Decoding method, speech coding processing unit and a network element
This invention is related to tandem free operation (TFO) in mobile cellular systems. The present invention implements a tandem free operation by using a special feedback loop which makes the decoded parameters available, performs the comfort noise insertion and bad frame handling operations, produces the parameter quantisation indices corresponding to the output of these operations, and synchronises the speech encoders and the speech decoders in the transmission path from the uplink mobile station to the downlink mobile station. This functionality is realized by partly decoding and re-encoding the parameters and synchronising and resetting the quantiser prediction memories in a specific manner. A basic idea of the invention is, that during BFH and CNI processes, a re-encoding block produces models of encoded speech parameters from the BFH/CNI processed speech parameters. These models of encoded speech parameters are then transmitted to the receiving end. The present invention provides a solution to the problem created by predictive, more generally non-stateless encoders in TFO operation.
1. A method for processing speech signal parameters encoded by a non-stateless encoder, characterized in that the method comprises steps of decoding the encoded speech parameters using a non-stateless decoder, processing the decoded speech parameters, re-encoding the processed decoded speech parameters using a second non-stateless encoder, updating the state of the non-stateless decoder at least in part with the state of the second non-stateless encoder, and replacing at least one encoded speech parameter with a re-encoded speech parameter to produce processed encoded speech parameters.
2. A method according to claim 1 , characterized in that before said re-encoding step, the state of said second non-stateless encoder is updated with the state of the non-stateless decoder.
3. A method according to claim 1 , characterized in that in said speech parameter processing step, comfort noise information is converted into the decoded speech parameters.
4. A method according to claim 1 , characterized in that in said speech parameter processing step, bad frame handling information is converted into the decoded speech parameters.
5. A speech coding processing unit for decoding encoded speech parameters and producing a decoded time domain speech signal and encoded speech parameters representing the signal, characterized in that the unit comprises
a non-stateless decoding block for decoding the encoded speech parameters,
a speech parameter processing block for processing the decoded speech parameters, and
a non-stateless encoding block for encoding the processed speech parameters for producing the encoded speech parameters representing the signal,
and in that, that the unit is arranged to update the state of the non-stateless decoding block at least in part with the state of the non-stateless encoding block.
6. A speech coding processing unit according to claim 5 , characterized in that said non-stateless decoding block is a predictive dequantiser, and said non-stateless encoding block is a predictive quantiser.
7. A speech coding processing unit according to claim 5 , characterized in that said speech parameter processing block is a comfort noise processing block.
8. A speech coding processing unit according to claim 5 , characterized in that said speech parameter processing block is a bad frame handling block.
9. A speech coding processing unit according to claim 5 , characterized in that the unit is a transcoder unit.
10. A speech coding processing unit according to claim 5 , characterized in that the unit is an uplink transcoder unit.
11. A telecommunications network element for receiving encoded speech parameters and transmitting a time-domain speech signal with embedded encoded speech parameters, characterized in that the network element comprises a speech coding processing unit having
a non-stateless decoding block for decoding the encoded speech parameters,
a speech parameter processing block for processing the decoded speech parameters, and
a non-stateless encoding block for encoding the processed speech parameters for producing the the embedded encoded speech parameters,
and in that, that said unit is arranged to update the state of the non-stateless decoding block at least in part with the state of the non-stateless encoding block.
12. A telecommunications network element according to claim 11 , characterized in that said non-stateless decoding block is a predictive dequantiser, and said non-stateless encoding block is a predictive quantiser.
13. A telecommunications network element according to claim 11 , characterized in that said speech parameter processing block is a comfort noise processing block.
14. A telecommunications network element according to claim 11 , characterized in that said speech parameter processing block is a bad frame handling block.
15. A telecommunications network element according to claim 11 , characterized in that said unit is a transcoder unit.
16. A telecommunications network element according to claim 11 , characterized in that said unit is an uplin transcoder unit.