Abstract: The invention relates, in particular, to a system for managing communication sessions between first terminals adapted to send audio and/or video and/or data streams during these sessions and second terminals accessible via an external data transmission network. The first terminals are connected to a stream concentrator itself connected to the external network. Since the streams sent by at least some of the first terminals are coded or intended to be coded by hierarchical encoders, the system includes means for processing said streams so that each of them is sent to the external network by the concentrator at a given hierarchy level and means for determining said hierarchy level as a function of parameters relating to the number and nature of the sessions set up between the first terminals and the second terminals.

Abstract: The invention relates, in particular, to a system for managing communication sessions between first terminals adapted to send audio and/or video and/or data streams during these sessions and second terminals accessible via an external data transmission network. The first terminals are connected to a stream concentrator itself connected to the external network. Since the streams sent by at least some of the first terminals are coded or intended to be coded by hierarchical encoders, the system includes means for processing said streams so that each of them is sent to the external network by the concentrator at a given hierarchy level and means for determining said hierarchy level as a function of parameters relating to the number and nature of the sessions set up between the first terminals and the second terminals.

Abstract: The messaging system (10) comprises means (14) for storing digital audio and/or video messages and/or data. It further comprises means for truncating (26f, 44) messages coded by hierarchical encoders in order to record messages in said storage means and/or to forward messages previously recorded in said storage means.

Abstract: At least one coded binary audio flux organized into frames is created from digital audio signals which were coded by transforming them from the time domain into the frequency domain. Transform coefficients of the signals in the frequency domain are quantized and coded according to a set of quantizers. The set is determined from a set of values extracted from the signals. The values make up selection parameters of the set of quantizers. The parameters are also present in the frames. A partial decoding state decodes then dequantizes transform coefficients produced by the coding based on a set of quantizers determined from the selection parameters contained in the frames of the coded binary audio flux or of each coded binary audio flux. The partially decoded frames are subjected to processing in the frequency domain. The thus-processed frames are then made available for use in a later utilization step.

Abstract: A video-conferencing system between participants located at distant sites, each site featuring a viewing screen. The viewing screen is a large screen of the projection, overhead projection, or direct projection type which has one or more viewing systems forming an image wall to simultaneously view all the distant participants within a space reproduced on the image wall formed by the screen. The system includes devices for each site that spatially distribute the sound to match the sound generated by each distant participant and the image of the participant projected on the screen at the site considered.

Abstract: In order to reduce acoustic echo and restore stereophonic sound in a digital transmission system of sound signals on multiple transmission channels, between a local site and one or several remote sites, comprising several microphones and loudspeakers, the method consists in the following steps: computing a weighting factor to be applied to each microphone signal before emitting, based on two attenuation factors computed from the microphone and loudspeaker signals; transmitting on each transmission channel the linear combination of weighted microphone signals; computing a weighting factor to be applied to each loudspeaker signal; and emitting on the loudspeaker the linear combination of transmitted, weighted signals.

Abstract: At least one coded binary audio flux organized into frames is created from digital audio signals which were coded by transforming them from the time domain into the frequency domain. Transform coefficients of the signals in the frequency domain are quantized and coded according to a set of quantizers. The set is determined from a set of values extracted from the signals. The values make up selection parameters of the set of quantizers. The parameters are also present in the frames. A partial decoding state decodes then dequantizes transform coefficients produced by the coding based on a set of quantizers determined from the selection parameters contained in the frames of the coded binary audio flux or of each coded binary audio flux. The partially decoded frames are subjected to processing in the frequency domain. The thus-processed frames are then made available for use in a later utilization step.

Abstract: An acoustic antenna for computer workstation, used for video conferencing that comprises a plurality of microphones connected to a summator circuit. The microphones are distributed in a layout so as to form a substantially straight line and are each spaced with respect to a reference microphone placed in the vicinity of the vertical axis of symmetry of the screen of the workstation according to a specified law. The layout exhibits a substantially cylindrical directivity pattern whose axis of revolution is formed by the straight line.

