EP1192621B1 - Audiokodierung mit harmonischen komponenten - Google Patents

Audiokodierung mit harmonischen komponenten Download PDF

Info

Publication number
EP1192621B1
EP1192621B1 EP00949623A EP00949623A EP1192621B1 EP 1192621 B1 EP1192621 B1 EP 1192621B1 EP 00949623 A EP00949623 A EP 00949623A EP 00949623 A EP00949623 A EP 00949623A EP 1192621 B1 EP1192621 B1 EP 1192621B1
Authority
EP
European Patent Office
Prior art keywords
frequency
cepstral
module
spectrum
audio signal
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Lifetime
Application number
EP00949623A
Other languages
English (en)
French (fr)
Other versions
EP1192621A1 (de
Inventor
François CAPMAN
Carlo Murgia
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nortel Networks France SAS
Original Assignee
Nortel Networks France SAS
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Nortel Networks France SAS filed Critical Nortel Networks France SAS
Publication of EP1192621A1 publication Critical patent/EP1192621A1/de
Application granted granted Critical
Publication of EP1192621B1 publication Critical patent/EP1192621B1/de
Anticipated expiration legal-status Critical
Expired - Lifetime legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
    • G10L19/02—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using spectral analysis, e.g. transform vocoders or subband vocoders
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
    • G10L19/04—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using predictive techniques
    • G10L19/08—Determination or coding of the excitation function; Determination or coding of the long-term prediction parameters
    • G10L19/10—Determination or coding of the excitation function; Determination or coding of the long-term prediction parameters the excitation function being a multipulse excitation

