Hello everyone I was wondering if they made other programs or VST's like Blue Cat's Re-Guitar
for other instruments ,,etc re-piano , re-bass , re-trumpet .. I can use the different A-I's
on line to take the music apart , then convert to midi ,and create a different piano sound
in Kontact with the same melody , but it would be much easier to send the piano track in
and have a different piano come out any ideas appreciated

Thank You
Korgrog

 

try the Bandmate feature (AI) in Toontrack EzKeys2. You can drop wav files onto it, and they don't always need to be already separated; but you could always do that first.

 

Martinic - Kee Bass is a String Bass Simulator --> www.martinic.com/en/products/kee-bass
GSi Solo Trumpet is a physical modeling emulation of a trumpet --> www.genuinesoundware.com/?a=showproduct&b=56
IK Multimedia - Modo Bass 2 is a physically modeled bass --> www.ikmultimedia.com/products/modobass2/

 

Thank you ,,all for your feedback

points me in the right direction

 

Look into tone transfer plugins. Melda Morph being one of them. They are not the same thing, not even the same concept, but you can achieve interesting results. Lets say you have a Piano sample, and you want to turn it into an organ. If you can provide the midi, and play an organ sound with the midi, you can then use that track as the sidechain input into the piano, morphing the piano sound to have the tonal quality of the organ. Its often times wonky, but with some fidling around, this sort of re-synthesis / morphing can yield really cool stuff.

 

Re-guitar is nothing but a bunch of EQ-curves (and perhaps some very soft dynamics processing and light saturation, I haven't plugin doctor'ed it) at the front of your chain. It kinda works because the electric guitar's sound is, welp, basically a bunch of eq's and amplification distortion.

An acoustic instrument's sound is much more complex than an electric guitar's. Too much stuff going on. You can hardly emulate different physical material with eq curves. AI will probably get us there, but for now, we can't.

And that's why, even though it does a more than decent job with different electric guitar sounds, Re-Guitar absolutely sucks ass in its acoustic guitar emulations.

 

When AI get's better at polyphony, you will be able to do this similar to how you can convert one person's voice into another. Because it's not just about EQ, it's about timbre, formant etc as well obviously. Not quite there today, but progress is being made.

 

physics says: there is stuff going on. Otherwise you wouldn't be able to record (capture) anything so "complicated. The resulting outcome of any instrument is always a continuos wave. You should know that. If it an electric guitar or acoustic guitar or piano or a tin whistle ... the physics is exactly the same. And so of course you can emulate / describe everything with an appropriate EQ curve. That is what analog tape and digital ausio does. Exactly that.

And when AI can do something, NI (Natural Intelligence) can do that too. Because we (NI) build AI after the model of us, after NI. A pupil studies a subject, so does AI.

 

Last edited by a moderator: Oct 30, 2024

To the OP: there are IRs of acoustic guitars. I bet you can find IRs of electric guitars.

And: go to youtube, watch the videos of Jim Lil. Strictly physics. DO NOT LISTEN TO ANY GURUS. (Paul Reed Smith is one the boldest and stupidest of them ...)

 

Are you saying "if there weren't stuff going on, you wouldn't be able to record" ?

Yeah, instruments produce soundwaves. How very deep.

Can't wait to hear you transform the sound of a guitar into a Gamelan orchestra using the appropriate EQ curve.

Major in physics, minor in CS. What a genius.

Now, what I meant is it's much easier to emulate electromagnetic induction than air pressure, sympathetic resonance and such.

Prove me wrong. Using only an EQ, have an alto sax sound like a french horn. Or, to make things easier. just have a Martin acoustic sound like an Ovation. I'll wait.

 

Stop! He's already dead!

To his (misguided) point tho, it is totally something an neural network could do in theory. It's going to be a hell of a lot more than an eq curve.

 

Intelligent people will build new machines, but only if they produce products that are actually used and needed will they make money with them.

Simple musicians will then buy these complex devices that A. do the thinking and research for them and B. can perhaps operate these machines but not understand them.

Yes, musicians who are alive today and those of tomorrow who are not yet born are the target group of market strategists and financiers. So it must be worthwhile or some intelligent pensioner who has time will make his knowledge and his working time or lifetime available to the open source community and then there would be AI for the small or growing minority of musicians.

I don't have a crystal ball either. But so far every new technical development has been pushed forward so that in the end practical and inexpensive products will be available for all of us to buy.