USRE38269E1 - Enhancement of speech coding in background noise for low-rate speech coder - Google Patents

Enhancement of speech coding in background noise for low-rate speech coder Download PDF

Info

Publication number
USRE38269E1
USRE38269E1 US09/422,820 US42282099A USRE38269E US RE38269 E1 USRE38269 E1 US RE38269E1 US 42282099 A US42282099 A US 42282099A US RE38269 E USRE38269 E US RE38269E
Authority
US
United States
Prior art keywords
speech
noise
low
vector
input
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
US09/422,820
Inventor
Yu-Jih Liu
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.)
Harris Corp
Original Assignee
ITT Manufacturing Enterprises LLC
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 ITT Manufacturing Enterprises LLC filed Critical ITT Manufacturing Enterprises LLC
Priority to US09/422,820 priority Critical patent/USRE38269E1/en
Application granted granted Critical
Publication of USRE38269E1 publication Critical patent/USRE38269E1/en
Assigned to Exelis Inc. reassignment Exelis Inc. ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS). Assignors: ITT CORPORATION
Anticipated expiration legal-status Critical
Assigned to HARRIS CORPORATION reassignment HARRIS CORPORATION MERGER (SEE DOCUMENT FOR DETAILS). Assignors: Exelis Inc.
Expired - Lifetime legal-status Critical Current

Links

Images

Classifications

    • GPHYSICS
    • G10MUSICAL INSTRUMENTS; ACOUSTICS
    • G10LSPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00Speech 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
    • G10L19/04Speech 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
    • GPHYSICS
    • G10MUSICAL INSTRUMENTS; ACOUSTICS
    • G10LSPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L21/00Processing of the speech or voice signal to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
    • G10L21/02Speech enhancement, e.g. noise reduction or echo cancellation
    • G10L21/0208Noise filtering
    • G10L21/0264Noise filtering characterised by the type of parameter measurement, e.g. correlation techniques, zero crossing techniques or predictive techniques
    • GPHYSICS
    • G10MUSICAL INSTRUMENTS; ACOUSTICS
    • G10LSPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L25/00Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
    • G10L25/03Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters
    • G10L25/09Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters the extracted parameters being zero crossing rates
    • GPHYSICS
    • G10MUSICAL INSTRUMENTS; ACOUSTICS
    • G10LSPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L25/00Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
    • G10L25/93Discriminating between voiced and unvoiced parts of speech signals

