APPARATUS FOR QUANTIZING SPEECH RECOGNITION PARAMETER VECTOR

摘要

The present invention relates to a device, it is used to quantify speech recognition parameter vector in a distributed speech recognition system. Pass through speech recognition parameter more specifically to vector quantization, its one of two various ways and predictive vector digital quantizer selection quantized result to non-memory vector quantization, it is if necessary speech recognition parameter vector to be divided into two or more subvectors by executing quantization, quantify speech recognition parameter vector less bits, it is similar or related with quantization device, a distributed speech recognition system of better performance can be implemented.