My experiments with HD Native

[Drew Mazurek]

I'm going to post my experiences here as I go, for what it's worth.

Picked up the card today from my dealer. It was a demo in a Mac Pro 4,1. I have the dreaded Harpertown 3,1.

It installed easily and PT9 on 10.6 saw it and it worked from the get go. PT10 on 10.7 asked for the firmware update and I had to cancel, fearing I'd blank my dealer's card. So no testing on 10 for me unfortunately.

The PT icon does bounce several times before launching with the Native card but everything else (opening sessions etc) seems faster.

I currently have the Native card in the Mac and the Magma chassis card as well. PT TDM seems to "trump" HD Native, in that I have to turn off/unplug the chassis if I want to use the Native card. Also, looks like I'll need to restart after switching the Digilink cables between the two systems. I plan to keep them side by side for a while I think.

I opened a few sessions and everything seems fine. It sees my interfaces no problem etc. I did some tests and found that the RTAS version of TL Space isn't very efficient so I had to swap that out for IR1.

Just doing some tests without musicians here:

My basic tracking template:

14 drum tracks, each with an API 550 and CLA 1176 (not what I actually do but just trying to stress the system)
s-gtr-PODfarm amp sim and a C4 multiband
s-voc-digi HPF and a CLA 1176
Bass DI and Bass mic into a subgroup with a 550 and 1176.

2 Waves IR1s (replacing the seemingly inefficient TL Spaces)
2 Echoboys
1 POD farm echo (trying it out as replacement to Echofarm)
1 POD farm Mods (replacing Pitchblender)

2 Headphone mixes.

No VIs.

Playback Engine:
Buffer is set to 64 Samples.
all 4 processors on (this is a single processor 2.8)
CPU set to 95%
Delay Comp is Long.

18 tracks armed and recording. Using 38 of 192 voices.
RTAS CPU is at 22%.

I can even change the buffer to 32 and the CPU only goes up to 25%.


It's weird to see RTAS plugs NOT bypass on armed tracks!!


That's it for now. So far so good.

[zakco]

I'll be keeping an eye on this thread...really interested in your verdict...

thanks Drew!

Z

[Dizzi45Z]

My concern about HD Native is always about how the project is going to handle when I am tracking my final lead vocals with a million virtual instruments and other tracks with a million plug-ins each . I'd hate to have to increase the buffer size at that point and track without plugs and effects and etc.

[Drew Mazurek]

I've opened a bunch of mixes and have had to go up to 128 for a few.

[acmost]

Hey thanks for posting...looking forward to the continuing story.

[Keith Robichaux]

Drew I would really like to see hard numbers on analog to analog throughput in samples at different buffers and at different points in the project.

Do you have access to running a UAD card along with the HD Native to see how that really works under load?

Thanks

[Drew Mazurek]

at a buffer of 32 samples-164 samples of delay.
at a buffer of 64 samples-228
at a buffer of 128 samples-356


No UAD card, sorry.

[Keith Robichaux]

Originally Posted by dmazurek

at a buffer of 32 samples-164 samples of delay.
at a buffer of 64 samples-228
at a buffer of 128 samples-356


No UAD card, sorry.

At what sample rate?

Thanks

[Drew Mazurek]

I do everything at 44.1

[cmaynes]

you might try using 2 processors at 99%- it works much better in my experience. I have a 3.1 with HD Native.

as well- the firmware issue only seems to be a problem under 10.6.8