I'm delving into geometry compression for a university seminar. But I still have some problems understanding the variation a the Huffman coding used.
1. Huffman coding takes a set of symbols as input to build up the tree. What are these symbols in geometry compression? Are these just the deltas between successive vertices/colors/normals that are compared? If so, are these XY(Z) or RGB(A) deltas histogrammed independently?
2. If I understand it correctly, the final bit stream is a combination of fixed-length geometry instruction opcodes plus variable length Huffman codes "inside" the setVertex, setColor and setNormal instructions??? <- THE question!
3. What is the variation in contrast to classic Huffman made up of? The header forwarding? The fact, that the Huffman codes represent Deltas inside instructions and not the whole instruction?
I'm referring to this article, chapter B.9: http://java.sun.com/products/java-media/3D/1_2_api/j3dguide/AppendixCompress.doc.html
Any help on this is highly appreciated!
Thanks in advance! Karsten
=========================================================================== To unsubscribe, send email to [EMAIL PROTECTED] and include in the body of the message "signoff JAVA3D-INTEREST". For general help, send email to [EMAIL PROTECTED] and include in the body of the message "help".
