EP0076234B1 - Method and apparatus for reduced redundancy digital speech processing - Google Patents

Method and apparatus for reduced redundancy digital speech processing Download PDF

Info

Publication number
EP0076234B1
EP0076234B1 EP82810391A EP82810391A EP0076234B1 EP 0076234 B1 EP0076234 B1 EP 0076234B1 EP 82810391 A EP82810391 A EP 82810391A EP 82810391 A EP82810391 A EP 82810391A EP 0076234 B1 EP0076234 B1 EP 0076234B1
Authority
EP
European Patent Office
Prior art keywords
speech
section
parameters
coded
sections
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
Application number
EP82810391A
Other languages
German (de)
French (fr)
Other versions
EP0076234A1 (en
Inventor
Stephan Dr. Horvath
Carlo Bernasconi
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.)
Omnisec AG Te Regensdorf Zwitserland
Original Assignee
Gretag AG
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 Gretag AG filed Critical Gretag AG
Priority to AT82810391T priority Critical patent/ATE15415T1/en
Publication of EP0076234A1 publication Critical patent/EP0076234A1/en
Application granted granted Critical
Publication of EP0076234B1 publication Critical patent/EP0076234B1/en
Expired 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
    • G10L19/04—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
    • G10L19/06—Determination or coding of the spectral characteristics, e.g. of the short-term prediction coefficients

Definitions

  • the invention relates to a method operating according to the method of linear predication and a corresponding device for redundancy-reducing digital speech processing according to the preamble of claim 1 and claim 13.
  • the LPC vocoders known and available today are not yet fully satisfactory. Although the language synthesized again after the analysis is usually still relatively understandable, it is distorted and sounds artificial. One of the reasons for this is with difficulty in making the decision as to whether there is a voiced or an unvoiced speech section with sufficient certainty. Other causes include poor determination of the pitch period and inaccurate determination of the sound formation filter parameters.
  • the data rate must in many cases be limited to a relatively low value. It is e.g. in the case of telephone networks, preferably only 2.4 kbit / sec.
  • the data rate is determined by the number of speech parameters analyzed in each speech section, by the number of bits required for these parameters and by the so-called frame rate, i.e. given the number of speech sections per second.
  • frame rate i.e. given the number of speech sections per second.
  • at least slightly more than 50 bits are required per speech section. This automatically sets the maximum frame rate, e.g. in a 2.4 kbit / sec system to around 45 / sec.
  • the voice quality at these relatively low frame rates is also correspondingly poor. It is not possible to increase the frame rate, which would in itself improve the voice quality, as this would exceed the specified data rate. To reduce the number of bits required per frame, on the other hand, a reduction in the number of parameters used or a coarsening of their quantization would be necessary, but this would automatically result in a deterioration in the quality of the speech reproduction.
  • the present invention now deals primarily with these difficulties caused by predetermined data rates and has in particular the aim of improving a method or a device of the type defined at the outset with regard to the quality of the speech reproduction without increasing the data rates.
  • the basic idea of the invention is therefore to save bits by improved coding of the speech parameters, so that the frame rate can be increased.
  • there is also an interrelation between the coding of the parameters and the frame rate since less bit-intensive coding, which reduces redundancy, is only possible or makes sense at higher frame rates.
  • this affects therefore, that the coding of the parameters according to the invention is based on the use of the correlation between adjacent voiced speech sections (interframe correlation), which of course becomes increasingly stronger with increasing frame rate.
  • FIG. 1 The general structure and mode of operation of the speech processing device according to the invention are shown in FIG. 1. That from any source, e.g. Analog voice signal originating from a microphone 1 is band-limited in a filter 2 and then sampled and digitized in an A / D converter 3. The sampling rate is about 6 to 16 kHz, preferably about 8 kHz.
  • the resolution is about 8 to 12 bit.
  • the pass band of the filter 2 usually extends from approximately 80 Hz to approximately 3.1-3.4 kHz in the case of so-called broadband speech, and from approximately 300 Hz to 3.1-3.4 kHz in the telephone language.
  • the speech section length is approximately 10 to 30 msec, preferably approximately 20 msec.
  • the frame rate ie the number of frames per second, is approximately 30 to 100, preferably 50 to 70.
  • sections as short as possible and correspondingly high frame rates are desirable, but there is one on the one hand, with real-time processing, the limited performance of the computer used and, on the other hand, the conclusion of the lowest possible bit rates during the transmission.
  • the analysis is therefore essentially divided into two main procedures, on the one hand in the calculation of the amplifier factor or volume parameter and the coefficients or filter parameters of the underlying vocal tract model filter and on the other hand in the voiced-unvoiced decision and in determining the pitch -Period in voiced case.
  • the filter coefficients are obtained in a parameter calculator 4 by solving the system of equations which is obtained when the energy of the prediction error, ie the energy of the difference between the actual samples and the samples estimated on the basis of the model assumption in the interval under consideration (speech section) is minimized as a function of the coefficients becomes.
  • the system of equations is preferably solved using the autocorrelation method using an algorithm according to Durbin (see, for example, LB Rabiner and RW Schafer, “Digital Processing of Speech Signals”, Prentice Hall Inc., Englewood Cliffs, NJ, 1978, pages 411-413).
  • the so-called reflection coefficients (k j ) also result, which are less sensitive transforms of the filter coefficients (a j ) to quantization.
  • the reflection coefficients are always smaller than 1 and, in addition, their amount decreases with an increasing atomic number. Because of these advantages, these reflection coefficients (k j ) are preferably transmitted instead of the filter coefficients (a j ).
  • the volume parameter G results from the algorithm as a by-product.
  • the digital voice - signal Sn stored in a buffer 5 first as long are calculated until the filter parameters (a j). The signal then passes through an inverse filter 6 set with the parameters (a j ), which has an inverse transfer function to the transfer function of the vocal tract model filter.
  • the result of this inverse filtering is a prediction error signal e n , which is similar to the excitation signal Xn multiplied by the gain factor G.
  • This prediction error signal e n is now supplied in the case of telephone speech directly or in the case of broadband speech via a low-pass filter 7 to an autocorrelation stage 8, which forms the autocorrelation function AKF standardized to the zero-order autocorrelation maximum, on the basis of which the pitch period p is determined in a pitch extraction stage 9. specifically in a known manner as the distance between the second autocorrelation maximum RXX and the first maximum (zero order), an adaptive search method preferably being used.
  • the language section under consideration is classified as voiced or unvoiced in a decision stage 11 according to certain criteria, which include also include the energy of the speech signal and the number of zero crossings in the section under consideration. These two values are determined in an energy determination stage 12 and a zero crossing determination stage 13.
  • the parameter calculator described above determines a set of filter parameters for each speech section (frame).
  • the filter parameters could also be determined differently, for example continuously by means of adaptive inverse filtering or another known method, the filter parameters being readjusted continuously with each sampling cycle, but only at the times determined by the frame rate for further processing or Transmission will be provided.
  • the invention is in no way restricted in this regard. It is only essential that there is a set of filter parameters for each language section.
  • the parameters (k j ), G and p obtained according to the method just described are then fed to a coding stage 14, where they are brought (formatted) and made available in a particularly bit-efficient form suitable for transmission, in a manner to be described in more detail below .
  • the speech signal is recovered or synthesized from the parameters in a known manner in that the parameters initially decoded in a decoder 15 are fed to a pulse-noise generator 16, an amplifier 17 and a vocal tract model filter 18 and the output signal of the model filter 18 by means of a D / A converter 19 brought into analog form and then after the usual filtering 20 by a playback device, for. B. a speaker 21 is made audible.
  • the volume parameter G controls the amplification factor of the amplifier 17, the filter parameters (k j ) define the transfer function of the sound formation or vocal tract model filter 18.
  • Fig. 2 An example of such a system is shown in Fig. 2 as a block diagram.
  • the multi-processor system shown essentially comprises four functional blocks, namely a main processor 50, two secondary processors 60 and 70 and an input / output unit 80. It implements both analysis and synthesis.
  • the input / output unit 80 contains the stages designated 81 for analog signal processing, such as amplifiers, filters and automatic gain control, as well as the A / D converter and the D / A converter.
  • the main processor 50 carries out the actual speech analysis or synthesis, for which purpose the determination of the filter parameters and the volume parameters (parameter calculator 4), the determination of energy and zero crossings of the speech signal (stages 13 and 12), the voiced-unvoiced decision (stage 11 ) and the determination of the pitch period (stage 9) or, on the synthesis side, the generation of the output signal (stage 16), its volume variation (stage 17) and its filtering in the speech model filter (filter 18).
  • the main processor 50 is supported by the secondary processor 60, which carries out the intermediate storage (buffer 5), inverse filtering (stage 6), optionally the low-pass filtering (stage 7) and the autocorrelation (stage 8).
  • the secondary processor 70 deals exclusively with the coding or decoding of the speech parameters and with the data traffic, e.g. a modem 90 or the like via an interface designated 71.
  • the data rate in an LPC vocoder system is determined by the so-called frame rate, i.e. the number of speech segments per second, the number of language parameters used and the number of bits required to encode the language parameters.
  • the basic principle of the invention consists in the consideration that if the speech signal is analyzed more often, that is to say the frame rate is increased, a better tracking of the transientities of the speech signal is possible. With stationary speech sections, a greater correlation between the parameters of the successive speech sections is thus achieved, which in turn leads to a more efficient, i.e. bit-saving coding can be used so that the overall data rate does not increase despite the increased frame rate, but the voice quality is significantly improved.
  • This special coding of the speech parameters according to the invention is explained in more detail below.
  • the basic idea of the parameter coding according to the invention is the so-called block coding principle, that is to say that the speech parameters are not coded independently of one another for each individual speech section, but rather two or three speech sections are combined to form a block and the parameters of all two or are coded within this block three language sections according to uniform rules and in such a way that in each case only the parameters of the first section are coded in full form, while the parameters of the other language section (s) are coded in differential form or possibly omitted or substituted entirely.
  • the coding within the block is also carried out differently, taking into account the typical properties of human speech, depending on whether it is a voiced or unvoiced block, the first speech section in each case determining the voiced character of the block.
  • Complete coding is understood to mean the usual coding of the parameters, for example 6 bit for the pitch parameter, 5 bit for the volume parameter and (for a ten-pole filter, for example) for the first four filter coefficients, each 5 bit, for the next four 4 bits each and reserved for the last two 3 or 2 bits.
  • the decreasing number of bits for the higher filter coefficients is explained from the fact that the reflection coefficients usually used decrease in magnitude with increasing atomic number and essentially only determine the fine structure of the short-term speech spectrum.
  • the coding according to the invention is different for the individual parameter types (filter coefficients, volume, pitch). It is explained below using the example of blocks consisting of three language sections each.
  • the filter parameters of the first section are encoded in full form, the filter parameters of the second and third sections, however, in differential form, ie only in the form of their difference compared to the corresponding parameters of the first or if necessary also of the second section.
  • the difference of a 5-bit parameter is e.g. represented by a 4-bit word, etc.
  • the last, only 2-bit parameter could be encoded in this way, but this would make little sense with only 2-bit.
  • the last filter parameter of the second and third sections is therefore either replaced by that of the first section or set to zero, which saves the transmission in both cases.
  • the filter coefficients of the second speech section can also be adopted immediately with those of the first section and therefore do not need to be coded or transmitted at all.
  • the bits released in this way can be used to encode the difference between the filter parameters of the third section and those of the first section with greater resolution.
  • the coding is done in a different way.
  • the filter parameters of the first section are full again, i.e. encoded in full form or full bit length, the filter parameters of the other two sections are not coded differentially, but also in full form.
  • bit reduction use is made of the fact that in the unvoiced case the higher filter coefficients make little contribution to the sound image, and accordingly the higher filter coefficients, e.g. from the seventh, not encoded or transmitted at all. On the synthesis side, they are then interpreted as zero.
  • This parameter encoding is performed in voiced and unvoiced case largely g e-based or even completely the same in a variant.
  • the parameters of the first and third sections are each fully coded, those of the middle section in the form of their difference from that of the first section.
  • the volume parameter of the middle speech section can also be assumed to be the same as that of the first section and therefore does not need to be coded or transmitted at all.
  • the synthesis-side decoder then automatically generates this parameter from the parameter of the first speech section.
  • the pitch parameter is coded the same for voiced and unvoiced blocks, just like that of the filter coefficients in the voiced case, i.e. full for the first language section (e.g. 7 bit) and differential for the other two sections.
  • the differences are preferably represented with 3 bits.
  • a change is indicated by a special code word, in that the difference to the pitch parameter of the first speech section, which in any case exceeds the representable difference range, is replaced by this code word.
  • the code word of course has the same format as the pitch parameter differences.
  • the running pitch parameter is a running average of the pitch parameters of a number, e.g. 2 to 7 previous language sections used.
  • the decoded pitch parameter is preferably synthesized on the synthesis side with a running average of the pitch parameters of a number, e.g. 2 to 7 previous language sections compared and replaced by the running average when a predetermined maximum deviation, for example about ⁇ 30% to ⁇ 60% is exceeded.
  • a predetermined maximum deviation for example about ⁇ 30% to ⁇ 60% is exceeded.
  • the “outlier” does not go into further averaging.
  • the coding is basically the same as for the blocks with three sections. All parameters of the first section are encoded in their entirety.
  • the filter parameters of the second speech section are either coded in differential form in voiced blocks or assumed to be the same as in the first section and accordingly not coded at all.
  • the filter coefficients of the second speech section are also encoded in their entirety, but the higher coefficients are omitted.
  • the pitch parameter of the second speech section is coded the same again in the voiced and in the unvoiced case, namely in the form of its difference to the pitch parameter of the first section.
  • a code word is used again.
  • the volume parameter of the second speech section is coded in the same way as in the case of blocks with three sections, that is to say in differential form or not at all.
  • the coding and decoding is preferably carried out by software using the computer system which is already available for the remaining speech processing.
  • the creation of a suitable program is within the skill of the average professional.
  • the coding rules A 1 , A 2 and A 3 and B 1 , B 2 and B 3 contained in FIG. 3 are shown in more detail in FIG. 4 and each indicate the format (bit assignments) of the parameters to be coded.
  • the programs for decoding are of course analog.

