|
(11) | EP 0 574 523 B1 |
| (12) | EUROPEAN PATENT SPECIFICATION |
|
|
| (54) |
VARIABLE BIT RATE SPEECH ENCODER SPRACHKODIERER MIT VERÄNDERLICHER BITRATE CODEUR VOCAL A DEBIT BINAIRE VARIABLE |
|
|
|||||||||||||||||||||||||||||||
| Note: Within nine months from the publication of the mention of the grant of the European patent, any person may give notice to the European Patent Office of opposition to the European patent granted. Notice of opposition shall be filed in a written reasoned statement. It shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention). |
Background of the Invention
[0001] The invention relates generally to communicating voice and video information over a channel having a fixed capacity, such as a telephone communication channel.
[0002] Video conferencing systems typically transmit both voice and video information over the same channel. A portion of the channel's bandwidth is typically dedicated to voice information and the remaining bandwidth is allocated to video information.
[0003] The amount of video and voice information varies with time. For example, at certain moments in time, a person at one end of the system may be silent. Thus, if the system includes a variable capacity voice encoder, little information needs to be transmitted during such moments of silence.
[0004] Similarly, the video signal may have little or no change between frames as, for example, when all objects within the field of view are still. If the system includes a variable capacity video encoder, little information needs to be transmitted during such moments of inactivity. At the other extreme, during times of great activity, the amount of video information may exceed the channel capacity allocated to video information. Accordingly, the system transmits as much video information as possible, discarding the remainder.
[0005] Typically, the video encoder accords priority to the most noticeable features of the video signal. Thus, high priority information is transmitted first and the less noticeable low priority information is temporarily discarded if the channel lacks sufficient capacity. Accordingly, it is desirable to have available as much video bandwidth as possible.
[0006] US-A-4949383 (Koh et al) discloses a bit allocation method in which bits are allocated to channels of a sub-band coder or to the coefficients of a transformer coder using a fixed number of bits. Only the selection of the channels to which the available bits are assigned is varied. Thus, different frequency bands can be encoded separately and differently.
[0007] US-A-4965830 (Burham et al) discloses a bit allocation method attempting to maintain a predetermined average quantization distortion level.
[0008] EP-A-309974 discloses a method for communicating a digital signal, in which data compression is obtained by encoding the sufficiently large samples of the signal, and then encoding a residual signal.
[0009] The invention relates to a method and apparatus for allocating transmission bits for use in transmitting samples of a digital signal. An aggregate allowable quantization distortion value is selected representing an allowable quantization distortion error for a frame of samples of the digital signal. A set of samples are selected from the frame of samples such that a plurality of the selected samples are greater than a noise threshold. For each sample of the set, a sample quantization distortion value is computed which represents an allowable quantization distortion error for the sample. The sum of all sample quantization distortion values is approximately equal to the aggregate allowable quantization distortion value. For each sample of the set, a quantization step size is selected which yields a quantization distortion error approximately equal to the sample's corresponding quantization distortion value. Each sample is then quantized using its quantization step size.
[0010] In preferred embodiments, the digital signal includes a noise component and a signal component. A signal index is prepared representing, for at least one sample of the frame, the magnitude of the signal component relative to the magnitude of the noise component. The aggregate allowable quantization distortion is selected based on the signal index.
[0011] The sample quantization distortion value is computed by dividing the aggregate allowable quantization distortion value by a number of samples in the frame of samples to form a first sample distortion value. A tentative set of samples of the digital signal is selected wherein each sample of the tentative set is greater than a noise threshold determined at least in part by the value of the first sample distortion value. The first sample distortion value is then adjusted by an amount determined by the difference between the first sample distortion value and at least one sample excluded from the tentative set (i.e, a "noisy sample"). Based on the adjusted distortion value, the process is repeated to identify any noisy samples of the tentative set; remove the noisy samples, if any, from the tentative set; and again adjust the first sample distortion value by an amount determined by the difference between the first sample distortion value and a noisy sample. The process is repeated until an adjusted first sample distortion value is reached for which no additional noisy samples of the tentative set are found or until the process has been repeated a maximum number of times.
[0012] After the adjustment is terminated, the number of bits required to transmit all samples of the tentative set is estimated. The estimated bit number is compared to a maximum bit number. If the estimated bit number is less than or equal to the maximum bit number, a final noise threshold is selected based on the adjusted first sample distortion value.
[0013] If the estimated bit number exceeds the maximum bit number, a second sample distortion value, is prepared. A second tentative set of samples of the digital signal is then selected wherein each of a plurality of samples of the second tentative set have a magnitude above the second sample distortion value. The number of bits required to transmit all samples of the second tentative set is then estimated. The estimated bit number is compared to the maximum bit number. If it is greater than the maximum bit number, the second sample distortion value is increased and the second tentative set of samples is re-selected based on the adjusted second sample distribution value. The number of bits required to transmit the second tentative set is again estimated. This process is repeated until a second sample distortion value is reached for which the estimated bit number is less than or equal to the maximum bit number.
