Forum Replies Created

Page 27 of 83
  • Just use the onboard GPU if you are waiting on a 480. The RX480 is great on power usage, but not the top in calc speeds. If you want to just stick in a temp card to enable OpenCL/GPU benefits, there is always ebay. I would recommend the following in order:

    AMD:
    5770 (1360 gflops)
    r9-270x
    6970
    7970-ghz
    RX480
    r9-290x
    R9-390x
    R9-FuryX (8600 gflops)

    The i7-6700 is around 113 gflops, so if you do any blurs or other GPU effects adding a GPU will help even a modern CPU. If you are planning on running a 4K monitor, then use the onboard GPU, and the add on GPU for compute only. 7970 and above are the only cards to support 4k. There is a difference between display resolution ability and computational performance.

  • NV does not have the best OpenCL support, AMD is much better, so try turning off your GPU under Options>Video>GPU. Restart vegas after the change. See if that stabilizes your edits. AVID, Premiere and Resolve will utilize your 760 better than Vegas.

    I posted this about a week ago, and variants of this about 6+ times in the past year.

    What version of Vegas are you running? The latest is build of Pro is 453.

    What plugins do you have loaded, and are using on the project timeline? Are they all Sony plugins or 3rd party?

    What is the workflow, from camera/source material, and what profile are you rendering to? “Media Info” (an application) info on source material helps others to know what you are dealing with.

    Are you running the latest video drivers for your GPU?

    Have you pushed through all the windows updates, or are there any updates pending a reboot to complete?

    Can you list a screen shot of “speccy” summary screen, and are your Temps on all measured devices staying below 55C or 130F?

    Have you tested your memory for 100% zero errors with Memtest86+ lately?

    Do you get a clean bill of health after running “sfc /scannow” from an admin command line?

    Have you run an AV scan with Malwarebytes, and cleaned anything found?

    Run CCleaner with Regclean and repair. Reboot after.

    Run the windows “disk cleanup” on C:, then select clean up system files, and check all boxes. Then reboot, and run windows update to check for additional updates. Install updates, and reboot.

    Test your system with DPC Latency Checker, https://www.thesycon.de/deu/latency_check.shtml , to verify that your system drivers are behaving.

    Test for proper OpenCL operation with a benchmark tool like Luxmark 2.0, enable for CPU+GPU render. Monitor heat levels are <60C in Speccy during render.

    Failing after editing awhile sounds like heat, driver issue, or memory errors to me off the top of my head.

  • Aaron Star

    July 12, 2016 at 5:40 pm in reply to: GPU ‘upgrade’ not an upgrade?

    You can lead a horse… ahh whatever.

  • Aaron Star

    July 12, 2016 at 5:22 pm in reply to: GPU ‘upgrade’ not an upgrade?

    “thumbs down for the R9 390 if you are a Vegas Pro user”

    That statement is based on what? You offer no insight into your testing process, media types or project settings, so there is no way for anyone to know what you are doing.

    I can assure you that the 390 will out perform the 570 with the proper use case. A use case like AVC, prores, cineform, xdcam, or HDCAM content with layers of timeline text, compositing, or effects like GBlur, Rays, SoftCon, Glow, cookie cutter (power windows,) min&max.

    A go to test for OpenCL enhancement is to use a 1 minute clip of AVC, prores, cineform, xdcam, or HDCAM encoded content, then apply a “Sony Min and Max” with a “Sony Defocus” effect. Test play back speed, and render times using both CPU only and Auto settings in the render template. Monitor GPU utilization with GPUz or “AMD System Monitor.” If you get any “…” after your Frame: counter, then your system is dropping massive amounts of frames to keep up even with the low speed it is showing.

    Most people also will render their finals in 32-bit Video Levels, this is another case where the enhanced Floating Point calculation speed will be noticed.

    There is more to the rendering pipeline than just MC encoder acceleration.

    I spent most my time editing and not rendering, so I would not give up an improvement in calculation speed that the GPU offers. If your use case for Vegas is just an elaborate video file converter, then you probably will not see much improvement with OpenCL.

  • Aaron Star

    July 11, 2016 at 8:54 pm in reply to: GPU ‘upgrade’ not an upgrade?

    For detailed help, posting a screen shot of Speccy overview, and details of your workflow.

    Workflow would mean

    Project settings screen shots.
    “Media info” (its an app) details on your source material
    Then render settings, along with what your final purpose is for the output.

