Why is GPU based audio processing not really a thing? (in: DSP and Plugin Devel…)
By: mEnZI1
22 Mar 09:35
In large-scale parallel computing tasks, GPUs have significant advantages. Currently, apart from FFT, convolution, and synthesizers with 1000 voices, I can’t think of many other computations in audio processing that would benefit sufficiently from parallelism.
In fact, I believe that audio processing dislikes parallel computation. Even simple IIR filters or feedback loops on an effects chain can destroy the entire parallel process. Yes, I know that IIRs can be processed in parallel; however, ad…
Go to first unread post in this topic
Read more here: Source link