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)
  • Transmission Systems Not Characterized By The Medium Used For Transmission (AREA)
  • Exchange Systems With Centralized Control (AREA)

Abstract

A digitized speech signal is divided into sections and each section is analyzed by the linear prediction method to determine the coefficients of a sound formation model, a sound volume parameter, information concerning voiced or unvoiced excitation and the period of the vocal band base frequency. In order to improve the quality of speech without increasing the data rate, redundance reducing coding of the speech parameters is effected. The coding of the speech parameters is performed in blocks of two or three adjacent speech sections. The parameters of the first speech section are coded in a complete form, and those of the other speech sections in a differential form or in part not at all. The average number of bits required per speech section is reduced to compensate for the increased section rate, so that the overall data rate is not increased.

Description

Die Erfindung betrifft ein nach der Methode der linearen Prädikation arbeitendes Verfahren und eine entsprechende Vorrichtung zur redundanzvermindernden digitalen Sprachverarbeitung gemäss dem Oberbegriff von Patentanspruch 1 bzw. Patentanspruch 13.The invention relates to a method operating according to the method of linear predication and a corresponding device for redundancy-reducing digital speech processing according to the preamble of claim 1 and claim 13.

Derartige Sprachverarbeitungssysteme, sogenannte LPC-Vocoder, erlauben eine erhebliche Redundanzreduktion bei der digitalen Übertragung von Sprachsignalen. Sie gewinnen heute immer mehr an Bedeutung und sind Gegenstand zahlreicher Veröffentlichungen und Patente, von denen hier nur einige repräsentative rein beispielsweise angeführt sind: B.S. Atal und S.L. Hanauer, Journal Acoust. Soc. Am., 50, S. 637-655, 1971

  • R.W. Schaferund L.R. Rabiner, Proc. IEEEVol. 63 Nr. 4, S. 622-667, 1975
  • L.R. Rabiner et al., Trans. Acoustics, Speech and Signal Proc., Vol. 24 Nr. 5, S. 399-418, 1976 B. Gold, Proc. IEEE Vol. 65 Nr. 12, S. 1636-1658, 1977
  • A. Kurematsu et al., Proc. IEEE, ICASSP, Washington 1979, S. 69-72
  • S. Horvath, «LPC-Vocoder, Entwicklungsstand und Perspectiven», Sammelband Kolloquiumvorträge «Krieg im Äther» XVII. Folge, Bern 1978
    • US-PS 3 624 302
    • US-PS 3 361 520
    • US-PS 3 909 533
    • US-PS 4 230 906
Such speech processing systems, so-called LPC vocoders, allow a considerable reduction in redundancy in the digital transmission of speech signals. They are becoming increasingly important today and are the subject of numerous publications and patents, of which only a few are listed here, for example: BS Atal and SL Hanauer, Journal Acoust. Soc. Am., 50, pp. 637-655, 1971
  • RW Schafer and LR Rabiner, Proc. IEEEVol. 63 No. 4, pp. 622-667, 1975
  • LR Rabiner et al., Trans. Acoustics, Speech and Signal Proc., Vol. 24 No. 5, pp. 399-418, 1976 B. Gold, Proc. IEEE Vol. 65 No. 12, pp. 1636-1658, 1977
  • A. Kurematsu et al., Proc. IEEE, ICASSP, Washington 1979, pp. 69-72
  • S. Horvath, “LPC Vocoder, Level of Development and Perspectives”, anthology colloquium lectures “War in the Aether” XVII. Episode, Bern 1978
    • U.S. Patent 3,624,302
    • U.S. Patent 3,361,520
    • U.S. Patent 3,909,533
    • U.S. Patent 4,230,906

Die heute bekannten und erhältlichen LPC-Vocoder arbeiten noch nicht voll zufriedenstellend. Zwar ist die nach der Analyse wieder synthetisierte Sprache meistens noch relativ verständlich, jedoch ist sie verzerrt und tönt künstlich. Eine der Ursachen dafür liegt u.a. in der Schwierigkeit, den Entscheid, ob ein stimmhafter oder ein stimmloser Sprachabschnitt vorliegt, mit ausreichender Sicherheit zu treffen. Weitere Ursachen sind mangelhafte Bestimmung der Pitchperiode und ungenaue Bestimmung der Klangbildungsfilterparameter.The LPC vocoders known and available today are not yet fully satisfactory. Although the language synthesized again after the analysis is usually still relatively understandable, it is distorted and sounds artificial. One of the reasons for this is with difficulty in making the decision as to whether there is a voiced or an unvoiced speech section with sufficient certainty. Other causes include poor determination of the pitch period and inaccurate determination of the sound formation filter parameters.

Neben diesen grundsätzlichen Schwierigkeiten ergibt sich eine weitere wesentliche Schwierigkeit daraus, dass die Datenrate in vielen Fällen auf einen relativ niedrigen Wert begrenzt sein muss. Sie beträgt z.B. bei Telefonnetzen vorzugsweise nur 2,4 kbit/ sec. Bei einem LPC-Vocoder ist die Datenrate durch die Anzahl der in jedem Sprachabschnitt analysierten Sprachparameter, durch die Anzahl der für diese Parameter benötigten Bits und durch die sog. Frame-Rate, d.h. die Anzahl Sprachabschnitte pro Sekunde gegeben. Bei den heute gebräuchlichen Systemen werden, damit überhaupt eine einigermassen brauchbare Sprachwiedergabe möglich ist, pro Sprachabschnitt minimal etwas über 50 Bit benötigt. Damit ist die maximale Frame-Rate automatisch festgelegt, bei einem 2,4 kbit/sec-System z.B. auf rund 45/sec. Die Sprachqualität bei diesen relativ geringen Frame-Raten ist auch entsprechend schlecht. Eine Erhöhung der Frame-Rate, die sich zur Verbesserung der Sprachqualität an sich anböte, ist nicht möglich, da dadurch die festgelegte Datenrate überschritten würde. Für die Erniedrigung der Anzahl der pro Frame benötigten Bits wäre anderseits eine Verminderung der Anzahl der verwendeten Parameter bzw. eine Vergröberung von deren Quantisierung nötig, was jedoch automatisch wieder auf eine Verschlechterung der Sprachwiedergabequalität hinauslaufen würde.In addition to these fundamental difficulties, another major difficulty arises from the fact that the data rate must in many cases be limited to a relatively low value. It is e.g. in the case of telephone networks, preferably only 2.4 kbit / sec. With an LPC vocoder, the data rate is determined by the number of speech parameters analyzed in each speech section, by the number of bits required for these parameters and by the so-called frame rate, i.e. given the number of speech sections per second. In the systems currently in use, in order for a reasonably usable speech reproduction to be possible at all, at least slightly more than 50 bits are required per speech section. This automatically sets the maximum frame rate, e.g. in a 2.4 kbit / sec system to around 45 / sec. The voice quality at these relatively low frame rates is also correspondingly poor. It is not possible to increase the frame rate, which would in itself improve the voice quality, as this would exceed the specified data rate. To reduce the number of bits required per frame, on the other hand, a reduction in the number of parameters used or a coarsening of their quantization would be necessary, but this would automatically result in a deterioration in the quality of the speech reproduction.

