Bit error concealment methods for speech coding
View Patent ↗A method of concealing bit errors in a signal is provided. The method includes decoding an encoded signal parameter based upon constraints placed on a signal parameter, comparing the decoded signal parameter against the constraints, and declaring the decoded signal parameter invalid when the constraints are violated.
1. A method of concealing bit errors in a speech signal within a signal parameter quantizer including a processor, the method comprising:
encoding a speech signal parameter based upon constraints placed on the signal parameter quantizer, the constraints being based upon naturally prohibitive speech parameter combinations in one or more speech parameter samples, the combinations including different types of parameters;
comparing the encoded speech signal parameter against the constraints, wherein the comparison includes comparing a logarithmic gain value of the speech signal with a threshold function determined based on a difference between a plurality of logarithmic gain values of the speech signal; and
modifying operation of the signal parameter quantizer when the constraints are violated.
2. The method of claim 1 , wherein the modifying includes declaring the encoded speech signal parameter invalid; and
wherein the method further comprises substituting the invalid speech signal parameter with a replacement value obtained from other signal parameter samples.
3. The method of claim 1 , wherein the comparing occurs within a decoder.
4. The method of claim 3 , wherein the different types of parameters include at least one from the group including (i) log gain, (ii) first normalized autocorrelation coefficient, and (iii) zero crossing rate.
5. The method of claim 4 , wherein the different types of parameters are conditioned on a previously decoded speech waveform.
6. The method of claim 5 , wherein the previously decoded speech waveform is a last sub-frame.
7. The method of claim 1 , wherein the logarithmic gain value of the speech signal includes a logarithmic gain value of a current sub-frame of the speech signal.
8. The method of claim 7 , wherein the threshold function is based on: (1) a logarithmic gain value of a proceeding sub-frame adjacent to the current sub-frame and (2) a difference between the logarithmic gain value of the proceeding sub-frame and a logarithmic gain value of a next proceeding sub-frame adjacent to the proceeding sub-frame.
9. The method of claim 1 , wherein the modifying is based on a quantization codebook search associated with the comparison of the logarithmic gain value of the speech signal and the threshold function.
10. An apparatus configured to conceal bit errors in a signal, comprising:
means for encoding a speech signal parameter based upon constraints placed on a signal parameter quantizer, the constraints being based upon naturally prohibitive speech parameter combinations in one or more speech parameter samples, the combinations including different types of parameters;
means for comparing the encoded speech signal parameter and the constraints, wherein the comparing includes comparing a logarithmic gain value of the speech signal with a threshold function determined based on a difference between a plurality of logarithmic gain value of the speech signal; and
means for modifying operation of the signal parameter quantizer when the constraints are violated.
11. The apparatus of claim 10 ,
wherein the modifying includes declaring the encoded speech signal parameter invalid; and
wherein the apparatus further comprises means for substituting the invalid speech signal parameter with a replacement value obtained from other signal parameter samples.
12. The apparatus of claim 10 , wherein the means for comparing includes a decoder.
13. The apparatus of claim 10 , wherein the different types of parameters include at least one from the group including (i) log gain, (ii) first normalized autocorrelation coefficient, and (iii) zero crossing rate.
14. The apparatus of claim 13 , wherein the different types of parameters are conditioned on a previously decoded speech waveform.
15. The apparatus of claim 14 , wherein the previously decoded speech waveform is a last sub-frame.
16. A non-transitory computer readable medium having stored thereon computer executable instructions that, if executed by a computing device, cause the computing device to perform a method of concealing bit errors in a speech signal within a signal parameter quantizer, the method comprising;
encoding a speech signal parameter based upon constraints placed on the signal parameter quantizer, the constraints being based upon naturally prohibitive speech parameter combinations in one or more speech parameter samples, the combinations including different types of parameters;
comparing the encoded speech signal parameter against the constraints within the computing device, wherein the comparing includes comparing a logarithmic gain value of the speech signal with a threshold function determined based on a difference between a plurality of logarithmic gain values of the speech signal; and
modifying operation of the signal parameter quantizer when the constraints are violated.
17. The computer readable medium of claim 16 , wherein the modifying includes declaring the encoded speech signal parameter invalid; and
wherein the method further comprises substituting the invalid speech signal parameter with a replacement value obtained from other signal parameter samples.
18. The computer readable medium of claim 16 , wherein the different types of parameters include at least one from the group including (i) log gain, (ii) first normalized autocorrelation coefficient, and (iii) zero crossing rate.
19. The computer readable medium of claim 18 , wherein the different types of parameters are conditioned on a previously decoded speech waveform.
20. The computer readable medium of claim 19 , wherein the previously decoded speech waveform is a last sub-frame.