Best source format/workflow for SD DVD?

[GregGerlach]

I have a Sony Z5 and was wondering which source format & workflow I should use to edit a video that will end up on an SD widescreen DVD. I can shoot either HDV 1080i or widescreen DV. Which workflow would give the best quality?

1. Shoot in HDV 1080i, edit in an HD timeline then downconvert it to SD using SE. Import AVI or wrapper into TMPGEnc to encode Mpg2. Or use TMPGEnc to downcovert and encode Mpg2?

2. Shoot in HDV 1080i, edit in a widescreen SD timeline. Import AVI or wrapper into TMPGEnc to encode Mpg2.

3. Shoot in widescreen DV, edit in a widescreen SD timeline. Import AVI or wrapper into TMPGEnc to encode Mpg2.

[KiloWatkins]

3. Shoot in widescreen DV, edit in a widescreen SD timeline. Import AVI or wrapper into TMPGEnc to encode Mpg2.

Save the project. Load the project.VTP into TMPG and set the 16x9 flag. No rendering.

TIA...EWK

[KiloWatkins]

fwiw, we use to have to set low bit rate to 2k, for older DVD players. I don't think that is true any more, maybe someone will comment. low=0 mid=5k high=8k but don't forget to edit audio to 192kbps in TMPG. It has been posted, those whom have the VFP installed from TMPG2.5, can still use import.VTP in other versions of TMPG4.0.

If you have a noisy section of video you want to use, send just that section to TMPG and run the noise filter. Expect 1min. of video to take an hour...guesstement...but you will love the results to put that back on the timeline. Mark Goodwrench put a revised/updated TMPG..EWK workflow in the yahoogroups VTNT files section.

[Dufusyte]

Try to keep the data as rich as possible as long as possible in your workflow pipeline. Down convert it at the very latest stage.

1. Shoot in HDV 1080i, edit in an HD timeline then use TMPGEnc to downcovert and encode Mpg2

The reason for keeping the data as rich as possible is so it will minimize math errors with any effects and adjustments you might be applying. For example, it will minimize color banding, and minimize jaggies, etc. And then, when you do your final down-conversion at the end, it will hide any imperfections.

The only reason to avoid a rich-data pipeline is if your hardware cannot keep up, e.g., things run too slow, or you run out of storage space.

In an ideal world, the best workflow would be to first upconvert the footage, making it, say, ten times larger resolution, and as deep a bit depth as possible. Then apply whatever effects/adjustments/composites (which are really all mathematical calculations), and then in the last stage down convert everything. By working with larger numbers, the rounding errors are minimized.

[GregGerlach]

But what if your project includes clips that were shot in SD? Wouldn't editing in SD be better in that situation?

[KiloWatkins]

But what if your project includes clips that were shot in SD? Wouldn't editing in SD be better in that situation?

EWK used a standard with his HD gear, if the project is to be SD, shoot in SD. He used Vegas alot, so he must have seen something to make him feel this way.

But VTSE or SE both conform footage to mix and match codec/format/resolutions on a timeline. That is why I've asked where folks are pointing their Cache Prefs. You can find some video, will not respond to a Control Tree aspect selection, so Z positioning is needed.

Color Correction is a perk of Cache, afaik.