Definitions

  • the present invention relates to the field of signal coding. audio. It applies in particular, but not exclusively, to the coding of the speech, narrowband or wideband, in various rate ranges coding.
  • the main purpose of designing an audio codec is to provide good compromise between the flow rate transmitted by the encoder and the signal quality audio that the decoder is able to reconstruct from this stream.
  • the coder estimates a fundamental frequency of the signal, representing its pitch pitch
  • spectral analysis consists of determining parameters representing the harmonic structure of the signal at frequencies which are integer multiples of this fundamental frequency.
  • a modeling of the non-harmonic, or non-voiced component can also be performed in the spectral domain.
  • the parameters transmitted to the decoder represent typically the modulus of the spectrum of the voiced and unvoiced components. he is added information representing either voiced / unvoiced decisions relating to different portions of spectrum, i.e. information on the probability of signal voicing, allowing the decoder to determine in which portions of the spectrum it should use the voiced component or the unvoiced component.
  • coder families include coders of the MBE type (“Multi-Band Excitation”), or STC (“Sinusoidal” type encoders) Transform Coder ”).
  • MBE Multi-Band Excitation
  • STC Small Cellular Transform Coder
  • An object of the present invention is to improve the modeling of the signal spectrum module, in a coding scheme using analysis in the spectral domain.
  • the invention thus proposes a method for coding an audio signal, according to claim 1.
  • the invention also provides an audio coder comprising means arranged to carry out the above processes.
  • the encoder and decoder described below are digital circuits which can, as is usual in the field of signal processing audio, be produced by programming a digital signal processor (DSP) or a specific application integrated circuit (ASIC).
  • DSP digital signal processor
  • ASIC application integrated circuit
  • the audio coder shown in FIG. 1 processes an input audio signal x which, in the nonlimiting example considered below, is a speech signal.
  • the signal x is available in digital form, for example at a sampling frequency F e of 8 kHz. It is for example delivered by an analog-digital converter processing the amplified output signal of a microphone.
  • the input signal x can also be formed from another version, analog or digital, coded or not, of the speech signal.
  • the encoder includes a module 1 which forms successive frames audio signal for the different treatments performed, and a multiplexer output 6 which delivers an output stream ⁇ containing games for each frame quantization parameters from which a decoder will be able to synthesize a decoded version of the audio signal.
  • Each frame 2 is composed of a number N of consecutive samples of the audio signal x.
  • N 256
  • the module 1 multiplies the samples of each frame 2 by a windowing function f A , preferably chosen for its good spectral properties.
  • the encoder in Figure 1 performs an analysis of the audio signal in the spectral range. It includes a module 3 which calculates the transform of Fast Fourier (TFR) of each signal frame.
  • TFR Fast Fourier
  • the TFR 3 module obtains the signal spectrum for each frame, the module and phase of which are respectively denoted
  • a fundamental frequency detector 4 estimates for each signal frame a value of the fundamental frequency F 0 .
  • the detector 4 can apply any known method of analysis of the speech signal of the frame to estimate the fundamental frequency F 0 , for example a method based on the autocorrelation function or the AMDF function, possibly preceded by a whitening module. by linear prediction.
  • the estimation can also be performed in the spectral domain or in the cepstral domain.
  • Another possibility is to evaluate the time intervals between the consecutive breaks in the speech signal attributable to closures of the glottis of the intervening speaker during the duration of the frame.
  • Well-known methods which can be used to detect such micro-ruptures are described in the following articles: M.
  • the estimated fundamental frequency F 0 is subject to quantization, for example scalar, by a module 5, which provides the output multiplexer 6 with an index iF for quantifying the fundamental frequency for each frame of the signal.
  • the coder uses cepstral parametric models to represent an upper envelope and a lower envelope of the spectrum of the audio signal.
  • the first step of the cepstral transformation consists in applying to the signal spectrum module a spectral compression function, which can be a logarithmic or root function.
  • the compressed LX spectrum of the audio signal is processed by a module 9 which extract of the spectral amplitudes associated with the harmonics of the signal corresponding to the multiples of the estimated fundamental frequency F0. These amplitudes are then interpolated by a module 10 in order to obtain a compressed upper envelope marked LX_sup.
  • spectral compression could equivalent to be performed after determining the associated amplitudes to harmonics. It could also be done after the interpolation, which would only change the form of the interpolation functions.
  • the module 9 for extracting the maxima takes account of the possible variation of the fundamental frequency on the analysis frame, of the errors that the detector 4 can make, as well as of the inaccuracies linked to the discrete nature of the frequency sampling.
  • the search for the amplitudes of the spectral peaks does not simply consist in taking the values LX (i) corresponding to the indexes i such that iF e / 2N is the frequency closest to a harmonic of frequency kF 0 (k ⁇ 1 ).
  • the spectral amplitude retained for a harmonic of order k is a local maximum of the spectrum module in the vicinity of the frequency kF 0 (this amplitude is obtained directly in compressed form when the spectral compression 8 is carried out before the extraction of the maxima 9 ).
  • FIGS. 4 and 5 show an example of the shape of the compressed spectrum LX, where it can be seen that the maximum amplitudes of the harmonic peaks do not necessarily coincide with the amplitudes corresponding to the integer multiples of the estimated fundamental frequency F 0 .
  • the sides of the peaks being quite steep, a small positioning error of the fundamental frequency F 0 , amplified by the harmonic index k, can strongly distort the estimated upper envelope of the spectrum and cause poor modeling of the formantic structure of the signal.
  • the interpolation is performed between points whose abscissa is the frequency corresponding to the maximum the amplitude of a spectral peak, and whose ordinate is this maximum, before or after compression.
  • the interpolation performed to calculate the upper envelope LX_sup is a simple linear interpolation.
  • Another form interpolation could be used (eg polynomial or spline).
  • the interpolation is carried out between points whose abscissa is a frequency kF 0 multiple of the fundamental frequency (in fact the closest frequency in the discrete spectrum) and whose ordinate is the maximum amplitude, before or after compression, of the spectrum in the vicinity of this multiple frequency.
  • the maximum amplitude search interval associated with a harmonic of rank k is centered on the index i of the frequency of the TFR closest to kF 0 , that is to say where ⁇ a ⁇ denotes the integer equal to or immediately less than the number a.
  • the width of this search interval depends on the sampling frequency F e , the size 2N of the TFR and the range of possible variation of the fundamental frequency. This width is typically of the order of ten frequencies with the examples of values previously considered. It can be made adjustable as a function of the value F 0 of the fundamental frequency and of the number k of the harmonic.
  • a non-linear distortion of the frequency scale is operated on the upper envelope compressed by a module 12 before the module 13 performs the inverse fast Fourier transform (TFRI) providing the cepstral coefficients cx_sup.
  • TFRI inverse fast Fourier transform
  • the non-linear distortion makes it possible to more effectively minimize the modeling error. It is for example performed according to a Mel or Bark type frequency scale. This distortion may possibly depend on the estimated fundamental frequency F 0 .
  • Figure 1 illustrates the case of the Mel scale. The relationship between the frequencies F of the linear spectrum, expressed in hertz, and the frequencies F 'of the Mel scale is as follows:
  • TFRI 13 module needs to calculate only a cepstral vector of NCS cepstral coefficients of orders 0 to NCS-1.
  • NCS can be equal to 16.
  • a post-filtering in the cepstral domain is applied by a module 15 to the compressed upper envelope LX_sup.
  • This post-liftrage corresponds to a manipulation of the cepstral coefficients cx_sup delivered by the module of TRFI 13, which corresponds approximately to a post-filtering of the harmonic part of the signal by a transfer function having the classical form:
  • a (z) is the transfer function of a linear prediction filter of the audio signal
