FORUMS: list search recent posts

RAID Drives Tested for 4K Editing

COW Forums : Avid Media Composer

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
Stefan Ilea
RAID Drives Tested for 4K Editing
on Sep 22, 2015 at 1:46:23 pm
Last Edited By Stefan Ilea on Sep 22, 2015 at 3:12:27 pm

How is it possible that the G-Tech I tried, performed amazing even though the AJA Test speeds were slower than the Promise drive I tried?

Here's what I tested:

- Playback of a 4K, 4 minute sequence in Media Composer 8.4
- Clips Transcoded to DNxHR HQ,

- Total size for transcoded clips was about 70GB (not sure if this is important)

- No transitions just a lot of cuts

- Proxy Off

- Video Quality Tested at Full Screen: Full Quality 10bit (green box +10bit), Full Quality (green box), Draft Quality (green/yellow) and Best Performance.


RAID Drives used for testing sequence playback:

PROMISE Pegasus2 R6 18TB, RAID 5, 6 Drives x 3TB
AJA Test Results @ Video Frame Size 4096x2160 10-bit RGB, File Size 4.0 GB:
Write: 861.6 MB/s
Read: 863.8 MB/s

G-Tech G-Speed Studio 24TB RAID 5, 4 Drives x 6TB
AJA Test Results @ Video Frame Size 4096x2160 10-bit RGB, File Size 4.0 GB:

Write: 529.1 MB/s
Read: 468.5 MB/s

Results
The G-Tech playback at:

- Best Performance (yellow box): Super Smooth. It felt like watching the smoothness of a DVD or movie at the theater.

- Draft Quality (green/yellow box): Looked pretty darn close to the smoothness of Best Performance, too close for me to see a difference in smoothness.

- Full Quality 10bit (green box +10bit): Looked VERY SMOOTH!!Yes, no distracting stutters. Maybe just the tiniest bit short of the super smoothness of Best Performance (yellow box). Any lag that is there is the least bit noticeable. To work at this resolution all the time would totally be worth sacrificing the lag, non-factoring lag, compared to Draft or Best Performance. Like I said it's so NOT distracting.

- Full Quality (green box): Strangely the playback was the worst of all of them, noticeable lagging! How is this possible when 10 bit playback looked AWESOME??

With the exception of Full Quality which seemed to be the worst, I'd like to screen cap/record my full screen playback at the various 3 other settings and overlay them with each other to see how much of a lag there actually is between them, because they all looked soooo close to each other in terms of smoothness when comparing in succession.

The Promise Drive Playback:
Playback was distractingly CHOPPY!!! at Full Quality 10bit (green box +10bit) and Full Quality (green box), and would only play smooth at Draft Quality (green/yellow box), and Best Performance (yellow box).

Wrap Up:

How is it possible the the G-Tech drive played WAAAY better than the Promise drive, especially since the AJA Test showed better speeds on the Promise Drive? Is it simply because it's a larger capacity? I thought that since the Promise has more physical drives, that would have made it faster, providing a smoother playback than the G-Tech.

I was hoping to make a final purchase decision and go with the Promise Drive, because I figured I'm not going to blow through 18TB so fast, but now that I see how much better the Media Composer playback performance is with the more expensive/larger G-Tech, I'm definitely leaning that way. Is it even worth testing a 24TB Promise to be on an equal playing field?

Thanks for any incite!!


Return to posts index

bryce arroyo
Re: RAID Drives Tested for 4K Editing
on Sep 22, 2015 at 8:11:09 pm

Stefan,

Im not surprised that the G-tech drives worked better during the playback. I think the more storage does help overall. Would love to talk about the G-tech drives more in depth. At Maxx Digital, we are one of the more premier resellers for G tech and work very very closely with them. Please, give me a call and would love to help you find a solution.

Bryce Arroyo
Maxx Digital
714-374-4944


Return to posts index

Bill Ravens
Re: RAID Drives Tested for 4K Editing
on Sep 24, 2015 at 1:23:09 pm

While not knowing, definitaiely, about the BMD Diskspeed Test, most diskspeed testers use a sequential read/write process. Video files, OTOH, are random read write. As you've discovered, that makes a significant difference to the actual thruput when real files are used, as opposed to test files.



Return to posts index

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
© 2017 CreativeCOW.net All Rights Reserved
[TOP]