For me creating the exact sound is not very interesting from sound designing perspective. You can always sample the real instrument.
Like physical modeling synthesis, the interesting part is to compress the sound to some parameters that you can tweak and generate new sounds
Another approach is VAE, which also you give your some latent embedding, you can tweak the embedding to generate new sound. However the meaning of this embedding is not explicit.
People who are interested in this application should check synplant[0]. It has a ML technology called "Genopatch" which gives you 2 functionality:
1. you can try to describe a sound with some tags and it will try to generate a sound to capture the feeling of these tags
2.you can feed it with a sound sample and it will try to re-synthesize the sound with its synth engine. Though the end result will usually be just a "re-imagined" version of your input sample.
My guess is the underlying model is not a "deep" model. The main benefit is that the end result is not a wave file, but a list of generated parameters that can be synthesized by the synthplant engine. And now it comes the interesting part: you can tweak these parameters to finetune the generated sound. These parameters have actual meanings (FM ratio, reverb etc.)
Try pair reading it with a good thinking LLM like GPT 5.5 or Claude Opus. I found it help me a lot.
I have started learning SBCL internal from the beginning of this year with the help of GPT, and I really want to contribute to SBCL compiler someday in future.
Some of these models are open weight. You can try hosting them and do the price calculation yourself.
They also publish papers talking about how to save kv cache and computation powers. Because currently they don't have the most powerful nvidia cards, training and inference efficiency is very import for them.
It’s true that American companies have benefited from sending jobs overseas. Instead of trying to stop them, perhaps we could explore ways to share this money more fairly across the country.
I'm very interesting in this field (realtime audio + GPU programming). How do you deal with the latency? Do you send or multiple single vectors/buffers to GPU?
Also I think because samples in one channel need to be processed sequentially, does that mean mono audio processing won't benefit a lot from GPU programming. Or maybe you are dealing with spectral signal processing?
> What features has Google killed, do you think?
The problem is more like google can implement whatever feature they want and force it into web standard
> How often has this been a problem, do you think?
very often
> What other browsers have gone ahead, with Google holding out?
Google is holding out in a sense that no other people can implement a feature-complete browser. Google is killing the "open standard" web by make the standard impossible.
I use rpi zero 2 as a coding server (paired with my iphone and a splitted keyboard) when I have slow or no internet. It's quite usable if you just mosh/ssh into it and use emacs/vim to code.
Like physical modeling synthesis, the interesting part is to compress the sound to some parameters that you can tweak and generate new sounds
Another approach is VAE, which also you give your some latent embedding, you can tweak the embedding to generate new sound. However the meaning of this embedding is not explicit.