Title :
Partial-Order Bit-Allocation Schemes for Lowrate Quantization
Author :
Ramprashad, Sean A.
Author_Institution :
DoCoMo Commun. Labs., San Jose, CA
fDate :
Oct. 29 2006-Nov. 1 2006
Abstract :
We consider quantizers that are structured using permuted patterns of bit-allocations. The designs assign bits to subvectors xk within a target vector x through a partial ordering of one or more "base" bit-allocations. The approach allows for a tradeoff in the bits spent explicitly on order information (position of each quantized subvector x circk in a vector) and bits spent on traditional quantization within each subvector (which represents a mix of order and value information). Prior work looked at designs for i.i.d. random variables and quantizers for speech/audio coders considering a limited range of ordering/permutation strategies. This paper investigates in more detail the balance of order and value information and looks at ways to improve performance by considering a wider range of strategies and design methods.
Keywords :
audio coding; speech coding; vector quantisation; audio coders; i.i.d. random variables; partial-order bit-allocation schemes; quantization; speech coders; subvector quantizers; Audio coding; Bit rate; Design methodology; Design optimization; Entropy; Image coding; Product codes; Quantization; Random variables; Speech coding;
Conference_Titel :
Signals, Systems and Computers, 2006. ACSSC '06. Fortieth Asilomar Conference on
Conference_Location :
Pacific Grove, CA
Print_ISBN :
1-4244-0784-2
Electronic_ISBN :
1058-6393
DOI :
10.1109/ACSSC.2006.354915