AU2777500A

AU2777500A - A process and circuit arangement for determining an item of quality information on the transmission quality of a voice signal in a digital transmission system

Info

Publication number: AU2777500A
Application number: AU27775/00A
Authority: AU
Inventors: Tim Dr Fingscheidt
Original assignee: Alcatel CIT SA; Alcatel SA
Current assignee: Alcatel Lucent SAS
Priority date: 1999-05-10
Filing date: 2000-04-14
Publication date: 2000-11-16
Also published as: DE19921504A1; EP1052621A3; EP1052621A2

Description

P/00/011 Regulation 3.2

AUSTRALIA

Patents Act 1990

ORIGINAL

COMPLETE

SPECIFICATION

Sn.

S

*55* STANDARD PATENT Invention Title: 'A process and circuit arrangement for determining an item of quality information on the transmission quality of a voice signal in a digital transmission system' The following statement is a full description of this invention, including the best method of performing it known to us: IP Australia i Docuiments received on: (n 14 PR 2000 ,t.ch N r- FHPSYDFA\NATPO20\fW6)69577 7 CE9p,0369562.9 2 A process and circuit arrangement for determining an item of quality information on the transmission quality of a voice signal in a digital transmission system Field of the invention The present invention relates to process for determining an item of quality information on the transmission quality of a voice signal and to a circuit for implementing the process.

Background of the invention In digital transmission systems, such as for example mobile telephone systems according to the GSM standard, the quality of the transmitted voice signals is substantially dependent upon the correction capability of the channel decoding process. However, in the channel decoding it is impossible to avoid the occurrence of residual bit errors which lead to a noticeably impaired voice quality. To improve upon this lack of voice quality, error °concealment measures are normally implemented before or after the voice decoding. The nature °of these error concealment measures is substantially dependent upon the accuracy of an item of 15 quality information which is a gauge of the transmission quality of the channel and ultimately of o othe voice quality.

A so-called soft-output Viterbi algorithm, abbreviated to SOVA, is known with which a quality value is supplied for each decoded bit, see Hagenauer Hoeher "A Viterbi Algorithm with Soft-Decision Outputs and its Applications", Proc. of GLOBECOM '89, p.

20 1680-1686, Dallas, Texas, November 1989. Although this algorithm can be implemented i °.relatively simply, the quality information which it supplies does not satisfactorily improve the voice quality.

Additionally, an error concealment method in a TDMA radio system is known, with which quality information is obtained from the channel decoding process, see US 5,502,713. In a so-called soft value calculator, weighting factors, with which instantaneously and previously received parameters of the voice coding process are weighted and interpolated, are calculated from quality information items.

The material disclosed in the foregoing discussion should not be considered to form part of the common general knowledge as at the priority date of this application.

CE00869562.9 3 Summary of the invention According to a first aspect of the present invention there is provided a process for determining an item of quality information on the transmission quality of a voice signal which in a digital transmission system reaches a subscriber via a channel decoder and via an error concealment unit and via a voice decoder, and wherein a first bit error rate is estimated from a first number of input bits by the channel decoder and wherein a second bit error rate is estimated from a second number of input bits by a receiver or equalizer, wherein by means of a transformation unit the first bit error rate and the second bit error rate are processed and combined in accordance with a transformation rule such that an item of quality information is determined as input parameter for the error concealment unit by which an improvement in the voice quality is effected.

According to a first aspect of the present invention there is provided a circuit arrangement for implementing the process, wherein the transformation unit is connected between the channel decoder and the error concealment unit such that the output of the channel 15 decoder, which supplies the first bit error rate, is connected to a first input of the transformation unit, that the second bit error rate supplied by a receiver or equalizer is applied to a second input o of the transformation unit, and that the outputs of the transformation unit are connected to the error concealment unit.

In basic concept the invention consists in that quality information before the channel 0* 20 decoding is combined with quality information after the channel decoding and then is processed by a transformation unit in accordance with different transformation rules adapted to the °°transmission conditions so as to form an improved item of quality information for a downstream error concealment unit.