Definitions

  • the present invention relates to enhanced speech coding techniques for low-rate speech coders, and particularly, to improved speech frame analysis and vector quantization methods.
  • a low-bit-rate speech coder is disclosed in U.S. Pat. No. 4,975,956, issued to Y. J. Liu and J. H. Rothweiler, entitled “Low-Bit-Rate Speech Coder Using LPC Data Reduction Processing”, which is incorporated herein by reference.
  • This speech coder employs linear predictive coding (LPC) analysis to generate reflection coefficients for the input speech frames and pitch and gain parameters. To obtain a low bit rate of 400 bps, these parameters are further compressed.
  • the reflection coefficients are first converted to line spectrum frequencies (LSFs) and formants. For even frames, these spectral parameters are vector-quantized into clean codeword indices. Odd frames are omitted, and are regenerated by interpolation at the decoder end.
  • LPC linear predictive coding
  • the vector quantization module compares the spectral parameters for an input word against a vocabulary of codewords for which vector indices have been generated and stored during a training sequence, and the optimally matching codeword is selected for transmission. Pitch and gain bits are quantized using trellis coding. Output speech is reconstructed from the regenerated vector-quantization indices using a matching codebook at the decoder end.
  • this 400-bps speech coder has a high intelligibility for a low-bit-rate transmission.
  • a background of high noise such as in a helicopter or jet
  • the encoded speech becomes unintelligible.
  • conversion of voicing and spectral parameters in the high-noise environment is the key to the loss of intelligibility.
  • the LPC conversion causes a majority of voiced frames to become unvoiced. The result is a whispering LPC speech and an almost inaudible low-rate voice.
  • Even if the voicing is correct, spectral distortion causes the low-rate voice to be significantly muffled and buzzy.
  • the pitch has no audible errors, the gain has a predominantly annoying effect.
  • a two-step approach to conversion of voicing and spectral parameters is taken.
  • robust speech frame features whose distributions are not strongly affected by noise levels are generated.
  • linear programming is used to determine an optimum combination of these features.
  • a technique of adaptive vector quantization is also used in which a clean codebook is updated based upon an estimate of the background noise levels, and the “noisy” codebook is then searched for the best match with an input speech vector. The corresponding clean codeword is then selected for transmission and for synthesis at the receiver end. The results are better spectral reproduction and significant intelligibility enhancement over the previous coding approach.
  • FIG. 1 is a block diagram of the component steps of the encoding side of a speech coder system in accordance with the invention
  • FIG. 2 is a block diagram of the component steps of the decoding side of the speech coder system
  • FIG. 3 is a spectral plot of a typical spectrum of a noisy background, i.e., helicopter noise;
  • FIG. 4 is a spectral plot of typical LPC spectrums comparing different orders of LPC analysis in a noisy environment to a quiet environment;
  • FIG. 5 is a block diagram of the steps for performing the robust feature extraction, voicing decisions, noise estimation, and updating of a noisy codebook in accordance with the invention
  • FIGS. 6, 7 and 8 are plots of the low-band energy for input in a noisy environment at a 400 Hz bandwidth, a quiet environment, and a noisy environment at 800 Hz bandwidth, which demonstrates selection of a robust feature for extraction in accordance with the invention
  • FIGS. 9 and 10 are plots of the distribution of zero-crossing counts for input with and without helicopter noise, which demonstrates selection of another robust feature for robust voicing decisions in the invention
  • FIGS. 11 and 12 are histograms demonstrating the performance of the AMDF ratio (speech periodicity) measure with helicopter noise and without helicopter noise, respectively, as another robust feature for robust voicing decisions;
  • FIGS. 13 and 14 are histograms demonstrating the performance of the low-pass filtered, backward correlations measure with helicopter noise and without helicopter noise, respectively, as another feature for robust voicing decisions;
  • FIGS. 15 and 16 are histograms demonstrating the performance of the low-pass filtered, forward correlations measure with helicopter noise and without helicopter noise, respectively, as another feature for robust voicing decisions;
  • FIGS. 17 and 18 are histograms demonstrating the performance of the inverse-filtered backward correlations measure with helicopter noise and without helicopter noise, respectively, as another feature for robust voicing decisions;
  • FIGS. 19 and 20 are histograms demonstrating the performance of the inverse-filtered pitch prediction gain measure with helicopter noise and without helicopter noise, respectively, as another feature for robust voicing decisions;
  • FIG. 21 is a plot of the voiced error percentage for voicing decisions obtained by the enhanced encoding techniques of the present invention as compared to the prior encoding method.
  • FIG. 1 a block diagram of an encoding sequence in accordance with the present invention illustrates the processing of input speech frames.
  • the encoding processing is basically similar to that used in the aforementioned U.S. Pat. No. 4,975,956.
  • the LPC features are generated for each speech frame as an input processing step ( 8 ).
  • the gain and pitch parameters are extracted ( 10 , 12 ) and converted to gain and pitch bits by trellis coding ( 11 , 13 ).
  • LPC spectral parameters are extracted ( 19 ) and converted to line spectrum frequencies (LSPs) and formants for the subsequent vector quantization and/or interpolation (VQ/I) step ( 18 ) in a low-bit-rate transmission.
  • LSPs line spectrum frequencies
  • VQ/I vector quantization and/or interpolation
  • the decoding sequence of the speech coder system is shown having the usual operations as disclosed in U.S. Pat. No. 4,975,956.
  • the gain and pitch bits are decoded ( 21 , 22 ) using the reverse process of the encoding method.
  • the transmitted spectral bits of the “clean” codewords are decoded to LSP parameters ( 23 ) using a “clean” codebook ( 24 ).
  • the decoded parameters are then converted to LPC format ( 25 ) and synthesized to output speech.
  • enhanced techniques for low-bit-rate coding are applied to a 400-bps speech coder in the environment of helicopter noise.
  • the principles of the invention illustrated herein are applicable for other low bit rates of transmission and to other types of noisy environments as well.
  • spectral parameters are not quantized with every speech frame.
  • vector quantization is performed for every even frame, while interpolation is performed for every odd frame.
  • interpolation bits are sent representing an interpolation factor used for the combination of the spectral codeword of its previous frame and future frame. Based upon a frame period of 22.5 msec used in a standard encoder, the preferred bit allocations are illustrated in Table I.
  • the major operations for obtaining robust voicing decisions include preliminary processing, robust feature extraction, voicing classification, and voicing smoothing.
  • the specific parameters of these processing steps depend upon the different applications and environments. In the described example, voicing decisions are made every half frame or 11.25 msec. To enable robust voicing decisions, feature distributions without strong dependence on noise levels are necessary. The selected features are then combined using optimum weights in a linear combination.
  • the preliminary processing includes high-pass filtering, voicing-window decisions, and low-pass filtering.
  • the low-pass filtering is particularly important for robust voicing decisions in a high noise environment. Even though real-world noise, such as helicopter noise, is usually distributed in characteristic patterns, the spectral strength is normally weak in the low frequency band. A typical spectrum of helicopter noise is shown in FIG. 3 with three salient formants. However, the noise components tend to be weaker below 500 Hz. Therefore, if the cut-off frequency of the low-pass filter is set below 500 Hz, a majority of noise energy is rejected.
  • the high-pass filter is set at a frequency cutoff, such as 100 Hz, which eliminates low frequency background transients and mechanical noise.
  • voicing decisions are the determination of fundamental periodicity in the input speech.
  • the fundamental frequency is usually below 400 Hz. Therefore, a good choice of the cut-off frequency is about 420 Hz.
  • a spectral plot of the effect of the low-pass filter is illustrated in FIG. 4 for various LPC orders (10th, 12th, 14th) for a helicopter noise environment, as compared to an input of 10th LPC order in a quiet environment.
  • the 10th order LPC analysis (solid line) usually generates a good spectral contour.
  • the 10th order analysis becomes insufficient for reliable spectral representation.
  • the peak from the helicopter noise in the high-frequency band is clearly visible.
  • three dominant formants are visible for the 14th and 12th order LPC analysis, whereas the third formant for the 10th order spectrum is missing. Based upon this evaluation, it is determined that higher-order LPC analysis is clearly preferred for a noisy environment, and therefore, a 14th order LPC analysis is selected herein.
  • Low-band energy distribution is a measure of energy in the low-frequency band.
  • voiced speech has higher low-band energy than unvoiced speech.
  • FIG. 7 illustrates low-band energy in a quiet background
  • FIGS. 6 and 7 show similar distributions. For unvoiced speech, the energy distributions are mainly at bin (frequency band) 1 . For voiced speech, the distributions are spread over all bins, but with little overlap with the unvoiced bins. A comparison of FIGS.
  • the zero-crossing rate is a count of the high frequency content in the signal.
  • unvoiced speech has a higher zero-crossing count than voiced speech.
  • the zero-crossing count is accumulated by counting changes in sign of 1 n , which is defined as positive if 1 n > ⁇ D, and negative if 1 n ⁇ D.
  • the low-band energy is computed according to the following equation:
  • the low-band noise energy is first estimated by assuming there are always available 16 frames without speech activity. Using these 16 frames, the average low-band noise energy E N is computed as:
  • E j N (1 ⁇ 1 /j ) E j ⁇ 1 N +(1 /j ) E j
  • the low-band noise energy is updated at frame k if three conditions are satisfied. First, this frame must be unvoiced. Second, there must already be an accumulation of 16 continuous unvoiced frames before this current frame. Third, the ratio of current low-band energy to average low-band noise energy is less than 1.6. If all three conditions are satisfied at frame k, the average low-band noise energy is updated as follows:
  • E N k ( ⁇ fraction (63/64) ⁇ ) E N k ⁇ 1 +( ⁇ fraction (1/64) ⁇ ) E k
  • a is defined as follows:
  • the newest value of D for frame k is then used to compute the sign of every low-pass filtered sample.
  • the zero-crossing count then follows the procedure mentioned above.
  • the performance of the zero-crossing count is indicated in FIG. 9 for input with helicopter noise and FIG. 10 without helicopter noise.
  • the distributions are mainly below bin 2 .
  • the distributions are mainly above bin 3 . Therefore, the zero-crossing feature has not only good discriminations but also robust distributions.
  • AMDF measure Another feature found to have robustness for speech coding in a noisy environment is a measure of periodicity of speech, referred to herein as AMDF measure.
  • voiced speech has smaller AMDF values than unvoiced speech.
  • the AMDF computation is done using inverse-filtered speech by passing the low-pass signal through a second-order LPC filter. If v i represents the inverse-filtered speech sample, the AMDF value is computed as follows:
  • a fourth robust feature for voicing decisions in speech coding is a measure of correlation strength at the pitch period, which is a low-pass filtered backward correlation.
  • voiced speech has higher correlation values than unvoiced speech.
  • a fifth robust feature for voicing decisions is a measure of correlation strength via low-pass filtered forward correlation using a positive pitch lag.
  • Another feature is an inverse-filtered backward correlation, which is also a measure of correlation strength at the pitch period using backward pitch lag.
  • the main difference from the two previous correlation measures is the use of inverse-filtered speech v i .
  • represents the pitch period. Normalization is done the same way as before with and without pitch lag. The performance of this feature is demonstrated in FIG. 17 with helicopter noise and in FIG. 18 without helicopter noise.
  • voiced speech the distributions concentrate mainly at bins 9 and 10 .
  • unvoiced speech the distributions are scattered throughout all bins but with very little overlap with voiced bins. Thus, this feature is also suitable for enhancing voicing decisions.
  • a 1 and a 2 are prediction coefficients.
  • E represents the zeroth-order autocorrelation coefficient
  • R represents the normalized autocorrelation coefficients.
  • the second term in the above equation is the prediction gain.
  • the feature used for voicing decisions is slightly modified by rearranging the above equation as follows:
  • g For voiced speech, g has a larger values than for unvoiced speech.
  • the performance of this feature is demonstrated in FIG. 19 with helicopter noise and in FIG. 20 without helicopter noise.
  • the distributions concentrate mainly at bins 10 and 11 .
  • the distributions are scattered throughout all bins but with very little overlap with voiced bins.
  • this feature is also suitable for enhancing voicing decisions.
  • f j represents the jth feature
  • w represents a weight assigned to the feature
  • c is a constant.
  • a frame is classified as unvoiced if the reverse inequality holds.
  • the optimum weights for the combination are determined using linear programming analysis of representative training patterns in which helicopter noise is mixed with clean speech. The correct voicing decisions are measured against LPC analysis of the clean speech.
  • the optimum weights are found when h is maximized for the training patterns.
  • voicing smoothing may also be used to desensitize the voicing decisions to rapid transitions in values. Factors considered in smoothing include the discriminant magnitude of the voiced/unvoiced decisions, the onset of a rapid transition (between half frames), and continuity (which requires no instantaneous change of voicing). The voicing is determined every half frame of 11.25 msec. In order to facilitate the smoothing decisions, the final voicing decisions may be delayed two frames.
  • vector quantization (VQ/I module) is used to quantize the speech-feature vector for each frame.
  • a codebook C has a vocabulary of model feature vectors mapped to the corresponding codeword indices in a low number of bits.
  • the distortion from each model vector in the codebook is computed.
  • the index of the word having the minimum distortion is then selected for transmission.
  • voiced codewords have indices ranging from 0 to 991 and unvoiced codewords have indices ranging from 992 to 1023. If the codebook is designed in the same environment as the input speech, the optimal speech reproduction can be expected.
  • adaptive vector quantization is used in the present invention. This refers to the updating of the original codebook C based upon an estimate of the background noise level to generate a “noisy” codebook C′.
  • the noisy codebook C′ is searched to find the best match with the input vector, then the index for the corresponding clean codeword is selected for transmission, and is also used at the receiver end for synthesis.
  • the noise estimate for frame i is only performed if two conditions are satisfied: the frame i is decided to be unvoiced; and there must be an accumulation of more than a given number L of continuous unvoiced frames.
  • a counter n is reset on each voiced frame and incremented on each unvoiced frame. For n>L, the following noise estimates are computed:
  • N i a (63/64) N i a +(1/64) E i ⁇ 15)
  • the average noise reflection coefficients B a are further converted to noise autocorrelation coefficients R N .
  • R N and N a are utilized. This greatly reduces the probability of including speech frames.
  • the noise estimate parameters R N and N a are then used to add noise parameters to the codebook vectors.
  • the LSFs are converted to autocorrelation coefficients for each codeword in the clean codebook.
  • the higher-order LPC vector can enhance discrimination of the formants in noise
  • R C kj represents the updated codeword vector
  • Q i represents the mixing ratio at the ith frame.
  • the mixing ratio is determined from the noise amplitude N a i , as follows:
  • f is a factor determined empirically, according to the level of noise amplitude, as follows:
  • the codebook update is performed only when the counter n is at a multiple factor of J frames, which is adjustable depending upon the processor speed. For a very fast processor, the codebook could be updated every frame.
  • the mixing ratio Q i is determined empirically to depend upon the signal-to-noise ratio, as follows:
  • each codeword is further converted to line-spectrum frequencies (LSFs) and formants.
  • LSFs line-spectrum frequencies
  • the input reflection coefficients are also converted to LSFs and formants.
  • each vector for a voiced frame consists of 14 LSFs and two lowest frequency formants, and each vector for an unvoiced frame consists of 14 LSFs and one highest frequency formant.
  • the N codewords of the codebook are then searched to find the codeword which has the best match with an input vector, and the corresponding index is transmitted to the receiver.
  • the previous speech coder techniques as described in U.S. Pat. No. 4,975,956 could be implemented for 400-bps transmission using a 100 nsec DSP processor (equivalent to 10 Mips).
  • the enhanced techniques can be implemented using two such DSPs, if tree searching for codeword matches and 32-frame codebook updates are used.
  • FIG. 21 the performance of the new voicing decision techniques is illustrated in FIG. 21 as a plot of error percentage versus signal-to-noise ratio by neglecting those frames with gain less than 5 in the quiet background. For the reference plot of the old voicing decisions, the error percentage is zero at a signal-to-noise ratio of 30 dB.
  • the error percentage climbs abruptly to 66% at a signal-to-noise ratio of ⁇ 10 dB.
  • the error percentage increases only about 1% as the signal-to-noise ratio drops from 30 dB to ⁇ 10 dB. If all voiced frames are considered regardless of gain, the error percentage increases from about 2% at S/N of 30 dB to 6% at S/N of ⁇ 10 dB. For unvoiced frames, the robustness remains about the same. The superiority of the enhanced speech coding techniques is thus clearly demonstrated.
  • the adaptive vector quantization was also tested using noisy speech samples of the same two types.
  • the listening tests showed that there is always an intelligibility improvement using codebook adaptation.
  • the degree of improvement depends upon three factors: signal-to-noise ratio; rate of codebook update; and the use of preemphasis.
  • Tests on the effect of S/N ratio showed that the intelligibility improvement is quite significant at very low S/N such as ⁇ 10 dB. for higher S/N, the improvement is less audible, which is expected since there is less noise corruption.
  • the intelligibility improvement seems to depend only a little on the rate of codebook update. Updating with every frame appeared only slightly better than updating every 32 frames.
  • preemphasis tests of mixed speech showed that the same factor as used in the clean codebook should be used, whereas for recorded speech, a smaller preemphasis factor can significantly improve intelligibility.

Abstract

A speech coding system employs measurements of robust features of speech frames whose distribution are not strongly affected by noise/levels to make voicing decisions for input speech occurring in a noisy environment. Linear programing analysis of the robust features and respective weights are used to determine an optimum linear combination of these features. The input speech vectors are matched to a vocabulary of codewords in order to select the corresponding, optimally matching codeword. Adaptive vector quantization is used in which a vocabulary of words obtained in a quiet environment is updated based upon a noise estimate of a noisy environment in which the input speech occurs, and the “noisy” vocabulary is then searched for the best match with an input speech vector. The corresponding clean codeword index is then selected for transmission and for synthesis at the receiver end. The results are better spectral reproduction and significant intelligibility enhancement over prior coding approaches. Robust features found to allow robust voicing decisions include: low-band energy; zero-crossing counts adapted for noise level; AMDF ratio (speech periodicity) measure; low-pass filtered backward correlation; low-pass filtered forward correlation; inverse-filtered backward correlation; and inverse-filtered pitch prediction gain measure.

