I have two laptops. One is a 2.5GHz dual-core w/ 4G ram. It's running Ubuntu 9.10 and only auto-installs Pitivi 0.13.3. The other laptop is a single-core 1.6GHz w/ 1.5G ram. It's running Ubuntu 10.4 and auto-installs the newer version of Pitivi (not handy or I'd tell you the exact version number).
Beware that the latest version is 0.13.5, and older versions are unsupported as we the project doesn't have the manpower to maintain multiple simultaneous versions.
OK. I used mencoder and a bunch of hand-waving to get that video out. Since then I added the ppa and upgraded to PiTiVi v0.13.5
First, it's a lot faster. I turned off rendering in the timeline, zipping around in the project with my 1.6GHz single-cpu was a breeze, exactly the way I'd expect it to be.
Even the rendering was faster. What used to take 24+hr was down to 1.5hr. Amazing.
There are still a couple of issues. Background: I'm combining several videos into one. For 3 of the videos I had someone do a voice-over, and I'm going to record the video in sync with the voice record. In between those 3 voiced-over segments there are two videos that are already shot that I'm inserting.
1. (preview cuts off mp3 tracks): When playing the voice-over (mp3) segment in the "preview" to line up the timing of the end of the voice-over with the next video, the voice-over gets cut off at the end, when I play the mp3 on any other play it's end is "normal". This concerned me and I figured we may be back to the same kind of problems I discussed in my last email, so I didn't produce the voice-over videos, I just wanted to test how the rendering would come out. I rendered the project with one white background image and just the video from the two segments that already had video. Here's a screenshot of the timeline: http://www.csrtechnologies.com/Screenshot.png
2. (rendering video speed varies): Well, everything "works" with the exception of the two videos that were inserted. When they go to play (in the final rendered video), their audio is the right speed but the video playback is too fast, it actually ends and sits on the white background for a little while before the next mp3-only segment picks up. I just noticed that I accidentally didn't use the original video when inserting those two video segments, so I'll be trying it with the original and see if it makes a difference. One thing to note, the final is 25fps and the two inserted videos were 15fps. Both are flv, my thought here is maybe the "render" code just copied the frames frame-for-frame out of the inserted videos into the final and naturally ran out of frames because of the frame rate difference.
I'll try it with the original video (ogv, also at 15fps) and let you know how that one turns out.
Bottom line. If I can't get this to work with compiling different sources I'll have to do the same as I did last time, video the entire thing, inserts et al, all as one video and then render it by hand w/ mencoder. It's a pain, but I'm not seeing another way around it.