HISE Meet Up

[d.healey]

@HISEnberg said in HISE Meet Up:

Ah, apologies, I meant spectral morphing with Loris in HISE realtime,

Ah ok, yeah I think probably some form of look up table, perhaps generated by Loris, would be the most efficient way to go. Actually it's kind of like what we can do with the wavetable synth isn't it.

[griffinboy]

@d-healey
@ustk

It's been done in real time recently using loris.
But it's what David said: you need to make your own engine for playback, using Loris for analysis and creating your own optimized way of synthesising the result from the loris partials array.

Check out Sumu.
The synth does exactly that.
Mind though, that it uses 30% cpu per instance right off the bat...

[d.healey]

@griffinboy said in HISE Meet Up:

Check out Sumu.

Oh nice, they are using Loris and their analysis tool is open source.

[griffinboy]

@d-healey

This is what spurred me to make my own additive spectral in c++. I couldn't beat Lori's sound wise yet. But I did beat Sumu in terms of efficiency.

I wish we had efficient additive but I think we are a ways off from that lol. I can never bring myself to use plugins that use more than 10% CPU. I run out of CPu headroom too fast. It's a real shame. Because Sumu sounds awesome.

[d.healey]

Here's the recording from yesterday. We lost a couple of bits in the middle when the recording stopped and had to be restarted, we also lost about 10 minutes from the end because the audio cut out.

I think in future we should use zoom as the recording system is better. @Simon as you know I also prefer foss but since we're using a remote service with either jisti and zoom the openess of jitsi makes little difference - unless you're self hosting. I always use the browser version of zoom so no need to install anything or even make an account. However if you want to keep searching for another foss solution, or something self hosted I'm happy to give it a try too.

[Simon]

@d-healey My FOSSy convictions drive me to figure out the nextcloud recording backend, which is probably gonna be the most reliable zoom replacement. Sorry for the audio :)

[Oli Ullmann]

@d-healey Thank you! :-)

[VirtualVirgin]

@d-healey said in HISE Meet Up:

@HISEnberg said in HISE Meet Up:

I don't think you could do spectral morphing in realtime

https://www.zynaptiq.com/morph/
https://web.archive.org/web/20180705061655/http://www.hakenaudio.com/RealTimeMorph/

I find the Kyma to be rather convincing for smooth morphing:

I would love to hear this kind of thing used to interpolate between sample sets for articulations that have capabilities for continuous control of timbre:
bow position, fluttertongue, growl, stick position on percussion etc.

There are a few different algorithms to choose from for morphing on the Kyma, and I don't know the details, but I think these are precomputed FFTs for resynthesis.

[VirtualVirgin]

It looks like the Kyma may use Loris, or some aspect of Loris in its system.
Kyma is mentioned in both of these papers on Loris:

https://quod.lib.umich.edu/i/icmc/bbp2372.2002.079?rgn=main;view=fulltext
https://www.hakenaudio.com/addsoundmorph

[Morphoice]

@d-healey sorry I couldn't make it in time to the meet, checking the youtube video now.
how did Kyma come into discussion? used this decades ago, didn't even know it still existed

[d.healey]

@Morphoice said in HISE Meet Up:

how did Kyma come into discussion?

We were discussing audio morphing/interpolation

[Morphoice]

@d-healey back in the days Kyma was the king of that, it's what they used for the podracers in Episode 1, I can't think of a better tool however back then you needed special hardware to use it