Description

This is a continuation of application Ser. No. 07/695,571 filed May 3, 1991 now abandoned.
The United States Government has rights in this invention pursuant to RADC Contract F30602-89-C-0118 awarded by the Department of the Air Force.
FIELD OF THE INVENTION
The present invention relates to enhanced speech coding techniques for low-rate speech coders, and particularly, to improved speech frame analysis and vector quantization methods.
BACKGROUND OF THE INVENTION
A low-bit-rate speech coder is disclosed in U.S. Pat. No. 4,975,956, issued to Y. J. Liu and J. H. Rothweiler, entitled “Low-Bit-Rate Speech Coder Using LPC Data Reduction Processing”, which is incorporated herein by reference. This speech coder employs linear predictive coding (LPC) analysis to generate reflection coefficients for the input speech frames and pitch and gain parameters. To obtain a low bit rate of 400 bps, these parameters are further compressed. The reflection coefficients are first converted to line spectrum frequencies (LSFs) and formants. For even frames, these spectral parameters are vector-quantized into clean codeword indices. Odd frames are omitted, and are regenerated by interpolation at the decoder end. The vector quantization module compares the spectral parameters for an input word against a vocabulary of codewords for which vector indices have been generated and stored during a training sequence, and the optimally matching codeword is selected for transmission. Pitch and gain bits are quantized using trellis coding. Output speech is reconstructed from the regenerated vector-quantization indices using a matching codebook at the decoder end.
In a quiet background, this 400-bps speech coder has a high intelligibility for a low-bit-rate transmission. However, in a background of high noise, such as in a helicopter or jet, the encoded speech becomes unintelligible. A detailed study has shown that conversion of voicing and spectral parameters in the high-noise environment is the key to the loss of intelligibility. The LPC conversion causes a majority of voiced frames to become unvoiced. The result is a whispering LPC speech and an almost inaudible low-rate voice. Even if the voicing is correct, spectral distortion causes the low-rate voice to be significantly muffled and buzzy. Although the pitch has no audible errors, the gain has a predominantly annoying effect.
SUMMARY OF INVENTION
It is therefore a principal object of the invention to provide an improved low-bit-rate speech coder capable of high quality speech coding in a high-noise environment. In accordance with the invention, a two-step approach to conversion of voicing and spectral parameters is taken. In the first step, robust speech frame features whose distributions are not strongly affected by noise levels are generated. In the second step, linear programming is used to determine an optimum combination of these features. A technique of adaptive vector quantization is also used in which a clean codebook is updated based upon an estimate of the background noise levels, and the “noisy” codebook is then searched for the best match with an input speech vector. The corresponding clean codeword is then selected for transmission and for synthesis at the receiver end. The results are better spectral reproduction and significant intelligibility enhancement over the previous coding approach.
In a preferred implementation of the system for the environment of helicopter, it is found that the following features are well distributed to allow good discrimination between voiced and unvoiced speech: (1) low-band energy; (2) zero-crossing counts adapted for noise level; (3) AMDF ratio (speech periodicity) measure; (4) low-pass filtered, backward correlation; (5) low-pass filtered, forward correlation; (6) inverse-filtered backward correlation; and (7) inverse-filtered pitch prediction gain measure. By linear programming analysis, five of these robust features are determined to significantly improve voicing decisions in the speech coder system. Adaptive vector quantization, using estimates of the average noise amplitude and average noise reflection coefficients to update codebook vectors, significantly improves input vector matching.
BRIEF DESCRIPTION OF DRAWINGS
The above principles and further features and advantages of the invention are described in detail below in conjunction with the drawings, of which:
FIG. 1 is a block diagram of the component steps of the encoding side of a speech coder system in accordance with the invention;
FIG. 2 is a block diagram of the component steps of the decoding side of the speech coder system;
FIG. 3 is a spectral plot of a typical spectrum of a noisy background, i.e., helicopter noise;
FIG. 4 is a spectral plot of typical LPC spectrums comparing different orders of LPC analysis in a noisy environment to a quiet environment;
FIG. 5 is a block diagram of the steps for performing the robust feature extraction, voicing decisions, noise estimation, and updating of a noisy codebook in accordance with the invention;
FIGS. 6, 7 and 8 are plots of the low-band energy for input in a noisy environment at a 400 Hz bandwidth, a quiet environment, and a noisy environment at 800 Hz bandwidth, which demonstrates selection of a robust feature for extraction in accordance with the invention;
FIGS. 9 and 10 are plots of the distribution of zero-crossing counts for input with and without helicopter noise, which demonstrates selection of another robust feature for robust voicing decisions in the invention;
FIGS. 11 and 12 are histograms demonstrating the performance of the AMDF ratio (speech periodicity) measure with helicopter noise and without helicopter noise, respectively, as another robust feature for robust voicing decisions;
FIGS. 13 and 14 are histograms demonstrating the performance of the low-pass filtered, backward correlations measure with helicopter noise and without helicopter noise, respectively, as another feature for robust voicing decisions;
FIGS. 15 and 16 are histograms demonstrating the performance of the low-pass filtered, forward correlations measure with helicopter noise and without helicopter noise, respectively, as another feature for robust voicing decisions;
FIGS. 17 and 18 are histograms demonstrating the performance of the inverse-filtered backward correlations measure with helicopter noise and without helicopter noise, respectively, as another feature for robust voicing decisions;
FIGS. 19 and 20 are histograms demonstrating the performance of the inverse-filtered pitch prediction gain measure with helicopter noise and without helicopter noise, respectively, as another feature for robust voicing decisions;
FIG. 21 is a plot of the voiced error percentage for voicing decisions obtained by the enhanced encoding techniques of the present invention as compared to the prior encoding method.
DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS
Referring to FIG. 1, a block diagram of an encoding sequence in accordance with the present invention illustrates the processing of input speech frames. The encoding processing is basically similar to that used in the aforementioned U.S. Pat. No. 4,975,956. The LPC features are generated for each speech frame as an input processing step (8). The gain and pitch parameters are extracted (10, 12) and converted to gain and pitch bits by trellis coding (11, 13). LPC spectral parameters are extracted (19) and converted to line spectrum frequencies (LSPs) and formants for the subsequent vector quantization and/or interpolation (VQ/I) step (18) in a low-bit-rate transmission. The main differences are in the employment of robust LPC feature extraction and voicing decision (14, 15), noise estimation (16), and updating of a clean codebook (17), in order to provide better spectral representation and codeword matching for input speech in a noisy environment. Upon optimal “noisy” codeword matching, the corresponding “clean” codeword indices are then transmitted (20).
In FIG. 2, the decoding sequence of the speech coder system is shown having the usual operations as disclosed in U.S. Pat. No. 4,975,956. The gain and pitch bits are decoded (21, 22) using the reverse process of the encoding method. The transmitted spectral bits of the “clean” codewords are decoded to LSP parameters (23) using a “clean” codebook (24). The decoded parameters are then converted to LPC format (25) and synthesized to output speech.
To identify speech parameters crucial for intelligibility in a high-noise environment, such as helicopter noise, several listening tests were performed comparing the performance of a clean speech file with a noisy speech file through LPC analysis. The listening tests showed that the voicing and spectrum parameters of LPC conversion must be enhanced to obtain intelligible speech coding. Also, the gain parameter requires correction to eliminate an annoying noise effect.
In the following preferred embodiments of the invention, enhanced techniques for low-bit-rate coding are applied to a 400-bps speech coder in the environment of helicopter noise. However, the principles of the invention illustrated herein are applicable for other low bit rates of transmission and to other types of noisy environments as well.
To achieve the low bit rate of 400 bps, spectral parameters are not quantized with every speech frame. As described in the aforementioned U.S. Pat. No. 4,975,956, vector quantization is performed for every even frame, while interpolation is performed for every odd frame. For the odd frame, interpolation bits are sent representing an interpolation factor used for the combination of the spectral codeword of its previous frame and future frame. Based upon a frame period of 22.5 msec used in a standard encoder, the preferred bit allocations are illustrated in Table I.
TABLE I
Parameter Even Frame Odd Frame Two Frames
Spectral
10 0 10
Gain 2 2 4
Pitch 1 1 2
Interpolation 0 2 2
Total: 13 5 18
For even frames, a total of 13 bits are allocated. For odd frames, only 5 bits are allocated. For every pair of even and odd frames, a total of 18 bits are used. Assuming a 45 msec period for every two frames, this bit allocation scheme fits within the 400 bits/second requirement.
The major operations for obtaining robust voicing decisions include preliminary processing, robust feature extraction, voicing classification, and voicing smoothing. The specific parameters of these processing steps depend upon the different applications and environments. In the described example, voicing decisions are made every half frame or 11.25 msec. To enable robust voicing decisions, feature distributions without strong dependence on noise levels are necessary. The selected features are then combined using optimum weights in a linear combination.
Following the usual operations in LPC analysis, the preliminary processing includes high-pass filtering, voicing-window decisions, and low-pass filtering. The low-pass filtering is particularly important for robust voicing decisions in a high noise environment. Even though real-world noise, such as helicopter noise, is usually distributed in characteristic patterns, the spectral strength is normally weak in the low frequency band. A typical spectrum of helicopter noise is shown in FIG. 3 with three salient formants. However, the noise components tend to be weaker below 500 Hz. Therefore, if the cut-off frequency of the low-pass filter is set below 500 Hz, a majority of noise energy is rejected. The high-pass filter is set at a frequency cutoff, such as 100 Hz, which eliminates low frequency background transients and mechanical noise.
Voicing decisions are the determination of fundamental periodicity in the input speech. For human speech, the fundamental frequency is usually below 400 Hz. Therefore, a good choice of the cut-off frequency is about 420 Hz. Using the Remetez exchange algorithm, a low-pass filter with cut-off frequency at 420 Hz and transition frequency at 650 Hz is used. This filter is selected to be even-symmetric with 40 taps. Typical values for the first 20 taps, hk, k=0, . . . 19, are illustrated in Table II.
