FR - Timestrech option in Obsidian

edited December 2018 in General chat

Would make obsidian a lot more flexible for sound design if there was an option for loaded sample/s to retain their true length when pitched up/down on keys.

Comments

  • Do they speed up/slow down with the pitch (like NS1) or is that the feature you're asking for?

  • I agree with this request. It’s kind of tricky to implement, but it’s worth it if it can be fit into the app.

    For non-beat based samples, a short term rigged solution is obviously a marked loop-point. But true time stretch is always better.

    That’s one feature I wouldn’t mind seeing.

  • Yeah would make a huge difference for sound design from samples that can’t be multisampled for whatever reason...

  • @flockz Easier said than done mate. Honestly. Getting anywhere near a half-decent timestretch algorithm is hard enough. Making it sound 'natural' is nigh on impossible, without some serious mathematics, time, space and relativity. Far better just to let Blip sort the AUv3 teething problems out first, and squash a few bugs before wading deep into that merde.

  • Caustic has 2 different algorithms that work well. We know the foundations of audio tracks are there in NS2. It’ll come in due time.

  • @tom_tm said:
    @flockz Easier said than done mate. Honestly. Getting anywhere near a half-decent timestretch algorithm is hard enough. Making it sound 'natural' is nigh on impossible, without some serious mathematics, time, space and relativity. Far better just to let Blip sort the AUv3 teething problems out first, and squash a few bugs before wading deep into that merde.

    Aren’t there open source timestrech algo/code options that could be added to obsidian sample playback for now? That would be better than no option in a lot of cases and matt could then work on adding a better quality, more ‘natural’ at later date..

  • @flockz said:

    @tom_tm said:
    @flockz Easier said than done mate. Honestly. Getting anywhere near a half-decent timestretch algorithm is hard enough. Making it sound 'natural' is nigh on impossible, without some serious mathematics, time, space and relativity. Far better just to let Blip sort the AUv3 teething problems out first, and squash a few bugs before wading deep into that merde.

    Aren’t there open source timestrech algo/code options that could be added to obsidian sample playback for now? That would be better than no option in a lot of cases and matt could then work on adding a better quality, more ‘natural’ at later date..

    Not sure about this, but I think it’s difficult to just take an existing algorithm and then “tweak” it to make it better over time. The base values affect all subsequent values in some algorithms, so any tweaks in the chain could send the other values off the edge.

    But I’m not a mathematician, so I might not have this straight in my mind.

  • @kleptolia said:

    @flockz said:

    @tom_tm said:
    @flockz Easier said than done mate. Honestly. Getting anywhere near a half-decent timestretch algorithm is hard enough. Making it sound 'natural' is nigh on impossible, without some serious mathematics, time, space and relativity. Far better just to let Blip sort the AUv3 teething problems out first, and squash a few bugs before wading deep into that merde.

    Aren’t there open source timestrech algo/code options that could be added to obsidian sample playback for now? That would be better than no option in a lot of cases and matt could then work on adding a better quality, more ‘natural’ at later date..

    Not sure about this, but I think it’s difficult to just take an existing algorithm and then “tweak” it to make it better over time. The base values affect all subsequent values in some algorithms, so any tweaks in the chain could send the other values off the edge.

    But I’m not a mathematician, so I might not have this straight in my mind.

    I’m thinking more like adding a ‘basic timestrech’ option for now, copy/paste code type job...and then in future adding a ‘pro’ option. So the two Algos would coexist but be totally separate options and would be no worries with backward compatibility. Eventually Just being able to select which algo you want.

  • @flockz said:

    @kleptolia said:

    @flockz said:

    @tom_tm said:
    @flockz Easier said than done mate. Honestly. Getting anywhere near a half-decent timestretch algorithm is hard enough. Making it sound 'natural' is nigh on impossible, without some serious mathematics, time, space and relativity. Far better just to let Blip sort the AUv3 teething problems out first, and squash a few bugs before wading deep into that merde.

    Aren’t there open source timestrech algo/code options that could be added to obsidian sample playback for now? That would be better than no option in a lot of cases and matt could then work on adding a better quality, more ‘natural’ at later date..

    Not sure about this, but I think it’s difficult to just take an existing algorithm and then “tweak” it to make it better over time. The base values affect all subsequent values in some algorithms, so any tweaks in the chain could send the other values off the edge.