Die vorliegende Erfindung befasst sich nun vornehmlich mit diesen durch vorgegebene Datenraten bedingten Schwierigkeiten und hat insbesondere zum Ziel, ein Verfahren bzw. eine Vorrichtung der eingangs definierten Art hinsichtlich der Sprachwiedergabequalität zu verbessern, ohne dabei die Datenraten zu erhöhen.The present invention now deals primarily with these difficulties caused by predetermined data rates and has in particular the aim of improving a method or a device of the type defined at the outset with regard to the quality of the speech reproduction without increasing the data rates.

Das erfindungsgemässe Verfahren und die erfindungsgemässe Vorrichtung sind in den Patentansprüchen 1 und 13 beschrieben. Bevorzugte Ausführungsformen ergeben sich aus den abhängigen Ansprüchen.The method according to the invention and the device according to the invention are described in patent claims 1 and 13. Preferred embodiments result from the dependent claims.

Der Grundgedanke der Erfindung besteht also darin, durch eine verbesserte Codierung der Sprachparameter Bits einzusparen, so dass die Frame-Rate erhöht werden kann. Anderseits besteht aber auch insofern eine Wechselbeziehung zwischen der Codierung der Parameter und der Frame-Rate, als eine weniger bit-intesive, redundanzvermindemde Codierung erst bei höheren Frame-Raten möglich bzw. sinnvoll ist. Dies rührt u.a. daher, dass die erfindungsgemässe Codierung der Parameter auf der Ausnützung der Korrelation zwischen benachbarten stimmhaften Sprachabschnitten (Interframe-Korrelation) basiert, welche mit zunehmender Frame-Rate natürlich immer stärker wird.The basic idea of the invention is therefore to save bits by improved coding of the speech parameters, so that the frame rate can be increased. On the other hand, there is also an interrelation between the coding of the parameters and the frame rate, since less bit-intensive coding, which reduces redundancy, is only possible or makes sense at higher frame rates. Among other things, this affects therefore, that the coding of the parameters according to the invention is based on the use of the correlation between adjacent voiced speech sections (interframe correlation), which of course becomes increasingly stronger with increasing frame rate.

Im folgenden wird die Erfindung anhand der Zeichnungen näher erläutert. Es zeigen:

  • Fig. 1 ein stark vereinfachtes Blockschaltbild eines LPC-Vocoders,
  • Fig. 2 ein Blockschaltbild eines entsprechenden Multi-Prozessor-Systems und
  • Fig. 3 und 4 ein Flussschema für ein Programm zur Durchführung einer Variante der erfindungsgemässen Codierung.
The invention is explained in more detail below with reference to the drawings. Show it:
  • 1 is a highly simplified block diagram of an LPC vocoder,
  • Fig. 2 is a block diagram of a corresponding multi-processor system and
  • 3 and 4 a flow diagram for a program for carrying out a variant of the coding according to the invention.

Der allgemeine Aufbau und die Funktionsweise der erfindungsgemässen Sprachverarbeitungsvorrichtung gehen aus Fig. 1 hervor. Das von irgendeiner Quelle, z.B. einem Mikrophon 1 stammende analoge Sprachsignal wird in einem Filter 2 bandbegrenzt und dann in einem A/D-Wandler 3 abgetastet und digitalisiert. Die Abtastrate beträgt bei etwa 6 bis 16 kHz, vorzugsweise etwa 8 kHz.The general structure and mode of operation of the speech processing device according to the invention are shown in FIG. 1. That from any source, e.g. Analog voice signal originating from a microphone 1 is band-limited in a filter 2 and then sampled and digitized in an A / D converter 3. The sampling rate is about 6 to 16 kHz, preferably about 8 kHz.

Die Auflösung ist etwa 8 bis 12 bit. Der Durchlassbereich des Filters 2 erstreckt sich bei sog. Breitbandsprache gewöhnlich von ca. 80 Hz bis etwa 3,1-3,4 kHz, bei Telefonsprache von etwa 300 Hz bis 3,1-3,4 kHz.The resolution is about 8 to 12 bit. The pass band of the filter 2 usually extends from approximately 80 Hz to approximately 3.1-3.4 kHz in the case of so-called broadband speech, and from approximately 300 Hz to 3.1-3.4 kHz in the telephone language.

Für die nun folgende digitale Verarbeitung des Sprachsignals wird dieses in aufeinanderfolgende, vorzugsweise überlappende Sprachabschnitte, sog. Frames, eingeteilt. Die Sprachabschnittslänge beträgt etwa 10 bis 30 msec, vorzugsweise etwa 20 msec. Die Frame-Rate, d.h. die Anzahl von Frames pro Sekunde, beträgt etwa 30 bis 100, vorzugsweise 50 bis 70. Im Interesse hoher Auflösung und damit Sprachqualität bei der Synthetisierung sind möglichst kurze Abschnitte und entsprechend hohe Frame-Raten erstrebenswert, jedoch stehen dem einerseits bei Echtzeit-Verarbeitung das begrenzte Leistungsvermögen des eingesetzten Computers und anderseits die Folgerung möglichst niedriger Bitraten bei der Übertragung entgegen.For the subsequent digital processing of the speech signal, it is divided into successive, preferably overlapping speech sections, so-called frames. The speech section length is approximately 10 to 30 msec, preferably approximately 20 msec. The frame rate, ie the number of frames per second, is approximately 30 to 100, preferably 50 to 70. In the interest of high resolution and thus speech quality in the synthesis, sections as short as possible and correspondingly high frame rates are desirable, but there is one on the one hand, with real-time processing, the limited performance of the computer used and, on the other hand, the conclusion of the lowest possible bit rates during the transmission.

Für jeden dieser Sprachabschnitte erfolgt nun eine Analyse des Sprachsignals nach den Prinzipien der linearen Prädikation, wie sie z.B. in den eingangs erwähnten Publikationen beschrieben sind. Grundlage der linearen Prädikation ist ein parametrisches Modell der Spracherzeugung. Ein zeitdiskretes Allpol-Digitalfilter modelliert die Klangformung durch Hals-und Mundtrakt (Vokaltrakt). Bei stimmhaften Lauten ist die Anregung dieses Filters eine periodische Pulsfolge, deren Frequenz, die sog. Pitchfrequenz, die periodische Anregung durch die Stimmbänder idealisiert. Bei stimmlosen Lauten ist die Anregung weisses Rauschen, idealisierend für die Luftturbulenz im Hals bei nicht angeregten Stimmbändern. Ein Verstärkungsfaktor schliesslich kontrolliert die Lautstärke. Auf der Grundlage dieses Modells ist somit das Sprachsignal durch die folgenden Parameter vollständig bestimmt:

  • 1. Die Information, ob der synthetisierende Laut stimmhaft oder stimmlos ist,
  • 2. die Pitch-Periode (bzw. die Pitchfrequenz) bei stimmmhaften Lauten (bei stimmlosen ist die Pitch- Periode per def. gleich 0),
  • 3. die Koeffizienten des zugrundegelegten Allpol-Digitalfilters (Vokaltraktmodells) und
  • 4. der Verstärkungsfaktor.
For each of these speech sections, the speech signal is now analyzed according to the principles of linear predication, as described, for example, in the publications mentioned at the beginning. The basis of linear predication is a parametric model of speech generation. A time-discrete all-pole digital filter models the sound shaping through the neck and mouth tract (vocal tract). In the case of voiced sounds, the excitation of this filter is a periodic pulse sequence whose frequency, the so-called pitch frequency, idealizes the periodic excitation by the vocal cords. With voiceless sounds, the excitation is white noise, idealizing the air turbulence in the throat when the vocal cords are not excited. Finally, an amplification factor controls the volume. On the basis of this model, the speech signal is therefore completely determined by the following parameters:
  • 1. The information as to whether the synthesizing sound is voiced or unvoiced
  • 2. the pitch period (or the pitch frequency) in the case of voiced sounds (in the case of voiceless ones, the pitch period by definition is 0),
  • 3. the coefficients of the underlying all-pole digital filter (vocal tract model) and
  • 4. the gain factor.

Die Analyse gliedert sich demnach im wesentlichen in zwei Hauptprozeduren, und zwar zum einen in die Berechnung des Verstärkerfaktors bzw. Lautstärkeparameters sowie der Koeffizienten bzw. Filterparameter des zugrundeliegenden Vokaltrakt-Modellfilters und zum anderen in den Stimmhaft-Stimmlos-Entscheid und in die Ermittlung der Pitch-Periode im stimmhaften Falle.The analysis is therefore essentially divided into two main procedures, on the one hand in the calculation of the amplifier factor or volume parameter and the coefficients or filter parameters of the underlying vocal tract model filter and on the other hand in the voiced-unvoiced decision and in determining the pitch -Period in voiced case.

