SpectraLayers not utilising all CPU Apple Silicon M1

SpectraLayers
1

I’m applying a process on a 40min piece of audio and it says 55mins to complete the process… the M1 CPU is at 49%
Why isn’t Spectralayers using all the CPU and extensions on an MacBook Pro M1? It’s like living in the middle ages with the slowest CPU…

2

I suspect your processing is thread dependent and SL is constraining real-time audio processing to specific threads. This is the behavior I would expect.

3

not quite sure what that means, and I shouldn’t need to, I’m not doing anything in real-time, its applying an offline process…
Running SL as an app… running other apps such as RX10 and perform a process and it utilises 100% of the CPU and most processes would at maximum take 5mins to achieve… obviously there is some serious issues with CPU utilisation with SL on Apple Silicon…
Not only that, if I click away from the SL window onto the browser window to while away the time whilst waiting 55mins for the unmix crowd noise process to complete, CPU use drops to 25%… clicking back onto the SL window it jumps back up to 50%… SL seriously needs to be optimised for use on Apple Silicon, the process time is from the dark ages

4

For speed in calculations of this kind you need GPU support , but Apple doesn’t have shared the necessary tools with external developers yet, as @Robin_Lobel explains in another thread:

5

I feel you. Currently processing a 4 GB file with Voice DeNoise from SL Pro 12 on my MacBook Pro M2 Max. Will be 24 h until finished.

6

Thanks for the info @bobmusic
But I’m wondering how every other developer has access to cpu optimisations except Steinberg. ATM it’s only using 50% of the main cpu… it should use 100% like literally 95% of the software that make intensive calculations (like machine learning / ai type stuff) are able to.
It feels like SL is code that is not native to Apple Silicon forced to run on it utilising hardly any of its power, it’s worse than non native software was running under Rosetta.

Of course ultilising gpu power would add even more power, but it doesn’t seem like they have the first step in place to correctly use and address all the cpu cores / threads.

And the new advert for SL12 is out saying how they have made all these enhancements to speed up productivity, literally the first thing they say in the advert… waiting for 30mins to 60mins for a process to complete in no way could be called ‘productive’. I can’t use SL in lots of situations especially if with a client or with a deadline to meet.

7

30mb .flac file took 30 mins on my MacBook Pro M1, a batch operation with 3 .flac 250mb file estimated time after 20mins was around 28 hrs and the base of the Mac was getting quite warm, something not experience with intense video processes in Davinci Resolve.