Chroma Keying - I need advice!

B

BNSFILMS

Member
Hi guys,

I'm going to be exporting keyed out footage of a family to be worked on in Flash. The idea is you can click on one of the family members to prompt a reaction from them.

What is the best way to keep the file size low for each keyed out family member? If there is no background does that mean there is no information in all the empty space? What I'm asking incase I'm not clear is if I keyed out 50% of the image, does that mean I'm reducing the file size by 50% as well? Or does that transparent background take up space?
 
will the file be smaller - yes. this of course depends on the codec used but in general, the less info, the smaller the file. in a typical jpeg/mpeg kind of setup, areas of repeating color (or no color) tend to get compressed very well. you probably won't see a 50% reduction but it all depends on the initial content
 
If you crop down to the size of the person, yes then your file size will be smaller. Imagine a box around the area where each member moves. If you have the whole frame - then depending on the codec, you may not reduce the file size at all. The "invisible" area still needs bits so that the computer knows it's invisible. Most codecs don't say, Hey at horizontal pixel 89 make it visible. They tend to work Pixel 1 - invis, Pixel 2 invis, Pixel 3 - vis
 
"less info, smaller the file" - I believe in most cases the file will be bigger - it's not like there's no pixels in the transparent parts - there's an additional channel (alpha) that is your mask.

The cool thing about putting video into flash is that you can crop it to whatever size & aspect ratio you need. Doing that in Squeeze (when you convert to FLV) is a piece of cake.
 
You must log in or register to reply here.
Back
Top