Die Filterkoeffizienten werden in einem Parameterrechner 4 durch Lösung des Gleichungssystems gewonnen, weiches erhalten wird, wenn die Energie des Prädikationsfehlers, d.h. die Energie der Differenz zwischen den tatsächlichen Abtastwerten und den aufgrund der Modellannahme geschätzten Abtastwerten im betrachteten Intervall (Sprachabschnitt) in Funktion der Koeffizienten minimiert wird. Die Auflösung des Gleichungssystems erfolgt vorzugsweise nach der Autokorrelationsmethode mittels eines Algorithmus nach Durbin (vgl. z.B. L.B. Rabiner and R.W. Schafer, «Digital Processing of Speech Signals», Prentice Hall Inc., Englewood Cliffs, N.J., 1978, Seiten 411-413). Dabei ergeben sich neben den Filterkoeffizienten bzw. -parametern (aj) gleichzeitig auch die sogenannten Reflexionskoeffizienten (kj), welche auf Quantisierung weniger empfindliche Transformierte der Filterkoeffizienten (aj) sind. Die Reflexionskoeffizienten sind bei stabilen Filtern dem Betrag nach stets kleiner als 1 und ausserdem nimmt ihr Betrag mit zunehmender Ordnungszahl ab. Wegen dieser Vorteile werden bevorzugt diese Reflexionskoeffizienten (kj) statt der Filterkoeffizienten (aj) übertragen. Der Lautstärkeparameter G ergibt sich aus dem Algorithmus als Nebenprodukt.The filter coefficients are obtained in a parameter calculator 4 by solving the system of equations which is obtained when the energy of the prediction error, ie the energy of the difference between the actual samples and the samples estimated on the basis of the model assumption in the interval under consideration (speech section) is minimized as a function of the coefficients becomes. The system of equations is preferably solved using the autocorrelation method using an algorithm according to Durbin (see, for example, LB Rabiner and RW Schafer, “Digital Processing of Speech Signals”, Prentice Hall Inc., Englewood Cliffs, NJ, 1978, pages 411-413). In addition to the filter coefficients or parameters (a j ), the so-called reflection coefficients (k j ) also result, which are less sensitive transforms of the filter coefficients (a j ) to quantization. In the case of stable filters, the reflection coefficients are always smaller than 1 and, in addition, their amount decreases with an increasing atomic number. Because of these advantages, these reflection coefficients (k j ) are preferably transmitted instead of the filter coefficients (a j ). The volume parameter G results from the algorithm as a by-product.

Zur Auffindung der Pitch-Periode p (Periode der Stimmbandgrundfrqeuenz) wird das digitale Sprach- signal Sn in einem Buffer 5 zunächst so lange zwischengespeichert, bis die Filterparameter (aj) berechnet sind. Dann passiert das Signal ein mit den Parametern (aj) eingestelltes Inversfilter 6, weiches eine zur Übertragungsfunktion des Vokaltraktmodellfilters inverse Übertragungsfunktion besitzt. Das Ergebnis dieser Invers-Filterung ist ein Prädiktionsfehlersignal en, welches dem mit dem Verstärkungsfaktor G multiplizierten Anregungssignal Xn ähnlich ist. Dieses Prädiktionsfehlersignal en wird nun im Falle von Telefonsprache direkt oder im Falle von Breitbandsprache über ein Tiefpassfilter 7 einer Autokorrelationsstufe 8 zugeführt, welches daraus die auf das Autokorrelationsmaximum nullter Ordnung normierte Autokorrelationsfunktion AKF bildet, anhand welcher in einer Pitchextraktionsstufe 9 die Pitchperiode p ermittelt wird, und zwar in bekannter Weise als Abstand des zweiten Autokorrelationsmaximums RXX vom ersten Maximum (nullter Ordnung), wobei vorzugsweise ein adaptives Suchverfahren angewandt wird.In order to find the pitch period p (period of Stimmbandgrundfrqeuenz), the digital voice - signal Sn stored in a buffer 5, first as long are calculated until the filter parameters (a j). The signal then passes through an inverse filter 6 set with the parameters (a j ), which has an inverse transfer function to the transfer function of the vocal tract model filter. The result of this inverse filtering is a prediction error signal e n , which is similar to the excitation signal Xn multiplied by the gain factor G. This prediction error signal e n is now supplied in the case of telephone speech directly or in the case of broadband speech via a low-pass filter 7 to an autocorrelation stage 8, which forms the autocorrelation function AKF standardized to the zero-order autocorrelation maximum, on the basis of which the pitch period p is determined in a pitch extraction stage 9. specifically in a known manner as the distance between the second autocorrelation maximum RXX and the first maximum (zero order), an adaptive search method preferably being used.

Die Klassifikation des betrachteten Sprachabschnitts als stimmhaft bzw. stimmlos erfolgt in einer Entscheidungsstufe 11 nach bestimmten Kriterien, welche u.a. auch die Energie des Sprachsignals und die Anzahl der Nulldurchgänge desselben im betrachteten Abschnitt beinhalten. Diese beiden Werte werden in einer Energiebestimmungsstufe 12 und einer Nulldurchgangsbestimmungsstufe 13 ermittelt.The language section under consideration is classified as voiced or unvoiced in a decision stage 11 according to certain criteria, which include also include the energy of the speech signal and the number of zero crossings in the section under consideration. These two values are determined in an energy determination stage 12 and a zero crossing determination stage 13.

Der vorstehend beschriebene Parameterrechner ermittelt pro Sprachabschnitt (Frame) je einen Satz Filterparameter. Selbstverständlich könnten die Filterparameter auch anders bestimmt werden, beispielsweise laufend mittels einer adaptiven inversen Filtrierung oder eines anderen bekannten Verfahrens, wobei die Filterparameter zwar mit jedem Abtasttakt laufend nachgeregelt, aber nur jeweils zu den durch die Frame-Rate festgelegten Zeitpunkten für die weitere Verarbeitung bzw. Übertragung bereitgestellt werden. Die Erfindung ist diesbezüglich in keiner Weise eingeschränkt. Wesentlich ist lediglich, dass für jeden Sprachabschnitt ein Satz Filterparameter vorliegt.The parameter calculator described above determines a set of filter parameters for each speech section (frame). Of course, the filter parameters could also be determined differently, for example continuously by means of adaptive inverse filtering or another known method, the filter parameters being readjusted continuously with each sampling cycle, but only at the times determined by the frame rate for further processing or Transmission will be provided. The invention is in no way restricted in this regard. It is only essential that there is a set of filter parameters for each language section.

Die nach der eben geschilderten Methode gewonnenen Parameter (kj), G und p werden dann einer Codierungsstufe 14 zugeführt, wo sie in noch näher zu beschreibender Weise in eine für die Übertragung geeignete, besonders bit-rationelle Form gebracht (formatiert) und bereitgestellt werden.The parameters (k j ), G and p obtained according to the method just described are then fed to a coding stage 14, where they are brought (formatted) and made available in a particularly bit-efficient form suitable for transmission, in a manner to be described in more detail below .

Die Rückgewinnung bzw. Synthese des Sprachsignals aus den Parametern erfolgt in bekannter Weise dadurch, dass die zunächst in einem Decoder 15 decodierten Parameter einem Puls-Rausch-Generator 16, einem Verstärker 17 und einem Vokaltraktmodellfilter 18 zugeführt werden und das Ausgangssignal des Modellfilters 18 mittels eines D/A-Wandlers 19 in analoge Form gebracht und dann nach der üblichen Filterung 20 durch ein Wiedergabegerät, z. B. einen Lautsprecher 21 hörbar gemacht wird. Der Puls-Rauschgenerator 16 erzeugt die durch den Verstärker 17 verstärkte Anregung Xn des Vokaltraktmodellfilters 18, und zwar im stimmlosen Falle (p = 0) weisses Rauschen und im stimmhaften Falle (p =i= 0) eine periodische Pulsfolge der durch die Pitchperiode p festgelegten Frequenz. Der Lautstärkeparameter G kontrolliert den Verstärkungsfaktor des Verstärkers 17, die Filterparameter (kj) definieren die Übertragungsfunktion des Klangbildungs- bzw. Vokaltraktmodellfilters 18.The speech signal is recovered or synthesized from the parameters in a known manner in that the parameters initially decoded in a decoder 15 are fed to a pulse-noise generator 16, an amplifier 17 and a vocal tract model filter 18 and the output signal of the model filter 18 by means of a D / A converter 19 brought into analog form and then after the usual filtering 20 by a playback device, for. B. a speaker 21 is made audible. The pulse-noise generator 16 generates the excitation X n of the vocal tract model filter 18 amplified by the amplifier 17, namely white noise in the unvoiced case (p = 0) and a periodic pulse sequence in the voiced case (p = i = 0) by the pitchpe period p set frequency. The volume parameter G controls the amplification factor of the amplifier 17, the filter parameters (k j ) define the transfer function of the sound formation or vocal tract model filter 18.

Vorstehend wurde der allgemeine Aufbau und die Funktion der erfindungsgemässen Sprachverarbeitungsvorrichtung der einfacheren Verständlichkeit halber anhand diskreter Funktionsstufen erläutert. Es ist für den Fachmann jedoch selbstverständlich, dass sämtliche Funktionen bzw. Funktionsstufen zwischen dem analyseseitigen A/D-Wandier 3 und dem syntheseseitigen D/A-Wandier 19; in denen also digitale Signale verarbeitet werden, in der Praxis vorzugsweise durch einen entsprechend programmierten Computer oder einen Mikroprozessor oder dergleichen implementiert sind. Die softwarenmässige Realisierung der einzelnen Funktionsstufen, wie z.B. Parameterrechner, die diversen Digitalfilter, Autokorrelation usw. ist für den mit der Datenverarbeitungstechnik vertrauten Fachmann Routine und in der Fachliteratur beschrieben (siehe z.B. IEEE Digital Signal Processing Committee: «Programs for Digital Signal Processing», IEEE Press Book 1980).The general structure and function of the speech processing device according to the invention has been explained above for the sake of clarity using discrete function levels. However, it is self-evident for the person skilled in the art that all functions or functional levels between the analysis-side A / D converter 3 and the synthesis-side D / A converter 19 ; in which digital signals are processed, preferably implemented in practice by a suitably programmed computer or a microprocessor or the like. The software implementation of the individual function levels, such as parameter computers, the various digital filters, autocorrelation, etc., is routine for the specialist familiar with data processing technology and is described in the specialist literature (see e.g. IEEE Digital Signal Processing Committee: "Programs for Digital Signal Processing", IEEE Press Book 1980).