    But I’m not a mathematician, so I might not have this straight in my mind.

    I’m thinking more like adding a ‘basic timestrech’ option for now, copy/paste code type job...and then in future adding a ‘pro’ option. So the two Algos would coexist but be totally separate options and would be no worries with backward compatibility. Eventually Just being able to select which algo you want.

    Ah, ok. That might be possible. I’m not sure how that would work in codeland, though.

  • edited December 2018

    @kleptolia
    Ah, ok. That might be possible. I’m not sure how that would work in codeland, though.

    not that easy at all :) First of all if is something open source it doesn’t mean you can “copy paste” it to your commercial app and use it freely :)) Most opensource licenses means that if you use some opensource library you must provide source code of your app as opensource too :) Opensource doesn’t mean freeware or free to use commercialy, even when lot people thinks it does :)

    Plus the way how those library are coded often needs significant amount of effort to incorporate them to your code base, its faaar away from just copy paste...

    Plus if you get any trouble/bug/whatever, it’s usually hell to fix someone’s else. code with usually minumum support of author...

    How this works is to uy licence fir some good comercial solution (like elastic audio), pay for this licence - then you get stable codebase, usually with updates, and support in case of issues...

  • @dendy said:

    @kleptolia
    Ah, ok. That might be possible. I’m not sure how that would work in codeland, though.

    not that easy at all :) First of all if is something open source it doesn’t mean you can “copy paste” it to your commercial app and use it freely :)) Most opensource licenses means that if you use some opensource library you must provide source code of your app as opensource too :) Opensource doesn’t mean freeware or free to use commercialy, even when lot people thinks it does :)

    Plus the way how those library are coded often needs significant amount of effort to incorporate them to your code base, its faaar away from just copy paste...

    Plus if you get any trouble/bug/whatever, it’s usually hell to fix someone’s else. code with usually minumum support of author...

    How this works is to uy licence fir some good comercial solution (like elastic audio), pay for this licence - then you get stable codebase, usually with updates, and support in case of issues...

    Hope Matt finds a way to add it. It’s something I miss a lot when using samples in obsidian. Not a dealbreaker, it’s not ‘always’ present in similar instruments, but definitely something I often utilise when it is available.

  • Seems like it would have to happen for audio tracks at least... I imagine.

  • I don’t see why it’s such a dealbreaker. BM3 has it, but then BM3 is geared towards hip-hop, whereas NS2 is more psychedelic IDM/EDM & synthesis. I think Audiogus has a more coherent take on this. Timestretch would be best addressed when audio tracks are eventually included, and it would be far better to get it right, opposed to just slapping some 'open source' in as a five minute fix. Getting it right is 100% Blip ethos, and if this takes a couple more years to implement, then so be it. Best to fix the most pressing stuff for the moment, getting AU's to work correctly, and adding some sort of basic Audiobus inclusion.

  • @tom_tm said:
    I don’t see why it’s such a dealbreaker. BM3 has it, but then BM3 is geared towards hip-hop, whereas NS2 is more psychedelic IDM/EDM & synthesis. I think Audiogus has a more coherent take on this. Timestretch would be best addressed when audio tracks are eventually included, and it would be far better to get it right, opposed to just slapping some 'open source' in as a five minute fix. Getting it right is 100% Blip ethos, and if this takes a couple more years to implement, then so be it. Best to fix the most pressing stuff for the moment, getting AU's to work correctly, and adding some sort of basic Audiobus inclusion.

    Not a dealbreaker. Instruments like synthscaper don’t have this feature either. And, similarly to obsidian, they’re far from redundant. It’s just a useful additional sound design option. A la kontakt and ableton etc. It’s not a ‘hip-hop’ focused feature at all imho. Just another tool that’s nice to have around when sculpting a sound involving any samples.

    From my POV it’s just a feature request, not a ‘add this urgently before anything else!’ type of thing. Definitely would put audio tracks, IAA support and aufx automation ahead of this kind of stuff.

    FWIW I wouldn’t say bm3 or ns2 feel focused toward a particular genre. Bm3 is a lot more ableton-like than ns2 in terms of features and philosophy and I wouldn’t call ableton a particularly hip hop focused daw?

  • I presume then you are a BM3 user?

  • @tom_tm said:
    I presume then you are a BM3 user?

    Your point?

Sign In or Register to comment.