International Business Machines Corporation

Avaya Technology LLC

Cisco Technology, Inc.

Gerard Richter

Bernard Y Pucci

Jean-Claude Abbiate

Maurice Duault

Claude Cesaro

Claude R Galand

Alain R Blanc

Patrick J Jeanniot

Gerald Lebizay

Jean E Menez

Gerard Orengo

Gerard Lebizay

Jean-Yves Solves

Aline Fichou

Michel F Ferry

Michele Rosso

Pascal Chauffour

Francois Kermarec

Charles Arnaud

Jean-Pierre Pantani

Walter Joseph Fant

Jean-Pierre Vaudaux

Serge Leroy

A method and an apparatus for managing signal-processing resources of a multimedia platform that is designed for applying signal-processing operations to multimedia signals defines multimedia functions each capable of monitoring the operation of a set of multimedia platform signal-processing resources, puts them in contact, adapts the contents of the signal-processing resource set depending on the multimedia signal to be processed, and uses the multimedia functions to apply the signal-processing operations to the multimedia signals.

Method for managing multimedia platform resources and multimedia platform for implementing it

An adaptive apparatus and method for managing a playout buffer (POB) in an edge node of a packet-based data communications network, such as an ATM network, in order to reduce the end-to-end communication delay introduced thereby and to allow the POB filling level to be corrected in the event of clock speed differences. A monitoring mechanism determines if the minimum filling level of the playout buffer over a time period whereby the average filling level of the playout buffer is reduced according to the minimum filling level of the playout buffer over the time period.

Adaptive playout buffer and method for improved data communication

A multimedia processing system architecture performs a plurality of multimedia tasks on multimedia data using a plurality of multimedia processing blocks, at least one communication bus to carry the flow of multimedia data between the multimedia processing blocks. An application interface manages the multimedia processing blocks. For each multimedia task, the application interface creates a subset of multimedia processing blocks to be assembled in order to run the multimedia task. In this architecture, each multimedia processing block comprises at least one input interface to connect it in receive mode to a communication bus and/or at least one output interface to connect it in send mode to a communication bus. To create a subset of multimedia processing blocks, the application interface selects multimedia processing blocks to form the subset, and then examines the input and output interfaces of the selected multimedia processing blocks in order to check whether all the multimedia data exchanges to perform the multimedia task are possible within the subset and determine, for these exchanges, the encoding format of the multimedia data to be exchanged and the communication buses to be used to connect the different multimedia processing blocks of the subset in pairs according to the complementarity of the ports of their respective input interface and output interface.

Multimedia processing system architecture

The present invention discloses a system and a method for transmitting DTMF signals over high speed digital networks using voice compression algorithms, and particularly to a system for ensuring the integrity of DTMF signals at the destination node of a network after compression and decompression of data on a voice connection. A DTMF Detector is placed in parallel with a voice compression unit performing the compression algorithm. When a candidate DTMF signal is detected, the signal component at the higher frequency in the candidate is filtered out to eliminate the possibility of double DTMF detection at end user equipment. When the candidate DTMF signal is finally validated by the source node as representing a true DTMF signal, only the data essential for reconstituting the DTMF signal are transferred to the destination node.

Method and system for transmitting a DTMF signal with compressed voice in a packet switching network

Process for transmitting compressed voice circuits over a packet switching network of the type comprising a plurality of switching nodes (SW-1 to SW-7) interconnected by connection lines and including at least an end switching node (SW-1) connected to a source exchange telephone device (PABX A) and another end switching node (SW-3) connected to a destination exchange telephone device (CX), and comprising the steps of receiving from the source exchange telephone device a sequence of uncompressed frames wherein each frame contains n slots containing each J bytes respectively associated to J voice circuits, compressing the data bits of each voice circuit in order to build a packet containing J compressed voice circuits, and providing this packet to the end switching node for transmission to the destination exchange telephone device.

Transmission system of telephony circuits over a packet switching network

Repetitive packets in a voice/data stream, are detected and suppressed in the transmitting side of a network, after a predefined number of consecutive repetitive packets have been transmitted. Then, at the receiving side of the network, suppressed repetitive packets are reconstituted by filling the resulting gap in the voice/data stream with the repetitive pattern contained in the last received repetitive packet. When the input voice/data stream is compressed, only non-repetitive packets are compressed so that repetitive patterns are not corrupted by compression.

Repetitive pattern removal in a voice channel of a communication network

A system for coding voice signal to optimize bandwidth occupation in a High Speed Packet Switching network while ensuring best voice transmission quality. The voice signal is first encoded using a conventional GSM like RPE/LTP coder providing first sub-frames of coded signal and tagging these first sub-frames as being non-discardable. In addition, a convenient difference between an RPE/LTP provided signal and a corresponding synthesized image is performed (see 36) and is also block encoded into second sub-frames which second sub-frames are tagged as being discardable sub-frames. Said second sub-frames when concatenated to corresponding first sub-frames provide so-called multirate frames. Then, when transmitting said multirate frames over the High Speed packet switching network, dropping discardable tagged data enables solution network congestion situations in any network node and at random with no significant disturbing effect over the voice communication operation.