Für Echtzeit-Anwendungen sind insbesondere bei hohen Abtastraten und kurzen Sprachabschnitten wegen der dann grossen Anzahl von in kürzester Zeit zu bewältigenden Operationen extrem leistungsfähige Rechner erforderlich. Für solche Zwecke werden dann am besten Multi-Prozessor-Systeme mit einer geeigneten Aufgabenteilung eingesetzt. Ein Beispiel für ein solches System ist in Fig. 2 als Blockschema dargestellt.Extremely powerful computers are required for real-time applications, in particular at high sampling rates and short speech sections, because of the large number of operations that can then be completed in a very short time. For such purposes it is best to use multi-processor systems with a suitable division of tasks. An example of such a system is shown in Fig. 2 as a block diagram.

Das dargestellte Multi-Prozessor-System umfasst im wesentlichen vier Funktionsblöcke, und zwar einen Hauptprozessor 50, zwei Nebenprozessoren 60 und 70 und eine Eingabe/Ausgabe-Einheit 80. Es implementiert sowohl Analyse als auch Synthese.The multi-processor system shown essentially comprises four functional blocks, namely a main processor 50, two secondary processors 60 and 70 and an input / output unit 80. It implements both analysis and synthesis.

Die Eingabe/Ausgabe-Einheit 80 enthält die mit 81 bezeichneten Stufen zur analogen Signalverarbeitung, wie Verstärker, Filter und automatische Verstärkungsregelung, sowie den A/D-Wandler und den D/A-Wandler.The input / output unit 80 contains the stages designated 81 for analog signal processing, such as amplifiers, filters and automatic gain control, as well as the A / D converter and the D / A converter.

Der Hauptprozessor 50 führt die eigentliche Sprachanalyse bzw. -synthese durch, wozu die Bestimmung der Filterparameter und der Lautstärkeparameter (Parameterrechner 4), die Bestimmung von Energie und Nulldurchgängen des Sprachsignals (Stufen 13 und 12), die Stimmhaft-Stimmlos-Entscheidung (Stufe 11) und die Bestimmung der Pitch- periode (Stufe 9) bzw. syntheseseitig die Erzeugung des Ausgangssignals (Stufe 16), dessen Lautstärkevariation (Stufe 17) und dessen Filtrierung im Sprachmodellfilter (Filter 18) gehören.The main processor 50 carries out the actual speech analysis or synthesis, for which purpose the determination of the filter parameters and the volume parameters (parameter calculator 4), the determination of energy and zero crossings of the speech signal (stages 13 and 12), the voiced-unvoiced decision (stage 11 ) and the determination of the pitch period (stage 9) or, on the synthesis side, the generation of the output signal (stage 16), its volume variation (stage 17) and its filtering in the speech model filter (filter 18).

Der Hauptprozessor 50 wird dabei vom Nebenprozessor 60 unterstützt, welcher die Zwischenspeicherung (Buffer 5), Inversfiltrierung (Stufe 6), gegebenenfalls die Tiefpassfiltrierung (Stufe 7) und die Autokorrelation (Stufe 8) durchführt.The main processor 50 is supported by the secondary processor 60, which carries out the intermediate storage (buffer 5), inverse filtering (stage 6), optionally the low-pass filtering (stage 7) and the autocorrelation (stage 8).

Der Nebenprozessor 70 schliesslich befasst sich ausschliesslich mit der Codierung bzw. Decodierung der Sprachparameter sowie mit dem Datenverkehr mitz.B. einem Modem 90 oder dgl. via eine mit 71 bezeichnete Schnittstelle.Finally, the secondary processor 70 deals exclusively with the coding or decoding of the speech parameters and with the data traffic, e.g. a modem 90 or the like via an interface designated 71.

Im folgenden wird auf die Codierung der Sprachparameter eingegangen.The coding of the speech parameters is discussed below.

Die Datenrate in einem LPC-Vocoder-System wird bekanntlich bestimmt durch die sog. Frame-Rate, i.e. die Anzahl Sprachabschnitte pro Sekunde, die Anzahl der verwendeten Sprachparameter und die Anzahl Bit, die zur Codierung der Sprachparameter benötigt werden.As is known, the data rate in an LPC vocoder system is determined by the so-called frame rate, i.e. the number of speech segments per second, the number of language parameters used and the number of bits required to encode the language parameters.

Bei den bisher bekannten Systemem werden gewöhnlich insgesamt etwa 10-14 Parameter verwendet, für deren Codierung pro Frame (Sprachabschnitt) in der Regel etwas über 50 bit benötigt werden. Bei einer auf 2,4 kbit/sec begrenzten Datenrate, wie sie bei Telefonnetzen üblich ist, führt dies zu einer maximalen Frame-Rate von rund 45. Wie die Praxis gezeigt hat, ist jedoch die Qualität der unter diesen Bedingungen verarbeitenden Sprache unbefriedigend.In the systems known hitherto, a total of about 10-14 parameters are usually used, for the coding of which a frame (speech section) generally requires just over 50 bits. With a data rate limited to 2.4 kbit / sec, as is customary in telephone networks, this leads to a maximum frame rate of around 45. However, as practice has shown, the quality of the speech processing under these conditions is unsatisfactory.

Dieses durch die Begrenzung der Datenrate auf 2,4 kbit/sec bedingte Dilemma wird nun durch die vorliegende Erfindung durch eine bessere Ausnützung der Redundanzeigenschaften der menschlichen Sprache gelöst. Das grundlegende Prinzip der Erfindung besteht in der Überlegung, dass,wenn das Sprachsignal öfter analysiert wird, also die Frame-Rate erhöht wird, eine bessere Verfolgung der Instationäritäten des Sprachsignals möglich ist. Damit wird bei stationären Sprachabschnitten eine grössere Korrelation zwischen den Parametern der aufeinanderfolgenden Sprachabschnitte erreicht, welche wiederum zu einer effizienteren, d.h. bitsparenden Codierung ausgenützt werden kann, so dass die Gesamtdatenrate trotz erhöhter Frame-Rate nicht erhöht, die Sprachqualität hingegen erheblich verbessert wird. Diese spezielle, erfindungsgemässe Codierung der Sprachparameter ist nachstehend näher erläutert.This dilemma caused by the limitation of the data rate to 2.4 kbit / sec is now solved by the present invention by better exploitation of the redundancy properties of human speech. The basic principle of the invention consists in the consideration that if the speech signal is analyzed more often, that is to say the frame rate is increased, a better tracking of the transientities of the speech signal is possible. With stationary speech sections, a greater correlation between the parameters of the successive speech sections is thus achieved, which in turn leads to a more efficient, i.e. bit-saving coding can be used so that the overall data rate does not increase despite the increased frame rate, but the voice quality is significantly improved. This special coding of the speech parameters according to the invention is explained in more detail below.

Der Grundgedanke der erfindungsgemässen Parameter-Codierung ist das sog. Blockcodierungsprinzip, d.h., die Sprachparameter werden nicht für jeden einzelnen Sprachabschnitt unabhängig voneinander codiert, sondern jeweils zwei oder drei Sprachabschnitte werden zu einem Block zusammengefasst und innerhalb dieses Blocks erfolgt die Codierung der Parameter aller zwei oder drei Sprachabschnitte nach einheitlichen Regeln und zwar derart, dass jeweils nur die Parameter des ersten Abschnitts in vollständiger Form codiert werden, während die Parameter des bzw. der übrigen Sprachabschnitte in differentieller Form codiert oder eventuell gänzlich weggelassen bzw. substituiert werden. Die Codierung innerhalb des Blocks wird ferner in Berücksichtigung der typischen Eigenschaften der menschlichen Sprache unterschiedlich vorgenommen je nachdem, ob es sich um einen stimmhaften oder einen stimmlosen Block handelt, wobei für den Stimmhaftigkeitscharakter des Blocks jeweils der erste Sprachabschnitt darin bestimmend ist.The basic idea of the parameter coding according to the invention is the so-called block coding principle, that is to say that the speech parameters are not coded independently of one another for each individual speech section, but rather two or three speech sections are combined to form a block and the parameters of all two or are coded within this block three language sections according to uniform rules and in such a way that in each case only the parameters of the first section are coded in full form, while the parameters of the other language section (s) are coded in differential form or possibly omitted or substituted entirely. The coding within the block is also carried out differently, taking into account the typical properties of human speech, depending on whether it is a voiced or unvoiced block, the first speech section in each case determining the voiced character of the block.

Unter Codierung in vollständiger Form wird die übliche Codierung der Parameter verstanden, bei der z.B. für den Pitch-Parameter 6 bit, für den Lautstärkeparameter 5 bit und (bei einem z.B. zehnpoligen Filter) für die ersten vier Filterkoeffizienten je 5 bit, für die nächsten vier je 4 bit und für die beiden letzten 3 bzw. 2 bit reserviert werden. (Die abnehmende Bitanzahl für die höheren Filterkoeffizienten erklärt sich daraus, dass die gewöhnlich verwendeten Reflexionskoeffizienten im Betrag mit steigender Ordnungszahl abnehmen und im wesentlichen nur die Feinstruktur des Kurzzeitsprachspektrums mitbestimmen.)Complete coding is understood to mean the usual coding of the parameters, for example 6 bit for the pitch parameter, 5 bit for the volume parameter and (for a ten-pole filter, for example) for the first four filter coefficients, each 5 bit, for the next four 4 bits each and reserved for the last two 3 or 2 bits. (The decreasing number of bits for the higher filter coefficients is explained from the fact that the reflection coefficients usually used decrease in magnitude with increasing atomic number and essentially only determine the fine structure of the short-term speech spectrum.)

