Using HD in an SD timeline

[jackfalbey]

OK, I finally got a chance to put this one to the test. Thanks to those of you who helped me in my other thread viewtopic.php?t=1878. I captured live from the HV20 into Adobe OnLocationon my m6300 laptop, generating 2 1-hour HDV/MPEG files on my eSATA external HDD. I imported these into a standard NTSC 4:3 timeline in PPro CS3, "Scale to Frame Size" off, didn't mess with Field Dominance. I did my pans & zooms, then exported as a DV-AVI. I imported that AVI into a new NTSC 4:3 project and all is well.

It's important to do your pans & zooms first and then export to AVI before adding any other effects, or the CPU really gets bogged down; the first time around I color-corrected the HDV footage first, rendered it, then tried to add the motion... forget about it! If you plan to do this, do the motion first, then reduce the data stream by exporting as AVI and color-correct it afterward.

It took almost 4 hours to render and export 1 hour of footage on my m6300 (see specs below) with both cores maxed at 100% the whole time... for 2 hours of concert footage with pan/zoom and color-correction I lost 16 hours of work-time while the computer chugged along. Do any of you guys with Quad-core rigs get much faster performance? I'm wondering if investing in a Quad-core desktop right now would be worth it for the increased speed and time saved?

BTW, the reason I didn't break it up into smaller segments first is that I still have to sync it up with 2 1-hour tapes of SD footage from my PD170 for multi-cam editing in PPro CS3. I figured it would be much easier to resize, color-correct, and synchronize 2 large clips rather than dozens of smaller ones...

[Chuck Engels]

I figured it would be much easier to resize, color-correct, and synchronize 2 large clips rather than dozens of smaller ones...

Excellent idea Jack, you saved yourself tons of time doing it that way. You can split it up after you get the others synced.

[jackfalbey]

That was my thinking, Chuck. The process seems likely to work best for me as follows:

1) Import/capture HDV in a NTSC project
2) Pan/zoom in HDV footage; export as AVI
3) Open a new NTSC project & import AVIs
4) Color-correct to match the PD170
5) Sync up the tracks for the whole concert
6) Split it up by artist/song into segments
7) Go to multi-cam mode for each segment
8) Add transitions/filters/graphics
9) Render each segment by Work Area Bar
10) Add titles/transitions between segments
11) Export entire timeline to Encore

Step 3 is necessary because the original project gets buggy after rendering (possibly because of the MPEGs).
I'm only at step 4, so I'll keep posting what I find as I go along.

[jeffterm]

Hey Jack,
I just finished doing this a month ago, but my program was only an hour and a half stage performance. I don’t remember the extent of my postings here of how I managed but I’ll sum it up again:
My master cam recorded HDV, captured directly into Premiere Elements loaded on my laptop directly to a USB external hard drive. (Yes, USB, I was BIG TIME paranoid about this, however after 3 days of rehearsals of capturing with out missing a frame, I ran with it.)
I had a Panasonic HD record the whole performance non-stop. (AVCHD converted to HDV — but this is another story!)
A third camera, SD — close to the stage, getting close-ups, reactions and the like — the operator kept turning on and off based upon scene changes. (She was paranoid about running out of tape — I had to sync up each clip.)
Project created in PPro CS2 SD widescreen format.
Import material without adjustments.
Laid the master track to determine all timing issues, synced other track to it.
Watched through master track as a camera operator — adding pans and zooms, etc. never exceeding 100% or less than 45%.
**on a sidebar — I fell to old habits for doing my multi cam switching at this point, rather than using Premiere’s I manually shrank my footage to watch all 3 together and choose what to cut and where. Some of this mutli-cam decisions helped to determine some of my pan & zooms so that the material worked well together and I was able to cut out a few camera walk thru’s.
I did a (relatively) quick run thru making quick cuts, then went back to fine tune the cuts, at which time I adjusted my work area around the scene, prerendered often to see the changes and scroll thru.
Not until I was happy with the cuts and trannys did I attempt to do color correction. Once I found a good balance for color, I copy pasted the settings across all of the clips from that camera. (In hind site, this may have only worked because of the limited nature of the camera footage.)
I am running a Q6600 at stock speeds, and did not notice any real annoying delays while working, nor did I see my 4 cores get maxed out in the task manager. But I did not do a complete render out to DVD with color corrections until the end. Even then, I recall prerendering most of the color corrections as I added them. Yes, I admit that does add a lot of time to the process, but only once do I remember leaving to get coffee.
In the end I outputted the whole thing in AVI to bring into Encore for DVD creation.
I don’t recall Encore taking longer than 2 hours to completely render out the whole thing when finished.

I will try your idea next time, It had not occurred to me to do it that way because of the way the project and footage otherwise came together. I was also distracted by the PITA method of converting the AVCHD footage to HDV (but again, don’t get me started on that tangent ) that I was just glad to be making progress on the project. The rough cuts also allowed me to run back and forth with the director to get her input and approval on sequences.

Jeff