[0014] The sample distortion value is then calculated from the adjusted first sample distortion value and the second sample distortion value. A final set of samples of the digital signal, is selected wherein each of a plurality of samples of the final set have a magnitude above a final threshold determined by the sample's corresponding sample distortion value.
[0015] In another aspect, the invention relates to a method and apparatus for communicating a digital signal which includes a noise component and a signal component. An estimation signal is prepared which is representative of the digital signal but has fewer samples than the digital signal. A signal index is prepared which represents, for at least one sample of the estimation signal, the magnitude of the signal component relative to the magnitude of the noise component. Based on the signal index, samples of the digital signal are selected which have a sufficiently large signal component. The selected samples of the digital signal and the samples of the digital estimation signal are both transmitted to a remote device. The remote device reconstructs the digital signal from the transmitted selected samples and estimation samples.
[0016] In preferred embodiments, the digital signal is a frequency domain speech signal representative of voice information to be communicated, and each sample of the estimation signal is a spectral estimate of the frequency domain signal in a corresponding band of frequencies. To reconstruct the frequency domain signal, a random number is generated for each nonselected sample of the frequency domain speech signal. A noise estimate of the magnitude of a noise component of the nonselected sample is prepared from at least one spectral estimate. Based on the noise estimate, a scaling factor is generated. The random number is then scaled according to the scaling factor to produce a reconstructed sample representative of the nonselected sample.
[0017] The estimation signal and the frequency domain speech signal each include a series of frames. Each frame represents the voice information over a specified window of time. To prepare a first noise estimate for a current frame, an initial noise estimate is first prepared for a prior frame of the estimation signal. The initial noise estimate is prepared from at least one spectral estimate representative of a band of frequencies of the prior frame. Based on the magnitude of the signal index, a rise time constant tr and a fall time constant tf are then selected. The selected rise time constant is added to the initial noise estimate to form an upper threshold, and the selected fall time constant is subtracted from the initial noise estimate to form a lower threshold.
[0018] A current spectral estimate of the current frame, representative of the same band of frequencies, is then compared to the upper and lower thresholds. If the current spectral estimate is between the thresholds, the current noise estimate is set equal to the current spectral estimate. If it is below the lower threshold, the current noise estimate is set equal to the lower threshold. If the current spectral estimate is above the upper threshold, the current noise estimate is set equal to the upper threshold.
[0019] To generate the scaling factor, a noise coefficient and a voice coefficient are first selected based on the value of the signal index. The current noise estimate is then multiplied by the noise index. Similarly, the current spectral estimate is multiplied by the signal index. The products of the multiplication are added together and the scaling factor is then formed from the sum.
[0020] The methods and apparatus described in this specification have as an objective, a reduction in the amount of channel bandwidth allocated to audio information whenever there is little audio information required to be sent. The remaining portion of bandwidth is allocated to video information. Thus, average, a lower bit rate is provided for audio information and a higher bit rate is provided for video information.
[0021] The invention will now be described by way of example with reference to the drawings, a brief description of which follows.
Brief Description of the Drawings
[0024] Figs. 2(a-c) are diagrams of a set of frequency coefficients and two types of estimates of the frequency coefficients.
[0028] Fig. 6 is a flow chart of a process for estimating the amount of voice in a frame of the microphone signal.
[0030] Figs. 8(a) and 8(b) are a flow chart of a process for computing a first estimate of the allowable distortion per spectral estimate.
[0031] Fig. 9 is a flow chart of a process for computing a second estimate of the allowable distortion per spectral estimate.
[0033] Fig. 11 is a diagram illustrating the interpolation between band quantization step sizes to form coefficient quantization step sizes.
System Overview
[0035] Referring to Figs. 1(a), and 1(b), a video conferencing system includes a near end microphone 12 which receives the voice of a person speaking at the near end and generates an electronic microphone signal m(t) representative of the voice, where t is a variable representing time. Similarly, a camera 14 is focused on the person speaking to generate video signal v(t) representative of the image of the person.
[0036] The microphone signal is applied to an audio encoder 16 which digitizes and encodes the microphone signal m(t). As will be explained more fully below, the encoder 16 generates a set of digitally encoded signals Fh(k), Sh(j), gi(F) and Ai(F) which collectively represent the microphone signal, where, as explained more fully below, k and j are integers representing values of frequency, and F is an integer identifying a "frame" of the microphone signal. Similarly, the video signal v(t) is applied to a video encoder 18 which digitizes and encodes the video signal.
