Announcement

Collapse
No announcement yet.

ADVC 300 Trouble Displaying Desktop Video in PPro CS3

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • ADVC 300 Trouble Displaying Desktop Video in PPro CS3

    Windows XP SP2, Adobe Premiere Pro CS3, Intel Quad Core @ 3.2Ghz, 4GB memory, video assets on G-RAID2 and previews on separate SATAII drive.

    I have a JVC TMH-150CGU monitor connected to the S-VHS out on my ADVC-300. I am trying to use it as an external device to view video from the Premiere timeline. I have Desktop video set to playback through the external DV device.

    Whenever I hit PLAY, the timeline does not play the video smoothly. I jerks and freezes, jerks and freezes. The audio may play fine without a hitch. If I disable playback through the external device, the timeline plays fine without issues.

    The DV controls are set to Generic NTSC, because I can't see a setting for the Canopus ADVC 300. I tried the Canopus ADVC 110 setting that was there, but same issue.

    Any ideas on what is happening? Is there some preset file I am missing? The machine is fast, and as I said I don't have any issues unless I try displaying through the ADVC-300.

  • #2
    What video format are you editing? Is it standard definition DV, or something else? The ADVC will only accept DV, and if you are editing in a different format, Premiere will have to encode it on-the-fly, which may be too much to do in realtime.

    Comment


    • #3
      Thanks, it's SD DV. The machine is a beast, and has no problem rendering on-the-fly. Definitely a hitch with the Canopus device.

      As it turns out, this looks to be a Firewire issue. Even though I have the ADVC connected to an add-in Firewire card, it appears with another device (M-Audio Firewire 410) connected to the motherboard firewire connector Premiere is having issues. On a hunch I pulled out the 410, and I am seeing no issues with the video playback through the Canopus.

      Adobe has said PPro has issues with multiple Firewire devices, but I had assumed that meant multiple DV devices. Either that, or Windows has the issue. I'm going to play around with which device is attached to which port, and see where things go from there. It'll be trial and error, and worst-to-worst, I pack up the 410 and use the on-board sound.

      Comment


      • #4
        Get another FireWire card.

        Two things go against you:
        1) DV requires a constant, low-latency data flow
        2) FireWire will run as fast as the slowest device connected to the bus

        Imagine you're drinking from a tube that's giving you a slow steady flow of liquid. No problem, you keep swallowing, breathing quickly in between. Now imagine the flow stops for a while - cool you get the breathe easy for a while. Then suddenly a bubble of liquid comes down the tube and dumps into your mouth - if it's too much, you'll choke or might even drown.

        Same thing with DV - it has to come at a steady rate or you lose frames, because time keeps marching onward.

        So what's probably happening is that the audio traffic is bursting and interrupting the DV traffic, causing hiccups in the DV stream.

        This is also why using a FireWire hard drive on the same bus as a DV device can cause problems capturing/outputting DV, as the hard drive traffic will interrupt the DV stream.

        Comment


        • #5
          Originally posted by GrassValley_BH View Post
          Get another FireWire card.

          Two things go against you:
          1) DV requires a constant, low-latency data flow
          2) FireWire will run as fast as the slowest device connected to the bus

          Imagine you're drinking from a tube that's giving you a slow steady flow of liquid. No problem, you keep swallowing, breathing quickly in between. Now imagine the flow stops for a while - cool you get the breathe easy for a while. Then suddenly a bubble of liquid comes down the tube and dumps into your mouth - if it's too much, you'll choke or might even drown.

          Same thing with DV - it has to come at a steady rate or you lose frames, because time keeps marching onward.

          So what's probably happening is that the audio traffic is bursting and interrupting the DV traffic, causing hiccups in the DV stream.

          This is also why using a FireWire hard drive on the same bus as a DV device can cause problems capturing/outputting DV, as the hard drive traffic will interrupt the DV stream.
          When you say get another Firewire card, do you mean a different brand, or physical card? Devices are connected as follows:

          1) M-Audio Firewire 410 connected to 6-pin Firewire 400 port on motherboard (only FW device attached to motherboard)
          2) Canopus ADVC-300 connected to 6-pin Firewire 400 port on separate PCI FW400 card installed in PC (only device attached to card)
          3) G-Tech GRAID-2 1.5TB connected to FW800 port on yet another, separate PCIe card installed in PC (only device attached to card)

          So each device is connected, physically to a different card (motherboard serving as one "card".

