Help - Search - Members - Calendar
Full Version: Huffman Decoding
Hydrogenaudio Forums > Lossy Audio Compression > AAC > AAC - Tech
itsapoorva
In the FAAD code he has used two methods for Huffman decoding the spectral values (2-step and binary) are there any other methods better than these two ? If no are these the only type of implementation possible ?
NumLOCK
by Better.. do you mean, faster ?
itsapoorva
Can you kindly refer some faster algorithms as well as some algorithms which require lesser memory(size of the tables)
xquake
If using a 3-step method instead of binary for some codebook. It maybe faster than binary and the table size is smaller than 2-step method.
menno
I'm thinking about replacing the binary tables with 3 step tables, I'm not sure if it will be a lot faster and the rewriting of the tables takes a lot of time.

The least memory consuming method is probably the one used in the aac reference code, but it is quite slow.

Menno
siddharth vaghela
QUOTE(menno @ Apr 11 2003, 11:48 PM) *

I'm thinking about replacing the binary tables with 3 step tables, I'm not sure if it will be a lot faster and the rewriting of the tables takes a lot of time.

The least memory consuming method is probably the one used in the aac reference code, but it is quite slow.

Menno


unsure.gif

im still struggling to understand the 2-step algorithm implemented in faad... can u direct me to a reference which describes the algorithm and explains the concept?
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.
Invision Power Board © 2001-2008 Invision Power Services, Inc.