TABLE II
Tap Value
h0 0.01787624
h1  0.02237480
h2  0.002685766
h3  0.01303141
h4  −0.0001381086
h5  −0.001044893
h6  −0.01218479
h7  −0.01683313
h8  −0.02370618
h9  −0.02454394
h10 −0.02252495
h11 −0.01385341
h12 −0.003387984
h13 0.01871256
h14 0.04112903
h15 0.0654924
h16 0.08902424
h17 0.109489
h18 0.124534
h19 0.132543
The next 20 tap values are determined from symmetry and are given as follows:
h39−n=hmn=0 1 . . . 19
All the features are extracted in the low-frequency band to minimize the noise corruption. The filtered speech can be computed as follows, where the input speech after high-pass filtering is sn: l n = i = 0 19 h k ( s n - k + s n + k - 39 )
Figure USRE038269-20031007-M00001
A spectral plot of the effect of the low-pass filter is illustrated in FIG. 4 for various LPC orders (10th, 12th, 14th) for a helicopter noise environment, as compared to an input of 10th LPC order in a quiet environment. In a quiet background, the 10th order LPC analysis (solid line) usually generates a good spectral contour. However, as the noise level increases, the 10th order analysis becomes insufficient for reliable spectral representation. The peak from the helicopter noise in the high-frequency band is clearly visible. In the low-frequency band, three dominant formants are visible for the 14th and 12th order LPC analysis, whereas the third formant for the 10th order spectrum is missing. Based upon this evaluation, it is determined that higher-order LPC analysis is clearly preferred for a noisy environment, and therefore, a 14th order LPC analysis is selected herein.
Two major criteria for good robust features are that their distributions must not strongly depend upon noise levels and that they must have good voiced/unvoiced discrimination. Speech samples were evaluated for male and female speakers in a quiet environment with a signal-to-noise ratio of 30 dB, and in a noisy environment with a signal-to-noise ratio of −10 dB. Robust features were then selected on the basis of both low-frequency distributions and voiced/unvoiced discriminations, using low-band energy measurements, zero-crossing rate, and selected correlation calculations as factors. The processing steps for the enhancement techniques of the present invention, including extraction of the robust features, their use for robust voicing decisions, noise estimation, and updating a clean codebook, are illustrated in the block diagram of FIG. 5.
Low-band energy distribution is a measure of energy in the low-frequency band. Typically, voiced speech has higher low-band energy than unvoiced speech. For normalization purposes, this energy is divided by the average voiced energy, as represented by the following equation, wherein 1 represents the speech signal after 100 Hz high-pass filtering and 420 Hz low-pass filtering, and LEA represents the average voiced energy in the low band: LE = Σ R i l LEA
Figure USRE038269-20031007-M00002
FIG. 6 illustrates a histogram of low-band energy with helicopter noise at S/N=−10 dB, FIG. 7 illustrates low-band energy in a quiet background, and FIG. 8 illustrates low-band energy with twice the bandwidth (i.e., increased to 800 Hz) with helicopter noise at S/N=−10 dB. FIGS. 6 and 7 show similar distributions. For unvoiced speech, the energy distributions are mainly at bin (frequency band) 1. For voiced speech, the distributions are spread over all bins, but with little overlap with the unvoiced bins. A comparison of FIGS. 6 and 8 shows that discrimination is clearly better using the lower bandwidth, since the voiced distribution is reduced at bin 1, where the unvoiced distribution dominates, and increased at bin 11, where the unvoiced distribution is minimal. On the basis of this evaluation, the lower band-width of 400 Hz is selected for robust feature extraction.
Another feature found to have robustness for good voicing decisions is measurement of the zero-crossing rate, i.e., the number of times the input signal crosses a zero (or reference) axis. In effect, it is a count of the high frequency content in the signal. Typically, unvoiced speech has a higher zero-crossing count than voiced speech. The zero-crossing count is accumulated by counting changes in sign of 1n, which is defined as positive if 1n>±D, and negative if 1n<±D.
To make the zero-crossing count robust in a noisy environment, it is counted in the low-frequency band, and the dither D is appropriately adjusted in noise. The low-band energy is computed according to the following equation:
E=Σ(1n)2
For the jth frame, this energy is indicated by Ej. The low-band noise energy is first estimated by assuming there are always available 16 frames without speech activity. Using these 16 frames, the average low-band noise energy EN is computed as:
E j N=(1−1/j)E j−1 N+(1/j)E j
j=1, . . . ,M=16
After these 16 frames, the low-band noise energy is updated at frame k if three conditions are satisfied. First, this frame must be unvoiced. Second, there must already be an accumulation of 16 continuous unvoiced frames before this current frame. Third, the ratio of current low-band energy to average low-band noise energy is less than 1.6. If all three conditions are satisfied at frame k, the average low-band noise energy is updated as follows:
E N k=({fraction (63/64)})E N k−1+({fraction (1/64)})E k
To adapt the coefficient D to noise, a quantity a is defined as follows:
a=E N k/7+1
After evaluating a, a minimum between a and 20 is selected. Next, the quantity b, which is the maximum between the selected minimum and 10 is obtained. Mathematically, b is given by the following equation:
b=max [min (a, 20),10]
where max represents the maximum and min represents the minimum. The adaptation coefficient D is updated as follows:
D=b, if E k /E N k<1.6
D=b/2, if E k /E N k>1.6
The newest value of D for frame k is then used to compute the sign of every low-pass filtered sample. The zero-crossing count then follows the procedure mentioned above. The performance of the zero-crossing count is indicated in FIG. 9 for input with helicopter noise and FIG. 10 without helicopter noise. For voiced speech, the distributions are mainly below bin 2. For unvoiced speech, the distributions are mainly above bin 3. Therefore, the zero-crossing feature has not only good discriminations but also robust distributions.
Another feature found to have robustness for speech coding in a noisy environment is a measure of periodicity of speech, referred to herein as AMDF measure. Typically, voiced speech has smaller AMDF values than unvoiced speech. The AMDF computation is done using inverse-filtered speech by passing the low-pass signal through a second-order LPC filter. If vi represents the inverse-filtered speech sample, the AMDF value is computed as follows:
AMDF=Σ|v i −v i+τ|
where τ represents the 60 possible pitch lags ranging from 20 samples to 156 samples. These 60 possible lags are searched to find a maximum and a minimum. This feature is then computed as the ratio of maximum AMDF to minimum AMDF, i.e., R=max(AMDF)/min(AMDF). The performance of the AMDF ratio measure is demonstrated in FIG. 11 with helicopter noise and in FIG. 12 without helicopter noise. For voiced speech, the distributions are scattered throughout all bins. There is only a slight overlap with unvoiced speech at bin 2. Both histograms are also quite similar without a strong dependence on noise, and thus demonstrates this to be another robust feature.
A fourth robust feature for voicing decisions in speech coding is a measure of correlation strength at the pitch period, which is a low-pass filtered backward correlation. Typically, voiced speech has higher correlation values than unvoiced speech. However, the correlation is done using negative pitch lags, and is defined mathematically as follows: R b = ( Σ l i l i - τ ) 2 Σ l i 2 Σ i - τ 2
Figure USRE038269-20031007-M00003
where τ represents the pitch period. The above equation shows this feature normalized with respect to low-pass energy with and without negative pitch lag. The performance of this feature is demonstrated in FIG. 13 with helicopter noise and in FIG. 14 without helicopter noise. For both figures, the voiced speech has values predominantly at bin 10 while the unvoiced speech has values below bin 6. Thus, the distributions in both figures are very similar and have good discrimination between voiced and unvoiced speech, and this feature demonstrates the necessary robustness for allowing enhanced voicing decisions.
A fifth robust feature for voicing decisions is a measure of correlation strength via low-pass filtered forward correlation using a positive pitch lag. Typically, the voiced speech has higher correlation values than unvoiced speech. It is defined mathematically as follows: R f = ( Σ l i l i + τ ) 2 Σ l i 2 Σ i + τ 2
Figure USRE038269-20031007-M00004
where τ represents the pitch period. The above equation shows this feature normalized with respect to low-pass energy with and without positive pitch lag. The performance of this feature is demonstrated in FIG. 15 with helicopter noise and in FIG. 16 without helicopter noise. Both distributions and discriminations show similar characteristics as the backward correlations.
Another feature is an inverse-filtered backward correlation, which is also a measure of correlation strength at the pitch period using backward pitch lag. The main difference from the two previous correlation measures is the use of inverse-filtered speech vi. Again, the voiced speech has higher correlation values than unvoiced speech. It is defined mathematically as follows: R 1 b = ( Σ v i v i - τ ) 2 Σ v i 2 Σ v i - τ
Figure USRE038269-20031007-M00005
where τ represents the pitch period. Normalization is done the same way as before with and without pitch lag. The performance of this feature is demonstrated in FIG. 17 with helicopter noise and in FIG. 18 without helicopter noise. For voiced speech, the distributions concentrate mainly at bins 9 and 10. For unvoiced speech, the distributions are scattered throughout all bins but with very little overlap with voiced bins. Thus, this feature is also suitable for enhancing voicing decisions.
Another feature found to have robustness for voicing decisions is the second-order pitch-prediction gain after inverse filtering, which is also a measure of speech periodicity). The pitch-prediction residual is given by the following equation:
δ=Σ(ξn −a 1 v n−τ+1 −a 2ξn−τ)2
where a1 and a2 are prediction coefficients. The optimum prediction coefficients can be found by differentiating δ with respect to both a1 and a2. Substituting these two optimum values into the above equation, the optimum prediction residual is expressed as follows: δ = E ( 1 - R e - 1 2 - R τ 2 - 2 R i R τ - 1 RC 1 - R 1 2 )
Figure USRE038269-20031007-M00006
where E represents the zeroth-order autocorrelation coefficient and R represents the normalized autocorrelation coefficients. The second term in the above equation is the prediction gain. The feature used for voicing decisions is slightly modified by rearranging the above equation as follows:
g=R 1 2 +R τ−1 2 +R τ 2−2R 1 R τ−1 R τ
For voiced speech, g has a larger values than for unvoiced speech. The performance of this feature is demonstrated in FIG. 19 with helicopter noise and in FIG. 20 without helicopter noise. For voiced speech, the distributions concentrate mainly at bins 10 and 11. For unvoiced speech, the distributions are scattered throughout all bins but with very little overlap with voiced bins. Thus, this feature is also suitable for enhancing voicing decisions.