  • ⁇ 1 and ⁇ 2 are coefficients between 0 and 1
  • ⁇ is a possibly zero pre-emphasis coefficient.
  • a normalization module 16 further modifies the cepstral coefficients by imposing the exact modeling constraint of a point on the initial spectrum, which is preferably the most energetic point among the spectral maxima extracted by the module 9 In practice, this normalization only modifies the value of the coefficient c p (0).
  • the normalization module 16 operates as follows: it recalculates a value of the synthesized spectrum at the frequency of the maximum indicated by the module 9, by Fourier transform of the truncated and post-liftral cepstral coefficients, taking into account the non-distortion frequency axis linear; it determines a normalization gain g N by the logarithmic difference between the value of the maximum provided by the module 9 and this recalculated value; and he adds the gain g N to the post-liftrated cepstral coefficient c p (0). This normalization can be seen as part of post-liftrage.
  • the post-liftralized and standardized cepstral coefficients are the subject of a quantification by a module 18 which transmits quantification indexes corresponding to the output multiplexer 6 of the encoder.
  • the module 18 can operate by vector quantization from cepstral vectors formed from post-liftred and normalized coefficients, denoted here cx [n] for the signal frame of rank n.
  • cx [n] 16 cepstral coefficients cx [n, 0], cx [n, 1], ..., cx [n, NCS-1] is distributed in four sub - cepstral vectors each containing four coefficients of consecutive orders.
  • the cepstral vector cx [n] can be processed by the means shown in FIG. 6, which are part of the quantization module 18.
  • rcx [n] denotes a residual prediction vector for the frame of rank n whose components are respectively denoted rcx [n, 0], rcx [n, 1], ..., rcx [n, NCS-1], and ⁇ (i) denotes a prediction coefficient chosen to be representative of an assumed inter-frame correlation.
  • the numerator of the relation (10) is obtained by a subtractor 20, whose components of the output vector are divided by the quantities 2- ⁇ (i) at 21.
  • the residual vector rcx [n] is subdivided into four sub-vectors, corresponding to the subdivision into four cepstral sub-vectors.
  • the unit 22 proceeds to the vector quantization of each sub-vector of the residual vector rcx [n].
  • This quantification can consist, for each sub-vector srcx [n], in selecting from the dictionary the quantized sub-vector srcx_q [n] which minimizes the quadratic error ⁇ srcx [n] -srcx_q [n] ⁇ 2 .
  • the set icxs of the quantization indices icx, corresponding to the addresses in the dictionary or dictionaries of the residual quantized sub-vectors srcx_q [n] is supplied to the output multiplexer 6.
  • Unit 22 also delivers the values of the residual sub-vectors which form the vector rcx_q [n]. This is delayed by one frame in 23, and its components are multiplied by the coefficients ⁇ (i) at 24 for supply the vector to the negative input of subtractor 20. This last vector is on the other hand supplied to an adder 25, the other input of which receives a vector formed by the components of the quantized residue rcx_q [n] respectively multiplied by the quantities 1- ⁇ (i) at 26. The adder 25 thus delivers the quantified cepstral vector cx_q [n] that will recover the decoder.
  • the prediction coefficient ⁇ (i) can be optimized separately for each of the cepstral coefficients.
  • Quantification dictionaries can also be optimized separately for each four cepstral sub-vectors.
  • the adaptation module 29 controls the post-lifter 15 so as to minimize a modulus gap between the spectrum of the audio signal and the values of corresponding module calculated in 28.
  • This module deviation can be expressed by a sum of absolute values of differences in amplitudes, compressed or not, corresponding to one or more of the frequencies harmonics. This sum can be weighted according to the amplitudes spectral associated with these frequencies.
  • the modulus difference taken into account in the adaptation of the post-liftring would take into account all the harmonics of the spectrum.
  • the module 28 can resynthesize the spectral amplitudes only for one or more frequencies multiple of the fundamental frequency F 0 , selected on the basis of the size of the spectrum module in absolute value .
  • the adaptation module 29 can for example consider the three most intense spectral peaks in the calculation of the module deviation to be minimized.
  • the adaptation module 29 estimates a spectral masking curve of the audio signal using a model psychoacoustics, and the frequencies taken into account in the calculation of the deviation of modules to be minimized are selected based on the importance of the spectrum modulus relative to the masking curve (we can by example take the three frequencies for which the spectrum module more than the masking curve).
  • Different conventional methods can be used to calculate the masking curve from the audio signal. We can for example use the one developed by J.D. Johnston ("Transform Coding of Audio Signals Using Perceptual Noise Criteria ", IEEE Journal on Selected Area in Communications, Vol. 6, No. 2, February 1988).
  • the module 29 can use a filter identification model.
  • a simpler method consists in predefining a set of post-liftring parameter sets, that is to say a set of couples ⁇ 1 , ⁇ 2 in the case of a post-liftring according to relations (8), to carry out the operations incumbent on the modules 15, 16, 18 and 28 for each of these sets of parameters, and to retain that of the sets of parameters which leads to the minimum module deviation between the signal spectrum and the recalculated values.
  • the quantization indexes provided by the module 18 are then those which relate to the best set of parameters.
  • the coder determines coefficients cx_inf representing a compressed lower envelope LX_inf.
  • a module 30 extracts from the compressed spectrum LX spectral amplitudes associated with frequencies located in regions of the intermediate spectrum with respect to the frequencies multiple of the estimated fundamental frequency F 0 .
  • F 0 simply corresponds to the modulus of the spectrum for the frequency (k + 1/2) .F 0 located in the middle of the interval separating the two harmonics.
  • this amplitude could be an average of the spectrum modulus over a small range surrounding this frequency (k + 1/2) .F 0 .
  • a module 31 performs an interpolation, for example linear, of the spectral amplitudes associated with the frequencies located in the zones intermediaries to obtain the compressed lower envelope LX_inf.
  • the cepstral transformation applied to this lower envelope compressed LX_inf is performed according to a resulting frequency scale of a non-linear distortion applied by a module 32.
  • the TFRI module 33 calculates a cepstral vector of NCI cepstral coefficients cx_inf of orders 0 to NCI-1 representing the lower envelope.
  • the non-linear transformation of the frequency scale for the cepstral transformation of the lower envelope can be achieved towards a finer scale at high frequencies than at low frequencies, which advantageously makes it possible to model the non-voiced components of the signal at high frequencies.
  • cepstral coefficients cx_inf representing the lower envelope compressed are quantified by a module 34, which can operate the same so that the module 18 for quantifying cepstral coefficients representing the compressed upper envelope.
  • the vector thus formed is subjected to a vector quantization of prediction residue, performed by means identical to those shown in Figure 6 but without subdivision into sub-vectors.
  • the encoder shown in Figure 1 has no device particular for coding the phases of the spectrum to the harmonics of the audio signal.
  • it includes means 36-40 for coding a temporal information related to the phase of the non-harmonic component represented by the lower envelope.
  • a spectral decompression module 36 and a TFRI module 37 form a temporal estimate of the frame of the non-harmonic component.
  • the module 36 applies a reciprocal decompression function of the compression function applied by the module 8 (that is to say an exponential or a power 1 / ⁇ function) to the compressed lower envelope LX_inf produced by the module interpolation 31. This provides the modulus of the estimated frame of the non-harmonic component, the phase of which is taken to be equal to that ⁇ X of the spectrum of the signal X over the frame.
  • the inverse Fourier transform performed by the module 37 provides the estimated frame of the non-harmonic component.
  • Module 38 subdivides this estimated frame of the non-harmonic component in several time segments.
  • module 38 calculates the energy equal to the sum of the squares of the samples, and forms a vector E1 formed of eight positive real components equal to the eight calculated energies. Most of these eight energies, denoted EM, is also determined to be supplied, with the vector E1, to a normalization module 39. This divides each component of the vector E1 by EM, so that the normalized vector Emix is made up of eight components between 0 and 1. It is this vector normalized Emix, or weighting vector, which is subject to quantification by module 40. This can perform vector quantization with a dictionary determined during a prior learning. The index of iEm quantification is provided by module 40 to output multiplexer 6 of the encoder.
