A method for testing the voice path in a communication network by using speech recognition. The method includes generating transmissible voice signals, preferably, computer generated voice signals which test predetermined voice path characteristics. The voice signals are transmitted in digitized packet form from a sending computer to a receiving computer at the other end of the voice path. The received voice signals are interpreted by a speech recognition engine which identifies speech patterns in the received voice signals. The speech patterns are then compared to the reference speech patterns of the voice signals sent from the sending computer and a voice path quality factor is generated as a result of this comparison. The voice path quality factor is a relative indicator of the quality of transmission and processing of the voice path. The voice quality factor can be used to fine tune and manage communication networks which transmit digitized voice data packets.

Method and apparatus for measuring voice path quality by means of speech recognition

A method for automatically and reproducibly rating the transmission quality of a speech transmission system, comprising the steps of: generating a test signal containing a predetermined recorded voice signal; transmitting the test signal by the speech transmission system from a transmitting end to a receiving end; extracting characteristic values from the received test signal; determining a differential pattern corresponding to the difference between the characteristic values and predetermined reference values; and feeding the differential pattern to a neural network that has been trained for classifying patterns according to a number of predetermined transmission quality classes.

Method and apparatus for automatically and reproducibly rating the transmission quality of a speech transmission system

A device for determining the quality of an ouput signal to be generated by a signal processing circuit with respect to a reference signal is provided with a first series circuit for receiving the output signal and with a second series circuit for receiving the reference signal and generates an objective quality signal by means of a combining circuit coupled to the two series circuits. The poor correlation between said objective quality signal and a subjective quality signal to be assessed by human observers can be considerably improved by reducing, in a differential arrangement, present in the combining circuit, for determining the difference between the two series circuit signals, said difference by a certain value, preferably as a function of a series circuit signal, and the poor correlation can be further improved by disposing a scaling circuit between the two series circuits for scaling at least one series circuit signal, it is furthermore also possible to scale the quality signal as a function of said scaling arrangement.

Signal quality determining device and method

The figure shows a speech transmission link. This leads from a telephone (11) of a private network (13) via exchanges (14, 17) to a mobile node (19) with an antenna (20) and via a radio link (22) to a mobile telephone (12). A mobile (31) and a stationary measuring unit (32) which are disposed near to the telephones (11, 12) are used to evaluate the given transmission quality. The one measuring unit (31, 32) transmits a pre-defined sequence of human speech, prefixed by a synchronisation pattern, as a test signal. The respective other measuring unit (32, 31) recognises the synchronisation pattern and thereby defines the starting time of the following sequence. This sequence is divided up into a multiplicity of respective intermediate values in that time intervals are formed and their sound characteristics are digitised. All of the intermediate values are finally evaluated by a neural network in such a way that an evaluation of "quality good, average or poor" understandable to humans is achieved at the output of said network. It is important here that the neural network is first trained using data evaluated by test persons.

Method of transmission quality estimation of a speech transmission link

A device for determining the quality of an output signal to be generated by a signal processing circuit including a radio link with respect to a reference signal is provided with a first series circuit for receiving the output signal and with a second series circuit for receiving the reference signal and generates an objective quality signal by means of a combining circuit coupled to the two series circuits, whereby a scaling circuit is disposed between the two series circuits for scaling at least one series circuit signal. The poor correlation between said objective quality signal and a subjective quality signal to be assessed by human observers can be considerably improved by disposing a discounting arrangement inside the combining circuit, which discounting arrangement should be coupled to the scaling circuit for receiving a comparison signal and for discounting said comparison signal while generating the quality signal.

Device and method for signal quality determination

In the measuring method according to the invention, speech samples and/or test signals, from which a speech recogniser or speaker recogniser has previously formed a reference pattern during a learning phase, are presented to this speech recogniser or speaker recogniser via a speech coder to be assessed or a transmission route to be tested Using an evaluation computer a comparison is carried out between the true meaning provided from a transmission memory and the meaning of a speech sample or a test signal recognised by the speech recogniser or speaker recogniser In this process, an error rate or a measure of the reliability of recognition is simultaneously calculated over a measurement cycle

Measuring method for assessing the quality of speech coders and/or transmission routes

Verfahren zum Feststellen der Identität des Inhabers einer Chipkarte, inbesondere Kreditkarte

The method is intended to be able to be used for supplementary identification of individuals by speech, in conjunction with chip cards or credit cards. For this purpose, targeted access to individual speech reference data should be possible. No memory capacity for speech reference data on the credit card should be necessary. No devices for speech recognition should be necessary in the terminal device which is used for input. A test computer is provided, and a connection to it is set up automatically by a terminal device if the identity of the credit card owner is to be tested by speech. A connection is then set up automatically by the test computer to a central computer, where reference pattern data are already stored, and these data are transmitted to the test computer. The test computer compares the reference pattern data received from the central computer with the corresponding words which the card owner is now supposed to speak. These words are transmitted by the terminal device to the test computer. Only an indication that speech pattern data exist for this credit card need be stored on the card. The terminal device can be in a relatively simple form, and does not need to contain any speech recognition and comparison devices.

Jens Dr Ing Weber

Papers

Measuring method for assessing the quality of speech coders and/or transmission routes

Verfahren zum Feststellen der Identität des Inhabers einer Chipkarte, inbesondere Kreditkarte

Method for verifying the identity of the owner of a chip card, especially a credit card