All of the seven features discussed above are found to have good discriminations and robust distributions. Further information on the features can be found in the references, “Voices/Unvoiced Classification of Speech with Applications to the U.S. Government LPC-10E Algorithm” by J. Campbell and T. Tremain, ICASSP'86 and “An Enhanced LPC Vocoder with No Voiced/Unvoiced Switch” by S. Y. Kwon and A. J. Goldberg, ASSP-32, 1984. Other robust features may be found using the same criteria. The histogram plots show the there are always some overlaps between voiced bins and unvoiced bins for all features. Therefore, no single feature should be relied upon to make voicing decisions. To minimize potential error, a combination of the features is utilized, as depicted in FIG. 5. A frame may be classified as being voiced if the following inequality of feature combination holds:
Σw j f j >c,
where fj represents the jth feature, w represents a weight assigned to the feature, and c is a constant. A frame is classified as unvoiced if the reverse inequality holds. The optimum weights for the combination are determined using linear programming analysis of representative training patterns in which helicopter noise is mixed with clean speech. The correct voicing decisions are measured against LPC analysis of the clean speech. The linear programming analysis solves the inequality equations using the well-known simplex method of linear optimization by first converting them to equalities using slack and surplus variables: j = 1 n = f ij w j + w n + i = ci = l , k j = 1 n = f ij w j - w n + i = ci = k + l , k + m
Figure USRE038269-20031007-M00007
The above equations are solved by maximizing a quantity h. A hyperplane is found separating the voiced region from the unvoiced region, and h is defined to be the average distance between the voiced region and the unvoiced region, given as follows: h = j = 1 n [ ( 1 m i = k + 1 k + m f ij ) - ( 1 k i = 1 k f ij ) ] w j
Figure USRE038269-20031007-M00008
The optimum weights are found when h is maximized for the training patterns.
The simplex method starts with an initial feasible solution. However, an initial solution is difficult to find if the number of equations becomes large. To simplify the initial solution, some artificial values are introduced, and the basic equations become as follows: j = 1 n = f ij w j + w n + i = ci = l , k j = 1 n = f ij w j - w n + i + w n + m + i = ci = k + l . k + m
Figure USRE038269-20031007-M00009
where the weights wj, j=n+m+k+1, . . . n+k+2m are artificial variables. All the artificial variables are also assigned the negative maximum weight. The quantity h is then given below: h = j = 1 n [ ( 1 m i = k + 1 k + m f ij ) - ( 1 k i = 1 k f ij ) ] w j - i = 1 m Mw n + i + m + l
Figure USRE038269-20031007-M00010
where M is an arbitrarily large number. The solutions are then iterated until all artificial variables are removed and the quantity h can no longer be increased. For a further discussion of this type of linear programming analysis, reference is made to “A Procedure For Using Pattern Classification Techniques To Obtain A Voiced/Unvoiced Classifier”, by L. Siegel, IEEE Trans., ASSP-27, February 1979, and Linear Programming, by G. Hadley, published by Addison Wesley, 1963.
Analyses performed by the above-described procedures showed that the five most useful features for the helicopter-noise patterns are low-band energy, zero-crossing rate, AMDF measure, low-pass filtered backward correlation, and inverse-filtered pitch-prediction gain. Therefore, these five features are combined in this example to make decisions as to when the input speech frames are voiced or unvoiced. Voicing smoothing may also be used to desensitize the voicing decisions to rapid transitions in values. Factors considered in smoothing include the discriminant magnitude of the voiced/unvoiced decisions, the onset of a rapid transition (between half frames), and continuity (which requires no instantaneous change of voicing). The voicing is determined every half frame of 11.25 msec. In order to facilitate the smoothing decisions, the final voicing decisions may be delayed two frames.
Referring again to FIG. 5, vector quantization (VQ/I module) is used to quantize the speech-feature vector for each frame. A codebook C has a vocabulary of model feature vectors mapped to the corresponding codeword indices in a low number of bits. For each input vector, the distortion from each model vector in the codebook is computed. The index of the word having the minimum distortion is then selected for transmission. For a 10-bit codebook used in the study, voiced codewords have indices ranging from 0 to 991 and unvoiced codewords have indices ranging from 992 to 1023. If the codebook is designed in the same environment as the input speech, the optimal speech reproduction can be expected. However, if the codebook is designed in a quiet background while the input speech comes from a noisy environment, selection of the optimum word becomes difficult. Noise cancellation is one conventional technique to remove the background noise from the input speech. However, if not done properly, spectral distortion is also introduced. To overcome this drawback, adaptive vector quantization is used in the present invention. This refers to the updating of the original codebook C based upon an estimate of the background noise level to generate a “noisy” codebook C′. The noisy codebook C′ is searched to find the best match with the input vector, then the index for the corresponding clean codeword is selected for transmission, and is also used at the receiver end for synthesis.
A background noise estimate can be performed in two One is the average noise amplitude Na i, and the other is the average noise reflection coefficients Ba ij, j=1, . . . P, where i represents the current frame number, j represents the coefficient number, and P is the LPC order. To prevent using voiced or unvoiced speech in the computation, the noise estimate for frame i is only performed if two conditions are satisfied: the frame i is decided to be unvoiced; and there must be an accumulation of more than a given number L of continuous unvoiced frames. To count continuous unvoiced frames, a counter n is reset on each voiced frame and incremented on each unvoiced frame. For n>L, the following noise estimates are computed:
N i a=(63/64)N i a+(1/64)E i−15)
B a ij=(63/64)B a ij=(63/64)B a ij+(1/64)B (i−15)j j=1,P
The average noise reflection coefficients Ba are further converted to noise autocorrelation coefficients RN. To compute RN and Na at frame i, the values at frame i-15 are utilized. This greatly reduces the probability of including speech frames. The noise estimate parameters RN and Na are then used to add noise parameters to the codebook vectors.
The LSFs are converted to autocorrelation coefficients for each codeword in the clean codebook. As described previously, the higher-order LPC vector can enhance discrimination of the formants in noise, and the codebook is preferably designed using a 14th-order LPC analysis, i.e. P=14. Assuming there are N codewords in the codebook, and each codeword has P autocorrelation coefficients, and RC kj represents the jth coefficient of the kth codeword, then the noise autocorrelation coefficients are added to each codeword as follows: R ij C = R ij C + Q l * R j N Q l + 1 j = 1 , P , k = 1 , N
Figure USRE038269-20031007-M00011
where RC kj represents the updated codeword vector and Qi represents the mixing ratio at the ith frame. The mixing ratio is determined from the noise amplitude Na i, as follows:
Q i=(N a i *f/70)2
where f is a factor determined empirically, according to the level of noise amplitude, as follows:
f=1.5, for N a i<10
f=1.2, for 10<N a i<24
f=1.0, for N a i>24
The codebook update is performed only when the counter n is at a multiple factor of J frames, which is adjustable depending upon the processor speed. For a very fast processor, the codebook could be updated every frame. In this case, the mixing ratio Qi is determined empirically to depend upon the signal-to-noise ratio, as follows:
Q i=(N a i /S i)2
where Si represents the speech amplitude at frame i. This mixing ratio is used in the same way as described above to compute the updated codewords.
After computing the updated codebook of autocorrelation coefficients, each codeword is further converted to line-spectrum frequencies (LSFs) and formants. The input reflection coefficients are also converted to LSFs and formants. For the 14th-order LPC analysis, each vector for a voiced frame consists of 14 LSFs and two lowest frequency formants, and each vector for an unvoiced frame consists of 14 LSFs and one highest frequency formant. The N codewords of the codebook are then searched to find the codeword which has the best match with an input vector, and the corresponding index is transmitted to the receiver.
In the receiver, only the clean codebook of N codewords is stored. The received index is used to select the corresponding clean codeword for synthesis. Thus, even though an updated (noisy) codebook is used to produce better matching, a clean codebook is used for synthesis of output speech in which spectral distortion is greatly reduced.
The previous speech coder techniques as described in U.S. Pat. No. 4,975,956 could be implemented for 400-bps transmission using a 100 nsec DSP processor (equivalent to 10 Mips). The enhanced techniques can be implemented using two such DSPs, if tree searching for codeword matches and 32-frame codebook updates are used. Using the voicing decisions from LPC analysis of clean speech via the prior techniques as a reference, the performance of the new voicing decision techniques is illustrated in FIG. 21 as a plot of error percentage versus signal-to-noise ratio by neglecting those frames with gain less than 5 in the quiet background. For the reference plot of the old voicing decisions, the error percentage is zero at a signal-to-noise ratio of 30 dB. However, the error percentage climbs abruptly to 66% at a signal-to-noise ratio of −10 dB. Using the new voicing decision techniques, the error percentage increases only about 1% as the signal-to-noise ratio drops from 30 dB to −10 dB. If all voiced frames are considered regardless of gain, the error percentage increases from about 2% at S/N of 30 dB to 6% at S/N of −10 dB. For unvoiced frames, the robustness remains about the same. The superiority of the enhanced speech coding techniques is thus clearly demonstrated.
Informal listening tests were also conducted both for speech samples in which noise was mixed with clean speech and those recorded in the actual helicopter noise environment. The listening tests showed none of the previous whispering LPC speech for either type of sample. The 400-bps speech in the noisy environment was reproduced as clearly audible but with some degradation in quality. To improve speech intelligibility, improved vector quantization can be applied.
The adaptive vector quantization was also tested using noisy speech samples of the same two types. The listening tests showed that there is always an intelligibility improvement using codebook adaptation. The degree of improvement depends upon three factors: signal-to-noise ratio; rate of codebook update; and the use of preemphasis. Tests on the effect of S/N ratio showed that the intelligibility improvement is quite significant at very low S/N such as −10 dB. for higher S/N, the improvement is less audible, which is expected since there is less noise corruption. The intelligibility improvement seems to depend only a little on the rate of codebook update. Updating with every frame appeared only slightly better than updating every 32 frames. As to preemphasis, tests of mixed speech showed that the same factor as used in the clean codebook should be used, whereas for recorded speech, a smaller preemphasis factor can significantly improve intelligibility.
The specific embodiments of the invention described herein are intended to be illustrative only, and many other variations and modifications may be made thereto in accordance with the principles of the invention. All such embodiments and variations and modifications thereof are considered to be within the scope of the invention, as defined in the following claims.