    Redering to Sony AVC, or what others have been doing the frame sever thing to handbrake would be aternatives. Mainly with SonyAVC, your GPU will help with decoding source material, timeline frame work, and then a very light amount of calculations with the encoder. The source frame decoding depends on the source codec however.

  • Aaron Star

    July 11, 2016 at 8:17 pm in reply to: Rendering Issue

    Whats with the odd frame size and FPS in the project settings? It this for web design or something?

    Trying doing 1:1 HD settings and see if the problem persists. Project setting 1920×1080@29.97FPS or match the media using the match media button.

    Try in both MC encoder and Sony AVC encoder.

  • Aaron Star

    July 11, 2016 at 8:12 pm in reply to: GPU ‘upgrade’ not an upgrade?

    Well you can go back to the 570, but you are trading a GPU cabable of 5000 GFLOPs vs 1400 GFLOPs. OpenCL in Vegas would like the 5000 over the 1400 for sure. But if your only goal is to use MC mp4 encoder with GPU, it makes sense.

  • You realize that the H110 chipset is only capable of Gen2 PCI and not PCI3.0.

    https://www.tomshardware.com/news/intel-skylake-broadwell-100-series,27244.html

    I would try the following in skylake terms:

    Z170 chipset board – Gigabyte
    16GB of DDR4
    6700K i7 CPU (113 GFLOPs)
    AMD 480 (5000 GFLOPs)(36 Compute Units)

    vs
    AMD 260x (1900 GFLOPs)(14 Compute Units)

    The 260 will work just fine with Vegas, but Sony used to spec the 5770 as a minimum at 1500 GFLOPs. A Vegas editing system without a GPU OpenCL compute units, is sort of like a unicycle vs a bicycle. Each will go, one will go faster.

  • Aaron Star

    July 10, 2016 at 6:02 am in reply to: 4K render in Sony vegas pro 12

    Here are the specs for YouTube: https://support.google.com/youtube/answer/1722171?hl=en

    I normally exceed those bitrate standards on upload by 25% to 50% to maintain quality due to the re-encode that YouTube does.

    You do not mention frame rate in your format specification, or duration of the material.

    XAVC-s LGOP at 60Mbs for a 4k image is pretty compressed already, unless they are expecting HEVC/x265 format.

  • Aaron Star

    July 10, 2016 at 3:27 am in reply to: GPU ‘upgrade’ not an upgrade?

    The Intel® Xeon® E3-1246 is more like a Gen3 i7 and pretty much meant to be a server chip for 1U rack systems. The 5960x is more like a dual proc xeon with the way the x99 chipset is laid out. The comparison is apples and oranges.

    Intel® Xeon® E3-1246
    16 lanes of PCI 3.0
    2 channels of DDR3

    vs

    5960x
    40 lanes of pci 3.0
    4 channels of DDR4

    Even with 32GB of ram on the Intel® Xeon® E3-1246 the memory bandwitdth is likely half or less of the 5960x/x99. The reduced amount of of PCI lanes on the E3 cpu is why you have to make sure the GPU has full bandwidth, and not sharing pci lanes with other motherboard devices. Latency on the system, like memory freq, x16 on the GPU, and no bad drivers causeing DPC issues is a must with reduced hardware.

    Keep in mind that Vegas operates in RGB space, so all footage is converted from .264 to uncompressed RGB, multicam and this is of course 4x the bandwidth needed. 4K uncompressed 8-bit RGB is somewhere around 750MB/s, times 4x streams, so your CPU/Mem/GPU need to be able to maintain those rates in a very solid fashion.

    “Winsat mem” from an admin command console will give you the memory bandwitdh that windows/vegas sees.

    My guess is your GPU driver set and Opencl may not be functioning correctly, and so not getting the benifits of the GPU for playback. Keep in mind that with the way OpenCL works, you are sharing the GPU with general display bandwidth capabilites. Here again the 5960x has more PCI lanes, and so a dual GPU with one for display and another dedicated for compute would be optimal.

    Also sounds like you may not be running the latest build of VP13-453.

    It is unlikely an i7-gen4 or less system hardware will ever push enough decompression of h.264 encoded content, then mux it, and stream the results to monitor in a smooth fashion. If you look under the hood, Vegas would like 8 threads for AVC decompression alone, and another 16 for rendering, vegas itself will run around 70+ threads. All that in contention with the other software you have running in the background on your system, and only 8 CPU threads to service them all in a time share fashion.

Page 27 of 83

We use anonymous cookies to give you the best experience we can.
Our Privacy policy | GDPR Policy