Avid Pro Tools 10.3.9 HD & 11.1.3 HD Video Performance Tests - Part 3
08-17-2014 09:11 AM
The Story So Far…
Unless you have been living under a rock you will be aware now from my previous articles that desktop video in Pro Tools 10 is like a drunk fish on a bike as it meanders in and out of sync on a single pass, so should be avoided like the plague. See my video here showing Syncheck 2 and Syncheck 3
We also discovered that using a peripheral like the Avid Mojo, even un-referenced with video syncs and a sync HD, actually gave quite good sync.
The no brainer was that Pro Tools 10 HD with a referenced Avid Mojo and a Sync HD was rock solid, but the surprise was that it was so accurate even when not using Avid MXF media.
Which is Better For Pro Tools - Blackmagic or AJA?
From some of my previous articles on video hardware, you will have noted that I have found that AJA hardware performs the best with Pro Tools 11 over and above the Blackmagic solutions. Blackmagic were having serious problems in Pro Tools 11.1.2 (and below) with their drivers, whereas AJA were not. Blackmagic have finally improved their game and released a more stable driver which narrows the field a bit. I still however maintain that I prefer the thunderbolt AJA IOXT over the PCie Blackmagic Decklink Studio for several reasons:
- I prefer the IO software. The AJA software tells me the frame and line rate of the reference input and it also tells me what sort of display the outputs are connected to. Compare that to the Blackmagic which just unhelpfully comments “reference present”.
- It also allows very easy downmixing of the HD to SD with a very easy to understand software interface, allowing me to choose what favour of down mix I would like.
- I find the responsiveness of the AJA hardware superior both in jog and in darting around the timeline. It just feels better - which is all important when you are cranking it for 12 hours a day!
- AJA support by phone is excellent and their email support is top drawer too.
- Thunderbolt is a future proof connection compared to the PCIe of the Blackmagic and gives a portability and laptop accessibility for the future.
Therefore I continue therefore to hold the AJA IO XT as my “gold standard” for testing.
Video Performance With Pro Tools 11 and the New Avid Video Engine
OK, so we get to the 64 million dollar question - What is Pro Tools 11 HD video playback like with the new Avid Video Engine (AVE)?
Pro Tools 11 has the new 64 bit Avid Video Engine from their flagship Media Composer editing software which can play Avid Video wrapped favours (MXF) like the DnXHD Avid codec (its allegedly preferred diet) out of non Avid peripherals (or even the desktop), as well as all flavours of QuickTime wrapped media such as Apple’s ProRes codec or even Avid’s own DnxHD codec. It promises to be all things to all men and end the hassle I was having to go through, transcoding media to Avid MXF media in my Media Composer to get guaranteed stable playback in Pro Tools 10 with my Mojo.
The Test Rig
- Pro Tools 11.1.2 HDX
- Sync HD
- Rosendahl Nanosyncs (SD sync)
- ESE DV 319 (HD sync)
- AJA 10.5.1 software
- Blackmagic 10.1.2 Software
- Pro-Tools-PC (3.5 Ghz quad i7, 32 GB 1600 MHz DDR3 ram, NVIDEA GeForce GT 610 1024 MB, OSX 10.8.5
- Slate Raven (1920x1080 from NVIDEA gfx card DVI)
- Samsung Sycmaster 940 fed from Intel on board gfx (same as mac min)
- Fresh PP3 Duracell batteries
- Fresh Sense Of Humour
The Test Results Explained
- Tests were measured at the Twickenham satellite lab of Pro Tools Expert using Pharaoh Audio’s Syncheck 3 and Syncheck 2 (>thank you Mike Wabro) and the results given are an average of at least three readings as long as the readings were very close, and listed separately if they were not. >
- A span is also quoted, which is the variance within a play pass or how “drunk or sober” the video is “walking down a white line in the road of sync”.
- Each pass is the harder variable timing random sync check test with at least 30 secs of repeated random sync flashes per pass.
- All tests are 25 fps PAL and results are quotes as frames difference (or part thereof) between audio and video (remember that 1 frame = 40 ms in PAL land).
- In each test we only changed one parameter to make comparisons between them.
Pro Tools 11 with SD Video (Video Hardware Comparison)
- AJA IO XT using DV MXF media (video referenced) - Average = 0.10f; Span = 0f (no test possible on Pro Tools 10 as AJA hardware not supported on PT10)
- Blackmagic Decklink using DV MXF media (video referenced) - Oops this one got missed off the test series.
