Video Compression

Question

This is what I want to do:
Read an AVI file, process each frame in it (i.e draw something onto them), save the processed frames/buffers to file.
My application implements the Effect interface and inserts this on the Processor's codec chain.
I've accomplished all of the tasks above, but I've stumbled upon some unacceptable shite..
1) The resulting file is HUGE. It seems like the frames are saved as RAW data. I.e: input: 1,5 MB AVI sequence (480x360, 10 seconds long, cvid (cinepak) encoded) results in a .mov file of 105 MB (!) (480x360, cvid encoded??)
2) I've not found any way to alter the output format of the video sequence read. I always get some exception when trying to specify the format manually. I've tried setting the format both at the processors track control-instances (and at the processed buffers setFormat()-methods.)
Do I really have to implement the compression and resizing of the output video myself? I got the impression that this can be defined by the setContentType() method of the processor. I.e. if I do the following:
processor.setContentDescriptor(new ContentDescriptor(FileTypeDescriptor.QUICKTIME));
...and then later when I've gotten hold of the video TrackControl object (vtc) I could do a:
vtc.setFormat(new VideoFile(videoFile.CINEPAK));
...to ensure that I use CINEPAK to encode the resulting buffers (that is later passed to a DataSink). The way I use it now, it seems like cvid (cinepak) is used, but the resulting filesize suggests that no compression applied.
I'd really appreciate any input and ideas on my case. I'm going crazy here!! ;-)
-- Regards Jennifer

Answer 1

Thanks alot for the suggestion. I've tried this out, but I'm still having problems.
I'm now using two processors (drawP and compP). The first (drawP) set with content descriptor RAW. This processor's got the effect/codec attached to it. The second processor (compP) gets its datasource from the first (drawP). Its content descriptor is set to QUICKTIME and a datasink is attached to it. This works fine, but the output still seems to be uncompressed (RAW?).
Does anyone have any suggestions on any other way to define the output format. The way you say "should be passed to another processor that defines that it accepts RAW input and produces compressed data"... Does this imply setting the first content descriptor to RAW and the second to i.e QUICKTIME, or is there a better way? Or maybe I've misunderstood..?
Appreciate views, ideas, comments, anything ;-) Thanks for your time!
-- Regards Jennifer