October 23, 2006
Behind the Scenes Fun
Personally, I do a lot of my learning through observation - seeing how other folks do something, figuring out why they did it that way, etc. So, for me, "Behind the Scenes" videos are really helpful. With that in mind, here are a couple I came upon this weekend.
2nd Unit TV: This group does on-set interviews and walkthroughs at major TV studios. The current episode is part two of a visit to The Office. Check out the "past episodes" as well. There's a bit of awkwardness present, but they're still pretty helpful.
Buzz Image: This is a production house which has done both major motion picture work as well as commercial work. Their making of section goes over some of their work in detail. Most of it is rotoscoping work, but that has been a subject of much discussion around the studios lately. The link goes to a piece of "Eternal Sunshine of the Spotless Mind"
October 17, 2006
One of the hardest thing to teach is video compression, because it's really difficult to visualize what's going on at a low level. One of the best tools for understanding what's going on inside a highly compressed video file is Textronix' MTS4EA software. It gives you a full visual analysis with motion vectors and macroblock coefficients and everything else a math geek could want. Unfortunately, even with the generous educational discount, it's still a bit out of most students' price range. About $20,000 out of their price range. Bummer.
However ... I've just become aware of a little application from Apple called MovieVideoChart. This program (described in depth in WWDC05 Session 208 for you ADC members) gives you a visual representation of each frame of your compressed video, showing you which frames are keyframes, which are intermediate, and how they're reordered in your video.
In the above screenshot, you can see a few things. The red frame (marked with the word "sync") is an I-Frame. It's the start of a GOP. In this case, this video has 150 frames between each keyframe, which is pretty extreme, but generally OK with modern codecs. Frames tagged "droppable" are are B frames (B frames can generally be ignored without breaking other frames, so they're the first to get dropped on a slow system). Everything else is a P frame.
The bottom row is showing you the order in which the frames are decoded. The middle row shows you order in which the frames are displayed. The top row would should you the effects of any edits made to the file itself.
This can teach you a lot about modern (h.264 in this case) video compression. Frame reordering is an important concept. If you want to understand why B frames are so important, just look at the "data size" entry for the B (droppable) frames. Most frames are running between 2000 and 4000 bytes (P and I frames respectively) but the B frame is between 130 and 700 bytes! That's a pretty huge reduction, and that's just one of the many things you can learn from this free program.
A full explanation of I P and B frames is out of the scope of this post (though I'd be happy to geek about it someday), but Apple has an OK blurb up if you're curious. Otherwise, take a look at MovieVideoChart. You might also be interested in Dumpster (at the bottom of the page) which shows you some good information about the internal constructs of your video file.
October 9, 2006
Production Assistant's Handbook
(leeching from HDForIndies)
There's a free PDF book called the Production Assistant's Handbook, available for download from NoEnd Press.
I haven't read through the whole thing, but I think there are some good tips in there. I think this might be a useful document for folks who are looking at moving from the sheltered world of the University into the scary, cold, Real World™.
The folding instructions are far too complicated for me though. Fold in half? Hu? I just keep ending up with a swan...
February 13, 2006
Excellent Student Work
January 5, 2006
Tech Vs Art - which wins?
Eventually I'm going to post some notes with my thoughts on how to best distill down the video recording and compression process. This post is a bit more general. I've been going back and forth about how much a person learning production actually needs to know about the nuts and bolts of video technology. Is it important that they understand what happens once the light hits the lens, or should they just understand that their video comes out the other end?
Along the same lines, I wonder if the types of people interested in the nitty-gritty are likely to be totally different from the types of people interested in production. If someone wants to be a freelance videographer, how much do they really need to know about their camera?
I like using the analogy (stolen from a friend) of teaching a painter to paint. You wouldn't hand someone a paintbrush and just say "go, do!" without explaining some basics of paint mixing and technique. Similarly, I personally feel it's important to understand how the light hitting the lens gets sampled, processed and laid down to tape. I think it helps yield better work.
If you agree that it's important to understand the medium, then you need to figure out how far to go. You can choose to stay fairly abstract - sort of like "assume the horse is a point" in physics. Or you can go ridiculously in-depth, with something like Tektronix' MTS4EA software suite. Is that just overkill?
As I mentioned at the top of this post, I will be adding some more thoughts about ways of distilling some of the information down. For now, I'd like to hear the opinions of others regarding this balance of teaching tech or teaching art.