Plug-In Development Frontiers

DSP, Plug-in and Host development discussion.
User avatar
vortico
KVRist
180 posts since 19 Jul, 2008

Post Thu Jul 26, 2018 1:28 am

Old reality (hallucinations of a data set) has a use in lots of audio applications. For example, I want to play some chords and a melody on my keyboard, run a command, and have a reasonable drum beat that follows the "intensity" of my progression with fills and variations in empty spots. It could be a NN that analyses a buffer filled with the STFT of each measure's audio and generates a vector for computing the locations of each channel's triggers. Stream-based machine learning is definitely a niche field compared to fixed-vector classification, but there are ways to cast the problem into more common terms. This can be used for lazy musicians like me, or for learning musicians that want to see how a professional might write a drum track for their song. I use randomization for inspiration a lot, and machine learning would simply give a more correlated random source, which would be a lot closer to what I usually want. The failured algorithms might even give rise to new genres of music. :)
VCV Rack open-source virtual modular synthesizer

User avatar
Delta Sign
KVRist
446 posts since 22 Jun, 2018

Re: Plug-In Development Frontiers

Post Thu Jul 26, 2018 5:28 am

Machine learning + additive synthesis could also be very interesting, like approximating an instrument by analyzing recordings of it. This probably wouldn't really work in any kind of realistic way, but the resulting sounds could be interesting nonetheless.

Making additive synthesis, which can theoretically produce any possible sound, controllable will be a big step in the future imo. We do have the processing power to do it now.

User avatar
Richard_Synapse
KVRian
847 posts since 20 Dec, 2010

Re: Plug-In Development Frontiers

Post Thu Jul 26, 2018 6:54 am

Benutzername wrote:I think one of the frontiers will be artificial intelligence. Maybe not in the DSP part of the software but in the creative, editing and mixing stages.
Actually it can be used successfully for realtime DSP algorithms, too. This is probably just not widely known, since AI is a fresh field with a lot of research activity, that is just beginning to take off. The core concepts may have been around for some time, yet there is a lot of new development in just the past year or two. Plus anyone can now rent a supercomputer to work on arbitrary problems, which makes AI particularly attractive.

Richard
Synapse Audio Software - www.synapse-audio.com

User avatar
fluffy_little_something
KVRAF
12042 posts since 5 Jun, 2012 from Portugal

Re: Plug-In Development Frontiers

Post Thu Jul 26, 2018 7:34 am

Delta Sign wrote:Machine learning + additive synthesis could also be very interesting, like approximating an instrument by analyzing recordings of it. This probably wouldn't really work in any kind of realistic way, but the resulting sounds could be interesting nonetheless.

Making additive synthesis, which can theoretically produce any possible sound, controllable will be a big step in the future imo. We do have the processing power to do it now.
Indeed because AI is particularly good at analyzing and then applying the results. Perfect for reverse engineering. No wonder the Chinese are investing so heavily in AI :hihi:

User avatar
fluffy_little_something
KVRAF
12042 posts since 5 Jun, 2012 from Portugal

Re: Plug-In Development Frontiers

Post Thu Jul 26, 2018 7:38 am

Urs wrote:In my opinion, academia does not often push new boundaries in our field.

As long as we live in a nostalgia-driven market, it's old farts with experience who figure it out, not young and ambitious lads who seek the new by all means. And that won't change because of the duality/dilemma of software vs. hardware. Unlike similar fields like Photography where digital was "merged into" the field with benefits, audio hardware is stubbornly setting the bar (and that's not remotely justified by any result).
But if I am not mistaken, professionals still prefer conventional, film-based, analog cams. Similar to vinyl vs CD's.

User avatar
Urs
u-he
22410 posts since 8 Aug, 2002 from Berlin

Re: Plug-In Development Frontiers

Post Thu Jul 26, 2018 8:07 am

fluffy_little_something wrote:
Urs wrote:In my opinion, academia does not often push new boundaries in our field.

As long as we live in a nostalgia-driven market, it's old farts with experience who figure it out, not young and ambitious lads who seek the new by all means. And that won't change because of the duality/dilemma of software vs. hardware. Unlike similar fields like Photography where digital was "merged into" the field with benefits, audio hardware is stubbornly setting the bar (and that's not remotely justified by any result).
But if I am not mistaken, professionals still prefer conventional, film-based, analog cams. Similar to vinyl vs CD's.
I think you're mistaken. Rich enough hobbyists won't part with their boutique analogue cams, but *professionals* who make a living from it use digital stuff all the way.

