I have a multimedia Win 11 PC with 20 CPU cores and NVME drives (very fast) yet today I tried running the EQ Match module and preview failed saying my CPU wasn’t fast enough for real time preview. What!?! I have mixed 100-track songs with full FX on Pro Tools but I can’t run ONE real time EQ Match module in SL12?
What is the problem here? “Unmix” functions, that use the GPU, run VERY fast - but functions that use CPU run slow - like it’s only using 1 core. I don’t see any settings anywhere for CPU core usage assignment, so??
Audio is a linear format in nature. How would you split up audio processing onto multiple CPU cores?
For example, modern game engines implement a highly distributable component-based architecture, which can utilize many CPU cores running different processes in parallel.
This doesn’t work the same way with audio, where everything needs to processed in a sequential order from beginning to end.
Wha wrote describes the current state of the art. This does not mean that algorithms for parallel processing of Audio for some purposes are theoretically fundamentally impossible.
Before zero loss compression was invented noone thought Audio could be compressed this much.
So basically we have to wait for some genius Scientists to come up with solutions.
Works for me,for the processes I need, but of course it takes a while before the preview is ready (I don’t think it ever purported to be “live”, or realtime). I select a small section only in order to see if my settings will be effective, before running the process across the whole file.
Can you provide a step-by-step, so others could try to reproduce what you are experiencing?
Select the Match EQ module - select a profile - run “Preview”.
I have other “EQ Matching” plugins from Sonible, Waves, Melda Production and IK Multimedia that all work JUST FINE in Pro Tools that is already running dozens of other tracks. But yet I can’t run ONE file with ONE “EQ Match” process in SpectraLayers because my “CPU isn’t fast enough”. Nonsense.
This is a coding problem in SL. Sloppy, inefficient, coding. Why is everyone here defending this?
Just tried it again this morning after a fresh startup and it is working. The day I wrote this thread it was NOT working. That was after I had done several separations and using SL for some time.
So it seems there is a memory leak in the coding that is not clearing RAM between tasks.
Nobody is defending anything here, as far as I can see. We’re almost exclusively fellow users here, and this is not Steinberg support.
As a fellow user, I am under no obligation to offer any help, or spend any time at all engaging with you, however if you were to provide enough information to enable a reproducible issue on anyone else’s machine, it might form the basis for you to make a bug report.