Really high CPU usage in 11.1.3

[albee1952]

I just changed from an Avid003 to HDN with an HD IO and a 192 IO. No major issues with setup. I had to open>close>re-open PT for it to find and recognize the HDN audio engine(now shows and works fine). I had to correctly assign all the IO, including telling the digital connections to use ADAT(or optical, depending) and all my IO(all preamps and 5 stereo headphone mixes) are working fine. The thing that is confusing me is why the cpu meter is so high. This session is 17 audio tracks, a few aux tracks, BF76(x4), EQ III(x12), 2 D-Verbs and 1 ReVibe. Got a click track and 5 sends per channel(to phones) at the 64 buffer, and cpu is sitting at over 50% with a moving line that touches 70-80%. Made all plugins inactive without much change. Made all sends inactive with no change at all. The session will record fine, but can I trust it with a client? The same hardware in 10.3.8 works fine and at a bit lower cpu meter(should it not be the other way around). The cpu meter with my 003 was much lower with the same session. Thoughts on this anyone?

Also, I was assured that HDN with a 128 buffer would have less perceived latency than FW at 64, but its not feeling that way to me(Delay Comp on, but no "bad" plugins(all tracks showing 0 for compensation in tracking mode(high-latency plugins are inactive).

[nst7]

Don't know about the computer usage, but regarding the latency, you may want to clarify which interface you're using. Part of the lower latency is because of the new interfaces, which literally convert faster. If you are testing this with the 192, or with other converters going Adat, etc., even into the HD IO, then it may not be the same. You'll want to test how the latency feels going in and out of just the HD IO, with no other converters in the path (this would include if you are using a monitor controller with digital ins, because again, the same speed might not be there with that particular converter).

Also, regarding the speed in general from my experience, I have generally described it as "128 feels more like 64, and 64 feels more like 32". To clarify what I mean by that, I would say it's like an imaginary buffer size that's almost like 64 but not quite. Like an imaginary buffer of 86 if you're using 128, or say, 48 if you're using 64. Meaning, it's not quite down a whole buffer size, but almost, and definitely lower than other setups at the same buffer.

At least that's how I would describe it, and I hope I haven't been misleading in my descriptions in the past.

[albee1952]

Hmmm, budget limits forces me to mix a new HD IO(8x8x8) and an older 192. I have my 8 best preamps going into the HD IO analog and my Focusrite Octopre Mk II Dynamic on that unit's lightpipe. Outputs are a mix of the HD IO analog out(1&2 to mains and 3&4 as a headphone mix. The Octopre is receiving lightpipe FROM the HD IO for 4 more stereo HP mixes. The 192 is handling 2 more outboard preamps via analog, and WILL receive lightpipe from a Midas XL48 that is on the way.

I am aware that there may be a slight difference in latency thru the various converters at work, but I expect that difference to be subtle, if audible. More testing is in order

[nst7]

Yes, I know there's been some discussion about converter latency and phase issues when miking drums, etc. But I know you're smart enough to not do something like have your left and right drum overheads going through two different converters, or something like that. I'm sure you'll get it all worked out.

I just wanted to emphasize that most people's experience with the "faster buffer" thing has been with the new HD interfaces, and that would be the most accurate test. Because of your mixture of equipment, that may not be practical for your recording, but you can at least test it and get an idea.

To take the converter latency out of the question, and simply test the HD Native speed vs. the firewire speed, I suppose the best way would be to use the Octopre via Adat into the 003, and then switch sytems and try the Octopre via Adat into the HD IO. That way it's the same converter latency in both scenarios. You may not want to record that way, but it might give you some idea of the differences.

[albee1952]

Appreciate the insight, but I am trying to NOT go back to the 003 at all, due to major rewiring. In any case, Here's the cpu meter with the same session opened in PT10. A head scratcher. I have a band session tomorrow so I will track in 10 and see what happens

[nst7]

Regarding the CPU, I don't know if you were using any version of PT11 before. You may want to try using HD11 without the hardware, for testing purposes. Perhaps the computer's built in output, or your 003 again (I know you don't want to hook it up).

This way you can find out if it's the HDN hardware, or if PT11 is simply taxing your computer in a different way, regardless of interface.

[albee1952]

I have run it with the 003 and performance was not so great Maybe gonna have to look at a new 'puter soon(like I didn't already spend enough.....).

[nst7]

It could be that PT11 is more picky than older versions of Protools. At least that tells you that it's the software and not the HD Native hardware.

[guitardom]

Dave, I don't have a lot if time right now, but run a DPC latency test. Something is quirky. Might try moving the card to another slot.

[albee1952]

3 things:
#1-where do I find this DPC latency checker?
#2- I am buried for the next several weeks, so I'm glad I can still work!
#3-changing slots may not be a good option as all are filled now

Ran all afternoon(in 10.3.8) without any glitches at all. Tracked at a 128 buffer and nobody ever commented on latency(and the players enjoyed having their own mixes..in stereo). Other than the PT11 weirdness, the move has been pretty painless