Author Topic: Trying to enhance the sound of super articulation voices using AI  (Read 3611 times)

0 Members and 1 Guest are viewing this topic.

Trying to enhance the sound of super articulation voices using AI
« on: September 08, 2023, 10:56:52 PM »
Hi guys, today I tried to enhance the SA2 Tenor Sax from the Genos/CVP809/ CVP909 using AI (Artificial intelligence) and here is the result:

https://drive.google.com/drive/folders/1LVq0uuK6GdL30dB3Phokg0vQpzCg8-r0?usp=sharing


What do you think? Would you like artificial intelligence to be present in Genos 2?

Offline Michael Trigoboff

Re: Trying to enhance the sound of super articulation voices using AI
« Reply #1 on: September 09, 2023, 04:26:43 AM »
What kind of AI did you use? The details would be interesting.
retired software developer and Computer Science instructor
Grateful Deadhead emeritus

"He had decided to live forever or die in the attempt."
-- Joseph Heller, Catch-22
 
The following users thanked this post: rodrigo.b

Offline RoyceM

Re: Trying to enhance the sound of super articulation voices using AI
« Reply #2 on: September 09, 2023, 05:31:37 AM »
I might not have heard everything I should have. I was distracted by the AI version being so "dry," lacking the beautiful reverb that was in the non-AI version.

That said, it's an interesting concept.

- Royce
DGX-305, Roland E60, S900, S910, S975, Center Point Stereo Spacestation V.3, Bose L1 Compact
PSR Performer page
 
The following users thanked this post: rodrigo.b

Re: Trying to enhance the sound of super articulation voices using AI
« Reply #3 on: September 09, 2023, 03:46:00 PM »
What kind of AI did you use? The details would be interesting.

Hi, thanks for comment. I use Musicfy. I uploaded the mp3 file in that website and then I chose the Sax option, it enhances the sound in less than one minute, and you can download the audio file.
« Last Edit: September 09, 2023, 03:51:07 PM by rodrigo.b »
 

Re: Trying to enhance the sound of super articulation voices using AI
« Reply #4 on: September 09, 2023, 03:48:23 PM »
I might not have heard everything I should have. I was distracted by the AI version being so "dry," lacking the beautiful reverb that was in the non-AI version.

That said, it's an interesting concept.

- Royce


Thank you for comment. Yes, I think that now the IAs are doing absolutely amazing things, there is of course room for improvement, but I think this is very impressive
 
The following users thanked this post: MIKEP

Offline Gleston

Re: Trying to enhance the sound of super articulation voices using AI
« Reply #5 on: September 10, 2023, 03:20:03 PM »
Dear Rodrigo.b

Despite of been impressive, whats is the point of use an AI to make a better instrumental or voice track, if we, as human, are the core point of making music? Why put a machine to do it?

Hire a better sax player so you will have a better sax track.  What we do in our  keyboards will always be a imitation/simulation. We can play it better with dedication.


Donīt get me wrong. Use all this computer power in music is good, until the point where we are replacing humans. i will always prefer a humam player than a software improving or replacing a human.

I fear the day when we will only listen to computer generated music. When we will not play music anymore.  The process of learning a song or learning how to play better is so beatiful and make us a more confident person because we reach a point of being a better musician and that struggle make us better humans too, that the use of AI has to be very carefull. I fear the day when  all music will be a FAKE NEWS  too.

This is the second time (as far as I know) that you came with this issue (AI)  and always show us the use of AI to replace a human player in the final result. IMHO thats not the way we have to use AI.
 

Re: Trying to enhance the sound of super articulation voices using AI
« Reply #6 on: September 10, 2023, 03:50:12 PM »
Dear Rodrigo.b

Despite of been impressive, whats is the point of use an AI to make a better instrumental or voice track, if we, as human, are the core point of making music? Why put a machine to do it?

Hire a better sax player so you will have a better sax track.  What we do in our  keyboards will always be a imitation/simulation. We can play it better with dedication.


Donīt get me wrong. Use all this computer power in music is good, until the point where we are replacing humans. i will always prefer a humam player than a software improving or replacing a human.