- Desktop Video using DV MXF media on Slate Raven (individual tests listed due to wide range of results)
0.36f; Span = 0.96f
0.10f; Span = 0.87f
0.46f; Span = 0.84f
0.19f; Span = 1.15f
- Desktop Video using DV MXF media on Samsung SyncMaster (individual tests listed due to wide range of results)
0.06f; Span = 0.85f
0.23f; Span = 1.08f
0.55f; Span = 0.78f
0.16f; Span = 0.82f
Pro Tools 10 with SD video (for comparison)
- Desktop video - DV Pal QT – no vid ref Av = 1.51f; Span = 4.70f (QT Offset = 36qf)
- Desktop video - h264 QT – no vid ref Av = 1.57f; Span = 5.26f (QT Offset = 36qf)
Pro Tools 11 with HD video (Video Hardware Comparison)
- AJA IO XT using DNxHD MXF media (video referenced) 0.26f span = 0f
- Blackmagic Decklink using DNxHD MXF media (video referenced)
0.20f; Span = 0.06f
0.05f; Span = 6.14f
0.20f; Span = 0.06f
0.12f; Span = 4.17f
0.12f; Span = 3.07f
0.24f; Span = 6.27f
- Desktop Video using DNxHD MXF media on Slate Raven
0.54f; Span = 0.94f
0.15f; Span = 0.88f
0.15f; Span = 0.85f
- Desktop Video using DNxHD MXF media on Samsung SyncMaster
0.51f; Span = 0.98f
0.02f; Span = 0.95f
0.07f; Span = 0.91f
Pro Tools 11 Video Codec Comparison
- AJA IO XT using DNxHD Quicktime media (video referenced) 0.07f; span = 0f
- AJA IO XT using ProRes HQ Quicktime media (video referenced) 0.02f; span = 0f
- AJA IO XT using H264 Quicktime media (video referenced) 0.12f; span = 0f
- AJA IO XT using DV Quicktime media (video referenced) 0.12f; span = 0f
- AJA IO XT using MP4 Quicktime media (video referenced) 0.11f; span = 0f
- AJA IO XT using MJPEG Quicktime media (video referenced) 0.11f; span = 5.49f
Pro Tools 11 HD Video Wrapper Comparison
- AJA IO XT using DNxHD Quicktime media (video referenced)
0.24f; Span = 0f
0.24f; Span = 0f
0.24f; Span = 0f
0.02f; Span = 3.03f
- AJA IO XT using DNxHD MXF media (video referenced)
0.26f; Span = 0f
0.19f; Span = 2.77f
0.25f; Span = 0f
0.25f; Span = 0f
Pro Tools 11 HD Video Reference vs Internal Sync Comparison
- AJA IO XT using DNxHD MXF media (video referenced) 0.26f; span = 0f
- AJA IO XT using DNxHD MXF media (internal reference)
2.19f; Span = 0f
0.46f; Span = 0f
0.16f; Span = 0.01f
0.44f; Span = 0f
0.48f; Span = 0f
0.02f; Span = 3.06f
0.08f; Span = 0f
Pro Tools 11 HD Vs SD Hardware Down-Conversion Comparison
- AJA IO XT using DNxHD MXF media (video referenced) - composite output
0.12f; Span = 6.22f
0.10f; Span = 5.44f
0.33f; Span = 5.70f
- AJA IO XT using DNxHD MXF media (video referenced) - HDMI output 0.26f span = 0.07f
- Blackmagic Decklink using DNxHD MXF media (video referenced) - composite output (Not tested).
- Blackmagic Decklink using DNxHD MXF media (video referenced) - HDMI output (Not tested).
Pro Tools 11 Tri-Level HD Sync vs SD Sync
- AJA IO XT using DNxHD MXF media (video referenced - 625i SD) from Nanosyncs 0.25f; span = 0f
- AJA IO XT using DNxHD MXF media (video referenced - 1080i HD from ESE DV 319 0.25f; span = 0f
If you are new to video in Pro Tools and don’t quite get all of this, panic ye not ,as we hope to run a “Video for Pro Tools 101 course” soon.
So there are the results, as I am off on holiday now I will leave you all to discuss and interpret. However as Fab Dupont of PureMix would say “No bickering please children”. Enjoy :-)
There are currently 1 users browsing this thread. (0 members and 1 guests)