AI everywhere…..

[Svama]

Could be the beginning of a great song

….to you I swear.

What’s goin on in Meldatown?

The new Sonible smart limit sounds nice. Cool finishing touch, which is not that easy to achieve like one two three with melda.
So, maybe 2022 is becoming the revolution year, it remains exciting.

Happy mixing

[Hexspa]

You know what would be cool? A machine learning (if that's correct) VST that lets you capture the effect of any plugin. You just set up a few LFO's to move the knobs, the capture plugin records the effects and - boom - instant clone. You just tell it which parameters you're capturing and you can set up an Easy Screen via multiparams.

I know that in MConvolutionMB, you can sample IRs so maybe this is the next step. We can already run sine waves to examine distortion and use meters to observe dynamics curves. There are also one or two plugins which let you see the effect of a particular filter. Something like this could streamline the process. Even if it doesn't make clones, it'd be cool to get a 'report' on the characteristics of the device. You could then use that information to better inform your own device and sound design. Better yet, you could save it as a preset and recall it for saturation, EQ and dynamics effects.

[vectorwarrior]

That's interesting... I made a post recently about doing dynamics matching, much like how the auto eqs do frequency matching, but for amplitude. If you could combine dynamics and frequency response matching in a single plugin, add in harmonic matching (e.g. capturing saturation) then you'd have a full cloner.

One thing about this, it would need to be non linear, as in, doing captures at different volume levels, e.g. auto eq captures at a bunch of volume levels.

[MeldaProduction]

Hehe the things is, most plugins would be too complex even for dynamic convolution. Convolution is just sooo very limiting.