Die erfindungsgemässe Codierung ist für die einzelnen Parameter-Typen Filterkoeffizienten, Lautstärke, Pitch) unterschiedlich. Sie wird im folgenden am Beispiel von aus jeweils drei Sprachabschnitten bestehenden Blöcken erläutert.The coding according to the invention is different for the individual parameter types (filter coefficients, volume, pitch). It is explained below using the example of blocks consisting of three language sections each.

Filterparameter (-koeffizienten):Filter parameters (coefficients):

Wenn der Block, d.g. der erste Sprachabschnitt darin stimmhaft (p # 0) ist, werden die Filterparameter des ersten Abschnitts in vollständiger Form codiert, die Filterparameter des zweiten und des dritten Abschnitts hingegen in differentieller Form, d.h., nur in Form ihrer Differenz gegenüber den entsprechenden Parametern des ersten bzw. gegebenenfalls auch des zweiten Abschnitts. Für die jeweilige Differenz wird z. B. um ein Bit weniger veranschlagt als für die vollständige Form, die Differenz eines 5-bit-Parameters wird also z.B. durch ein 4-bit-Wort dargestellt, u.s.f. Im Prinzip könnte so auch der letzte, nur 2 bit umfassende Parameter codiert werden, allerdings wäre dies bei nur 2 bit wenig sinnvoll. Der letzte Filterparameter des zweiten und des dritten Abschnitts wird daher entweder durch den des ersten Abschnitts ersetzt oder gleich Null gesetzt, was in beiden Fällen die Übertragung erspart.If the block, i.e. If the first speech section is voiced (p # 0), the filter parameters of the first section are encoded in full form, the filter parameters of the second and third sections, however, in differential form, ie only in the form of their difference compared to the corresponding parameters of the first or if necessary also of the second section. For the respective difference z. B. estimated by one bit less than for the complete form, the difference of a 5-bit parameter is e.g. represented by a 4-bit word, etc. In principle, the last, only 2-bit parameter could be encoded in this way, but this would make little sense with only 2-bit. The last filter parameter of the second and third sections is therefore either replaced by that of the first section or set to zero, which saves the transmission in both cases.

Gemäss einer ebenfalls bewährten Variante können die Filterkoeffizienten des zweiten Sprachabschnitts auch gleich mit denen des ersten Abschnitts angenommen werden und brauchen demzufolge überhaupt nicht codiert bzw. übertragen zu werden. Die dabei freiwerdenden Bits können dazu verwendet werden, die Differenz der Filterparameter des dritten Abschnitts zu denen des ersten Abschnitts mit grösserer Auflösung zu codieren.According to a variant which has also been tried and tested, the filter coefficients of the second speech section can also be adopted immediately with those of the first section and therefore do not need to be coded or transmitted at all. The bits released in this way can be used to encode the difference between the filter parameters of the third section and those of the first section with greater resolution.

Im stimmlosen Fall, d.h. also wenn der erste Sprachabschnitt des Blocks stimmlos ist (p = 0), erfolgt die Codierung in anderer Weise. Zwar werden die Filterparameter des ersten Abschnitts wieder voll, d.h. in vollständiger Form bzw. voller Bitlänge codiert, die Filterparameter der beiden übrigen Abschnitte werden jedoch nicht differentiell, sondern ebenso in vollständiger Form codiert. Damit dennoch eine Bitreduktion möglich ist, wird von der Tatsache Gebrauch gemacht, dass im stimmlosen Fall die höheren Filterkoeffizienten wenig zum Klangbild beitragen, und dementsprechend werden die höheren Filterkoeffizienten, z.B. ab dem siebenten, überhaupt nicht codiert bzw. übertragen. Syntheseseitig werden sie dann als Null interpretiert.In the unvoiced case, i.e. So if the first speech section of the block is unvoiced (p = 0), the coding is done in a different way. The filter parameters of the first section are full again, i.e. encoded in full form or full bit length, the filter parameters of the other two sections are not coded differentially, but also in full form. In order that bit reduction is nevertheless possible, use is made of the fact that in the unvoiced case the higher filter coefficients make little contribution to the sound image, and accordingly the higher filter coefficients, e.g. from the seventh, not encoded or transmitted at all. On the synthesis side, they are then interpreted as zero.

Lautstärkeparameter (Verstärkungsfaktor):Volume p arameters (gain):

Bei diesem Parameter erfolgt die Codierung im stimmhaften und im stimmlosen Falle weitestge-hend oder in einer Variante sogar vollständig gleich. Der Parameter des ersten und des dritten Abschnitts wird jeweils voll codiert, der des mittleren Abschnitts in Form seiner Differenz zu dem des ersten Abschnitts. Im stimmhaften Falle kann der Lautstärkeparameter des mittleren Sprachabschnitts auch gleich wie der des ersten Abschnitts angenommen werden und braucht demzufolge überhaupt nicht codiert bzw. übertragen zu werden. Der syntheseseitige Decoder erzeugt dann diesen Parameter selbsttätig aus dem Parameter des ersten Sprachabschnitts.This parameter encoding is performed in voiced and unvoiced case largely g e-based or even completely the same in a variant. The parameters of the first and third sections are each fully coded, those of the middle section in the form of their difference from that of the first section. In the voiced case, the volume parameter of the middle speech section can also be assumed to be the same as that of the first section and therefore does not need to be coded or transmitted at all. The synthesis-side decoder then automatically generates this parameter from the parameter of the first speech section.

Pitch-Parameter:Pitch parameters:

Die Codierung des Pitch-Parameters erfolgt für stimmhafte und für stimmlose Blöcke gleich, und zwar so wie die der Filterkoeffizienten im stimmhaften Falle, d.h. für den ersten Sprachabschnitt (z.B. 7 bit) voll und für die beiden übrigen Abschnitte differentiell. Die Differenzen werden dabei vorzugsweise mit 3 bit dargestellt.The pitch parameter is coded the same for voiced and unvoiced blocks, just like that of the filter coefficients in the voiced case, i.e. full for the first language section (e.g. 7 bit) and differential for the other two sections. The differences are preferably represented with 3 bits.

Eine Schwierigkeit ergibt sich jedoch, wenn innerhalb eines Bocks nicht alle Sprachabschnitte stimmlos oder stimmhaft sind, der Stimmhaftigkeitscharakter also wechselt. Zur Behebung dieser Schwierigkeit wird gemäss einem weiteren Gedanken der Erfindung ein solcher Wechsel durch ein spezielles Codewort angezeigt, indem die anstatt der dann den darstellbaren Differenzbereich in der Regel ohnehin übersteigende Differenz zum Pitch-Parameter des ersten Sprachabschnitts durch dieses Codewort ersetzt wird. Das Codewort hat dabei natürlich dasselbe Format wie die Pitch-Parameter-Differenzen.A difficulty arises, however, if not all speech sections within a goat are unvoiced or voiced, i.e. the voicing character changes. To remedy this difficulty, according to a further idea of the invention, such a change is indicated by a special code word, in that the difference to the pitch parameter of the first speech section, which in any case exceeds the representable difference range, is replaced by this code word. The code word of course has the same format as the pitch parameter differences.

Im Falle eines Wechsels von stimmhaft zu stimmlos, also p * 0 zu p = 0, ist klar, wie das Codewort syntheseseitig decodiert werden muss - es braucht dann lediglich der betreffende Pitch-Parameter gleich Null gesetzt zu werden. Im umgekehrten Falle weiss man jedoch lediglich, dass ein Wechsel stattgefunden hat, aber nicht, wie gross der betreffende Pitch-Parameter ist. Aus diesem Grunde wird syntheseseitig in diesem Falle als betreffender Pitch- Parameter ein laufender Mittelwert aus den Pitch-Parametern einer Anzahl, z.B. 2 bis 7 vorangegangener Sprachabschnitte verwendet.In the event of a change from voiced to unvoiced, i.e. p * 0 to p = 0, it is clear how the codeword has to be decoded on the synthesis side - it is then only necessary to set the relevant pitch parameter to zero. In the opposite case, however, you only know that a change has taken place, but not how large the pitch parameter in question is. For this reason, in the case of the synthesis, in this case the running pitch parameter is a running average of the pitch parameters of a number, e.g. 2 to 7 previous language sections used.

Als weitere Sicherung gegen Fehlcodierungen und Fehlübertragungen und auch gegen Fehlberechnungen der Pitch-Parameter wird syntheseseitig vorzugsweise der decodierte Pitch-Parameter mit einem laufenden Mittelwert aus den Pitch-Parametern einer Anzahl, z.B. 2 bis 7 vorangegangener Sprachabschnitte verglichen und beim Überschreiten einer vorgegebenen Maximalabweichung, beispielsweise etwa ± 30% bis ± 60%, durch den laufenden Mittelwert ersetzt. Der «Ausreisser» geht dann natürlich auch nicht in die weitere Mittelwertbildung ein.As a further safeguard against incorrect coding and incorrect transmissions and also against incorrect calculations of the pitch parameters, the decoded pitch parameter is preferably synthesized on the synthesis side with a running average of the pitch parameters of a number, e.g. 2 to 7 previous language sections compared and replaced by the running average when a predetermined maximum deviation, for example about ± 30% to ± 60% is exceeded. Of course, the “outlier” does not go into further averaging.

Bei Blöcken mit nur zwei Sprachabschnitten erfolgt die Codierung im Prinzip gleich wie bei den Blöcken mit drei Abschnitten. Sämtliche Parameter des ersten Abschnitts werden in vollständiger Form codiert. Die Filterparameter des zweiten Sprachabschnitts werden bei stimmhaften Blöcken entweder in differentieller Form codiert oder als gleich wie beim ersten Abschnitt angenommen und dementsprechend überhaupt nicht codiert. Bei stimmlosen Blöcken werden wiederum auch die Filterkoeffizienten des zweiten Sprachabschnitts in vollständiger Form codiert, dafür werden aber die höheren Koeffizienten weggelassen.In the case of blocks with only two language sections, the coding is basically the same as for the blocks with three sections. All parameters of the first section are encoded in their entirety. The filter parameters of the second speech section are either coded in differential form in voiced blocks or assumed to be the same as in the first section and accordingly not coded at all. In the case of unvoiced blocks, the filter coefficients of the second speech section are also encoded in their entirety, but the higher coefficients are omitted.

