Transcode Time Update
 

Tim from Tech support has responded to my ticket and apparently there is an issue with the firewire port on the Asus P4c 800E motherboard I am using.

He is using the firewire port on his SB Audigy card without problems (he has the same board as me.)

I'll pick up a new firewire card and see if that improves, all the other possibilities he pointed out are covered, i.e. updates, drive controllers, etc.

I'll update as soon as I try out the new configuration.

Jay

OK I installed a new SB card tonight and captured two clips to one of my SCSI U160 drives.

The captured clips are 2 min 42 sec. HDLink transcoded the two clips in 7 min 7 sec.

Are these normal transcode times for a P4 3.2 w 2G RAM?

Additionally, my preview looks good in the preview window, but scrubbing the timeline does have dropped frames, but I am assuming that to be normal for HDV on this system.

Can I expect better performance from a SATA Raid?

And as far as video preview out, I'm leaning towards the Matrox Parhelia 8X AGP. Does anyone know if this configuration will preview out in After Effects 6.5? (I know off topic, but while I have your attention...) I am very dependent on AE and need to keep my preview out.

David, keep up the good work, as I look forward to the time when we have an affordable (Cineform) true RT HDV solution as we have in DV now.

I was seriously considering Canopus Edius NX, but really want to stay in the Adobe software, so I'm counting on you guys.

Jay

Jay,

Are you say the you converted 2 x 2min 42sec clip (totaling 5min 24) in 7min 7 sec. That sounds about correct. 1.3 : 1 is a pretty good conversion time for a P4. If you say the source clips are a total length of 2min 42sec (your post was ambigous) took 7m7s then something is still wrong.

When you say scrubbing has dropped frames -- what do you mean? No dropped frames are to be expected or acceptable upon capture. If any frames are missing, there is still a conflict in your system. The main design of the buffered capture conversion is to prevent this. Please discuss this further with Tim as it is great that you both have the same motherboard, so we know the ASUS P4c 800E works.

Have you tried the same test using HDLink, capture to M2T first, then converting in a second step. Are the conversion times and the output results the same? When a system is working correctly the results are the same, with no dropped frame and faster conversion -- clearly making the one pass capture with auto conversion the preferred mode.

Idea: Try capturing to another drive -- non SCSI U160 -- any single IDE or SATA will do fine. If your SCSI controller is a hog of PCI bus time you could have packet loss over firewire. We don't recommened you need SCSI or PCI based RAID controllers as they all can add unnecessary system loads. Motherboard based RAID controllers are typically very good if you would a faster disk system.

Thanks for your reply David, I'll try to answer your questions here from the office, my edit machine is in our home office.

Are you say the you converted 2 x 2min 42sec clip (totaling 5min 24) in 7min 7 sec. That sounds about correct. 1.3 : 1 is a pretty good conversion time for a P4. If you say the source clips are a total length of 2min 42sec (your post was ambigous) took 7m7s then something is still wrong.

HDLink saw the capture as two scenes. The scenes combined length was 2:42. I assume it saw them as two scenes because the during recording the captured footage was shot in two different sessions. So it sounds like I still have something amiss there.


When you say scrubbing has dropped frames -- what do you mean? No dropped frames are to be expected or acceptable upon capture. If any frames are missing, there is still a conflict in your system. The main design of the buffered capture conversion is to prevent this. Please discuss this further with Tim as it is great that you both have the same motherboard, so we know the ASUS P4c 800E works.

There are no dropped frames, just that the playback when scrubbing the timeline does not show all frames, or it is jerky, like a frame drop. (Isn't this part of the degraded preview, or is that avoided with Aspect HD?) Should I expect to see the same preview quality as I have with Premiere and DV preview?

Have you tried the same test using HDLink, capture to M2T first, then converting in a second step. Are the conversion times and the output results the same? When a system is working correctly the results are the same, with no dropped frame and faster conversion -- clearly making the one pass capture with auto conversion the preferred mode.

This capture test was done outside of Premiere with HDlink. The conversion was automatic after capture (incidently, I really like the log feature.)

Idea: Try capturing to another drive -- non SCSI U160 -- any single IDE or SATA will do fine. If your SCSI controller is a hog of PCI bus time you could have packet loss over firewire. We don't recommened you need SCSI or PCI based RAID controllers as they all can add unnecessary system loads. Motherboard based RAID controllers are typically very good if you would a faster disk system.

My first test from the first post was from an IDE. I thought the SCSI may give improved performance. I do have three SATA drives being shipped to me now that should be here this week, and at that point I can do a SATA test. It would be great if I could pull the SCSI controller out of the system and go with SATA only.

Thanks,

Jay

>>There are no dropped frames, just that the playback when scrubbing the timeline does not show all frames, or it is jerky, like a frame drop. (Isn't this part of the degraded preview, or is that avoided with Aspect HD?) Should I expect to see the same preview quality as I have with Premiere and DV preview?

Playback should be completely smooth just like DV, on your PC you should be able to constantly playback two streams at full frame rate, with a third steam for short intervals + titles. If you can't play a single stream you are either playing in the wrong edit mode (make sure you are using the Aspect HD presets) or something is still wrong in your setup. Sorry you are experiencing this -- this is plug-n-play for 99% of users.

David,

I'll check that out tonight.

And don't worry about me being one of the 1%. After all, it is not only the product that people look for, but the support when necessary, and your team has been outstanding so far.

Jay

Would the aspect capture settings affect transcode times? Are you using the recommended medium setting?

Capture settings only effect the timing small amount 10-15% -- so that is not the cause.