Abstract: For sound analysis, especially for speaker identification, a long-term spectrum is often formed from the recorded signal and compared to a long-term spectrum of a previously recorded signal. When the signal is, for example, received over telephone lines the transfer function thereof may distort the received signal to such an extent that identification is impossible. It is now proposed not to use the long-term spectrum for identification but rather its change, that is to say to accumulate the difference between two consecutive short-term spectra or the difference between such short-term spectrum and the long-term spectrum and to divide this accumulated value by the long-term spectrum in a Divider. Since the numerator and denominator each contain the frequency-weighting factor (a.sub.J) of the telephone line transfer function, this factor is effectively cancelled out by virtue of the division operation performed in the Divider thereby eliminating the effect of the unknown transfer function.

Abstract: Sounds or speech signals are converted into an electrical sound and broken down into separate spectral components in a filter bank. These components are integrated over a short period of time and produce a short-time spectrum of the signal. The spectral components are applied, via a corresponding number of parallel leads, to a number of pattern detectors which supply an output signal only if the short-time spectrum corresponds to the pattern adjusted in the relevant pattern detector. To this end, each pattern detector includes two threshold value (window) detectors which supply a signal if the applied input signal lies between two adjustable thresholds. Each threshold value detector is connected to a different output of the filter bank for another spectral range and is generally adjusted to different threshold values. The outputs of the threshold value detectors are combined via an AND-element, so that a signal is produced only if all threshold value detectors of the pattern detector are activated.

Abstract: The characteristic values are obtained from the signal to be verified, e.g. in the case of a voice signal the total energy in the individual frequency spectrum ranges, and these characteristic values are quantified. The frequency with which the individual partial ranges of the quantified characteristic values occur is found during a sampling phase, and these frequencies are stored. In the test phase, the signal is analyzed in the same way and the stored frequencies of the partial ranges into which the characteristic values of the signal to be verified have fallen are multiplied and compared with a threshold. In an arrangement for implementing the method, there is a characteristic store with a store address for a multi-digit binary number for each partial range of each characteristic.

Abstract: An arrangement for signal analysis in which the function to be examined is split into its spectral components in a filter bank. The signal values of the individual components are successively applied to a comparison device and to the input of a store. The output of the store feeds the other input of the comparison device which generates a signal whenever the instantaneous signal values of all spectral components correspond to the stored, corresponding signal values, within predetermined tolerance limits. The successful comparison of each time a pair of signal values is counted for all spectral components, and the output signal is generated only if a predetermined counter position is reached. As a result, complete correspondence is not required. An additional pulse generator can also generate pulses for given spectral components which correspond to a successful comparison.

Abstract: In practice it is often desirable to represent only a given, selectable dynamic range of a measuring signal, preferably in logarithmic form. This is achieved in accordance with the invention in that from the measured value in the form of a binary number only the positions corresponding to the desired dynamic range are selected for forming the base number, the weight of the most significant bits having the value "1" being converted into a binary number. The mantissa is derived from the following positions, i.e. independent of the selected dynamic range, so that even when the most significant bit is in the last position, a mantissa comprising the full number of positions is formed. The magnitude of the dynamic range is chosen by omitting the correspondingly more significant positions of the base number formed for the maximum range.

Abstract: Method and apparatus for producing a grey-tone image having areas of different average density values which are given by an n-digit binary number, all the bits of the binary number being interrogated in parallel by clock patterns. For the area to be printed an elemental area containing 2.sup.n pulses is selected which is printed in a total cycle of 2.sup.n .sup.- 1 clock pulses of the clock pattern. In this total cycle, at each clock pulse the drive of each pair of dots is determined in a manner such that the most-significant bit of the binary number is always used for driving while the remaining bits are used a number of times which depends upon their positional values. For this purpose, for each position of the binary number the clock pattern has a number of clock pulses which corresponds to the value of this position. The clock pulses for the individual positions are non-uniformly distributed in time, however, no clock pulses for the second and any further positions of the binary number coincide in time.