Timestretching/Pitchshifting

[d.healey]

@BWSounds

This would be the scripting for it

const var Sampler1 = Synth.getSampler("Sampler1");

inline function onKnob1Control(component, value)
{
	Sampler1.setAttribute(Sampler1.Timestretching, value);
};

Content.getComponent("Knob1").setControlCallback(onKnob1Control);

But it returns an error that implies the attribute hasn't been setup in the HISE source. I'll take a look and see if it's a simple fix I can do.

[BWSounds]

@d-healey
Ahh ok makes sense now, thank you sir!

[d.healey]

Oh actually I think you're meant to do it through the JSON thingy - https://forum.hise.audio/topic/7303/timestretching-pitchshifting/77?_=1692482537361

But then I guess it shouldn't show up as an option in the parameter ID dropdown

[BWSounds]

@d-healey
I can script the ratio but the timestretch mode still does nothing.

[d.healey]

You have to use the names of the modes not a number.

I took this from the docs - https://docs.hise.audio/hise-modules/sound-generators/list/streamingsampler.html#timestretching

HiseSnippet 1325.3ocyX8+aSbCE2WacGIvPCDSZ+xjNEseHHwpRfBLooIBMsEUAkF0qflDBgbuyI0q9rOY6qqYS7+z9SX+Iw+AaO66tbNgnRZFvH+PUdO+9xG+7yedNcfRFS0ZoBEz3nwYTTv0vQiElS5eBgIP6sMJ3awFVJUaTTS7I6KSnGABnsFmQzZZBJHX0mXMMnwZH2m28nsHbhHlVqBgdojESeFKkYp0Nn2SYb9tDHhPBp0uYu8hkh9RtLGf0p3NnLR7ojQzmSrlsBFEr9NILiTEYHFpFrYKYx3nSj+tnv9Wxzri4TqPWTDDnB0n9mv3ICp1xZDJ.Ont.rZQA3V38YIrI5qKDeiagvZO7qAAqbQPp6k.RAdPZsBHcCbTrhkYpWwhmqh2SXnpgDnT6CkBaQqzdEbeIXgvrQJ4T5tJPXhGsePmN2ID9ys+4lMgxs1DdFQEFQRy3TU2veIz0FrwHpoTW6VUK1x5CSvYBZ3vbQrgIEgRwSExi6ZynRxaGKSyjBH22AhKOmd6l+YyFbYLgGJO92rguLX1LbTcC1AY1noaConz7TnkSCN7pVayzDnFlz5NgsbMTPwVYrR1.7RhhQDEhTH4.9iAaeMDHHiaX6bgn3h1qbPBVoYiIvPOWX.dBH4sfkUkR.t8q1asa41ysts06xcdeBmeLzu1d5BBDEuR0ykF5Ah1thRy21Lb1kFNbtqUFK6gw7V1he0E4XaQd5wTk+Qh0PnOa5l20Wrl23hRhmgRwdBl4fLZo7tRdhsoz982uUGUVSsM8kHDL03Z4udYKeDmkPUHFDjuB6JnHG38IWPuXusIFRUXfHBYIipLL6lIXa5YPuRwMoF3so5SMxLmskGi.o2GLomCbEAWM.gFa+xtPR0lw1Rx5NCPorjDNcfTyrkbe1PsglEw9CvTLt6Fc.KE9fOkbds3U5oyGNjctikaHimBsirr8RAxOTvMwuPSCSnCI4bSn9THNvUvSoIGIc1idehCfxRljyIlo4wrD1kKXAqO4gkfP.ahw9agKA4VmKjbaQg3MvCXESalCFWYNXDN6+TfwxQBeMdmgCowlZ.tFd2ecY4+W7z+8Eo+l3HfUh.MNiJoqbXnIth75iwz3rEdZ7.EkKIIE8zUl8281B5bopo09W8bo5woxbwT4pD4GRynDikZ1+NvgG9DkLOaVud2ibMEGo.1UnT3uvABJTqM96w9JoVOD1NtXo8WZPtZjsVUq4P5YTkdZcOOOENuEBJWauxGX4Yz1QNFV79D3d449V6xRmiHtyX3xnSt6Lx2cF46Mi7lyHe+YjevLxOrVtnftOIy1YXehzLT5WYwnzg.Lf.aYOK2hMZRvKzLM8gmoSp4Nb48frHpw.GYZOamyKeV+S0iwVawtL9AlV8kO85TucrQEFiX1SucDmAWaAMVLdSXLnaLRk1ooP1WJjYmHEr3ouj.M8iFQU9XetanGaLvczZM2p2gTNkn8tk+C8dF77QhZ569WpZwk+czy8756vEvMzx6E9e4b6yZa1ENhb0EZD4Gc7t7iK+rV5td4fyP2ue3K7y4qgcT8g1Iy+eA0Ck4V56podXXzXD7bfXp+.xUrCcKj6Xks.JhJRbB+C7obwtUSSsK1sZQTbUnf8V42cOpOXUpv8q9ru7iCyo49szdOW9yBJSIwJ4ahK9wJVrdEmF.kB2+bhF38sxgS9UJXbG2a9SXuIN1dQ3G6hPy2m6tD9bukvmMWBet+R3yCVBed3R3yOcg9XelviyMxzBRRPwfcJZrB1YRi0pn+E7KkaDE