  • FIG. 7 shows an alternative embodiment of the means used by the coder of FIG. 1 to determine the vector Emix of energy weighting of the frame of the non-harmonic component.
  • the spectral decompression and TFRI modules 36, 37 operate like those which have the same references in FIG. 1.
  • a selection module 42 is added to determine the value of the spectrum module subjected to the inverse Fourier transform 37. On the based on the estimated fundamental frequency F 0 , the module 42 identifies harmonic regions and non-harmonic regions of the spectrum of the audio signal.
  • a frequency will be considered to belong to a harmonic region if it is in a frequency interval centered on a harmonic kF 0 and of width corresponding to a width of spectral line synthesized, and to a non-harmonic region otherwise.
  • the complex signal subjected to TFRI 37 is equal to the value of the spectrum, that is to say that its modulus and its phase correspond to the values
  • this complex signal has the same phase ⁇ X as the spectrum and a module given by the lower envelope after spectral decompression 36. This procedure according to Figure 7 provides more precise modeling of non-harmonic regions.
  • the decoder represented in FIG. 8 comprises an input demultiplexer 45 which extracts from the bit stream ⁇ , coming from an encoder according to FIG. 1, the indexes iF, icxs, icxi, iEm for quantifying the fundamental frequency F 0 , cepstral coefficients representing the compressed upper envelope, coefficients representing the compressed lower envelope, and the weighting vector Emix, and distributes them respectively to modules 46, 47, 48 and 49.
  • These modules 46-49 include quantization dictionaries similar to those of modules 5, 18, 34 and 40 of FIG. 1, in order to restore the values of the quantized parameters.
  • the modules 47 and 48 have dictionaries to form the quantized prediction residues rcx_q [n], and they deduce therefrom the quantized cepstral vectors cx_q [n] with elements identical to the elements 23-26 of FIG. 6. These quantified cepstral vectors cx_q [n] provide the cepstral coefficients cx_sup_q and cx_inf_q processed by the decoder.
  • a module 51 calculates the fast Fourier transform of the cepstral coefficients cx_sup for each signal frame.
  • the frequency scale of the resulting compressed spectrum is modified non-linearly by a module 52 applying the reciprocal non-linear transformation of that of module 12 of FIG. 1, and which provides the estimate LX_sup of the compressed upper envelope .
  • a spectral decompression of LX_sup operated by a module 53, provides the upper envelope X_sup comprising the estimated values of the module of the spectrum at frequencies multiple of the fundamental frequency F 0 .
  • the module 54 synthesizes the spectral estimate X v of the harmonic component of the audio signal, by a sum of spectral lines centered on the frequencies multiple of the fundamental frequency F 0 and whose amplitudes (in module) are those given by the envelope superior X_sup.
  • the decoder of figure 8 is capable of extracting information on this phase from the cepstral coefficients cx_sup_q representing the compressed upper envelope. This phase information is used to assign a phase ⁇ (k) to each of the spectral lines determined by the module 54 in the estimation of the harmonic component of the signal.
  • the speech signal can be considered to be at minimum phase.
  • the minimum phase information can easily be deduced from a cepstral modeling. This minimum phase information is therefore calculated for each harmonic frequency.
  • the minimum phase assumption means that the energy of the synthesized signal is localized at the start of each period of the fundamental frequency F 0 .
  • module 56 deduces post-liftrated cepstral coefficients and smoothed the phase minimum assigned to each spectral line representing a harmonic peak of the spectrum.
  • the operations performed by the modules 56, 57 for smoothing and extracting the minimum phase are illustrated by the flowchart in FIG. 9.
  • the module 56 examines the variations of the cepstral coefficients to apply a lesser smoothing in the presence of sudden variations only in the presence of slow variations. For this, it performs the smoothing of cepstral coefficients by means of a forgetting factor ⁇ c chosen as a function of a comparison between a threshold d th and a distance d between two successive sets of post-liftrated cepstral coefficients.
  • the threshold d th is itself adapted as a function of variations in cepstral coefficients.
  • the first step 60 consists in calculating the distance d between the two successive vectors relating to frames n-1 and n. These vectors, noted here cxp [n-1] and cxp [n], correspond for each frame to all the NCS post-liftrated cepstral coefficients representing the upper envelope compressed.
  • the distance used can in particular be the Euclidean distance between the two vectors or a quadratic distance.
  • Two smoothings are first carried out, respectively by means of forgetting factors ⁇ min and ⁇ max , to determine a minimum distance d min and a maximum distance d max .
  • the forgetting factors ⁇ min and ⁇ max are themselves selected from two distinct values, respectively ⁇ min1 , ⁇ min2 and ⁇ max1 , ⁇ max2 between 0 and 1, the indices ⁇ min1 , ⁇ max1 each being substantially closer to 0 than the indices ⁇ min2 , ⁇ max2 . If d> d min (test 61), the forget factor ⁇ min is equal to ⁇ min1 (step 62); otherwise it is taken equal to ⁇ min2 (step 63). In step 64, the minimum distance d min is taken equal to ⁇ min .d min + (1- ⁇ min ) .d.
  • step 66 the forgetting factor ⁇ max is equal to ⁇ max1 (step 66); otherwise it is taken equal to ⁇ max2 (step 67).
  • step 68 the minimum distance d max is taken equal to ⁇ max .d max + (1- ⁇ max ) .d.
  • step 72 If the distance d between the two consecutive cepstral vectors is greater than the threshold d th (test 71), a value ⁇ c1 relatively close to 0 is adopted for the forget factor ⁇ c (step 72). In this case, the corresponding signal is considered to be of the non-stationary type, so that there is no need to keep a large memory of the previous cepstral coefficients. If d ⁇ d th , in step 73 we adopt for the forget factor ⁇ c a value ⁇ c2 less close to 0 in order to further smooth the cepstral coefficients.
  • the module 57 then calculates the minimum phases ⁇ (k) associated with the harmonics kF 0 .
  • the minimum phase for a harmonic of order k is given by: where cxl [n, m] denotes the smooth cepstral coefficient of order m for the frame n.
  • step 75 the harmonic index k is initialized to 1.
  • the phase ⁇ (k) and the cepstral index m are initialized respectively at 0 and 1 in step 76.
  • the module 57 adds to phase ⁇ (k) the quantity -2.cxl [n, m] .sin (2 ⁇ mk.F 0 / F e ).
  • the cepstral index m is incremented in step 78 and compared to NCS in step 79. Steps 77 and 78 are repeated as long as m ⁇ NCS.
  • m NCS
  • the calculation of the minimum phase is completed for the harmonic k, and the index k is incremented in step 80.
  • the calculation of minimum phases 76-79 is repeated for the following harmonic as long as kF 0 ⁇ F e / 2 (test 81).
  • the module 54 holds account of a constant phase over the width of each spectral line, equal to the minimum phase ⁇ (k) supplied for the corresponding harmonic k by the module 57.
  • the estimate X v of the harmonic component is synthesized by summing spectral lines positioned at the harmonic frequencies of the fundamental frequency F 0 .
  • the spectral lines can be positioned on the frequency axis with a resolution greater than the resolution of the Fourier transform. For that, one precalculates once and for all a reference spectral line according to the higher resolution. This calculation can consist of a Fourier transform of the analysis window f A with a transform size of 16384 points, providing a resolution of 0.5 Hz per point.
  • each harmonic line is then carried out by the module 54 by positioning the reference line at high resolution on the frequency axis, and by sub-sampling this reference spectral line to reduce to the resolution of 16.625 Hz of the Fourier transform on 512 points. This allows to precisely position the spectral line.
  • the TFR 85 module of the decoder of figure 8 receives the NCI quantified cepstral coefficients cx_inf_q of orders 0 to NCI - 1, and it supplements them advantageously by the NCS - NCI cepstral coefficients cx_sup_q of order NCI to NCS - 1 representing the upper envelope. Indeed, we can estimate first approximation as the rapid variations of the compressed lower envelope are well reproduced by those of the compressed upper envelope. In another embodiment, the TFR 85 module could only consider NCIs cepstral parameters cx_inf_q.
  • Module 86 converts frequency scale reciprocally of the conversion carried out by the module 32 of the coder, in order to restore the LX_inf estimate of the compressed lower envelope, submitted to the spectral decompression 87.
  • the decoder has of a lower envelope X_inf comprising the values of the spectrum module in the valleys between the harmonic peaks.