Brief description of the drawings The invention will now be explained in detail in the form of an exemplary embodiment.

In the associated drawing: Figure 1 is a block circuit diagram of the circuit arrangement according to the invention and Figure 2 is a schematic illustration of data blocks at the input of an error concealment unit.

CEOG369562.9 4 Detailed description of the embodiments In accordance with Figure 1, a circuit arrangement for voice transmission in the direction of a subscriber terminal (upstream) in a digital mobile telephone system comprises a channel decoder 1, an error concealment unit 2 and a voice decoder 3 and, in accordance with the invention, a transformation unit 4. The channel decoder I is supplied with an input signal y' from which the channel decoder I decodes the error protection code, generated in the transmitter for reasons of error protection, and makes this available as output signal A part of the transmitted data can also be transmitted with no error protection code and is represented in Figure 1 as data signal The error protection code can be generated either continuously or in each case for a data block of the length bl. In the case of block-oriented processing, in accordance with Figure 2 the number bI of input bits of the error concealment unit 2 is composed of a number M of input bits of the data signal z' with no error protection code and a number L of input bits of the output signal x' of the channel decoder 1. A bit error rate BER(x') of the output signal x' of the channel decoder 1 is estimated by the channel decoder 1. The bit error rate is also referred to as soft-output or in general as quality information. A bit error rate BER(z') of the data signal z' with no error protection code can be estimated for example by a receiver or equalizer. The thus available, estimated quality information items BER(x') and BER(z') are now supplied as input variables to the transformation unit 4 and are further processed in accordance with the following rules such that the error concealment unit 2 is supplied with improved quality information to increase the quality of the output signal of the voice decoder 3.

Transformation rules: a) Separate mean value formation from quality information items: i L-1 BERn(xi') Y, BER(Xk') L k=O 0, L A I M-1 BERn(zj') Y, BER(zj') M k=O j 0, M -I CE00869562.9 Here it is assumed that L values xi' and M values z i exist per block. The values of the bit error rate BER(x') and BER(z') improved by the transformation are designated BERn(x') and BERn(z').

b) Combined mean value formation from available quality information items: BERn(x') BERn(z') L ML BER(x k BER(z,' L+M t=0 i= 0, L-1 j= 0, M-1 Here again, the assumptions described under a) apply.

c) Mean value formation with selection criterion variant 1: BERn(x,') max[BER(x i BERm, where b L B*ER,, mina+- -IBER(xk'),0.5 (1) L k=0 i= 0, L-1 The improved estimate value BERn(x') of the bit error rate BER(x') estimated by the channel decoder 1 is determined from the maximum of the individual bit error rates BER(xi') estimated by the channel decoder 1 and a bit error probability BERtmp. The bit error probability BERtmp is derived in accordance with equation from the mean value previously described 15 under a).

d) Mean value formation with selection criterion variant 2: BERn(x' max[BER(x' BERp where F d L-1 BER,P max c BER(xk'),0.0 S k=o i= 0, L-1 The transformation rule described here differs from that described under c) by the determination of the bit error probability BERtmp.

CE00869562.9 6 The parameters a, b, c and d referred to under c) and d) in the transformation rules are selected as a function of the relevant transmission modes, such as GSM full rate or GSM enhanced full rate.

It is particularly advantageous to render the execution of the individual process steps dependent upon a so-called bad frame indicator BFI. Depending upon the quality of the transmission conditions, the signal BFI is set at in the case of poor transmission conditions and at in the case of good transmission conditions. In accordance with Figure 1, the signal BFI is supplied when necessary to the error concealment unit 2, the voice decoder 3 and the transmission unit 4, whereby the assemblies can be activated or deactivated. In addition to activating/deactivating the forementioned assemblies, the signal BF1 can be linked with the transformation rules. This ensures that the reliability of the quality information is better adapted to the transmission conditions. Thus for example the estimated bit error rate BERn(x') can be increased with the transformation rule described under c) whereby in the case of poor transmission conditions, thus when BFI=I, it is possible to compensate for an underestimation 15 of the bit error rate by the soft-output Viterbi algorithm. In the case of good transmission conditions, thus when BFI=0, use of the transformation rule stated under d) is more suitable to reduce the bit error rate and thus to avoid the effects of model faults of the transmission system.