I fear the day when we will only listen to computer generated music. When we will not play music anymore.  The process of learning a song or learning how to play better is so beatiful and make us a more confident person because we reach a point of being a better musician and that struggle make us better humans too, that the use of AI has to be very carefull. I fear the day when  all music will be a FAKE NEWS  too.

This is the second time (as far as I know) that you came with this issue (AI)  and always show us the use of AI to replace a human player in the final result. IMHO thats not the way we have to use AI.

Hello. I don't want to replace a real musician, but I'm always looking for ways to get a more realistic sound in my productions, with virtual instruments and musical keyboards.
 

Offline BogdanH

Re: Trying to enhance the sound of super articulation voices using AI
« Reply #7 on: September 10, 2023, 05:28:33 PM »
My opinion...
There's no such thing as "AI" music/sound creation! I can notice that there's a big hype about AI "whatever" lately -simply because it sounds smart and superior.
Those who know something about programming know that it's impossible for a program (which must be written by human), to invent/create something out of nothing. Everything that happens inside every program is predefined by human and so is the result -no magic here.

Improving sax (or any) voice... what does that even mean? It can't be improved, it can only sound differently. And that can be easy achieved by using well known existing techniques: EQ, sound effects, compressors, cutoff filters, etc. And that's exactly what AI tools are doing. The only difference between human and AI is, human will apply settings depending on what he wish to get. AI can't know what sounds best in particular case and so random settings are applied, where randomness is defined by human's programming according to certain criteria -and that criteria is (again) defined by human.
In short: AI sound improvement program is a lottery of sound settings.

I checked both files that rodrigo posted and I personally prefer original. Pay attention on short low tone at about 10 sec from start and you will notice that in processed file, that tone just doesn't sound right. AI can't know that, of course.. AI doesn't really care about that.

Just my 2c,
Bogdan
PSR-SX700 on K&M-18820 stand
Playing for myself on Youtube
 

Re: Trying to enhance the sound of super articulation voices using AI
« Reply #8 on: September 10, 2023, 05:58:15 PM »
My opinion...
There's no such thing as "AI" music/sound creation! I can notice that there's a big hype about AI "whatever" lately -simply because it sounds smart and superior.
Those who know something about programming know that it's impossible for a program (which must be written by human), to invent/create something out of nothing. Everything that happens inside every program is predefined by human and so is the result -no magic here.

Improving sax (or any) voice... what does that even mean? It can't be improved, it can only sound differently. And that can be easy achieved by using well known existing techniques: EQ, sound effects, compressors, cutoff filters, etc. And that's exactly what AI tools are doing. The only difference between human and AI is, human will apply settings depending on what he wish to get. AI can't know what sounds best in particular case and so random settings are applied, where randomness is defined by human's programming according to certain criteria -and that criteria is (again) defined by human.
In short: AI sound improvement program is a lottery of sound settings.

I checked both files that rodrigo posted and I personally prefer original. Pay attention on short low tone at about 10 sec from start and you will notice that in processed file, that tone just doesn't sound right. AI can't know that, of course.. AI doesn't really care about that.

Just my 2c,
Bogdan

Thank you for comment
 

Offline Oxford1035

Re: Trying to enhance the sound of super articulation voices using AI
« Reply #9 on: September 11, 2023, 02:13:50 AM »
l listened to both files before reading any comments and l agree with both Royce and Bogdan

The AI just sounds dead and lifeless. It's as if the sax player has been shut in a very small broom cupboard. The original is way more realistic and life like in my opinion. l don't know why anyone would want to go to the bother of changing the original and ending up with the end result you did sorry.

Kind regards,

Russ

Offline ugawoga

Re: Trying to enhance the sound of super articulation voices using AI
« Reply #10 on: September 17, 2023, 04:28:57 PM »
Hi

I can manage to improve the sax sound or to match the sound to what i need  using Izotope Neutron in Cubase.
Also Strings and Brass benefit in Neutron by AI listening.
I can say it does make the sound more pro sounding but with a little EQ to get of resonances. :)
This is my opinion only. :)
Genos, I7 computer 32 gig ram, Focusrite 6i6, Cubase controller, Focal Alpha Monitors, Yamaha DXR8 Speakers
Cubase 10, Sonarworks, Izotope.  Sampletank, Arturia and Korg software.  Now IK Mixbox
 
The following users thanked this post: rodrigo.b