Using Chat GPT with your software Synths

[Winstontaneous]

I am not sure who names them but they must also name race horses, boats, and guitar pedals

Haha that is a great observation!

[ROTMetro]

Thanks for the info. I had no idea all of these customization options were available. Were these added recently?

Regardless, I'm going to try your Reddit suggestion. Anything to stop the obsequious praise, chipper acknowledgement of its shortcomings, and evasiveness when refusing to answer certain questions straightforwardly.

I think they've been there a while. It'll still be too quick to praise, etc, but it makes it better. You can always use the Monday voice (it's a take on the depressed robot Marvin from Hitchhiker's Guide to the Galaxy) in the voice interface to make it sound like it's making fun of you when it gives praise.

[Touch The Universe]

Would like to see someone make a bridge so chat GTP can interface with any arbitrary synth or even a DAW.

Heck all you need to do is prompt GTP to actually make it

Yes, check GTP is willing and able to make it happen. That should be a solid business idea if I've ever heard one, at least the potential.

Before we are talking about it's the idea that matters not so much the coding skill but I would add to that even if you have a good idea it's actually the will itself to see it through to completion that's key.

[korgscrew]

Playing Devil's Advocate here.

99% of people who listen to the music you/we make don't care if you made the sounds from scratch, got them from a preset library or if Chatgpt gave you the synth settings.

As long as it sounds good, brah.

[Touch The Universe]

Sound design is a crucial and critical aspect of one's quote unquote sound. Depending on the genre, yeah that actually does sound like good advice, but not as potentially fun or rewarding.

[osofaux]

I'm one of those that view AI with a suspicious and skeptical eye, but if it could take the Emulator II Factory Library images I have on my hard drive, bypass Arturia Emulator II V's half baked "just the samples" .eii import, and convert them into native Emulator II V patches with proper filter settings and envelopes that sound just like they do on the hardware, I think I'd become an instant proponent.

[IvyBirds]

Sound design is a crucial and critical aspect of one's quote unquote sound. Depending on the genre, yeah that actually does sound like good advice, but not as potentially fun or rewarding.

To each his own, I love sound design, using AI tools enhances my sound design experience and makes it more fun and rewarding

[rewthless]

Remember to always say "please" and "thank you".

[Hyperbole]

I'm one of those that view AI with a suspicious and skeptical eye, but if it could take the Emulator II Factory Library images I have on my hard drive, bypass Arturia Emulator II V's half baked "just the samples" .eii import, and convert them into native Emulator II V patches with proper filter settings and envelopes that sound just like they do on the hardware, I think I'd become an instant proponent.

I think that would be a pretty tall order. Converting patch and sample libraries from one format to another requires much more than mathematically mapping parameter values from source to destination formats, although that is usually a necessary first step. It then requires using one's ears to fine-tune the conversion, compensating for linear vs. exponential envelope curves, as just one example. That involves a lot of detailed listening.

But ChatGPT can't "hear" anything. There is currently no way for it to "listen" to the sound of the filter envelope decaying, for instance, within the source composite sound and match that in the destination patch/instrument. Or at least I don't think there is.

Human hearing, perception, and mental processing are extraordinary when you think about it. They are the product of millions, if not billions, of years of evolution. ChatGPT has a ways to go to match what we can do with the small, low-power "meat computer" running in our heads.

[husker37]

I've been playing around with this with Arturia synths. However, though it spits out files that I can import, the files never actaully seem to work. The plugin states the file(s) import successfully, but they then will not show up in the browser.
I asked ChatGPT about this, and it said "yeah, sorry, these files won't work, I'd have to do it in the actualy plugin which I can't do." So, I'm curious, how have you folks been able to get it to generate legitimately importable samples. (I've tried the Dx7, Obxa, and Pigments - same issue with all three).

[Coriolis Lab]

I feel it could be really worth for Csound or SuperCollider stuff... We all use chatGPT as a coding assistant at my developer job (web oriented). So far I tried it for Max/MSP and it was nowhere as good, giving me only non-sense patching structure... but let's see.

[vurt]

Remember to always say "please" and "thank you".

and eat your vegetables, especially broccoli...

[jdoo]

...
But ChatGPT can't "hear" anything. There is currently no way for it to "listen" to the sound of the filter envelope decaying, for instance, within the source composite sound and match that in the destination patch/instrument. Or at least I don't think there is.

Human hearing, perception, ...

Pattern matching has been around for a while now. Use SoundHound to get what song you are whistling, or hearing on the television. Yikes. Breaking down frequencies, timbres, etc.. has to just be around the corner?

To this conversation. If you have used ChatGPT for any length of time... Type the following:
I am xxx yyyyyy ( replace with your name ). What so you know about me. What is your honest opinion of me?

Now sit back and see what it comes up with. It knew details about me that floored me.

I subsequently had to type "Forget my xxxx investments, projected xxx income, detail about yyy and zzz"

to which ChatGPT replied, "Got it, JDoo—those details are forgotten. If there’s anything else you’d like to update or keep private, just let me know anytime."

so... be careful out there!