User avatar
fluffy_little_something
KVRAF
12042 posts since 5 Jun, 2012 from Portugal

Re: Plug-In Development Frontiers

Post Thu Jul 26, 2018 8:49 am

Yes, depends on what the profession is :hihi:

Interesting...
https://www.theguardian.com/uk-news/201 ... enaissance

User avatar
Ichad.c
KVRian
1039 posts since 8 Feb, 2012 from South - Africa

Re: Plug-In Development Frontiers

Post Thu Jul 26, 2018 10:54 am

CPU side -> better materials for better thermals, heat has always been enemy no1 IMHO. Better memory bandwidth/latency would help too.

Interactivity -> in the mid-term, I'm really hoping that the upcoming MIDI-CI spec lives up to its promises. Better hardware/software intergration will be beneficial to both camps.

User avatar
Urs
u-he
22410 posts since 8 Aug, 2002 from Berlin

Re: Plug-In Development Frontiers

Post Fri Jul 27, 2018 2:23 am

fluffy_little_something wrote:profession
"young fans", obviously.

stratum
KVRAF
1843 posts since 29 May, 2012

Re: Plug-In Development Frontiers

Post Fri Jul 27, 2018 3:29 am

Urs wrote:
fluffy_little_something wrote:profession
"young fans", obviously.
So the question here seems to be, who buys more plugins? Young fans or professionals?
~stratum~

User avatar
Urs
u-he
22410 posts since 8 Aug, 2002 from Berlin

Re: Plug-In Development Frontiers

Post Fri Jul 27, 2018 3:49 am

stratum wrote:
Urs wrote:
fluffy_little_something wrote:profession
"young fans", obviously.
So the question here seems to be, who buys more plugins? Young fans or professionals?
Well, if you want to be strict on the comparison, it's "young fans" who still buy analogue cameras.

Thing is, in the world of photography, digital sensors have made analogue film obsolete, and likewise did digital image development make analogue image development obsolete. Cameras, lenses and light are still physical though.

In the world of audio production, nostalgia prevails. Analogue stuff is the shnitzle, even though its limitations are often dreadful compared to what is possible in software (think compressors, eqs, mixing desks). I.e. it's possible to make digital compressors which surpass any analogue one in sound quality, flexibility, musicality, efficiency, price and what have you. But people still get teary eyes in front of a Fairchild, while a softcomp is always "just a tool" for them. :bang:

bbtr
Banned
410 posts since 21 Nov, 2005

Re: Plug-In Development Frontiers

Post Fri Jul 27, 2018 4:00 am

Urs wrote:digital compressors which surpass
Hey, I often wonder - where are those?! I know of two which are great, but come on, that's nothing compared to the hundreds of 'analog' plug-ins. And they are recommending sonic degradation as a feature, as a selling point! And people like that! ?!?! :dog: :dog: :dog:

So who is going to make THE experimental digital compressor from the future? Any volunteers? :oops:

User avatar
Urs
u-he
22410 posts since 8 Aug, 2002 from Berlin

Re: Plug-In Development Frontiers

Post Fri Jul 27, 2018 5:00 am

bbtr wrote:
Urs wrote:digital compressors which surpass
Hey, I often wonder - where are those?! I know of two which are great, but come on, that's nothing compared to the hundreds of 'analog' plug-ins. And they are recommending sonic degradation as a feature, as a selling point! And people like that! ?!?! :dog: :dog: :dog:

So who is going to make THE experimental digital compressor from the future? Any volunteers? :oops:
Let's put it that way: I'm extremely biased. But also hats off to Dave Gamble, if you need any example without my biased opinion. But there's more, certainly, I just don't know them well enough.

User avatar
Urs
u-he
22410 posts since 8 Aug, 2002 from Berlin

Re: Plug-In Development Frontiers

Post Fri Jul 27, 2018 5:06 am

To elaborate: With the parts available in nineteen-fifty-something, you could only model such and such compression curves. Add a few tricks, subtract any prior experience from competition. With software we can do *any* curve. Any. We can say "this works shit on drums, let's try something else". The obsession with "we want what (we know)/(always worked)" is mind boggling.

Translate to any other kind of audio processor, still be right. Just look at Cytomic's The Scream. You're free to use parts they never had. THAT is a perfect example of where we can go.

User avatar
EvilDragon
KVRAF
16925 posts since 7 Jan, 2009 from Croatia

Re: Plug-In Development Frontiers

Post Fri Jul 27, 2018 5:07 am

bbtr wrote:
Urs wrote:digital compressors which surpass
Hey, I often wonder - where are those?!
Tokyo Dawn Labs. Klanghelm. Presswerk.

Return to “DSP and Plug-in Development”