI pitch wheel commands.
But for a while now I though that it would be great to have a pattern command that would essentially directly translate to MIDI pitch wheel commands (so that you can directly record them from MIDI and don't have to set up Zxx MIDI macros yourself). So what if we combine this fine-tuning command and MIDI pitch bend command into one new command? Their functionality is pretty much the same. The command should work identically for plugins and samples, so its effective pitch depth would depend on the pitch wheel depth configured in instrument settings (in sample mode, we would assume +/-1). So if you set up a sample-based instrument with a pitch wheel depth of 1, you would get exactly what was previously proposed with 1/128th of a semitone. If you set up a pitch wheel depth of 2, the resolution would be reduced to 1/64th of a semitone but with a wider pitch range (and when using #xx command extensions, there would be 8 more bits of precision of course).

9

Development Corner / Re: High-precision pitch offset command

« Last post by Saga Musix on January 24, 2021, 00:21:00 »

I'm getting the command ready to be commited (for OpenMPT 1.30), and so far it's working nicely.
I'm still a bit torn with regards to the resolution, though. Of course using 128th of a semitone plays nicely with the hexadecimal nature of the pattern notation, so that e.g. a parameter of 0x40 is half a semitone. But would it maybe be easier to use actual cent values? They don't translate nicely to hex (e.g. E4 would be +1 semitone), but they are much easier to transfer from real-world scale definitions, which may outweigh the niceness of hexadecial magic constants. What do you think? You probably have a better idea how you would like to use this feature.
Currently I'm leaning more towards using cents.

BTW, all of this is always in the sense of a semitone/cents of a 12TET scale, even when using custom tunings.

Edit: We could also make this command even more precise using the #xx parameter extension: A second command would be put in the following row to define a 16-bit instead of 8-bit value, making the effective resolution a 256th of a cent. Do you think it's worth implementing that?

10

Development Corner / Re: libopenmpt emscripten/wasm into AudioWorkletProcessor

« Last post by jllodra on January 22, 2021, 15:10:25 »

r14059 (libopenmpt 0.6 development) or r14060 (libopenmpt 0.5) might work without requiring any polyfill when using make CONFIG=emscripten EMSCRIPTEN_TARGET=audioworkletprocessor. Could you please check that?

You can either get unreleased sources from svn or git, or from tarballs at https://builds.openmpt.org/builds/auto/libopenmpt/src.makefile/ .

Excellent, it works without polyfills now.

@Saga Musix: Cool, that helped. Thanks!

I think all set with this issue.