Claims (24)

I claim:
1. In a method of low-bit-rate speech coding of input speech occurring in a noisy environment, for a system which employs linear predictive coding (LPC) analysis of input speech frames to generate reflection coefficients, conversion of the reflection coefficients to vectors representing spectral parameters of the input speech frames, and matching of the spectral parameter vectors against reference vectors of a vocabulary of codewords generated in a training sequence in order to select the corresponding index of an optimally matching codeword for transmission,
the improvement comprising the steps of:
selecting a set of at least two features which are characterized by a probability distribution which is not strongly affected in the noisy environment and which allow discrimination between voiced and unvoiced input speech, wherein said selected features include the feature of zero-crossing counts which are based on average noise energy;
measuring the selected features for input speech frames; and
using said feature measurements to make voiced/unvoiced speech decisions in order to select the voice/unvoiced excitation for speech synthesis in the receiver;
using noise estimates to update the reference vectors of the vocabulary of codewords, wherein new reference vectors are generated corresponding to said vocabulary of codewords in the noisy environment, said noise estimates including noise amplitude and noise reflection coefficients, wherein said noise estimate for speech frame I is performed only if the ith speech frame is unvoiced and more than a given number L of continuous unvoiced speech frames are accumulated, in order to prevent using voiced or unvoiced speech in the noise estimate.
2. A low-bit-rate speech coding method according to claim 1, wherein said voicing decision step includes the substep of determining a linear combination of said features which provides a high voiced/unvoiced discrimination capability; and determining respective weights to be applied to said features in order to obtain an optimal linear combination of said features.
3. A low-bit-rate speech coding method according to claim 2, wherein said weights determining substep of said voicing decision step is performed using the simplex method for obtaining a maximum quantity h for an average distance between voiced and unvoiced regions of the input speech.
4. A low-bit-rate speech coding method according to claim 1, wherein said selected features include the feature of low-band energy.
5. A low-bit-rate speech coding method according to claim 1, wherein said selected features include an AMDF ratio (speech periodicity) measure.
6. A low-bit-rate speech coding method according to claim 1, wherein said selected features include a backward correlations measure responsive to low-pass-filtered speech energy.
7. A low-bit-rate speech coding method according to claim 1, wherein said selected features include a forward correlations measure responsive to low-pass-filtered speech energy.
8. A low-bit-rate speech coding method according to claim 1, wherein said selected features include a backward correlations measure responsive to inverse-filtered speech energy.
9. A low-bit-rate speech coding method according to claim 1, wherein said selected features include a pitch prediction gain measure responsive to inverse-filtered speech energy.
10. A low-bit-rate speech coding method according to claim 1, adapted for the environment of helicopter noise, and further comprising the step of low-pass filtering of speech energy at a cutoff frequency of about 420 Hz.
11. A low-bit-rate speech coding method according to claim 10, wherein said LPC analysis is conducted as 14th-order LPC analysis.
12. In a method of low-bit-rate speech coding of input speech occurring in a noisy environment, for a system which employs linear predictive coding (LPC) analysis of input speech frames to generate reflection coefficients, conversion of the reflection coefficients to vectors representing spectral parameters of the input speech frames, and matching of the spectral parameter vectors against reference vectors of a vocabulary of codewords generated in a training sequence in order to select the corresponding index of an optimally matching codeword for transmission,
the improvement comprising the steps of:
selecting a set of features which are characterized by a probability distribution which is not strongly affected in the noisy environment and which allow discrimination between voiced and unvoiced input speech;
measuring the selected features for input speech frames; and
using said feature measurements to make voiced/unvoiced speech decisions in order to select the voice/unvoiced excitation for speech synthesis in the receiver;
using noise estimates to update the reference vectors of the vocabulary of codewords, wherein new reference vectors are generated corresponding to said vocabulary of codewords in the noisy environment, said noise estimates including noise amplitude and noise reflection coefficients, wherein said noise estimate for speech frame I is performed only if the ith speech frame is unvoiced and more than a given number L of continuous unvoiced speech frames are accumulated, in order to prevent using voiced or unvoiced speech in the noise estimate.
13. A low-bit-rate speech coding method according to claim 12, wherein the vocabulary of codewords is generated for speech in a quiet environment, said quiet environment vocabulary is updated with noise estimates to obtain a vocabulary of codewords corresponding to the noisy environment, said noisy environment vocabulary constituting said reference vectors against which said spectral parameter vectors are matched, and speech is synthesized at a received end of the speech coding system using said quiet environment vocabulary.
14. A method for speech coding for transmission comprising:
providing a first vector quantization codebook of reference vectors, each reference vector representing spectral characteristics of a corresponding time frame of a reference speech signal;
accepting an input audio signal;
updating the reference vectors of the first vector quantization codebook to produce a corresponding second vector quantization codebook, including updating each reference vector to represent spectral characteristics of a combination of the corresponding time frame of the reference speech signal and a background noise present in the input audio signal; and
quantizing a time frame of the input audio signal according to the second vector quantization codebook, selecting the most similar updated reference vector to spectral characteristics of said time frame of the input audio signal.
15. The method of claim 14 further comprising:
transmitting an index of a reference vector of the first vector quantization codebook that corresponds to the selected updated reference vector of the second vector quantization codebook;
receiving the transmitted index; and
synthesizing a time frame of an output signal, including selecting a reference vector of the first vector quantization codebook according to the received index.
16. The method of claim 14 further comprising estimating a characteristic of the background noise present in the input audio signal.
17. The method of claim 16 further comprising repeating the steps of estimating the background noise of the input audio signal, updating the reference vectors of the first vector quantization codebook and quantizing the input signal according to the second vector quantization codebook for each of a sequence of time frames of the input audio signal.
18. The method of claim 16 wherein estimating a characteristic of the background noise includes:
determining any of a consecutive series of a predetermined number of time frames of the input audio signal includes voiced speech; and
if none of the series of time frames includes voiced speech, estimating the characteristic of the background noise in at least one of said series of time frames.
19. The method of claim 18 wherein estimating the characteristic of the background noise in at least one of said series of time frames includes determining an average noise amplitude in said time frame.
20. The method of claim 18 wherein estimating the characteristic of the background noise in at least one of said series of time frames includes determining a spectral representation of the background noise.
21. The method of claim 20 wherein determining the spectral representation of the background noise includes determining autocorrelation coefficients of the background noise.
22. The method of claim 16 wherein:
estimating the characteristic of the background noise includes determining autocorrelation coefficients of the background noise; and
updating each reference vector includes determining an autocorrelation coefficient representation of the spectral characteristics associated with the reference vector, and combining said autocorrelation coefficients with the autocorrelation coefficients of the background noise to produce an autocorrelation coefficient representation of the spectral characteristics represented by the updated reference vector.
23. The method of claim 22 wherein updating each reference vector further includes computing a line spectral frequency (LSF) representation from the autocorrelation coefficients associated with each of the updated reference vectors, and quantizing the time frame of the input audio signal includes comparing a LSF representation of the time frame of the input signal with the LSF representation of the updated reference vectors.
24. A speech encoder comprising:
an input processor for accepting and input audio signal;
a background noise estimator coupled to the input processor for estimating a characteristic of a background noise present in the input audio signal;
a first vector quantization codebook including reference vectors, each reference vector representing spectral characteristics of a corresponding time frame of a reference speech signal;
a second vector quantization codebook of updated reference vectors, each updated reference vector corresponding to a different one of the reference vectors of the first vector quantization codebook, and each reference vector representing spectral characteristics of a combination of the corresponding time frame of the reference speech signal and the background noise present in the input audio signal;
a codebook updater coupled to the first vector quantization codebook, the background noise estimator, and the second vector quantization codebook, configured to accept reference vectors of the first vector quantization codebook and the characteristic of the background noise and produce the reference vectors of the second vector quantization codebook;
a vector quantizer coupled to the second quantization codebook and the input processor configured to quantize a time frame of the input audio signal according to the second vector quantization codebook by selecting the most similar reference vector of said second codebook to spectral characteristics of said time frame of the input audio signal; and
a transmitter coupled to the vector quantizer to sending an index of the reference vector of the first vector quantization codebook that corresponds to the selected most similar reference vector of the second vector quantization codebook.
US09/422,820 1991-05-03 1999-10-21 Enhancement of speech coding in background noise for low-rate speech coder Expired - Lifetime USRE38269E1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US09/422,820 USRE38269E1 (en) 1991-05-03 1999-10-21 Enhancement of speech coding in background noise for low-rate speech coder

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
US69557191A 1991-05-03 1991-05-03
US08/060,710 US5680508A (en) 1991-05-03 1993-05-12 Enhancement of speech coding in background noise for low-rate speech coder
US09/422,820 USRE38269E1 (en) 1991-05-03 1999-10-21 Enhancement of speech coding in background noise for low-rate speech coder

Related Parent Applications (1)

Application Number Title Priority Date Filing Date
US08/060,710 Reissue US5680508A (en) 1991-05-03 1993-05-12 Enhancement of speech coding in background noise for low-rate speech coder

Publications (1)

Publication Number Publication Date
USRE38269E1 true USRE38269E1 (en) 2003-10-07

Family

ID=24793564

Family Applications (2)

Application Number Title Priority Date Filing Date
US08/060,710 Ceased US5680508A (en) 1991-05-03 1993-05-12 Enhancement of speech coding in background noise for low-rate speech coder
US09/422,820 Expired - Lifetime USRE38269E1 (en) 1991-05-03 1999-10-21 Enhancement of speech coding in background noise for low-rate speech coder

Family Applications Before (1)

Application Number Title Priority Date Filing Date
US08/060,710 Ceased US5680508A (en) 1991-05-03 1993-05-12 Enhancement of speech coding in background noise for low-rate speech coder

Country Status (1)

Country Link
US (2) US5680508A (en)

Cited By (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20020150264A1 (en) * 2001-04-11 2002-10-17 Silvia Allegro Method for eliminating spurious signal components in an input signal of an auditory system, application of the method, and a hearing aid
US20030139923A1 (en) * 2001-12-25 2003-07-24 Jhing-Fa Wang Method and apparatus for speech coding and decoding
US20040158461A1 (en) * 2003-02-07 2004-08-12 Motorola, Inc. Class quantization for distributed speech recognition
US20050091066A1 (en) * 2003-10-28 2005-04-28 Manoj Singhal Classification of speech and music using zero crossing
US20050131680A1 (en) * 2002-09-13 2005-06-16 International Business Machines Corporation Speech synthesis using complex spectral modeling
US20050273447A1 (en) * 2004-06-04 2005-12-08 Jinbo Bi Support vector classification with bounded uncertainties in input data
US20070250312A1 (en) * 2006-04-25 2007-10-25 Philip Garner Signal processing apparatus and method thereof
US20080049647A1 (en) * 1999-12-09 2008-02-28 Broadcom Corporation Voice-activity detection based on far-end and near-end statistics
US20090248407A1 (en) * 2006-03-31 2009-10-01 Panasonic Corporation Sound encoder, sound decoder, and their methods
US20090254340A1 (en) * 2008-04-07 2009-10-08 Cambridge Silicon Radio Limited Noise Reduction
US20100070272A1 (en) * 2008-03-04 2010-03-18 Lg Electronics Inc. method and an apparatus for processing a signal
US20110222574A1 (en) * 2010-03-09 2011-09-15 Massachusetts Institute Of Technology Two-dimensional wavelength-beam-combining of lasers using first-order grating stack
US20140244247A1 (en) * 2013-02-28 2014-08-28 Google Inc. Keyboard typing detection and suppression
US20160111105A1 (en) * 2013-07-04 2016-04-21 Huawei Technologies Co.,Ltd. Frequency envelope vector quantization method and apparatus

