No announcement yet.

How many video tracks can be output in RT before rendering?

  • Filter
  • Time
  • Show
Clear All
new posts

  • How many video tracks can be output in RT before rendering?

    I'd like to know from you editors a simple benchmark in Edius.

    How many stacked video tracks with 3D PiP on each are you able to play in RT by pressing Space-Bar from Timeline and without losing any frame?

    I tried this test on my old P4 2,8 ghz and I can reach 3 video tracks with full RT and smooth playback and no frames lost.
    From the fourth one my system starts losing frames and image preview quality is decreased.

    Is this kind of performance due to IDE or SATA drives or to the amount and kind of RAM or to CPU speed or to VGA card and amout of video ram? Or a combination of these?

    Also, I'd like to know if a Dual or Quad core CPU can help to increase the limit.

    If you don't feel bored, I'd like to know your feedback.
    Thanks mates.

  • #2
    I have achieved nine in SD, but that really was the limit. System spec as per my signature.
    Andrew Pinder
    Edius 9 with Blackmagic Intensity Pro 4k; Windows 10 (64 bit Pro); Gigabyte Z390 Aorus Pro; i9-9900K CPU; 32GB RAM;
    Asus GTX1060 graphics; RME Fireface800 audio; SATA RAID


    • #3
      3 streams sounds about right for a lowly P4 2.8

      A new CPU and more RAM will definitely make a difference - but it'd probably also mean a new motherboard, power supply, etc.

      Virtually a new system.


      • #4
        I think I got 8 layers with my 2.13 GHZ Core 2 Dual. I think I could have gotten even more layers with a RAID system. I say this because I could add filters to my 8 layers. I think a SATA drive might max out at 8 layers of DV-25. I was using Edius 3.5. I know Edius 4.6 has a bigger buffer so I can only imagine what Core 2 Dual at 3.2 GHZ could do. You could try a Quad Core at 2.4 GHZ. With my 2.8 GHZ Celeron I think I go about the same results as you did.


        • #5
          EDIUS real-time performance depends on a number of factors...
          1. Data rate (file size) of the source material
            If the disks can't read the data fast enough, no real-time.
          2. Decompression complexity of the source material (compression type - DV, Canopus HQ, etc)
            If the CPU(s) can't decode the frames fast enough, no real-time.
          3. Disk subsystem speed
            The faster you can read the data, the better
          4. CPU speed
            The faster you can decode the data, the better
          5. Complexity of additional processing - effects, transitions, framerate conversions, aspect ratio conversions, up/down scaling
          So... in the end it's usually a balance between disk speed and CPU speed.
          More-compressed formats like AVCHD and HDV put less strain on the disk subsystem, but put more strain on the CPU(s).

          Likewise, less-compressed formats like Canopus HQ put more strain on the disk subsystem, but put less strain on the CPU(s).


          • #6
            I managed 7 in SD without pre-buffering, just pressing the space bar.
            No raid, just a normal SATA 7200 HDD

            My system spec's below. . .