International Business Machines Corporation

Michele Rosso

Claude R Galand

Jean E Menez

Paolo Scotton

Philippe Elie

Francoise Bottau

Gerald Lebizay

Gerard Richter

Bernard Y Pucci

Charles Arnaud

Emmanuel Lancon

Hubert Crepy

Thierry Liethoudt

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

Technique for coding video images by filtering the original input image into subband images using successive high-pass and low-pass filterings over a tree-shaped filter bank. Each subband image is then submitted to activity detection and coded with a different number of levels depending on the activity detected and on the subband involved.

Method and apparatus for subband coding video signals

The echo cancellation method includes starting with an arbitrary set of echo filter coefficients. The coefficients are first sequentially re-estimated using a fast converging method. These operations are repeated a predefined number of times, and then the coefficients estimation is switched to a low converging but more precise method. In addition, an adaptive threshold is being defined and readjusted and all filter coefficients lower than said threshold are being discarded. Finally, sampling time boundaries are adaptively defined based on non-discarded filter coefficients with reference to said threshold, and in such a way the final echo cancellation algorithm is adaptive to echo variations (backtracking algorithm).

Adaptive echo cancellation method and device for implementing said method

Method and apparatus allows accurate detection of voice presence on a digitized communication line even in the presence of tone signals. The method includes analyzing digitized communication input signals by current blocks of 20ms, to define the stationary or non stationary state of any current block. Then an analysis on the states of the M last 20ms blocks allows the final decision on voice presence.

Method for detecting voice presence on a communication line

In a voice coding system, the baseband or residual signal is encoded at a lower rate by finding a best estimate at a lower rate. The voice terminal signal x(n) is split into a low-pass filtered band signal y1(n) and a high-pass filtered band signal y2(n). Both y1(n) and y2(n) signals are coded into lower-rate sub-sequences of samples x1(n), x2(n) and x3(n), x4(n) respectively. The sequence of samples to be representative of x(n) is selected among x1(n), x2(n), x3(n) and x4(n) for being the closest to x(n).

Low bit rate voice coding method and device

A process for deriving voice pitch related delay values M to tune a Long-Term Prediction (LTP) filter to be used in an LTP based speech coder converting a speech derived digital signal r(n) into a lower bit rate signal, said filter being provided with a variable length delay line y(n) fed with a reconstructed signal r'(n). The process includes splitting r(n) into segments and each segment into sub-segments; then cross-correlating the first current r(n) sub-segment with a previously reconstructed segment and sorting the cross-correlation values for peak location, whereby a first delay value M1 is derived and used to tune the filter. Then, said M1 is used to compute sample indexes n for a predefined number of samples located about M1/p, . . . , M1, 2M1, . . . , pM1 and repeating cross-correlation and sorting operations to derive M2 and so on up to a full segment length (e.g. 160 samples). Then the process is started all over again.

Fast pitch tracking process for LTP-based speech coders

Some signal coding techniques involve quantizing the signal samples using a signal derived parameter to set the quantizer step. This parameter is then used in the decoding process to synthesize the original signal back to its analog form. Any error affecting the integrity of said parameter, between coding and decoding, would affect the synthesized signal. The method proposed here for protecting data integrity involves, at the encoding site, forcing the parameter to a fixed parity. Then, before synthesizing operations are performed, any parameter failing a parity control test is discarded and replaced with a parameter derived from available valid parameter(s).

Method for eliminating errors in block parameters

A signal vector quantizing coder (CELP) is provided with an adaptive codebook originally loaded with preselected codewords. The codebook is split into a fixed contents portion and a fixed length adaptive contents portion. During coding operations, the codewords dynamically selected for coding the coder input signal are shifted into the fixed length adaptive codebook section for codebook contents updating purposes.

Method and apparatus for dynamically adapting a vector-quantizing coder

The voice signal s(n) is filtered through a short-term predictive filter (13) tuned with PARCOR derived coefficients computed over a pre-emphasized s(n), said filter (13) providing a short-term residual r(n). Said r(n) signal is then processed through a first Cod-Excited/Long-Term Predicative coder providing first couples of table address and gain data (k1, gl)'s. An error signal r'(n) is then derived by subtracting coded/decoded data from uncoded data. Then said error signal is processed through a second Code-Excited/Long-Term Predictive coder providing second couples of data (k2, g2)'s. Full rate coding is achieved by multiplexing both couples (k1, gl)'s and (k2, g2)'s into a multi-rate frame; while switching to a lower rate is achieved through a mere delation of (g2, k2)'s from the full rate frame.

Multi-rate voice encoding method and device

This low bit rate voice encoding involves short-term predictive filtering the voice signal s(n) using partial correlation related coefficients derived from pre-emphasized s(n), and deriving a short term signal r(n); then deriving a long-term residual signal e(n) by subtracting a delayed synthesized short term b.r'(n-31 M) from said r(n); and code excited encoding e(n) into codeword references k's and associated gains G's.

Low bit rate voice coding method and system

A pitch detector to adjust long term prediction in a pulse excitation speech coder. A residual signal r(n) is first derived from the speech signal s(n) by short term filtering. Then, r(n) is processed to calculate a prediction error signal e(n) which is subsequently pulse excitation encoded. The processing of e(n) entails prediction of a residual by measuring a pitch related factor M, employing two steps. First calculating a coarse M value through peak clipping and sign transition detection, and then adjusting the M value by autocorrelation--calculations about the roughly spaced peaks.

Pitch detection for use in a predictive speech coder

A process for multirate subband encoding voice signals. At least a portion of the original signal bandwidth is split into p sub-bands the contents of which are to be coded using dynamic allocation of quantizing levels throughout the sub-bands. Let's assume the signal is to be coded for a set of coding rates R(1) &lt;R(2)&lt;. . .&lt;R(k) . . . &lt;R(K). First the number of quantizing levels 1(i, R(1)) for the lowest rate is determined. Then the levels for 1(i, R(k)), k=2, . . . , p are determined using: ##EQU1## Then the sub-band samples S(i,j) are quantized according to 1(i, R(K). Finally, the quantized sub-band samples are embedded into a multirate frame.

Growing community of inventors

Nice, France

Michele Rosso