Families Citing this family (78)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
IN184794B (en) * 1993-09-14 2000-09-30 British Telecomm
TW271524B (en) * 1994-08-05 1996-03-01 Qualcomm Inc
US5794199A (en) * 1996-01-29 1998-08-11 Texas Instruments Incorporated Method and system for improved discontinuous speech transmission
US6029129A (en) * 1996-05-24 2000-02-22 Narrative Communications Corporation Quantizing audio data using amplitude histogram
KR100198476B1 (en) * 1997-04-23 1999-06-15 윤종용 Quantizer and the method of spectrum without noise
EP2154681A3 (en) 1997-12-24 2011-12-21 Mitsubishi Electric Corporation Method and apparatus for speech decoding
US6226606B1 (en) 1998-11-24 2001-05-01 Microsoft Corporation Method and apparatus for pitch tracking
US6519260B1 (en) 1999-03-17 2003-02-11 Telefonaktiebolaget Lm Ericsson (Publ) Reduced delay priority for comfort noise
US6850559B1 (en) 1999-06-28 2005-02-01 At&T Corp. System and methods for transmitting data
EP1190495A1 (en) * 1999-07-02 2002-03-27 Tellabs Operations, Inc. Coded domain echo control
US6910011B1 (en) * 1999-08-16 2005-06-21 Haman Becker Automotive Systems - Wavemakers, Inc. Noisy acoustic signal enhancement
US7117149B1 (en) 1999-08-30 2006-10-03 Harman Becker Automotive Systems-Wavemakers, Inc. Sound source classification
US7236640B2 (en) * 2000-08-18 2007-06-26 The Regents Of The University Of California Fixed, variable and adaptive bit rate data source encoding (compression) method
US7472059B2 (en) * 2000-12-08 2008-12-30 Qualcomm Incorporated Method and apparatus for robust speech classification
US20030028386A1 (en) * 2001-04-02 2003-02-06 Zinser Richard L. Compressed domain universal transcoder
WO2003019533A1 (en) * 2001-08-24 2003-03-06 Kabushiki Kaisha Kenwood Device and method for interpolating frequency components of signal adaptively
US6959276B2 (en) * 2001-09-27 2005-10-25 Microsoft Corporation Including the category of environmental noise when processing speech signals
US7478043B1 (en) * 2002-06-05 2009-01-13 Verizon Corporate Services Group, Inc. Estimation of speech spectral parameters in the presence of noise
US7343283B2 (en) * 2002-10-23 2008-03-11 Motorola, Inc. Method and apparatus for coding a noise-suppressed audio signal
US20060004567A1 (en) * 2002-11-27 2006-01-05 Visual Pronunciation Software Limited Method, system and software for teaching pronunciation
US7895036B2 (en) * 2003-02-21 2011-02-22 Qnx Software Systems Co. System for suppressing wind noise
US8271279B2 (en) * 2003-02-21 2012-09-18 Qnx Software Systems Limited Signature noise removal
US8326621B2 (en) 2003-02-21 2012-12-04 Qnx Software Systems Limited Repetitive transient noise removal
US7885420B2 (en) 2003-02-21 2011-02-08 Qnx Software Systems Co. Wind noise suppression system
US7725315B2 (en) * 2003-02-21 2010-05-25 Qnx Software Systems (Wavemakers), Inc. Minimization of transient noises in a voice signal
US8073689B2 (en) 2003-02-21 2011-12-06 Qnx Software Systems Co. Repetitive transient noise removal
US7949522B2 (en) 2003-02-21 2011-05-24 Qnx Software Systems Co. System for suppressing rain noise
US7103371B1 (en) 2003-10-22 2006-09-05 Itt Manufacturing Enterprises, Inc. Method and apparatus for dynamic voice reservation within wireless networks
EP1569200A1 (en) * 2004-02-26 2005-08-31 Sony International (Europe) GmbH Identification of the presence of speech in digital audio data
US7716046B2 (en) * 2004-10-26 2010-05-11 Qnx Software Systems (Wavemakers), Inc. Advanced periodic signal enhancement
US7680652B2 (en) * 2004-10-26 2010-03-16 Qnx Software Systems (Wavemakers), Inc. Periodic signal enhancement system
US8543390B2 (en) 2004-10-26 2013-09-24 Qnx Software Systems Limited Multi-channel periodic signal enhancement system
US8306821B2 (en) * 2004-10-26 2012-11-06 Qnx Software Systems Limited Sub-band periodic signal enhancement system
US7610196B2 (en) * 2004-10-26 2009-10-27 Qnx Software Systems (Wavemakers), Inc. Periodic signal enhancement system
US8170879B2 (en) * 2004-10-26 2012-05-01 Qnx Software Systems Limited Periodic signal enhancement system
US7949520B2 (en) 2004-10-26 2011-05-24 QNX Software Sytems Co. Adaptive filter pitch extraction
US8284947B2 (en) * 2004-12-01 2012-10-09 Qnx Software Systems Limited Reverberation estimation and suppression system
WO2006089055A1 (en) * 2005-02-15 2006-08-24 Bbn Technologies Corp. Speech analyzing system with adaptive noise codebook
US8219391B2 (en) * 2005-02-15 2012-07-10 Raytheon Bbn Technologies Corp. Speech analyzing system with speech codebook
US8027833B2 (en) 2005-05-09 2011-09-27 Qnx Software Systems Co. System for suppressing passing tire hiss
US8296134B2 (en) * 2005-05-13 2012-10-23 Panasonic Corporation Audio encoding apparatus and spectrum modifying method
US8170875B2 (en) * 2005-06-15 2012-05-01 Qnx Software Systems Limited Speech end-pointer
US8311819B2 (en) * 2005-06-15 2012-11-13 Qnx Software Systems Limited System for detecting speech with background voice estimates and noise estimates
JP5092748B2 (en) 2005-09-02 2012-12-05 日本電気株式会社 Noise suppression method and apparatus, and computer program
US7844453B2 (en) 2006-05-12 2010-11-30 Qnx Software Systems Co. Robust noise estimation
US20080147385A1 (en) * 2006-12-15 2008-06-19 Nokia Corporation Memory-efficient method for high-quality codebook based voice conversion
US8335685B2 (en) 2006-12-22 2012-12-18 Qnx Software Systems Limited Ambient noise compensation system robust to high excitation noise
US8326620B2 (en) 2008-04-30 2012-12-04 Qnx Software Systems Limited Robust downlink speech and noise detector
US8850154B2 (en) 2007-09-11 2014-09-30 2236008 Ontario Inc. Processing system having memory partitioning
US8904400B2 (en) 2007-09-11 2014-12-02 2236008 Ontario Inc. Processing system having a partitioning component for resource partitioning
US8694310B2 (en) 2007-09-17 2014-04-08 Qnx Software Systems Limited Remote control server protocol system
US8209514B2 (en) 2008-02-04 2012-06-26 Qnx Software Systems Limited Media processing system having resource partitioning
CN101771417B (en) 2008-12-30 2012-04-18 华为技术有限公司 Methods, devices and systems for coding and decoding signals
GB2466671B (en) * 2009-01-06 2013-03-27 Skype Speech encoding
GB2466673B (en) * 2009-01-06 2012-11-07 Skype Quantization
GB2466670B (en) * 2009-01-06 2012-11-14 Skype Speech encoding
GB2466669B (en) * 2009-01-06 2013-03-06 Skype Speech coding
US20100174539A1 (en) * 2009-01-06 2010-07-08 Qualcomm Incorporated Method and apparatus for vector quantization codebook search
GB2466674B (en) 2009-01-06 2013-11-13 Skype Speech coding
GB2466675B (en) 2009-01-06 2013-03-06 Skype Speech coding
GB2466672B (en) * 2009-01-06 2013-03-13 Skype Speech coding
US8260220B2 (en) * 2009-09-28 2012-09-04 Broadcom Corporation Communication device with reduced noise speech coding
US8452606B2 (en) * 2009-09-29 2013-05-28 Skype Speech encoding using multiple bit rates
US9015044B2 (en) 2012-03-05 2015-04-21 Malaspina Labs (Barbados) Inc. Formant based speech reconstruction from noisy signals
JP2014123011A (en) * 2012-12-21 2014-07-03 Sony Corp Noise detector, method, and program
US10255903B2 (en) * 2014-05-28 2019-04-09 Interactive Intelligence Group, Inc. Method for forming the excitation signal for a glottal pulse model based parametric speech synthesis system
US9858922B2 (en) 2014-06-23 2018-01-02 Google Inc. Caching speech recognition scores
US9361899B2 (en) * 2014-07-02 2016-06-07 Nuance Communications, Inc. System and method for compressed domain estimation of the signal to noise ratio of a coded speech signal
CN105336344B (en) 2014-07-10 2019-08-20 华为技术有限公司 Noise detection method and device
US9299347B1 (en) 2014-10-22 2016-03-29 Google Inc. Speech recognition using associative mapping
EP3079151A1 (en) * 2015-04-09 2016-10-12 Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V. Audio encoder and method for encoding an audio signal
US9786270B2 (en) 2015-07-09 2017-10-10 Google Inc. Generating acoustic models
US10373608B2 (en) 2015-10-22 2019-08-06 Texas Instruments Incorporated Time-based frequency tuning of analog-to-information feature extraction
US10229672B1 (en) 2015-12-31 2019-03-12 Google Llc Training acoustic models using connectionist temporal classification
US20180018973A1 (en) 2016-07-15 2018-01-18 Google Inc. Speaker verification
GB201617016D0 (en) 2016-09-09 2016-11-23 Continental automotive systems inc Robust noise estimation for speech enhancement in variable noise conditions
GB2563687B (en) * 2017-06-19 2019-11-20 Cirrus Logic Int Semiconductor Ltd Audio test mode
US10706840B2 (en) 2017-08-18 2020-07-07 Google Llc Encoder-decoder models for sequence to sequence mapping

Citations (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4074069A (en) * 1975-06-18 1978-02-14 Nippon Telegraph & Telephone Public Corporation Method and apparatus for judging voiced and unvoiced conditions of speech signal
US4091237A (en) * 1975-10-06 1978-05-23 Lockheed Missiles & Space Company, Inc. Bi-Phase harmonic histogram pitch extractor
US4296279A (en) * 1980-01-31 1981-10-20 Speech Technology Corporation Speech synthesizer
US4589131A (en) * 1981-09-24 1986-05-13 Gretag Aktiengesellschaft Voiced/unvoiced decision using sequential decisions
US4630304A (en) * 1985-07-01 1986-12-16 Motorola, Inc. Automatic background noise estimator for a noise suppression system
US4696038A (en) * 1983-04-13 1987-09-22 Texas Instruments Incorporated Voice messaging system with unified pitch and voice tracking
US4720802A (en) * 1983-07-26 1988-01-19 Lear Siegler Noise compensation arrangement
US4933973A (en) * 1988-02-29 1990-06-12 Itt Corporation Apparatus and methods for the selective addition of noise to templates employed in automatic speech recognition systems
US4975956A (en) * 1989-07-26 1990-12-04 Itt Corporation Low-bit-rate speech coder using LPC data reduction processing
US5073940A (en) * 1989-11-24 1991-12-17 General Electric Company Method for protecting multi-pulse coders from fading and random pattern bit errors
US5127053A (en) * 1990-12-24 1992-06-30 General Electric Company Low-complexity method for improving the performance of autocorrelation-based pitch detectors
US5459814A (en) * 1993-03-26 1995-10-17 Hughes Aircraft Company Voice activity detector for speech signals in variable background noise
US5806024A (en) * 1995-12-23 1998-09-08 Nec Corporation Coding of a speech or music signal with quantization of harmonics components specifically and then residue components
US6018707A (en) * 1996-09-24 2000-01-25 Sony Corporation Vector quantization method, speech encoding method and apparatus
US6081776A (en) * 1998-07-13 2000-06-27 Lockheed Martin Corp. Speech coding system and method including adaptive finite impulse response filter

