Well done to Steinberg there Mac Suport is epic ..this guy does great videos on new Macs and DAWs.
I think he does a decent job in general, but it would be nice if these creators took a few minutes to expand upon what really are “real world use-cases” and not just “how many duplicate tracks can I have before the DAW introduces artifacts and dropouts.”
There are innumerable threads here about “Cubase vs Reaper vs Ableton vs ProTools” and demands for “Cubase performance to match Reaper” but they all seem to ignore one the most critical of “real-world” requirements: monitored and armed tracks.
All modern DAWs basically maintain a single (as in “one”) priority, real-time, low-latency DSP processing thread. And this is (mostly) obvious because of the inefficiencies and complexities inherent in trying to deserialize a real-time audio stream for parallel processing split into multiple threads, only to have to serialize them all again. Even Reaper with its “anticipative processing” (borne of pre-scheduling massive numbers of threads) ultimately relies on a single thread for its real-time DSP. Just like all of them.
What would really help, in my opinion, in both on-line materials and posts outlining performance issues is some manner of indication of what tracks people have armed/monitored, and what the DSP requirements are (meaning, what effects chains exist for each) when they see that “only 3 of my 12 cores are being used by Cubase.” I could have no issue at all playing back 100 highly processed tracks when each track/bus/submix/group is processed by its own thread, but if I then monitor several and record-arm a few others, it could immediately saturate the core handling the single DSP thread and create drop-outs and artifacts.
So yes, I agree dude’s got decent videos, but I think it would be a service to others to better highlight where (in my experience, anyway) the actual issues are, which has nothing at all to do with “x-tracks on y-DAW with z-CPU” but more “are you paying attention to what are you forcing onto a single thread?” That could really help people identify when it’s time to bounce a track. Just my opinion, of course.
I actually do my own tests when I get a new CPU same Test I’ve been doing forever needy track running Kontakt with a massive chord like 40 notes all hitting on bar 1 , then start dup the tracks , it hammer the CPU as all notes hit at once each , M1 Ultra then M2 I now have M3 Ultra and each time its a big leap forward …
Andy
Hi Thor ! If you one day do videos I would LOVE to watch ![]()
Now wondering if you sound like Dan Worrall ![]()
As much as it is great to see great performance on OSX, Windows users are left behind in terms of efficiency.
Great news is great news, but we really should push for better efficiency on Windows.