  • This envelope X_inf will modulate the spectrum of a noise frame whose phase is processed as a function of the quantized weighting vector Emix extracted by the module 49.
  • a generator 88 delivers a normalized noise frame whose 4 ms segments are weighted in a module 89 in accordance with the normalized components of the Emix vector provided by the module 49 for the current frame.
  • This noise is a high-pass filtered white noise to take account of the low level which in principle the unvoiced component has at low frequencies.
  • the Fourier transform of the resulting frame is calculated by the TFR module 91.
  • the spectral estimate X uv of the non-harmonic component is determined by the spectral synthesis module 92 which performs frequency-by-frequency weighting. This weighting consists in multiplying each complex spectral value supplied by the TFR module 91 by the value of the lower envelope X_inf obtained for the same frequency by the spectral decompression module 87.
  • the spectral estimates X v , X uv of the harmonic (voiced in the case of a speech signal) and non-harmonic (or non-voiced) components are combined by a mixing module 95 controlled by a module 96 for analyzing the degree of harmony (or voicing) of the signal.
  • the analysis module 96 comprises a unit 97 for estimating a degree of voicing W dependent on the frequency, from which four gains dependent on are calculated.
  • the frequency namely two gains g v , g uv controlling the relative importance of the harmonic and non-harmonic components in the synthesized signal, and two gains g v_ ⁇ , g uv_ ⁇ used to noise the phase of the harmonic component.
  • the degree of voicing W (i) is a continuously variable value between 0 and 1 determined for each frequency index i (0 ⁇ i ⁇ N) as a function of the upper envelope X_sup (i) and the lower envelope X_inf (i) obtained for this frequency i by the decompression modules 53, 87.
  • the threshold Vth (F 0 ) corresponds to the average dynamics calculated on a synthetic spectrum purely voiced at the fundamental frequency. It is advantageously chosen depending on the fundamental frequency F 0 .
  • the degree of voicing W (i) for a frequency other than the harmonic frequencies is obtained simply as being equal to that estimated for the nearest harmonic.
  • the gain g v (i), which depends on the frequency, is obtained by applying a non-linear function to the degree of voicing W (i) (block 98).
  • This non-linear function has for example the form represented in FIG. 11: the thresholds W1, W2 being such that 0 ⁇ W1 ⁇ W2 ⁇ 1.
  • phase ⁇ ' v of the mixed harmonic component is the result of a linear combination of the phases ⁇ v , ⁇ uv of the harmonic and non-harmonic components X v , X uv synthesized by the modules 54, 92.
  • the gains g v_ ⁇ , g uv_ ⁇ respectively applied to these phases are calculated from the degree of voicing W and also weighted according to the frequency index i, since the sound effects of the phase are only really useful beyond of a certain frequency.
  • a first gain g v1_ ⁇ is calculated by applying a non-linear function to the degree of voicing W (i), as shown diagrammatically by block 100 in FIG. 10.
  • This non-linear function can have the form represented in FIG. 12: the thresholds W3 and W4 being such that 0 ⁇ W3 ⁇ W4 ⁇ 1, and the minimum gain G1 being between 0 and 1.
  • a multiplier 101 multiplies for each index frequency i the gain g v1_ ⁇ by another gain g v2_ ⁇ depending only on the frequency index i, to form the gain g v_ ⁇ (i).
  • the gain g v2_ ⁇ (i) depends non-linearly on the frequency index i, for example as shown in Figure 13: the indices i1 and i2 being such that 0 ⁇ i1 ⁇ i2 ⁇ N, and the minimum gain G2 being between 0 and 1.
  • . exp [j ⁇ ' v (i)] + g uv (I). X uv (I) with ⁇ ' v (i) g v_ ⁇ (I). ⁇ v (i) + g uv_ ⁇ (I). ⁇ uv (I) where ⁇ v (i) denotes the argument of the complex number X v (i) supplied by the module 54 for the frequency of index i (block 104 of FIG.
  • ⁇ uv (i) denotes the argument of complex number X uv (i) supplied by the module 92 (block 105 of FIG. 10). This combination is carried out by the multipliers 106-110 and the adders 111-112 shown in FIG. 10.
  • the frames successively obtained in this way are finally processed by the time synthesis module 116 which forms the audio signal decoded x and.
  • the temporal synthesis module 116 performs a sum at overlapping of frames modified with respect to those successively evaluated at the output of module 115.
  • the modification can be seen in two steps illustrated respectively in FIGS. 14 and 15.
  • the first step (FIG. 14) consists in multiplying each frame 2 ′ delivered by the TFRI module 115 by a window 1 / f A opposite to the analysis window f A used by the module 1 of the coder.
  • the resulting 2 "frame samples are therefore weighted uniformly.
  • each sample of the decoded audio signal x and thus obtained is assigned a uniform overall weight, equal to A.
  • This overall weight comes from the contribution of a single frame if the sample has in this frame a rank i such that L ⁇ i ⁇ N - L, and comprises the summed contributions of two successive frames if 0 ⁇ i ⁇ L where N - L ⁇ i ⁇ N.
  • FIG. 16 shows the appearance of the compound window f C in the case where the analysis window f A is a Hamming window and the synthesis window f S has the form given by the relations (19) to (21) .
  • the encoder of the Figure 1 can increase the rate of training and analysis of the frames, so transmit more quantization parameters to the decoder.
  • a frame of N 256 samples (32 ms) is formed every 20 ms.
  • the notations cx_q [n-1] and cx_q [n] denote determined quantified cepstral vectors, for two successive frames of entire rank, by the quantization module 18 and / or by the quantization module 34. These vectors include for example four consecutive cepstral coefficients each. They could also understand more cepstral coefficients.
  • a module 120 performs an interpolation of these two vectors cepstraux cx_q [n-1] and cx_q [n], in order to estimate an intermediate value cx_i [n-1/2].
  • the interpolation performed by the module 120 can be a simple arithmetic mean of the vectors cx_q [n-1] and cx_q [n].
  • the module 120 could apply a more sophisticated interpolation formula, for example polynomial, also based on cepstral vectors obtained for frames prior to frame n-1.
  • the interpolation takes into account the relative position of each interpolated frame.
  • the coder calculates also the cepstral coefficients cx [n-1/2] relative to the half-integer row frame.
  • these cepstral coefficients are those provided by the TFRI module 13 after post-liftrage 15 (for example with the same post-lifter coefficients as for the previous frame n-1) and normalization 16.
  • the coefficients cepstraux cx [n-1/2] are those delivered by the TFRI 33 module.
  • a subtractor 121 forms the difference ecx [n-1/2] between the cepstral coefficients cx [n-1/2] calculated for the half-integer row frame and the coefficients cx_i [n-1/2] estimated by interpolation.
  • This difference is supplied to a quantization module 122 which addresses indexes of quantization icx [n-1/2] at the output multiplexer 6 of the encoder.
  • Module 122 works for example by vector quantization of interpolation errors ecx [n-1/2] successively determined for half-integer rank frames.
  • This quantification of the interpolation error can be carried out by the coder for each of the NCS + NCI cepstral coefficients used by the decoder, or only for some of them, typically those of orders the smallest.
  • the corresponding means of the decoder are illustrated in FIG. 19.
  • the decoder essentially functions as that described with reference to the Figure 8 to determine whole rank signal frames.
  • a module interpolation 124 identical to the module 120 of the coder estimates the coefficients intermediaries cx_i [n-1/2] from the quantized coefficients cx_q [n-1] and cx_q [n] provided by module 47 and / or module 48 from the icxs, icxi indexes extracts from the flow ⁇ .
  • a parameter extraction module 125 receives the index of quantization icx [n-1/2] from the input demultiplexer 45 of the decoder, and deduces the quantized interpolation error ecx_q [n-1/2] from the same quantization dictionary as that used by the module 122 of the coder.
  • a adder 126 sums the cepstral vectors cx_i [n-1/2] and ecx_q [n-1/2] in order to provide the cepstral coefficients cx [n-1/2] which will be used by the decoder (modules 51-57, 95, 96, 115 and / or modules 85-87, 92, 95, 96, 115) to form the interpolated frame of rank n-1/2.
  • the decoder can also interpolate the other parameters F 0 , Emix used to synthesize the signal frames.
  • the fundamental frequency F 0 can be interpolated linearly, either in the time domain, or (preferably) directly in the frequency domain.
  • the interpolation should be carried out after denormalization and of course taking account of the time offsets between frames.
  • the coder uses the cepstral vectors cx_q [n], cx_q [n-1], ..., cx_q [n-r] and cx_q [n-1/2] calculated for the last frames passed (r ⁇ 1) to identify an interpolator filter optimal which, when subject to the quantified cepstral vectors cx_q [n-r], ..., cx_q [n] relative to the frames of whole rank, delivers a cepstral vector interpolated cx_i [n-1/2] which has a minimum distance from the vector cx [n-1/2] calculated for the last frame of half-integer rank.
