| George Wilson 2005-11-30, 7:55 am |
| Thanks for the replies.
My question was "How base values are generated for the code length
generated using Huffman trees and extra bits"
Let me explain to make my question clear.
The length/literal and distance codes from LZ 77 are mapped to [0-285]
and [0-29] respectively. Those frequencies are computed and a Huffman
tree is generated for both individually. The Huffman tree is optimized
to restrict the code length to be only 15 bits.
Now from this optimized codelength for each symbol "[1] how smallest
code for each code length is computed" and finally how next values for
the same code length and extra bits are added, any resource will be
appreciated, "RFC explains but still not clear".
Question the other way
======================
Say from the point of the puff.c, it expands the compressed code
lengths and generates two huffman table namely lencode and distcode,
it doesn't generate any prefix codes but "[2] how is it able to match
the bits from the bits loaded using bits(s,n) ?"
Clear explanation will be appreciated.
George Wilson
|