Data processing: speech signal processing – linguistics – language – Speech signal processing – For storage or transmission
Reexamination Certificate
1999-04-21
2001-05-08
Knepper, David D. (Department: 2748)
Data processing: speech signal processing, linguistics, language
Speech signal processing
For storage or transmission
C704S201000, C714S716000, C370S249000
Reexamination Certificate
active
06230120
ABSTRACT:
The present invention relates to digital telecommunication systems, and particularly to the detection of speech channel back-looping therein.
In conventional digital telecommunication systems, speech is transferred in a digital pulse code modulated (PCM) transmission channel as standard (A-law encoded or mu-law encoded) PCM samples, typically at a rate of 64 kbit/s, i.e. 8,000 samples/second. This provides excellent speech quality.
In some cases it has been possible to combine with digital speech samples other supplementary information, such as signalling messages or speech coding parameters of speech coded to a low bit rate. This supplementary information transfer can be based on what is known as the bit stealing technique, in which one or more bits (usually the least significant LSB) of the speech sample is chosen for this purpose. Since the LSB of a speech sample contains very little speech information, there is no detectable deterioration of speech quality. Such supplementary information is usually transmitted between the transmission devices or speech processing devices of the network.
In call switching, the following situation may arise. The speech processing device is activated before a call is switched forward in the telephone switching centre. In this case the speech processing device transmits PCM speech samples towards the telephone switching centre. It is the property of some telephone switching centres that speech channels are looped directly back to the sender if the switching forward is not ready. This results in the speech samples returning to the speech processing device which sent them. Usually this speech channel back-looping does not cause problems since call switching is not ready up to the user and consequently the back-looped speech/silence is not heard by anyone.
In contrast, in cases where supplementary information is sent in a speech sample, speech channel back-looping may cause problems. When the speech processing unit gets back supplementary information (e.g. a signalling message) it has sent, it may assume that the received information originates from some other device, not itself. Particularly if the signalling message is of a fixed format, it is impossible to distinguish the sender solely from said message.
In the following some examples will be given of cases involving the above problem.
In digital mobile communication systems, for example, the most limited resource is the radio path between mobile stations and base stations. To reduce the bandwidth required by one radio connection on the radio path, speech transfer utilizes speech coding for achieving a lower, e.g. 16 or 8 kbit/s, transfer rate than the 64 kbit/s transfer typically employed in telephone networks. The mobile station and the fixed network side must naturally have a speech encoder and decoder for the purposes of speech coding. On the network side the speech coding functions may be located in a plurality of alternative places, such as in a base station or in association with a mobile exchange. The speech encoder and decoder are often far away from the base station in the system as what is known as a remote transcoder unit, whereby speech coding parameters are transferred between the base station and the transcoder unit in the network in specific frames.
In each mobile terminating or originating speech call a transcoder is connected to the speech connection on the network side. The transcoder interface towards the mobile exchange is 64 kbit/s. The transcoder decodes a speech signal vocoded to a transmission channel of an 8/16 kbit/s rate from a mobile station (uplink) to a 64 kbit/s rate and encodes a 64 kbit/s speech signal to the mobile station (downlink) and from the mobile exchange to an 8/16 bit/s rate. Hence speech quality is lower than in a normal telephone network. This arrangement is trouble-free as long as one party of the call is a mobile station and the other e.g. a subscriber in the public switched telephone network (PSTN).
In the case of a mobile to mobile call MMC, the operation of the mobile communications network causes there to be one transcoder on a connection between a calling mobile station and a mobile exchange, and similarly another transcoder between a called mobile subscriber and (the same or another) mobile exchange. These transcoders are then coupled together via the mobile exchange(s) as a result of normal call switching. In other words, two transcoder units are coupled in tandem for each MMC call and the call is subjected twice to speech encoding and decoding. This is called tandem coding. Tandem coding is a problem in mobile communication networks since it impairs speech quality owing to extra speech encoding and decoding. Up to now tandem coding has not caused very much trouble since relatively few calls have been MMC calls. However, the number of MMC calls will continue to increase with an increasing number of mobile stations.
The applicant's Finnish patent application FI951807 discloses a transcoder having what is known as tandem coding prevention. An MMC call is switched as usual with the connection having two transcoders in a tandem configuration. The speech to be transferred between a transcoder and a mobile station has been coded by the vocoding method which decreases transfer rate. Both transcoders perform normal transcoding operations on the speech such that the speech is decoded in one transcoder into normal digital pulse code modulated (PCM) speech samples which are transferred to the other transcoder and encoded therein by said vocoding method. Speech information received from the mobile station and complying with said vocoding method, i.e. speech parameters, which are not subjected to transcoding operations (encoding and decoding) in either tandem connected transcoder, is transferred at the same time in a subchannel formed by one or two least significant bits of the PCM speech samples. The receiving transcoder selects the speech information complying primarily with this vocoding method for transmission across the interface to the receiving mobile station. As a result, vocoding is principally performed only in mobile stations and the vocoded speech information, i.e. speech parameters, are transferred through the mobile communication network without tandem coding, resulting in improved speech quality. When the receiving transcoder does not find vocoded speech information in the least significant bits of the PCM speech samples, the speech information to be transmitted over the radio interface is encoded as usual from the PCM speech samples.
The applicant's Finnish patent application FI960590 discloses a transmission equipment for optimizing the use of transmission resources on a transmission connection between telecommunication network elements, such as exchanges or base station controllers. Both ends of the connection are provided with a transmission equipment which is connected to a number of PCM channels originating from the switching centre. Between the transmission equipments is a lower-capacity PCM link where the bits of the PCM samples of each channel form subchannels in which lower-rate vocoded speech or data can be transferred. If a PCM coded speech signal in which one or more least significant bits of the PCM samples form a lower-rate subchannel is also received from the switching centre, the contents of this subchannel are multiplexed to one subchannel of the PCM link. If only a PCM coded speech signal is received from the switching centre, it is encoded into a lower-rate vocoded speech signal and the vocoded speech signal is multiplexed into one subchannel of the PCM link. At the other end of the connection the transmission equipment decodes the vocoded speech signal back to PCM samples, into whose least significant bits are placed the contents of the subchannel without decoding. This transmission equipment is suitable for use particularly in association with the tandem coding described in patent application FI951807.
EP application 0,333,345 describes tandem speech coding in a fixed telephone network using digital switc
Knepper David D.
Nokia Communications Oy
Pillsbury Madison & Sutro LLP
LandOfFree
Detection of speech channel back-looping does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Detection of speech channel back-looping, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Detection of speech channel back-looping will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2452841