Methods for speech transmission
First Claim
1. A method of encoding speech, the method comprising the steps of:
- breaking the speech into a plurality of time segments each representing a time segment of the speech;
estimating a set of model parameters for each of the time segments;
quantizing the set of model parameters to produce quantized model parameters;
encoding some of the quantized model parameters using error correction codes, the encoding comprising the steps ofusing first and second types of error correction codes, the first type possessing a higher level of redundancy than the second type;
encoding a first group and a second group of quantized model parameters, the first group of quantized model parameters being more sensitive to bit errors than the second group of quantized model parameters;
using the first type of error correction code to encode the first group of quantized model parameters; and
using the second type of error correction code to encode the second group of quantized model parameters; and
interleaving the encoded first group of quantized model parameters with the encoded second group of quantized model parameters and a third group of quantized model parameters, the third group of quantized model parameters being less sensitive to bit errors than the second group of quantized model parameters.
0 Assignments
0 Petitions
Accused Products
Abstract
The quantized parameter bits are grouped into several categories according to their sensitivity to bit errors. More effective error correction codes are used to encode the most sensitive parameter bits, while less effective error correction codes are used to encode the less sensitive parameter bits. This method improves the efficiency of the error correction and improves the performance if the total bit rate is limited. The perceived quality of coded speech is improved. A smoothed spectral envelope is created in the frequency domain. The ratio between the actual spectral envelope and the smoothed spectral envelope is used to enhance the spectral envelope. This reduces distortion which is contained in the spectral envelope.
278 Citations
4 Claims
-
1. A method of encoding speech, the method comprising the steps of:
-
breaking the speech into a plurality of time segments each representing a time segment of the speech; estimating a set of model parameters for each of the time segments; quantizing the set of model parameters to produce quantized model parameters; encoding some of the quantized model parameters using error correction codes, the encoding comprising the steps of using first and second types of error correction codes, the first type possessing a higher level of redundancy than the second type; encoding a first group and a second group of quantized model parameters, the first group of quantized model parameters being more sensitive to bit errors than the second group of quantized model parameters; using the first type of error correction code to encode the first group of quantized model parameters; and using the second type of error correction code to encode the second group of quantized model parameters; and interleaving the encoded first group of quantized model parameters with the encoded second group of quantized model parameters and a third group of quantized model parameters, the third group of quantized model parameters being less sensitive to bit errors than the second group of quantized model parameters. - View Dependent Claims (2, 3, 4)
-
Specification