[BWSounds]

@d-healey
got it, I have something like that earlier but I had numbers instead of the actual names... I appreciate your help

[d.healey]

@BWSounds Yeah I tried numbers first as well :)

[Lumi283]

@d-healey I tried, but it's not working... I tried with Tonality = 0.01, it changes the pitch but impossible to define which one.

[d.healey]

@Lumi283 I just looked at the docs and I don't think the tonality parameter does what you want - https://docs.hise.audio/hise-modules/sound-generators/list/streamingsampler.html#timestretching

Looks like you can just use any standard pitch modulator as long as timestretching is enabled for the sampler

If the timestretching is enabled, any pitch modulation (using the modulation or root note detune) will not change the pitch using resampling but with the pitch transposition of the timestretch algorithm which retains the duration of the sample.

[d.healey]

@Christoph-Hart Did you say there was a way to get a stretch ratio below 50%?

Oh never mind, I can just set it in the function and it works.

[Casmat]

@d-healey Can't get timestretch working properly on my end, I have three samplers and setting 2x timevariant timestretch to all three causes a lot of audio issues and clips, keeping only one sampler activated works better, but theres still audio issues, doesnt seem to be really cpu since my task manager cpu chart shows some resources left... the video shows the issues:

timestretchissue.mp4

In the vid, the first part is showing the timestretch enabled on all three samplers with half and double speed stretches. Then I turn off the first two samplers so that only one is on half speed stretching and it still makes some audio issues. Any ideas?

[d.healey]

@Casmat I can't assess what's going on behind the scenes in that video. Make a test project, add 3 samples, load some samples, play around with the settings in the sampler module UI. See what result you get. If it works without issue then the problem is caused by something else in your main project.

Are you using IPP?

[Casmat]

@d-healey yup, test project does the same, ipp is enabled and just recompiled today..

[d.healey]

@Casmat Send me the test project (with some of the samples you're using if possible)

[Casmat]

@d-healey https://we.tl/t-bDHJ1vv0Jg Try pressing down chords of 3/4/5

[Christoph Hart]

@Casmat you can‘t play 10 voices of timestretched sounds. One voice is about 10-15% depending on your system.

[d.healey]

I'm not noticing an issue

https://filedn.eu/larUQgXOwVjQdvpPaD96lHH/2023-08-28 12-09-59.mkv

[Casmat]

@Christoph-Hart @d-healey so i guess its time to get a new computer:grinning_face_with_sweat:

[d.healey]

@Christoph-Hart Do you think there is anything that can be done to improve the CPU usage, especially when in low latency mode? At the moment I think the applications for the time stretching in virtual instruments are quite limited because of these two factors.

[Christoph Hart]

@d-healey no it‘s a pretty complex algorithm. For me it‘s about as performant as the complex pro algorithm in Ableton.

You might add a HQ button that toggles the low latency that people can use when exporting to bring down the spikes a bit.