Patent Citations (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4074069A (en) * 1975-06-18 1978-02-14 Nippon Telegraph & Telephone Public Corporation Method and apparatus for judging voiced and unvoiced conditions of speech signal
US4091237A (en) * 1975-10-06 1978-05-23 Lockheed Missiles & Space Company, Inc. Bi-Phase harmonic histogram pitch extractor
US4296279A (en) * 1980-01-31 1981-10-20 Speech Technology Corporation Speech synthesizer
US4589131A (en) * 1981-09-24 1986-05-13 Gretag Aktiengesellschaft Voiced/unvoiced decision using sequential decisions
US4696038A (en) * 1983-04-13 1987-09-22 Texas Instruments Incorporated Voice messaging system with unified pitch and voice tracking
US4720802A (en) * 1983-07-26 1988-01-19 Lear Siegler Noise compensation arrangement
US4630304A (en) * 1985-07-01 1986-12-16 Motorola, Inc. Automatic background noise estimator for a noise suppression system
US4933973A (en) * 1988-02-29 1990-06-12 Itt Corporation Apparatus and methods for the selective addition of noise to templates employed in automatic speech recognition systems
US4975956A (en) * 1989-07-26 1990-12-04 Itt Corporation Low-bit-rate speech coder using LPC data reduction processing
US5073940A (en) * 1989-11-24 1991-12-17 General Electric Company Method for protecting multi-pulse coders from fading and random pattern bit errors
US5127053A (en) * 1990-12-24 1992-06-30 General Electric Company Low-complexity method for improving the performance of autocorrelation-based pitch detectors
US5459814A (en) * 1993-03-26 1995-10-17 Hughes Aircraft Company Voice activity detector for speech signals in variable background noise
US5806024A (en) * 1995-12-23 1998-09-08 Nec Corporation Coding of a speech or music signal with quantization of harmonics components specifically and then residue components
US6018707A (en) * 1996-09-24 2000-01-25 Sony Corporation Vector quantization method, speech encoding method and apparatus
US6081776A (en) * 1998-07-13 2000-06-27 Lockheed Martin Corp. Speech coding system and method including adaptive finite impulse response filter

Non-Patent Citations (6)

* Cited by examiner, † Cited by third party
Title
Chen et al., "Robust Vector Quantization Based on Spectral Mapping with a Noise Estimate", IEEE Pacific RIM Conference on Communications, Conference date: Jun. 1, 1989.
Deller et al., ("Discrete-Time Processing of Speech Signals," Prentice Hall, Upper Saddle River, NJ, pp. 244-251, 471-473, Dec. 1987).* *
Kou et al., "Vector-Adaptive Vector Quantization with Application to Speech Coding", IEEE Transactions on Communications, vol. 39, No. 6, Jun. 1991.
Rabiner et al., ("Digital Processing of Speech Signals," Prentice Hall, Upper Saddle River, NJ, pp. 130-133, 451-452, Dec. 1978).* *
Siegel, ("A Procedure for Using Pattern Classification Techniques to Obtain a Voiced/Unvoiced Classifier", IEEE Transactions on Acoustics, Speech, and Signal Processing, vol. ASSP-27, No. 1, Feb. 1979).* *
Wolfgang Hess, ("Pitch Determination of Speech Signals", pp. 373-383, Springer-Verlag, NY, 1983).* *

Cited By (27)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110058496A1 (en) * 1999-12-09 2011-03-10 Leblanc Wilfrid Voice-activity detection based on far-end and near-end statistics
US7835311B2 (en) * 1999-12-09 2010-11-16 Broadcom Corporation Voice-activity detection based on far-end and near-end statistics
US8565127B2 (en) 1999-12-09 2013-10-22 Broadcom Corporation Voice-activity detection based on far-end and near-end statistics
US20080049647A1 (en) * 1999-12-09 2008-02-28 Broadcom Corporation Voice-activity detection based on far-end and near-end statistics
US20020150264A1 (en) * 2001-04-11 2002-10-17 Silvia Allegro Method for eliminating spurious signal components in an input signal of an auditory system, application of the method, and a hearing aid
US7305337B2 (en) * 2001-12-25 2007-12-04 National Cheng Kung University Method and apparatus for speech coding and decoding
US20030139923A1 (en) * 2001-12-25 2003-07-24 Jhing-Fa Wang Method and apparatus for speech coding and decoding
US20050131680A1 (en) * 2002-09-13 2005-06-16 International Business Machines Corporation Speech synthesis using complex spectral modeling
US8280724B2 (en) * 2002-09-13 2012-10-02 Nuance Communications, Inc. Speech synthesis using complex spectral modeling
US6961696B2 (en) * 2003-02-07 2005-11-01 Motorola, Inc. Class quantization for distributed speech recognition
US20040158461A1 (en) * 2003-02-07 2004-08-12 Motorola, Inc. Class quantization for distributed speech recognition
US20050091066A1 (en) * 2003-10-28 2005-04-28 Manoj Singhal Classification of speech and music using zero crossing
US20050273447A1 (en) * 2004-06-04 2005-12-08 Jinbo Bi Support vector classification with bounded uncertainties in input data
US7480639B2 (en) * 2004-06-04 2009-01-20 Siemens Medical Solution Usa, Inc. Support vector classification with bounded uncertainties in input data
US20090248407A1 (en) * 2006-03-31 2009-10-01 Panasonic Corporation Sound encoder, sound decoder, and their methods
US20070250312A1 (en) * 2006-04-25 2007-10-25 Philip Garner Signal processing apparatus and method thereof
US7890319B2 (en) * 2006-04-25 2011-02-15 Canon Kabushiki Kaisha Signal processing apparatus and method thereof
US8135585B2 (en) * 2008-03-04 2012-03-13 Lg Electronics Inc. Method and an apparatus for processing a signal
US20100070272A1 (en) * 2008-03-04 2010-03-18 Lg Electronics Inc. method and an apparatus for processing a signal
US20090254340A1 (en) * 2008-04-07 2009-10-08 Cambridge Silicon Radio Limited Noise Reduction
US9142221B2 (en) * 2008-04-07 2015-09-22 Cambridge Silicon Radio Limited Noise reduction
US20110222574A1 (en) * 2010-03-09 2011-09-15 Massachusetts Institute Of Technology Two-dimensional wavelength-beam-combining of lasers using first-order grating stack
US20140244247A1 (en) * 2013-02-28 2014-08-28 Google Inc. Keyboard typing detection and suppression
US9520141B2 (en) * 2013-02-28 2016-12-13 Google Inc. Keyboard typing detection and suppression
US20160111105A1 (en) * 2013-07-04 2016-04-21 Huawei Technologies Co.,Ltd. Frequency envelope vector quantization method and apparatus
US9805732B2 (en) * 2013-07-04 2017-10-31 Huawei Technologies Co., Ltd. Frequency envelope vector quantization method and apparatus
US10032460B2 (en) 2013-07-04 2018-07-24 Huawei Technologies Co., Ltd. Frequency envelope vector quantization method and apparatus

Also Published As

Publication number Publication date
US5680508A (en) 1997-10-21

Similar Documents

Publication Publication Date Title
USRE38269E1 (en) Enhancement of speech coding in background noise for low-rate speech coder
US6691092B1 (en) Voicing measure as an estimate of signal periodicity for a frequency domain interpolative speech codec system
US7693710B2 (en) Method and device for efficient frame erasure concealment in linear predictive based speech codecs
KR100742443B1 (en) A speech communication system and method for handling lost frames
US5751903A (en) Low rate multi-mode CELP codec that encodes line SPECTRAL frequencies utilizing an offset
EP1338003B1 (en) Gains quantization for a celp speech coder
US6694293B2 (en) Speech coding system with a music classifier
US6269333B1 (en) Codebook population using centroid pairs
KR100798668B1 (en) Method and apparatus for coding of unvoiced speech
US6205423B1 (en) Method for coding speech containing noise-like speech periods and/or having background noise
EP1597721B1 (en) 600 bps mixed excitation linear prediction transcoding
EP0849724A2 (en) High quality speech coder and coding method
Madane et al. Speech compression using Linear predictive coding
Özaydın et al. Matrix quantization and mixed excitation based linear predictive speech coding at very low bit rates
EP0534442B1 (en) Vocoder device for encoding and decoding speech signals
JPH09508479A (en) Burst excitation linear prediction
Oh et al. Output Recursively Adaptive (ORA) Tree Coding of Speech with VAD/CNG
Rebolledo et al. A multirate voice digitizer based upon vector quantization
Ozaydin et al. A 1200 bps speech coder with LSF matrix quantization
Sharma Speech Compression Using Linear Predictive Coding (LPC)
Sharma et al. Advanced Speech Compression VIA Voice Excited Linear Predictive Coding Using Discrete Cosine Transform (DCT)
Rodríguez Fonollosa et al. Robust LPC vector quantization based on Kohonen's design algorithm
Ozaydin et al. Matrix quantization based speech coder at 1200 bps
Miseki et al. Adaptive bit-allocation between the pole-zero synthesis filter and excitation in CELP
Sadek et al. An enhanced variable bit-rate CELP speech coder

Legal Events

Date Code Title Description
FPAY Fee payment

Year of fee payment: 8

FPAY Fee payment

Year of fee payment: 12

AS Assignment

Owner name: EXELIS INC., VIRGINIA

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:ITT CORPORATION;REEL/FRAME:027567/0311

Effective date: 20111221

AS Assignment

Owner name: HARRIS CORPORATION, FLORIDA

Free format text: MERGER;ASSIGNOR:EXELIS INC.;REEL/FRAME:039362/0534

Effective date: 20151223