Rendering in VMS or DVDA

HaroldC wrote on 1/6/2007, 8:03 AM

I have just made a couple of dvds of the same video. The movie is about 1 hour 15 minutes long. First after capturing I edited and rendered in vms into avi. Then I used dvda to render into mpeg2, prepare and burn. The time it took to make the dvd in this manner was 3 hours and 35 minutes.

Being curious, I then rendered in vms the avi into mpeg2. Then I prepared and burned the dvd in dvda. The time it took to make the dvd in this manner was 2 hours and 25 minutes.

VMS has a much more efficient rendering engine. Which is what you would expect. The only time that you would want to render in dvda is when you have a video that has a length longer than 1 1/4 hour, give or take a couple of minutes. If it is longer you would need to use dvd shrink or a similar program to fit it to a dvd. Of course this would be another step and more time. But still considering the inefficiency of dvda rendering engine, it could still be quicker.

I assume that everyone else has pretty much the same experience. These times were just bare bones rendering. I did not add any effects or anything else to the video. I only cut out the extraneous parts.

Comments

IanG wrote on 1/6/2007, 8:40 AM

I've always maintained it's better to produce the MPEG2 from VMS, but I'm still surprised the difference was so large. Did you use a lot of effects which would have slowed the avi rendering down, or use settings that would have forced DVDAS to re-encode the MPEG2?

Ian G.

Paul Mead wrote on 1/6/2007, 9:01 AM

I've pondered this myself. I sure hope that the Madison folks use the same rendering engine in both components. I suspect it is more to do with how many times you shuffle / process the data.

If you simply have VMS produce AVI output then VMS can basically copy the incoming AVI clips to the output file. That in itself takes a fair amount of time. But, if you did any effects, crops, pans, media generator, etc., then those have to be processed first, adding lots of CPU overhead and elapsed time. Then DVDA has to take all that data and do the render, which means all the I/O overhead again and the CPU overhead to process each frame. I suspect if you have VMS produce the MPEG2 then you gain the efficiency of only having to process each frame and write it to disk once instead of multiple times like you do when you first generate AVI from VMS and then have DVDA render to MPEG2.

That's my guess at least.

HaroldC wrote on 1/6/2007, 12:32 PM

Those times are accurate, give or take a couple of minutes. There weren't any effects, crops, pans, etc done. They are just bare bone put in dvd and play. I haven't watched either movie yet to determine if there is a quality difference. My usual process is to render to mpeg2 in vms. It had always seemed the quicker way to produce a dvd, but I wasn't sure.

No, in creating the second dvd, dvda do not re-render the mpeg2. That would have taken about 5 hours.

braze wrote on 1/8/2007, 11:52 AM

HaroldC, I saw similar results to what you encountered. I posted them
back on november 16 - "MPEG2 Rendering questions"
http://www.sonymediasoftware.com/forums/ShowMessage.asp?Forum=12&MessageID=495421
In summary, VMS was considerably faster, but produced a smaller file.
The file size differences concerned me that there might be a huge quality
difference. But my source video was not of high quality in the first place,
so I couldnt tell the difference. Check out my post for the actual render
times and file sizes.