Der Pitch-Parameter des zweiten Sprachabschnitts wird im stimmhaften und im stimmlosen Fall wieder gleich codiert, und zwar in Form seiner Differenz zum Pitch-Parameter des ersten Abschnitts. Für den Fall eines Stimmhaft-Stimmlos-Wechsels innerhalb eines Blocks wird wiederum ein Codewort verwendet.The pitch parameter of the second speech section is coded the same again in the voiced and in the unvoiced case, namely in the form of its difference to the pitch parameter of the first section. For in the case of a voiced-unvoiced change within a block, a code word is used again.

Der Lautstärkeparameter des zweiten Sprachabschnitts wird gleich codiert wie im Falle von Blöcken mit drei Abschnitten, also in differentieller Form oder gar nicht.The volume parameter of the second speech section is coded in the same way as in the case of blocks with three sections, that is to say in differential form or not at all.

Vorstehend wurde bis auf einige Ausnahmen lediglich von der Codierung der Sprachparameter auf der Analyseseite des kompletten Sprachverarbeitungssystems gesprochen. Es versteht sich jedoch von selbst, dass auf der Syntheseseite eine entsprechende Decodierung der Parameter erfolgen muss, welche Decodierung auch die Erzeugung (vorvereinbarter Werte) der nicht codierten Parameter mit einschliesst.With a few exceptions, only the coding of the speech parameters on the analysis side of the complete speech processing system was mentioned above. However, it goes without saying that a corresponding decoding of the parameters must take place on the synthesis side, which decoding also includes the generation (pre-agreed values) of the uncoded parameters.

Ferner versteht es sich, dass die Codierung und die Decodierung vorzugsweise per Software mittels des für die übrige Sprachverarbeitung ohnehin vorhandenen Computersystems durchgeführt wird. Die Erstellung eines geeigneten Programms liegt im Bereich des Könnens des durchschnittlichen Fachmanns. Ein Beispiel für ein Flussschema eines solchen Programms, und zwar für den Fall von Blöcken mit je drei Sprachabschnitten, ist in den Fig. 3 und 4 dargestellt. Die Flussschemen sind aus sich heraus verständlich, es sei lediglich erwähnt, dass der Index i laufend die einzelnen Sprachabschnitte numeriert und zählt, während der Index N = i mod 3 die Nummer der Abschnitte innerhalb jedes einzelnen Blocks angibt. Die in Fig. 3 enthaltenen Codierungsvorschriften A1, A2 und A3 sowie B1, B2 und B3 sind in Fig. 4 detaillierter dargestellt und geben jeweils das Format (Bitzuteilungen) der zu codierenden Parameter an.Furthermore, it goes without saying that the coding and decoding is preferably carried out by software using the computer system which is already available for the remaining speech processing. The creation of a suitable program is within the skill of the average professional. An example of a flow diagram of such a program, specifically for the case of blocks with three language sections each, is shown in FIGS. 3 and 4. The flow diagrams are self-explanatory, it should only be mentioned that the index i numbers and counts the individual language sections, while the index N = i mod 3 indicates the number of sections within each individual block. The coding rules A 1 , A 2 and A 3 and B 1 , B 2 and B 3 contained in FIG. 3 are shown in more detail in FIG. 4 and each indicate the format (bit assignments) of the parameters to be coded.

Die Programme für die Decodierung sind natürlich analog.The programs for decoding are of course analog.

Claims (13)

1. Redundance reducing speech processing process according to the method of linear prediction, wherein on the analysis side, the digital speech signal obtained by the scanning of the optionally band limeted analog speech signal is divided into sections, and for each speech section the prarameters of a speech model filter, a volume parameter and the pitch parameter (period of the voice band base frequency) are determined and prepared for transmission in a coded form or transmitted, and wherein on the synthesis side, the filter parameters, the volume parameter and the pitch parameter are decoded and with these parameters a synthesis stage consisting essentially of an excitation generator and a speech model filter, is actuated for the recovery of the speech signal, characterized in that the coding of the parameters is effected in blocks over two or three successive speech sections, wherein the parameters of the first speech section are coded in a complete form and at least a part of the remaining sections is coded in a differential form or eliminated.
2. Process according to Claim 1, characterized in that the coding of the parameters is effected in a differential manner depending on whether the first speech section of a block of speech sections is voiced or unvoiced.
3. Process according to Claim 2, characterized in that in the case of blocks comprising three speech sections each, with a voiced first speech section, the filter and the pitch parameters of the first speech section are coded in their complete form and the filter and pitch parameters of the two other speech sections are coded in the form of their differences with respect to the corresponding parameters of the first and the second sections, respectively, and that in case of an unvoiced first speech section, the filter parameters of a higher order are eliminated ant the remaining filter parameters of all three speech sections are coded in their complete form and the pitch parameters are coded as in the voiced case.
4. Process according to Claim 2, characterized in that in the case of blocks comprising three speech sections each, with a voiced first speech section, the filter and pitch parameters of the first section are coded in their complete form, the filter parameters of the intermediate speech section are not coded at all and the pitch parameter of this section is coded in the form of its difference with respect to the pitch parameter of the first section, and the filter and the pitch parameters of the last section are coded in the form of their differences with respect to the corresponding parameters of the first section, and that in the case of an unvoiced first speech section, the filter parameters of a higher order are eliminated and the remaining filter parameters of all three of the speech sections are coded in their complete form and the pitch parameter is coded as in the voiced case.
5. Process according to Claim 1, characterized in that in the case of blocks comprising two speech sections each, with a voiced first speech section, the fitter parameters and the pitch parameter of the first speech section are coded in their complete form and the filter parameters of the second section are not coded at all or is coded in the form of their differences with respect to the prevailing parameters of the first section and the pitch parameter of the second section is coded in the form of its difference with respect to the pitch parameter of the first section, and that in the case of an unvoiced first speech section, the filter parameters of a higher order are eliminated and the remaining filter parameters of both sections are coded in their complete form and the pitch parameters are coded as in the voiced case.
6. Process according to Claims 3 or 4, characterized in that in the case of a voiced first speech section, the volume parameters of the first and the last speech sections are coded in their complete form and that of the intermediate section is not coded at all or is coded in the form of its difference with respect to the volume parameter of the first section, and that in the case of an unvoiced first speech section, the volume parameters of the first and last speech sections are coded in their complete form of its difference with respect to the volume parameter of the first section.
7. Process according to Claim 5, characterized in that with a voiced first speech section, the volume parameter of the first section is coded in its complete form and that of the second speech section is not coded at all or is coded in the form of its difference with respect to the volume parameter of the first section, and that in case of an unvoiced first speech section, the volume parameter of the first section is coded in its complete form and that of the second section is coded in the form of its difference with respect to the volume parameter of the first section.
8. Process according to one of Claims 3 to 7, characterized in that in case of a change from voiced to unvoiced, or vice versa, within a block of speech sections the pitch parameter of the section involved is replaced by a special code word.
9. Process according to Claim 8, characterized in that on the synthesis side in case of the occurrence of the code word and if the preceding speech section was unvoiced, a continuing average value of the pitch parameters of a plurality of the preceding speech sections is used as the corresponding pitch parameter.
10. Process according to one of the preceding claims, characterized in that on the synthesis side the decoded pitch parameter is compared with a continuing average value of the pitch parameters of a plurality of preceding speech sections and is replaced by the continuing average value if a predetermined maximum deviation is exceeded.
11. Process according to one of the preceding claims, characterized in that the length of the individual speech sections, for which the speech parameters are determined, is no greater than approximately 30 msec, and is preferably about 20 msec.
12. Process according to one of the preceding claims, characterized in that the number of speech sections per second is at least 55, and preferably at least 60.
13. Apparatus for the realization of the process according to one of the preceding claims, including a signal preparation part for cyclically scanning the analog speech signal and digitizing the scanned values, an analysis part for analyzing the digitized speech signal by sections and comprising a parameter computer, a pitch decision stage and a pitch computing stage, and a coding stage for coding the speech parameters determined by the analysis part, characterized in that the analysis part is a multiprocessor system with a principal processor (50) and two secondary processors (60, 70), wherein one secondary processor (60) intermediately stores the speech signal, produces the prediction error signal from the intermediately stored speech signal by inverse filtering and forms from said error signal, optionally after deep pass filtering, the standardized autocorrelation function, while the principal processor (50) performs the analysis of the speech signal and the other secondary processor (70) is responsible for the coding of the speech parameters determined by the principal processor in combination with the speech parameters determined by the first secondary processor.
EP82810391A 1981-09-24 1982-09-20 Method and apparatus for reduced redundancy digital speech processing Expired EP0076234B1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
AT82810391T ATE15415T1 (en) 1981-09-24 1982-09-20 METHOD AND DEVICE FOR REDUNDANCY-REDUCING DIGITAL SPEECH PROCESSING.