System for coding voice signals to optimize bandwidth occupation in high

A method and apparatus for discriminating between voice and voiceband data (fax/modem data) in an input signal from a voiceband channel, which is available by blocks (packets) of samples. Said discrimination is based upon the computation of two characteristics of the input signal: an autocorrelation function and a power variation function, the combination of which provides a discrimination factor which is highly accurate while requiring a low computing power.

Voice/data discriminator

This invention deals with a method and system for dynamically adjusting the communication bandwidth assigned to an audio channel connection in a high speed digital network. More particularly, the invention is made to track the activity of a voice assigned connection (e.g. a PBX or PABX entry to the network), define a so-called activity bit for each block of audio channel signal and then dynamically adjust the assigned network communication bandwidth accordingly. Adjustment of the communication bandwidth is accomplished by integrating the audio channel activity bits through a predefined integration function, the result of which is then compared against predefined threshold values in order to determine an appropriate bandwidth setting for the audio channel.

Method and system for monitoring traffic to optimize the bandwidth

A process for computing the coefficients C(i) of an adaptive filter (206) used in an echo-cancellation structure for a telecommunication apparatus. The process is based on the Least Mean Squared (L.M.S.) or gradient algorithm for computing the coefficients of the filter in accordance with the formula: ##EQU1## where e(i) is the estimation error and .alpha. is the step size used for performing an adjustement of the convergence of the filtering process. The process uses two different estimations of the power of the analog signal, a former short-term estimation and a latter long-term estimation in order to derive the appropriate value of the stepsize .alpha. which is used for computing the new coefficients. In the preferred embodiment the value of .alpha. is determined by dividing a constant .alpha..sub.0 being slightly lower than the maximum value ensuring stability of the convergence process by the greater of the two different estimations of the power of the analog signals.

Process for computing the coefficients of an adaptive filter in an

A method and an apparatus for removing the silence from the digitalized voice signals conveyed through packets or cells switching networks. The silence samples are neither packetized nor sent over the network but are regenerated at the output of the network. The silence samples generated are white noise samples, where the level is adapted to the background noise of the silence samples received at the input node of the network. For long periods of silence, the white noise level is periodically refreshed to be adapted to the last silence samples received at the input node of the network. The method provides also a control of packet or cell loss. The method uses are not control packets; in the later case, it can be used for ATM networks with AAL1. The method is implemented as a program executed in a Digital Signal Processor located on adapter cards dedicated to voice processing in the network access nodes.

Method and a system for silence removal in a voice signal transported

A processing unit for performing convolution computation according to the HARVARD architecture which includes a first and second input register for receiving a first and second operand, a multiplier for multiplying the operand and a Arithmetic and Logic Unit (ALU) circuit. The unit further includes a coefficient storage memory which is used for loading at least one set of coefficients allowing the convolution computation. The memory storage is addressed either from an internal address generator or directly from the internal data bus thereby allowing the possibility to store either coefficients or data into the memory. The flexibility is still increased by the use of a particular set of multiplexing circuits allowing multiple configurations. An internal address generation circuit is used for performing a partial addressing of the set of coefficients thereby providing decimation capability.

Processing circuit for performing a convolution computation

A method for detecting the presence of voice, single tone and dual tone signals compares a ratio r which equals the square of the maximum value of the received signal (A max) during a sampling period divided by a measure E of the energy to three different thresholds. If r is less than the first threshold (3), a single tone is indicated. If it is greater than the first but less than the second, the received signal level is compared to the third threshold (-43 dB) and receipt of a voice signal is indicted if the level of the received signal exceeds this threshold.

Voice activity detection method and apparatus using the same

A method for detecting the presence of voice, single tone and dual tone signals compares a ratio r which equals the square of the maximum value of the received signal (A max) during a sampling period divided by a measure E of the energy to three different thresholds. If r is less than the first threshold (3) a single tone is indicated. If it is greater than the first but less than the second (5.2) a dual tone is indicated. If it is greater than the second, the received signal level is compared to the third threshold (-43 dB) and receipt of a voice signal is indicated if the level of the received signal exceeds this threshold.

A method for detecting the presence of single tone and dual tone signals compares a ratio r which equals the square of the maximum value of the received signal (A max) during a sampling period divided by a measure E of the energy to three different thresholds. If r is less than the first threshold (3) a single tone is indicated. If it is greater than the first but less than the second (5.2) a dual tone is indicated. When a single tone is indicated the received signal is subjected to a second order autoregressive process to determine the frequency of the single tone. If a dual tone is indicated the received signal is subjected to a fourth order autoregressive process to determine the frequencies of the dual tones.

Growing community of inventors

Saint Jeannet, France

Gerard Richter