          Comment


          • #6
            Oh... You do have all yoru devices connected to separate FireWire buses. Interesting...

            Try swapping the M-Audio and ADVC300's FireWire connection (ADVC300 to motherboard FireWire) - that might work if it has something to do with bus bandwidth to/from the PCI FW400 card.

            Comment


            • #7
              Yeah, purposely bought all the dang cards because I thought that would be the best way to keep everything happy. Seems to have worked in terms of FW400 & FW800, but not for the FW400 connections on different cards. It seems like Windows XP may treat this all like one FW400 bus. I had swapped around the connections yesterday (moved the 410 to the PCI card and the ADVC to the mobo) but had the same issue, and then even some obvious sound issues. So there is definitely a conflict on the bus. Could be a problem with Windows, the M-Audio device or drivers, or Premiere.

              I'm going to experiment with a different brand Firewire PCI card (right now the PCI card is SIIG, and I happen to have an Adaptec lying around) to see if that changes anything, but I suspect that this is either a Windows Firewire support issue, or a bad driver implementation with M-Audio. I'll keep you posted.

              Short term solution is to use the on-board sound, which isn't bad, but is forcing me to find a MUCH longer digital audio cable. Long term will be to get a USB 2.0 external sound solution instead of the Firewire.

              Comment


              • #8
                Interesting. Make sure you installed the FireWire800-related hotfix (sorry, don't remember the KB # offhand). It's the one that made everything go slow on FireWire800.

                Also make sure 1394 Network adapters are disabled.

                I have an old machine that has 4 FireWire400 cards in it (all PCI) and they run pretty well, though I had to tweak the PCI latency to get them concurrently transferring (I have a PC Geiger for measuring PCI bus bandwidth - made it easier to see what was going on) because it's only a 1GHz Pentium III.

                Back on topic though - maybe connect the ADVC and the M-Audio device on the same bus and see if performance is the same, better, or worse. It's worth a try at least.

                Comment


                • #9
                  Interesting. Didn't know about the hotfix, and never even crossed my mind to disable to network instances. I'll do that as soon as I get back home. Is this an XP hotfix directly related to FW800?

                  Really, though, this seems like a device conflict rather than a performance thing. I'm just using a small sample DV clip on the timeline, no effects, and the timeline is rendered. I think that 410 is getting in the way of the Canopus or vice versa.

                  I'll play around with it tonight and let you know what happens.

                  Comment


                  • #10
                    Here's the hotifx...
                    http://support.microsoft.com/kb/885222

                    Comment


                    • #11
                      Thanks. I checked my xfer speeds with EZDVtest and it looks like I'm getting at least 75mbps on read with somewhat less for write, which is comparable to the speed on the on-board SATAII drives I have, so I think the FW800 is operating at appropriate speed. I did want to run some of the more thorough tests I have, as they are destructive and I'd need to move some data around. This seemed like a good troubleshooting benchmark just to see if there was any evidence of a problem. I'm going to play around tomorrow with swapping in the Adaptec card and trying out some other ideas. Thanks again for all the help.

                      Comment


                      • #12
                        Roadkil's Disk Speed test is a much better benchmark than EZDVtest.

                        Comment


                        • #13
                          Well, long road but after spending about 8 hours of trial and error this weekend, I managed to solve my problem. Hopefully this will stay long term. Much resource shifting, slot swapping and disabling of devices allowed me to finally isolate the Firewire port for the M-Audio 410 to an IRQ not shared by anything else at all. This appears to have solved the solution. The driver for this device is not very sophisticated and can't handle the way Windows XP shares resources above IRQ 15. I was just about to junk it, and still may for a MOTU or Edirol solution.

                          Anyway, thanks again for all the help. Now I can finally concentrate on capturing and editing.

                          Comment


                          • #14
                            About that Firewire quickfix...

                            I have XP SP2 and added the FW800 card after SP2, so I assume that this is required to make the MS driver perform well at FW800. However, I have a switching utility that replaces the MS driver for the 800 card with a unibrain driver that is supposed to perform better than the MS one, so I assume that the fix is therefore moot. Do you have any specific knowledge regarding this?

                            Comment


                            • #15
                              If you're using the Unibrain driver, then I don't think the hotfix applies.

                              Comment

                              Working...
                              X