I know you hate it - They hate us too

[Innermost]

Low audio quality, artifacts and generates loops poorer than a human can. Fine if you enjoy low bit depth creativity over high definition man made ideas

Completly agree with you about low quality and artifacts. I wonder what creative minds will make with what is currently a flaw

[Innermost]

Pretty cool. I actually like the lo-fi sound. I'll see what I can come up with.

Nice! Can't wait to listen to what you are going to release with

[wagtunes]

Question: How do you change the key signature these loops play in? Can you even do that?

[Innermost]

Question: How do you change the key signature these loops play in? Can you even do that?

Arf when they are already generated you can't, I already tried to add a pitch shifter but I had some latency and some clicks. Then, if you want to regenerate a sound with the same prompt, use the dropdown menu on the right, at the right of the master volume and update the key. If you update pitch knob it updates the speed too.
But even with the same prompt you will not have the same generation because I don't give access to the seed. Would that be something you would like to have access? With diffusion models as Stable Audio 3 (or Stable diffusion models for example) with the same prompt and the same seed (a random big number) you have the same result.

[wagtunes]

Question: How do you change the key signature these loops play in? Can you even do that?

Arf when they are already generated you can't, I already tried to add a pitch shifter but I had some latency and some clicks. Then, if you want to regenerate a sound with the same prompt, use the dropdown menu on the right, at the right of the master volume and update the key. If you update pitch knob it updates the speed too.
But even with the same prompt you will not have the same generation because I don't give access to the seed. Would that be something you would like to have access? With diffusion models as Stable Audio 3 (or Stable diffusion models for example) with the same prompt and the same seed (a random big number) you have the same result.

What I would like to be able to do is say "I want these loops to play in C major". Right now it seems random. Is it?

[Innermost]

What I would like to be able to do is say "I want these loops to play in C major". Right now it seems random. Is it?

Ah ok, then just update the key on the dropdown menu on the right then generate and normaly that's it. But keep in mind that as it's AI it will not always be perfect.

[BBFG#]

What I would like to be able to do is say "I want these loops to play in C major". Right now it seems random. Is it?

Ah ok, then just update the key on the dropdown menu on the right then generate and normaly that's it. But keep in mind that as it's AI it will not always be perfect.

Wouldn't it be better to ask it to transpose loops by half or full steps than by key and mode?
That should help in keeping any accidentals intact rather than shifting them into a specific mode IMO.

[Innermost]

Yep totally agree. I still have the pitch shifter code, I could try to redevelop that feature but I had some artefacts, I wasn't happy with. Let me think about that. I have to think about how to implement that technicaly and on the UI.
How would you see that on your side?

[BBFG#]

Maybe limit it to five half steps up and six half steps down. ICBWH, but I tend to think of generalized artefacts occuring more in higher registers.

[Innermost]

Maybe limit it to five half steps up and six half steps down. ICBWH, but I tend to think of generalized artefacts occuring more in higher registers.

Ok. I'll try to implement that again then. Hope it will work correctly.

[wagtunes]

I've started working on the song. It's called "Do The Neural"

I hope to have it finished by tomorrow or Monday the latest.