  • this interpolator filter 128 is present in the encoder, and a subtractor 129 subtracts its output cx_i [n-1/2] of the calculated cepstral vector cx [n-1/2].
  • a minimization module 130 determines the parameter set ⁇ P ⁇ of the interpolator filter 128, for which the ecx interpolation error [n-1/2] delivered by the subtractor 129 presents a minimum standard.
  • This set of parameters ⁇ P ⁇ is addressed to a module of quantization 131 which provides a corresponding quantization index iP to encoder output multiplexer 6.
  • the decoder From the iP quantification indices of the parameters ⁇ P ⁇ obtained in the bit stream ⁇ , the decoder reconstructs the interpolator filter 128 (at errors quantization), and processes the spectral vectors cx_q [n-r], ..., cx_q [n] in order to estimate the cepstral coefficients cx [n-1/2] used to synthesize the half-full row frames.
  • the decoder can use a method simple interpolation (without transmission of parameters from the encoder for half-integer rows), an interpolation method with tap taking into account a quantized interpolation error (according to FIGS. 17 and 18), or an interpolation method with an optimal interpolator filter (according to the figure 19) to evaluate half-full row frames in addition to row frames integer evaluated directly as explained with reference to Figures 8 to 13.
  • the time synthesis module 116 can then combine all of these frames evaluated to form the synthesized signal x and in the manner explained below with reference to Figures 14, 21 and 22.
  • the module 116 performs an overlap sum of modified frames with respect to those successively evaluated at the output of the module 115, and this modification can be seen in two stages, the first of which is identical to that previously described with reference to FIG. 14 (divide the samples of the frame 2 'by the analysis window f A ).
  • the synthesis window f ' S (i) increases progressively for i going from N / 2 - M / p to N / 2. It is for example a raised sinusoid over the interval N / 2 - M / p ⁇ i ⁇ N / 2 + M / p.
  • the synthesis window f ′ S can be, over this interval, a Hamming window (as shown in FIG. 21) or a Hanning window.
  • FIG. 21 shows the successive 2 "frames repositioned in the time by module 116.
  • the hatching indicates the portions eliminated from the frames (summary window at 0).
  • property (25) ensures a uniform weighting of the samples of the synthesized signal.
  • FIG. 22 shows the shape of the compound window f ' C in the case where the windows f A and f' S are of the Hamming type.
  • Interpolated frames may be subject to reduced transmission coding parameters, as described above, but this is not mandatory.
  • This embodiment makes it possible to keep an interval M relatively large between two analysis frames, and therefore limit the throughput of transmission required, while limiting the discontinuities likely to appear due to the size of this interval compared to time scales typical of variations in audio signal parameters, including cepstral coefficients and the fundamental frequency.
  • Figures 23 to 25 show other embodiments of the means used to process the cepstral coefficients cx_sup delivered by the TFRI module 13 of FIG. 1, representing the upper envelope.
  • post-lifter 15 standardization 16, quantization 18 and calculation of the spectral amplitudes 28 are essentially identical to those previously described with reference to the Figure 1.
  • post-lifter 140, smoothing 141 and minimum phase extraction 142 are provided to process the coefficients post-liftrated and quantified cepstrals cx_sup_q delivered by the module quantization 18.
  • These 140-142 modules essentially operate like the corresponding modules 55-57 of the decoder of figure 8.
  • the adaptation module 144 performs a function similar to that of the module 29 in FIG. 1. But the adaptation is not carried out solely on the basis of the spectrum module.
  • the module 144 determines the best set of coefficients for the post-lifter 15 by minimizing the difference between the spectrum of the audio signal, in module
  • the modules of these latter complex values are given by the calculation module 28, and their phases correspond to the minimum phases ⁇ (k) supplied by the extraction module 142.
  • the module 144 can take into account any appropriate distance in the complex plane, for example the Euclidean distance.
  • the adaptation of the post-lifter 15 by the module 144 takes account in a combined manner of frequency aspects of the signal, reflected by the spectrum module, and temporal aspects, reflected by the phase of the spectrum.
  • the post-lifter 140 can also be adaptive, the adaptation carried out by the module 144 bearing jointly on the two post-lifters 15, 140.
  • the post-lifter 55 of the decoder (FIG. 8) is adapted, like the post-lifter 140, as a function of iLif parameters that the adaptation module 144 supplies to the multiplexer 6 so that it includes them in the digital stream ⁇ .
  • a few sets of coefficients ⁇ 1 , ⁇ 2 are provided for the post-lifter 140 and 55, and the module 144 performs an exhaustive test of these different sets of coefficients to retain the one which minimizes the difference in the complex plane.
  • the adaptation module 29 of the post-lifter 15 is identical to that of FIG. 1.
  • FIG. 24 shows a module 145 for estimating a masking curve allowing module 29 to select, for the minimization of the difference in module, the frequency or frequencies harmonics that extend more than the masking curve calculated on the base of the spectrum in module
  • the post-lifter 140 in FIG. 24 is adapted separately by a module 146 which minimizes the differences between the phase ⁇ X of the signal spectrum and the minimum phase ⁇ (k) calculated by the module 142 for one or more of the harmonics .
  • the harmonics selected for the calculation of the minimized phase difference can be chosen as a function of the masking curve estimated by the module 145.
  • the module 146 supplies the output multiplexer 6 of the coder with the iLif parameters which represent the optimal post-lifter 140, so that they are used in the post-filter 55 of the decoder.
  • the post-lifter 140 used in the calculation of the minimum phases is not adaptive.
  • the minimum phases ⁇ (k) calculated by the module 142 for the harmonics of the fundamental frequency are compared with the phases ⁇ X of the spectrum of the audio signal, and the phase difference is the subject of a quantification by a module 148.
  • the corresponding quantization indices i ⁇ are supplied by the module 148 to the output multiplexer 6 of the encoder.
  • a module 149 uses these quantization indices i ⁇ supplied by the demultiplexer 45 to obtain the values of the quantized phase differences, which an adder 150 adds to the minimum phases ⁇ (k) calculated by module 57 (the post-lifter 140 and 55 being identical). The phases provided by the adder 150 are then used by the module 54 which synthesizes the spectral lines of the harmonic component X v .
  • the peak (s) for which the phase difference is can be chosen according to the spectral energy represented by the upper envelope, which is available to the coder and the decoder, which allows the decoder to determine at which spectral line the deviations must be applied.
  • the phase differences can be the subject of a scalar, or vector quantization if they are grouped for several peaks.
  • the variation of the phase ⁇ X around the minimum phase ⁇ (k) over the width of a harmonic peak can be represented simply by the slope of a linear segment selected as being that which presents a minimum quadratic distance with the phase variation curve of the spectrum over the width of the line, and possibly by an offset at the origin.
  • slopes can be scaled, or vector if we group them for several peaks.
  • Quantification of phase variations on harmonic peaks may relate to all harmonic frequencies. Another possibility is to quantify several slopes each obtained by averaging the harmonic slopes on one or more sub-bands of the spectrum. This averaging can be weighted to take into account relative energies at different harmonic frequencies, represented by the envelope higher.
  • the module 148 can also model the phase variation on the width of a peak by a curve more complex than a linear segment, by example a spline, whose parameters are quantified to be transmitted to the decoder.
  • harmonic phase models representative of phase variations across the width of the peaks, observed on a corpus of reference signals. These models are stored in a dictionary memorized by modules 148 and 149.
  • the encoder module 148 determines the i ⁇ indexes corresponding to the addresses of models closest to phase variations in the vicinity harmonic peaks considered, and the decoder module 149 recovers these models for the synthesis of the phase of the harmonic component.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Computational Linguistics (AREA)
  • Signal Processing (AREA)
  • Health & Medical Sciences (AREA)
  • Spectroscopy & Molecular Physics (AREA)
  • Human Computer Interaction (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Compression, Expansion, Code Conversion, And Decoders (AREA)
  • Stereophonic System (AREA)
  • Cereal-Derived Products (AREA)
  • Reduction Or Emphasis Of Bandwidth Of Signals (AREA)