Abstract: A system for selective sound capture for hands-free telephony and multimedia workstations of the acoustic antenna type with an aerial formed by a member with a concave cylindrical surface. The system includes a plurality of electro-acoustic transducers which are distributed over and in the vicinity of the cylindrical surface. Circuits for partial summation of a plurality of analog sound signals make it possible to form elementary acoustic antennas and a summation circuit produces a resultant analog signal representative of the sound signal arising from a useful speaking area.

Abstract: A method and a system for processing digital audio signal preechoes. Samples x(n) are frequency transform block-coded as coefficients y(k) then transmitted as coefficients Cy(k), decoded as coefficients y(k), and reverse frequency transform processed to recreate the original post-transmission signal x'(n). During the coding process and after the coding of coefficients for the transmission of coded coefficients Cy(k), the coded coefficients Cy(k) are decoded and reverse frequency transform processed to recreate the decoded signal x"(n), which is substantially identical to signal x'(n), and the optimal filtering mask is determined in order to recreate the original signal x(n). The coded coefficients Cy(k) and the gain and filtering mask parameters are simultaneously transmitted to the decoding process. Suitable filtering of the original post-transmission signal x'(n) is performed during the decoding process in order to recreate the original post-transmission and post-processing signal xt(n).

Abstract: After a windowing operation h(n) performed by space-time weighting of the samples, a method and an apparatus for encoding-decoding a digital signal comprising a sequence of samples x(n) consists in performing a modified discrete cosine transform of the samples to calculate the even order transformation coefficients: ##EQU1## for k .epsilon.[0, . . . , N/2-1] with YN-k=-Yk-1 The coefficient Y2k are expressed in the form of an invertible complex transformation: ##EQU2## for k=0, . . . , N/4-1 with y'n=x2n.h2n y"n xN-2n-1.hN-2N-1W4N=cos(2.pi./4N)+j.sin(2.pi./4N) The invertible comples transformation is calculated using an auxiliary calculation equation: ##EQU3## with: zn=(y2n-yN/2-1-2n)+j(yN-1-2n+yN/2+2n) The invention is applicable to encoding and decoding digital audio or video signals.

Abstract: A method and a system process the pre-echoes of an audio-digital signal coded by frequency transformation. The method and the system, after transforming the signal by a frequency transform into transform coefficients y(k), over weighting windows formed over blocks of samples x(n) including N successive samples, and then quantizing the transform coefficients y(k), make it possible to perform, after decoding the transform coefficients y(k), an optimal filtering of KALMAN filtering type. When a transition in amplitude of the signal is present in the sample block, this KALMAN filter reduces the resulting additive noise arising from the quantization coding of the transform coefficients. The method and system may be used in the processing, storage and transmission of high-fidelity quality audio digital signals.

Abstract: A device for compressing a digital audio signal, includes a device for allocating bits available for the transmission or storage of the signal, controlling means for the adaptive quantization of the signal, in order to enable a major reduction in the bit rate while at the same time preserving the quality of the starting signal to the maximum extent. The device includes means for allocating a specific number of bits for the expression of the coefficients of each frequency band of a transformed digital audio signal, as a function of a piece of auxiliary information corresponding to a description of the spectrum of the signal, said device being informed by means for the prior elimination of spectral components of said transformed signal as a function of a psychoauditive criterion.

Abstract: A sound signal coding method and installation are disclosed, which method, usable particularly for transmitting or storing high quality audio signals in digital form, involves chopping the sound signal into blocks of samples and subjecting the samples of each block to time-frequency transform coding whose coefficients are transmitted in digital form, the reverse operations taking place at reception. At least the most significant coefficients of the transform of each block are subjected to predictive and adaptive coding using the stationary state of the signal over a duration greater than that of a block.