Thanks for posting this! Anyone interested in GPU parallel audio processing will enjoy this 1 hour technical discussion.
The main thing for me is that they define “low latency” for their technology to be either 1ms or 150us depending on the tier of use. Also, it works on ~any NVIDIA GPU sold in the past 2-3 years and AMD support is only slightly lagging the NVIDIA support. Apple M1 has progressed to early proof of concept with very good results. GPU Audio worked with both Nvidia, AMD and Apple on this development.
The biggest thing there for me is how they did what many experts say is “impossible”. Sasha says ~“We did not know it was impossible, so we did it.”.
I’ve been doing Nvidia GP-GPU parallel development since the initial CUDA SDK release in 2007. These folks know what they are talking about and have done what they claim.
Pay attention and stay tuned for the revolution