Claims (22)

  1. Verfahren zum Codieren eines Audiosignals (x), bei dem eine Grundfrequenz (F0) des Audiosignals geschätzt wird, ein Spektrum des Audiosignals durch eine Transformation eines Blocks des Audiosignals in den Frequenzbereich bestimmt wird und einem digitalen Ausgangsdatenstrom (Φ) Daten zum Codieren einer harmonischen Komponente des Audiosignals hinzugefügt werden, die Daten umfassen, die für Spektralamplituden, die Vielfachfrequenzen der Grundfrequenz zugeordnet sind, repräsentativ sind, bei dem die Spektralamplitude, die einer der Vielfachfrequenzen der Grundfrequenz zugeordnet ist, ein lokales Maximum des Moduls des Spektrums in der Nähe der Vielfachfrequenz ist, und bei dem die Daten, die für Spektralamplituden, die Vielfachfrequenzen der Grundfrequenz (F0) zugeordnet sind, repräsentativ sind, mittels Cepstralkoeffizienten (cx_sup) erzielt werden, die durch Transformieren einer komprimierten oberen Einhüllenden (LX_sup) des Spektrums des Audiosignals in den Cepstralbereich berechnet werden.
  2. Verfahren nach Anspruch 1,
    bei dem die komprimierte obere Einhüllende (LX_sup) durch Interpolation der den Vielfachfrequenzen der Grundfrequenz (F0) zugeordneten Spektralamplituden unter Anwendung einer Funktion zur Spektralkompression bestimmt wird.
  3. Verfahren nach Anspruch 2,
    bei dem die Interpolation zwischen Punkten durchgeführt wird, deren Abszisse eine Vielfachfrequenz der Grundfrequenz (F0) ist und deren Ordinate die komprimierte oder nicht komprimierte Spektralamplitude ist, die dieser Vielfachfrequenz zugeordnet ist,.
  4. Verfahren nach einem der vorhergehenden Ansprüche,
    bei dem die Transformation der komprimierten oberen Einhüllenden (LX_sup) in den Cepstralbereich gemäß einer nicht linearen Frequenzskala durchgeführt wird.
  5. Verfahren nach einem der vorhergehenden Ansprüche,
    bei dem die Cepstralkoeffizienten (cx_sup) quantisiert werden, um die Daten zu bilden, die für die den Vielfachfrequenzen der Grundfrequenz (F0) zugeordneten Spektralamplituden repräsentativ sind.
  6. Verfahren nach Anspruch 5,
    bei dem die Quantisierung der Cepstralkoeffizienten (cx_sup) einen Prädiktionsrestwert für jeden der Cepstralkoeffizienten betrifft.
  7. Verfahren nach Anspruch 6,
    bei dem der Prädiktionsrestwert für einen Cepstralkoeffizienten die Form (cx[n,i] - α(i).rcx_q[n - 1,i])/[2 - α(i)] hat, wobei cx[n,i] einen laufenden Wert des Cepstralkoeffizienten bezeichnet, rcx_q[n-1,i] einen vorhergehenden Wert des quantisierten Prädiktionsrestwerts und α(i) einen Prädiktionskoeffizienten bezeichnet.
  8. Verfahren nach Anspruch 6 oder 7,
    bei dem unterschiedliche Prädiktoren verwendet werden, um die Prädiktionsrestwerte für wenigstens zwei der Cepstralkoeffizienten zu bestimmen.
  9. Verfahren nach einem der Ansprüche 5 bis 8,
    bei dem die Cepstralkoeffizienten (cx_sup) in mehreren cepstralen Untervektoren aufgeteilt sind, die durch eine Vektorquantisierung getrennt quantisiert werden, die einen Prädiktionsrestwert der Cepstralkoeffizienten betrifft.
  10. Verfahren nach einem der Ansprüche 5 bis 9,
    bei dem die Cepstralkoeffizienten (cx_sup) vor der Quantisierung normiert werden, indem der Cepstralkoeffizient der Größe 0 derart verändert wird, dass die einer Vielfachfrequenz der Grundfrequenz (F0) zugeordnete Spektralamplitude genau durch die normierten Cepstralkoeffizienten gebildet wird.
  11. Verfahren einem der Ansprüche 5 bis 10,
    bei dem die Cepstralkoeffizienten (cx_sup) vor ihrer Quantisierung durch Lifterung in den Cepstralbereich transformiert werden.
  12. Verfahren nach Anspruch 11,
    bei dem die Lifterung die Form cp(i) = [1 + y2 i - y1 i].c(i) - (µi/i) hat, wobei cp(i) und c(i) den Cepstralkoeffizienten mit i>0 vor bzw. nach Lifterung bezeichnen, y1 und y2 Koeffizienten zwischen 0 und 1 sind und µ ein Vorentzerrungskoeffizient ist.
  13. Verfahren nach Anspruch 12,
    bei dem µ = (y2 - y1).c(1).
  14. Verfahren nach einem der Ansprüche 11 bis 13,
    bei dem ein Wert des Moduls des Audiosignalspektrums auf der Basis der transformierten und quantisierten Cepstralkoeffizienten (cx_sup_q) auf wenigstens eine Vielfachfrequenz der Grundfrequenz (F0) umgerechnet wird, und die Lifterung so angepasst wird, dass eine Modulabweichung zwischen dem Spektrum des Audiosignals und wenigstens einem Wert des umgerechneten Moduls minimiert wird.
  15. Verfahren nach einem der Ansprüche 11 bis 13,
    bei dem ein Wert des Moduls des Audiosignalspektrums auf der Basis der transformierten und quantisierten Cepstralkoeffizienten (cx_sup_q) auf wenigstens eine Vielfachfrequenz der Grundfrequenz (F0) umgerechnet wird, die Cepstralkoeffizienten durch Lifterung und Glättung in den Cepstralbereich retransformiert werden, Minimalphasen (ϕ(k)) des Audiosignals auf der Basis der retransformierten Cepstralkoeffizienten (cxl[n]) auf Vielfachfrequenzen der Grundfrequenz berechnet werden, und die vor der Quantisierung durchgeführte Lifterung so angepasst wird, dass eine Abweichung zwischen dem Spektrum des Audiosignals und wenigstens einem komplexen Wert minimiert wird, dessen Modul einen auf eine Vielfachfrequenz der Grundfrequenz umgerechneten Wert hat und dessen Phase durch die für diese Vielfachfrequenz berechnete Minimalphase gegeben ist.
  16. Verfahren nach Anspruch 15,
    bei dem die vor und nach der Quantisierung durchgeführten Lifterungen zusammen derart angepasst werden, dass diese Abweichung minimiert wird, und bei dem den Daten zur Codierung der harmonischen Komponente Parameter (iLif) hinzugefügt werden, die für die nach der Quantisierung durchgeführte angepasste Lifterung repräsentativ sind.
  17. Verfahren nach einem der Ansprüche 14 bis 16,
    bei dem die zur Anpassung der Lifterung minimierte Abweichung sich auf wenigstens eine Vielfachfrequenz der Grundfrequenz (F0) bezieht, die auf der Basis der Bedeutung des Moduls des Spektrums in Absolutzahlen ausgewählt ist.
  18. Verfahren nach einem der Ansprüche 14 bis 16,
    bei dem mittels eines psychoakustischen Modells eine Kurve der Spektralmaskierung des Audiosignals geschätzt wird, und sich die zur Anpassung der Lifterung minimierte Abweichung auf wenigstens eine Vielfachfrequenz der Grundfrequenz (F0) bezieht, die auf der Basis der Bedeutung des Moduls des Spektrums in Bezug auf die Maskierungskurve ausgewählt ist.
  19. Verfahren nach Anspruch 1,
    bei dem das Spektrum des Audiosignals und die aus der Transformation der komprimierten oberen Einhüllenden resultierenden Cepstralkoeffizienten (cx_sup) für aufeinander folgende Blöcke von N Abtastwerten des Audiosignals bestimmt werden, die gegenseitige Überlappungen aufweisen, und bei dem die Daten, die für Spekralamplituden, die den Vielfachfrequenzen der geschätzten Grundfrequenz (F0) zugeordnet sind, repräsentativ sind und mittels der durch Transformieren der komprimierten oberen Einhüllenden berechneten Cepstralkoeffizienten erzielt werden, dem digitalen Ausgangsdatenstrom (Φ) bei nur einer einzigen Untergruppe der Blöcke hinzugefügt werden.
  20. Verfahren nach Anspruch 19,
    bei dem für die nicht zu dieser Untergruppe gehörenden Blöcke dem digitalen Ausgangsdatenstrom (Φ) Daten (icx[n-1/2]) zur Quantisierung eines Interpolationsfehlers (ecx[n-1/2]) der aus der Transformation der komprimierten oberen Einhüllenden (LX_sup) resultierenden Cepstralkoeffizienten hinzugefügt werden.
  21. Verfahren nach Anspruch 19,
    bei dem für die nicht zu dieser Untergruppe gehörenden Blöcke ein optimaler Interpolationsfilter (128) für die aus der Transformation der komprimierten oberen Einhüllenden (LX_sup) resultierenden Cepstralkoeffizienten bestimmt wird und dem digitalen Ausgangsdatenstrom (Φ) Daten (iP) hinzugefügt werden, die diesen optimalen Interpolationsfilter darstellen.
  22. Audio-Codierer mit Mitteln, die zum Durchführen eines Verfahrens nach einem der vorhergehenden Ansprüche angeordnet sind.
EP00949623A 1999-07-05 2000-07-04 Audiokodierung mit harmonischen komponenten Expired - Lifetime EP1192621B1 (de)