More reliable quality information is obtained by the process according to the invention compared to the known prior art, thereby facilitating an improved parameter estimation for error 20 concealment measures. An improved voice quality downstream of the voice decoder 3 is achieved in this way. The process can be implemented with a smaller outlay compared to known, more complex algorithms, as it is only necessary to interpose a transformation unit 4 between the channel decoder 1 and the error concealment unit 2.

The circuit arrangement illustrated in Figure 1 can be implemented as an integrated circuit, in which case the following integration options can be selected if required: channel decoder 1 and transformation unit 4 error concealment unit 2 and voice decoder 3 channel decoder 1, transformation unit 4 and error concealment unit 2 transformation unit 4 and error concealment unit 2 CE00369562.9 7 It will be understood that the invention disclosed and defined herein extends to all alternative combinations of two or more of the individual features mentioned or evident from the text or drawings. All of these different combinations constitute various alternative aspects of the invention.

The foregoing describes embodiments of the present invention and modifications, obvious to those skilled in the art can be made thereto, without departing from the scope of the present invention.

Claims

1. A process for determining an item of quality information on the transmission quality of a voice signal which in a digital transmission system reaches a subscriber via a channel decoder and via an error concealment unit and via a voice decoder, and wherein a first bit error rate is estimated from a first number of input bits by the channel decoder and wherein a second bit error rate is estimated from a second number of input bits by a receiver or equalizer, characterised in that by means of a transformation unit the first bit error rate and the second bit error rate are processed and combined in accordance with a transformation rule such that an item of quality information is determined as input parameter for the error concealment unit by which an improvement in the voice quality is effected.

2. A process according to Claim 1, characterised in that by means of the transformation unit a first mean value (BERn(x')) is formed from the first bit error rates determined block by block, and a second mean value (BERn(z')) is formed from the second bit error rates determined block by block.

3. A process according to Claim 1, characterised in that by means of the transformation unit a combined mean value is formed from the first bit error rates and the second bit error rates in accordance with eooe 1 1 L-1 M -1 9* BERn(x,') BERn(z, L BER(xk') BER(z' L+M[L=o t=o i= 0, L-1 j= 0, M 1

4. A process according to Claim 1, characterised in that by means of the transformation unit a quality gauge is determined from the first bit error rates according to a first selection criterion in accordance with BERn(x,') max[BER(x,' ),BERmp] where SCE00369562.9 9 BER b L= 1 BER,, mina -I BER(xk ),0.5 I L k=o i= 0, L- 1 A process according to Claim 1, characterised in that by means of the transformation unit a quality gauge is determined from the first bit error rates according to a second selection criterion in accordance with BERn(x,') max[BER(x,' BER,,] where d L-I BER,,, max c BER(x' ),0.0 L k=O i= 0, L- 0

6. A process according to Claim 1, characterised in that the transformation rule of the transformation unit is selected as a function of the value of a so-called bad frame indicator. S" 7. A circuit arrangement for implementing the process, characterised in that the transformation unit is connected between the channel decoder and the error concealment unit :such that the output of the channel decoder, which supplies the first bit error rate is connected to a first input of the transformation unit, that the second bit error rate (BER(z')) supplied by a receiver or equalizer is applied to a second input of the transformation unit, and *o that the outputs of the transformation unit are connected to the error concealment unit.

8. A circuit arrangement according to Claim 7, wherein a third input of the transformation unit and the error concealment unit are connected to a line supplying the bad frame indicator.

9. A process for determining an item of quality information on the transmission quality of a signal substantially as hereinbefore described with reference to the accompanying drawings. CE00369562.9 A circuit substantially as hereinbefore described with reference to the accompanying drawings. Dated this 14th day of April 2000 ALCATEL by its attorneys Freehills Patent Attorneys