I was wondering coming across this google project:
Would it be even possible to have this in some sort of in daw/vst version?
Cause i would love to use this inside my daw so i can more easily transfer sounds instead of having to go to their website, upload, see the result etc.
Is this anything somebody is already working on? and would it even be a good thing?
Cause in my eye’s i can see using these technology’s more closer into a working flow would be an enrichment for everyone involved in music.
Let me know! and sorry if this is the wrong place to ask about this.
machine learning things are not always realtime algorithms, but even if they aren’t you could just use ARA to implement it i guess
We’ve (Magenta) actually got a prototype DDSP working in JUCE at the moment, and are just bug squashing before we open source it, hopefully in the next couple months. We’ll let you know on the magenta webpage g.co/magenta.
I was actually thinking about magenta DDSP last night for some reason and lo and behold. Super excited for this.
Wow that’s great to hear! I can’t wait
hi @MagentaDevices / Magenta Team,
Any update on this?
I’m quite excited on this VST
Yup, sorry for the delay, it is definitely still in the works, and we’re just tuning it up to try and make sure it still sounds right while being efficient. We’ll definitely post when we release (hopefully soon, cautiously before the end of the year)
Seems like someone has made a version of this already:
wow awesome! is there any download link to the vst3 version? or i just can’t find it.
Their JUCE project is available here (link is in the paper): https://github.com/SMC704/juce-ddsp
Got it, thanks! I think this is still very prototype but cool to see where it’s heading.
Did someone get the Audio part for the vst to work? i can only place it on midi channels and can’t seem to route audio into it somehow.
Hi @MagentaDevices ,
would you mind sharing an update on this? I’ll begin writing my Master’s Thesis in the next semester here at ETH Zürich and I would like to extend DDSP/Tone Transfer in some way. Creating a VST and/or making Tone Transfer work in real-time is a natural choice. I would like to avoid reinventing the wheel and so it would be great to know how your VST is coming along. If you have extension ideas that you don’t have time to implement, I’d be happy to hear them too. Feel free to email me at “vvolhejn at student.ethz.ch” as well.
I think the cautious estimate from before is on track. We’re in the internal testing stages now and should hopefully have the open source release before the end of the year. Again, sorry for the delay.
awesome! any known date yet?
Thanks for being patient, time lines definitely have a way of shifting :). We’re now shooting for April 15th release.
Is the great day here now ?
yikes, that’s what I get for promising exact dates
We’ve actually made we’ve made a lot of progress and are about ready to beta test (still adding new features / training before full release).
If you’d like to try it out, just let me know here https://forms.gle/HoZUDPmTCiMDSHgB6 and we’ll try to get it out to you soon. We’d be happy to hear feedback
No worry’s! rather have a good working product and thanks for sharing that link, ive signed up so fingers crossed!!
Definitely no worries !
Its great to hear that its still in working hands.
Better a “some day in the near future” then a “sudden never”
Thanks for the update !