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
FR9908634 1999-07-05
FR9908634A FR2796190B1 (fr) 1999-07-05 1999-07-05 Procede et dispositif de codage audio
PCT/FR2000/001908 WO2001003120A1 (fr) 1999-07-05 2000-07-04 Codage audio avec composants harmoniques

Publications (2)

Publication Number Publication Date
EP1192621A1 EP1192621A1 (de) 2002-04-03
EP1192621B1 true EP1192621B1 (de) 2004-09-22

Family

ID=9547703

Family Applications (1)

Application Number Title Priority Date Filing Date
EP00949623A Expired - Lifetime EP1192621B1 (de) 1999-07-05 2000-07-04 Audiokodierung mit harmonischen komponenten

Country Status (6)

Country Link
EP (1) EP1192621B1 (de)
AT (1) ATE277404T1 (de)
AU (1) AU6292200A (de)
DE (1) DE60014086D1 (de)
FR (1) FR2796190B1 (de)
WO (1) WO2001003120A1 (de)

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5054072A (en) * 1987-04-02 1991-10-01 Massachusetts Institute Of Technology Coding of acoustic waveforms
US5630011A (en) * 1990-12-05 1997-05-13 Digital Voice Systems, Inc. Quantization of harmonic amplitudes representing speech
US5369724A (en) * 1992-01-17 1994-11-29 Massachusetts Institute Of Technology Method and apparatus for encoding, decoding and compression of audio-type data using reference coefficients located within a band of coefficients
US5787387A (en) * 1994-07-11 1998-07-28 Voxware, Inc. Harmonic adaptive speech coding method and system
JP3653826B2 (ja) * 1995-10-26 2005-06-02 ソニー株式会社 音声復号化方法及び装置

Also Published As

Publication number Publication date
FR2796190B1 (fr) 2002-05-03
WO2001003120A1 (fr) 2001-01-11
ATE277404T1 (de) 2004-10-15
EP1192621A1 (de) 2002-04-03
AU6292200A (en) 2001-01-22
FR2796190A1 (fr) 2001-01-12
DE60014086D1 (de) 2004-10-28

Similar Documents

Publication Publication Date Title
EP0782128B1 (de) Verfahren zur Analyse eines Audiofrequenzsignals durch lineare Prädiktion, und Anwendung auf ein Verfahren zur Kodierung und Dekodierung eines Audiofrequenzsignals
EP1593116B1 (de) Verfahren zur differenzierten digitalen Sprach- und Musikbearbeitung, Rauschfilterung, Erzeugung von Spezialeffekten und Einrichtung zum Ausführen des Verfahrens
EP1692689B1 (de) Optimiertes mehrfach-codierungsverfahren
FR2596936A1 (fr) Systeme de transmission d'un signal vocal
EP0801790A1 (de) Verfahren zur sprachkodierung mittels analyse durch synthese
EP0428445B1 (de) Verfahren und Einrichtung zur Codierung von Prädiktionsfiltern in Vocodern mit sehr niedriger Datenrate
FR2784218A1 (fr) Procede de codage de la parole a bas debit
FR2670313A1 (fr) Procede et dispositif pour l'evaluation de la periodicite et du voisement du signal de parole dans les vocodeurs a tres bas debit.
EP1192619B1 (de) Audio-kodierung, dekodierung zur interpolation
WO2023165946A1 (fr) Codage et décodage optimisé d'un signal audio utilisant un auto-encodeur à base de réseau de neurones
EP1192621B1 (de) Audiokodierung mit harmonischen komponenten
EP1192618B1 (de) Audiokodierung mit adaptiver lifterung
WO2001003121A1 (fr) Codage et decodage audio avec composants harmoniques et phase minimale
EP1194923B1 (de) Verfahren und system für audio analyse und synthese
WO2001003119A1 (fr) Codage et decodage audio incluant des composantes non harmoniques du signal
FR2773653A1 (fr) Dispositifs de codage/decodage de donnees, et supports d'enregistrement memorisant un programme de codage/decodage de donnees au moyen d'un filtre de ponderation frequentielle
FR2980620A1 (fr) Traitement d'amelioration de la qualite des signaux audiofrequences decodes
WO2002029786A1 (fr) Procede et dispositif de codage segmental d'un signal audio
FR2739482A1 (fr) Procede et dispositif pour l'evaluation du voisement du signal de parole par sous bandes dans des vocodeurs
FR2737360A1 (fr) Procedes de codage et de decodage de signaux audiofrequence, codeur et decodeur pour la mise en oeuvre de tels procedes
WO2013135997A1 (fr) Modification des caractéristiques spectrales d'un filtre de prédiction linéaire d'un signal audionumérique représenté par ses coefficients lsf ou isf

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

17P Request for examination filed

Effective date: 20020107

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): AT BE CH CY DE DK ES FI FR GB GR IE IT LI LU MC NL

AX Request for extension of the european patent

Free format text: AL;LT;LV;MK;RO;SI

RAP1 Party data changed (applicant data changed or rights of an application transferred)

Owner name: NORTEL NETWORKS FRANCE

GRAP Despatch of communication of intention to grant a patent

Free format text: ORIGINAL CODE: EPIDOSNIGR1

GRAS Grant fee paid

Free format text: ORIGINAL CODE: EPIDOSNIGR3

GRAA (expected) grant

Free format text: ORIGINAL CODE: 0009210

AK Designated contracting states

Kind code of ref document: B1

Designated state(s): AT BE CH CY DE DK ES FI FR GB GR IE IT LI LU MC NL PT SE

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: NL

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20040922

Ref country code: IT

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRE;WARNING: LAPSES OF ITALIAN PATENTS WITH EFFECTIVE DATE BEFORE 2007 MAY HAVE OCCURRED AT ANY TIME BEFORE 2007. THE CORRECT EFFECTIVE DATE MAY BE DIFFERENT FROM THE ONE RECORDED.SCRIBED TIME-LIMIT

Effective date: 20040922

Ref country code: IE

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20040922

Ref country code: FI

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20040922

Ref country code: AT

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20040922

REG Reference to a national code

Ref country code: GB

Ref legal event code: FG4D

Free format text: NOT ENGLISH

REG Reference to a national code

Ref country code: CH

Ref legal event code: EP

REG Reference to a national code

Ref country code: IE

Ref legal event code: FG4D

Free format text: FRENCH

REF Corresponds to:

Ref document number: 60014086

Country of ref document: DE

Date of ref document: 20041028

Kind code of ref document: P

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: SE

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20041222

Ref country code: GR

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20041222

Ref country code: DK

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20041222

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: DE

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20041223

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: ES

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20050102

GBT Gb: translation of ep patent filed (gb section 77(6)(a)/1977)

Effective date: 20050105

LTIE Lt: invalidation of european patent or patent extension

Effective date: 20040922

NLV1 Nl: lapsed or annulled due to failure to fulfill the requirements of art. 29p and 29m of the patents act
REG Reference to a national code

Ref country code: IE

Ref legal event code: FD4D

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: GB

Payment date: 20050614

Year of fee payment: 6

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: LU

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20050704

Ref country code: CY

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20050704

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: FR

Payment date: 20050706

Year of fee payment: 6

PLBE No opposition filed within time limit

Free format text: ORIGINAL CODE: 0009261

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: NO OPPOSITION FILED WITHIN TIME LIMIT

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: MC

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20050731

Ref country code: LI

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20050731

Ref country code: CH

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20050731

Ref country code: BE

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20050731

26N No opposition filed

Effective date: 20050623

REG Reference to a national code

Ref country code: CH

Ref legal event code: PL

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: GB

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20060704

GBPC Gb: european patent ceased through non-payment of renewal fee

Effective date: 20060704

REG Reference to a national code

Ref country code: FR

Ref legal event code: ST

Effective date: 20070330

BERE Be: lapsed

Owner name: *NORTEL NETWORKS FRANCE

Effective date: 20050731

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: PT

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20050222

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: FR

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20060731