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CH616881 1981-09-24
CH6168/81 1981-09-24

Publications (2)

Publication Number Publication Date
EP0076234A1 EP0076234A1 (en) 1983-04-06
EP0076234B1 true EP0076234B1 (en) 1985-09-04

Family

ID=4305342

Family Applications (1)

Application Number Title Priority Date Filing Date
EP82810391A Expired EP0076234B1 (en) 1981-09-24 1982-09-20 Method and apparatus for reduced redundancy digital speech processing

Country Status (6)

Country Link
US (1) US4618982A (en)
EP (1) EP0076234B1 (en)
JP (1) JPS5870300A (en)
AT (1) ATE15415T1 (en)
CA (1) CA1184656A (en)
DE (1) DE3266042D1 (en)

Families Citing this family (30)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CA1229681A (en) * 1984-03-06 1987-11-24 Kazunori Ozawa Method and apparatus for speech-band signal coding
CA1255802A (en) * 1984-07-05 1989-06-13 Kazunori Ozawa Low bit-rate pattern encoding and decoding with a reduced number of excitation pulses
CA1252568A (en) * 1984-12-24 1989-04-11 Kazunori Ozawa Low bit-rate pattern encoding and decoding capable of reducing an information transmission rate
US4890328A (en) * 1985-08-28 1989-12-26 American Telephone And Telegraph Company Voice synthesis utilizing multi-level filter excitation
US4912764A (en) * 1985-08-28 1990-03-27 American Telephone And Telegraph Company, At&T Bell Laboratories Digital speech coder with different excitation types
EP0245531A1 (en) * 1986-05-14 1987-11-19 Deutsche ITT Industries GmbH Application of a semiconductor read only memory
US5018200A (en) * 1988-09-21 1991-05-21 Nec Corporation Communication system capable of improving a speech quality by classifying speech signals
US4972474A (en) * 1989-05-01 1990-11-20 Cylink Corporation Integer encryptor
JPH03136100A (en) * 1989-10-20 1991-06-10 Canon Inc Method and device for voice processing
US6006174A (en) * 1990-10-03 1999-12-21 Interdigital Technology Coporation Multiple impulse excitation speech encoder and decoder
JP2810252B2 (en) * 1991-05-22 1998-10-15 シャープ株式会社 Audio playback device
US5317567A (en) * 1991-09-12 1994-05-31 The United States Of America As Represented By The Secretary Of The Air Force Multi-speaker conferencing over narrowband channels
US5272698A (en) * 1991-09-12 1993-12-21 The United States Of America As Represented By The Secretary Of The Air Force Multi-speaker conferencing over narrowband channels
FI95086C (en) * 1992-11-26 1995-12-11 Nokia Mobile Phones Ltd Method for efficient coding of a speech signal
US5517511A (en) * 1992-11-30 1996-05-14 Digital Voice Systems, Inc. Digital transmission of acoustic signals over a noisy communication channel
FI96248C (en) * 1993-05-06 1996-05-27 Nokia Mobile Phones Ltd Method for providing a synthetic filter for long-term interval and synthesis filter for speech coder
US5457685A (en) * 1993-11-05 1995-10-10 The United States Of America As Represented By The Secretary Of The Air Force Multi-speaker conferencing over narrowband channels
PL174216B1 (en) * 1993-11-30 1998-06-30 At And T Corp Transmission noise reduction in telecommunication systems
US5715365A (en) * 1994-04-04 1998-02-03 Digital Voice Systems, Inc. Estimation of excitation parameters
AU696092B2 (en) * 1995-01-12 1998-09-03 Digital Voice Systems, Inc. Estimation of excitation parameters
US5754974A (en) * 1995-02-22 1998-05-19 Digital Voice Systems, Inc Spectral magnitude representation for multi-band excitation speech coders
US5701390A (en) * 1995-02-22 1997-12-23 Digital Voice Systems, Inc. Synthesis of MBE-based coded speech using regenerated phase information
US6240384B1 (en) * 1995-12-04 2001-05-29 Kabushiki Kaisha Toshiba Speech synthesis method
SE506034C2 (en) * 1996-02-01 1997-11-03 Ericsson Telefon Ab L M Method and apparatus for improving parameters representing noise speech
US6131084A (en) * 1997-03-14 2000-10-10 Digital Voice Systems, Inc. Dual subframe quantization of spectral magnitudes
US6161089A (en) * 1997-03-14 2000-12-12 Digital Voice Systems, Inc. Multi-subframe quantization of spectral parameters
US6199037B1 (en) 1997-12-04 2001-03-06 Digital Voice Systems, Inc. Joint quantization of speech subframe voicing metrics and fundamental frequencies
US6377916B1 (en) 1999-11-29 2002-04-23 Digital Voice Systems, Inc. Multiband harmonic transform coder
US7080009B2 (en) * 2000-05-01 2006-07-18 Motorola, Inc. Method and apparatus for reducing rate determination errors and their artifacts
DE102004001293A1 (en) * 2004-01-07 2005-08-11 Deutsche Thomson-Brandt Gmbh Device and method for data transmission with a reduced amount of data

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US3017456A (en) * 1958-03-24 1962-01-16 Technicolor Corp Bandwidth reduction system for television signals
DE1162399B (en) * 1961-10-24 1964-02-06 Ibm Compressor for data whose statistical distribution fluctuates very strongly
US3236947A (en) * 1961-12-21 1966-02-22 Ibm Word code generator
US3439753A (en) * 1966-04-19 1969-04-22 Bell Telephone Labor Inc Reduced bandwidth pulse modulation scheme using dual mode encoding in selected sub-block sampling periods
US4053712A (en) * 1976-08-24 1977-10-11 The United States Of America As Represented By The Secretary Of The Army Adaptive digital coder and decoder
CA1123955A (en) * 1978-03-30 1982-05-18 Tetsu Taguchi Speech analysis and synthesis apparatus
US4335277A (en) * 1979-05-07 1982-06-15 Texas Instruments Incorporated Control interface system for use with a memory device executing variable length instructions

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
IEEE TRANSACTIONS ON COMMUNICATIONS, Band COM-23, Nr. 12, Dezember 1975, Seiten 1466-1474, New York, USA, C.K. UN et al.: "The residual-excited linear prediction vocoder with transmission rate below 9.6 kbits/s" *

Also Published As

Publication number Publication date
JPS5870300A (en) 1983-04-26
DE3266042D1 (en) 1985-10-10
ATE15415T1 (en) 1985-09-15
CA1184656A (en) 1985-03-26
US4618982A (en) 1986-10-21
EP0076234A1 (en) 1983-04-06

Similar Documents

Publication Publication Date Title
EP0076234B1 (en) Method and apparatus for reduced redundancy digital speech processing
DE3244476C2 (en)
DE69915830T2 (en) IMPROVED METHODS FOR RECOVERING LOST DATA FRAMES FOR AN LPC BASED, PARAMETRIC LANGUAGE CODING SYSTEM.
DE69412913T2 (en) Method and device for digital speech coding with speech signal height estimation and classification in digital speech coders
EP0076233B1 (en) Method and apparatus for redundancy-reducing digital speech processing
DE60219351T2 (en) SIGNAL MODIFICATION METHOD FOR EFFICIENT CODING OF LANGUAGE SIGNALS
DE68912692T2 (en) Transmission system suitable for voice quality modification by classifying the voice signals.
DE69230324T2 (en) Process for time scale modification of signals
DE60123651T2 (en) METHOD AND DEVICE FOR ROBUST LANGUAGE CLASSIFICATION
DE60209861T2 (en) Adaptive postfiltering for speech decoding
DE69529356T2 (en) Waveform interpolation by breaking it down into noise and periodic signal components
DE69910240T2 (en) DEVICE AND METHOD FOR RESTORING THE HIGH FREQUENCY PART OF AN OVER-SAMPLE SYNTHETIZED BROADBAND SIGNAL
EP1979901B1 (en) Method and arrangements for audio signal encoding
DE68916944T2 (en) Procedure for the rapid determination of the basic frequency in speech coders with long-term prediction.
DE60214358T2 (en) TIME CALENDAR MODIFICATION OF SIGNALS WITH SPECIFIC PROCEDURE ACCORDING TO DETERMINED SIGNAL TYPE
DE102008042579B4 (en) Procedure for masking errors in the event of incorrect transmission of voice data
DE60133757T2 (en) METHOD AND DEVICE FOR CODING VOTING LANGUAGE
DE4237563A1 (en)
DE69613611T2 (en) System for storing and accessing voice information
DE60034429T2 (en) METHOD AND DEVICE FOR DETERMINING LANGUAGE CODING PARAMETERS
DE69902480T2 (en) METHOD FOR QUANTIZING THE PARAMETERS OF A LANGUAGE CODIER
DE69807806T2 (en) METHOD AND DEVICE FOR CODING AN AUDIO SIGNAL BY MEANS OF "FORWARD" AND "REVERSE" LPC ANALYSIS
DE68917584T2 (en) Encoder suitable for improving speech quality using a double system for pulse generation.
DE69612958T2 (en) METHOD AND DEVICE FOR RESYNTHETIZING A VOICE SIGNAL
DE69629485T2 (en) COMPRESSION SYSTEM FOR REPEATING TONES

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: 19820922

AK Designated contracting states

Designated state(s): AT CH DE FR GB IT LI NL SE

ITF It: translation for a ep patent filed
GRAA (expected) grant

Free format text: ORIGINAL CODE: 0009210

AK Designated contracting states

Designated state(s): AT CH DE FR GB IT LI NL SE

REF Corresponds to:

Ref document number: 15415

Country of ref document: AT

Date of ref document: 19850915

Kind code of ref document: T

REF Corresponds to:

Ref document number: 3266042

Country of ref document: DE

Date of ref document: 19851010

ET Fr: translation filed
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

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

Ref country code: AT

Payment date: 19860825

Year of fee payment: 5

26N No opposition filed
PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: NL

Payment date: 19870930

Year of fee payment: 6

REG Reference to a national code

Ref country code: CH

Ref legal event code: PUE

Owner name: OMNISEC AG

REG Reference to a national code

Ref country code: FR

Ref legal event code: TP

REG Reference to a national code

Ref country code: GB

Ref legal event code: 732

ITPR It: changes in ownership of a european patent

Owner name: CESSIONE;OMNISEC AG

NLS Nl: assignments of ep-patents

Owner name: OMNISEC AG TE REGENSDORF, ZWITSERLAND.

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

Ref country code: GB

Effective date: 19880920

Ref country code: AT

Effective date: 19880920

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

Ref country code: LI

Effective date: 19880930

Ref country code: CH

Effective date: 19880930

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

Ref country code: NL

Effective date: 19890401

NLV4 Nl: lapsed or anulled due to non-payment of the annual fee
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: 19890531

REG Reference to a national code

Ref country code: CH

Ref legal event code: PL

GBPC Gb: european patent ceased through non-payment of renewal fee
PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: DE

Effective date: 19890601

REG Reference to a national code

Ref country code: FR

Ref legal event code: ST

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

Ref country code: SE

Payment date: 19890921

Year of fee payment: 8

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

Ref country code: SE

Effective date: 19900921

EUG Se: european patent has lapsed

Ref document number: 82810391.1

Effective date: 19910527