Glitchy 96k playback w/only 10%CPU usage on 12 core 5.1

[pencilina]

Hi,

I’m using a 12 core 2.66 5,1 mac pro and am getting processor spikes mixing songs with only 6 tracks at 96K/24bit. Plenty of plugins but none over 40% in effect performance RT, most under 20% PG and real time bar is at 50% during playback. Project SSD lives in a PCI slot. 32 gig ram. Activity monitor says user CPU use is around 10%. Glitches happen regardless of audio interface. Happens at highest buffer (2048) and changing prime ms doesn’t seem to do much. Other projects with plenty of tracks play fine at 48K/24 with oodles of plugins (system gets a little sluggish but doesn’t get audibly glitch on playback). Any ideas on how to improve performance or why this is spiking? MOTU tech support didn't have a solution. Thanks in advance.

[Henry Robinett]

Yes. I think this is a serious DP bug. I record and mix almost exclusively at 96k. Glitchy. AND I'm on a brand new 2019 Mac Pro 16 core, 160 gb 4TB. Apple is aware of it. My interface is Metric Halo ULN-8. They are trying to fix it but it is apparently MOTU's problem and I don't think they are responsive on it.

It doesn't always happen and is not 100% reproducible, so it's hard to find.I'm not sure this is the same problem I have had, but it sounds similar enough.

[Gravity Jim]

I record at 96K all the time on my 12-core 5,1, and don't experience any "glitchiness" in DP 10.11. Not to say this statement is helpful, but it might let you know that the problem is not inherent in DP.

[Henry Robinett]

Yeah, I know. But that's what I'm saying. A lot of folk have not experienced this. A hell of a lot of troubleshooting and analysis has gone on with Apple and MH since January. Not sure whether it's a conflict with the new MP or Catalina but BOTH issues have to do with DP. Most people don't get it. Both say they can probably find a workaround, which MH did for the most part for me, but the PROBLEM they both insist is probably MOTUs. The bug can be ferreted out with enough work. Most people called me crazy in January. But by February I was taken seriously.

[pencilina]

So its not just an issue with my system. I did 24/96 remote recording with my 2012 MBP with DP9 and my RME FF800 a few years back and had no issues and I was running off the internal drive. I suspect its something with DP10. Good thing bounces work and I'm not mixing with clients in my studio. Henry and Jim, do you guys have any issues playing back through internal audio? Maybe the issue has to do with clocking or playback on non-MOTU HW.

[MIDI Life Crisis]

I know people don't like hearing it (especially those with 12 core systems) but observing posts over the years there seems to be a fair number of people experiencing these kinds of issues using those super powerful systems.

I continue to wonder if the app(s) can't keep up with the machines. As dumb as that sounds, as penicillin states: " I did 24/96 remote recording with my 2012 MBP with DP9 and my RME FF800 a few years back and had no issues and I was running off the internal drive."

[Henry Robinett]

I haven’t tried internal audio. This is only something as far as I know introduced on Catalina and DP perhaps on the New Mac Pro. January of this year. Apple loaned the developer my same 2019 Mac Pro 16 core machine. After a while he was able to duplicate the problem. I think it’s only at higher sample rates. I’m thinking it might have something to do with the way DP does it’s background processing. Even when I turn it off it’s better but still I get some.

To be clear it’s less glitching than the audio engine overloads. It used to just stop audio altogether. Now it just spikes and j get the notice. Sometimes then i hear the glitching.

But this is ridiculous on a 7,1 16 core, 160 ram. Sometimes with not too many VIs or plugins.


Sent from my iPhone using Tapatalk

[Henry Robinett]

Well I’ve done a TON of 96/24 sessions for years with Metric Halo and DP and never had any issues. That’s why I know there’s a bug. Growing pains always is part of the process. Frustrating though it is I’d rather move forward. The bleeding edge is normally not for me. But I knew this was likely to happen. It always does in the edge. Always.


Sent from my iPhone using Tapatalk

[Henry Robinett]

I know people don't like hearing it (especially those with 12 core systems) but observing posts over the years there seems to be a fair number of people experiencing these kinds of issues using those super powerful systems.

I continue to wonder if the app(s) can't keep up with the machines. As dumb as that sounds, as penicillin states: " I did 24/96 remote recording with my 2012 MBP with DP9 and my RME FF800 a few years back and had no issues and I was running off the internal drive."

And yes. That always happens. When i was doing 96/24 remote with my ULN-8s there was never an issue. I probably did 30 or more of those.

Software always has to catch up with hardware when it’s new and pushing the envelope.


Sent from my iPhone using Tapatalk

[MIDI Life Crisis]

Not that I don't have CPU envy... lol

[pencilina]

My l tweaked 11 year old 12 core is hardly pushing any envelopes or cutting edge. Audible spikes are happening. It just might be a SW issue. Again, all is cool (but slow) on my system at 48K with tons of tracks and plugins but 96K ain't working too well with only a few tracks regardless of load or buffer settings on the same system as Jim with a slightly slower processor. I'm going to try playback on my windows laptop and see what happens. If anyone else has issues or a fix I'd be very interested.