[0037] The encoded video signal Ve and the encoded set of microphone signals are applied to a bit stream controller 20 which merges the signals into a serial bit stream for transmission over a communication channel 22.
[0038] Referring to Fig. 1(b), a bit stream controller 24 at the far end receives the bit stream from communication channel 22 and separates it into the video and audio components. The received video signal Ve is decoded by a video decoder 26 and applied to a display device 27 to recreate a representation of the near end camera image. At the same time, an audio decoder 28 decodes the set of received microphone signals to generate a loudspeaker signal L(t) representative of the near end microphone signal m(t). In response to the loudspeaker signal, a loudspeaker 29 reproduces the near end voice.
[0039] Referring to Fig. 1(a), encoder 16 includes an input signal conditioner 30 for digitizing and filtering the microphone signal m(t) to produce a digitized microphone signal m(n), where n is an integer representing a moment in time. The digitized microphone signal, m(n), is provided to a windowing module 32 which separates m(n) into groups mF(n), in the illustrated embodiment, of 512 consecutive samples referred to as a frame. The frames are selected to overlap. More specifically, each frame includes the last 16 samples from the previous group and 496 new samples.
[0040] Each frame of samples is then applied to a normalization module 34 which calculates the average energy Eav of all microphone samples in the frame:
It then computes a normalizing gain g(F), for each frame F, equal to the square root of the frame's average energy Eav:
[0041] As explained more fully below, the normalizing gain g is used at the near end to scale the microphone samples in each frame. At the far end, the normalizing gain is used to restore the decoded microphone samples to their original scale. Accordingly, the same gain used to scale the samples at the near end must be transmitted to the far end for use in rescaling the microphone signal. Since the normalization gain g calculated by module 34 has a relatively large number of bits, normalization module 34 supplies the gain g to a quantizer 35 which generates a gain quantization index gi which represents the gain using fewer bits than the number of bits specifying gain g. The gain quantization index gi is then provided to bit steam controller 20 for transmission to the far end.
[0042] The far end audio decoder 28 reconstructs the gain from the transmitted gain quantization index gi and uses the reconstructed gain gq to restore the microphone signal to its original scale. Since the reconstructed gain gq typically differs slightly from the original gain g, audio encoder 16 at the near end normalizes the microphone signal using the same gain gq as used at the far end. More specifically, an inverse quantizer 37 reconstructs the gain gq from the gain quantization index gi in the same manner as the far end audio decoder 28.
[0043] The quantized gain gq and the frame of microphone signals mF(n) are forwarded to a discrete cosine transform module (DCT) 36 which divides each microphone sample mF(n) by the gain gq to yield a normalized sample m'(n). It then converts the group of normalized microphone signals m'(n) to the frequency domain using the well known discrete cosine transform algorithm. DCT 36 thus generates 512 frequency coefficients F(k) representing samples of the frame's frequency spectrum, where k is an integer representing a discrete frequency in the frame's spectrum.
[0044] The frequency coefficients F(k) are encoded (using an entropy adaptive transfer coder described below) and transmitted to the far end. To reduce the number of bits necessary to transmit these coefficients, the encoder 16 estimates the relative amounts of signal (e.g., voice) and noise in various regions of the frequency spectrum, and chooses not to transmit coefficients for frequency bands having a relatively large amount of noise.
Further, for each coefficient selected for transmission, encoder 16 selects the number of bits required to represent the coefficient based on the amount of noise present in the frequency band which includes the coefficient's frequency. More specifically, it is known that humans can tolerate more corrupting noise in regions of the audio spectrum having relatively large amounts of audio signal, because the audio signal tends to mask the noise. Accordingly, encoder 16 coarsely quantizes coefficients having relatively large amounts of audio signal since the audio signal masks the quantization distortion introduced through the coarse quantization. Thus, the encoder minimizes the number of bits needed to represent each coefficient by selecting a quantization step size tailored to the amount of audio signal represented by the coefficient.
[0045] To estimate the relative amounts of noise and signal in various regions of the spectrum, the frequency coefficients F(k) are first supplied to a spectrum estimation module 38. As will be explained more fully below, estimation module 38 reduces the frequency coefficients F(k) to a smaller set of spectral estimates S(j) which represent the frequency spectrum of the frame with less detail, wherein j an integer representing a band of frequencies in the spectrum. A speech detection module 40 processes the spectral estimates of each frame to estimate the energy component of the microphone signal due to noise in the near end room. It then provides a signal index Ai(F) for each frame F which approximates the percentage of the microphone signal in the frame which is attributable to voice information.
[0046] For each frame F, the signal index Ai(F) and the spectral estimates S(j) are both used in quantizing and encoding the frequency coefficients F(k) for transmission to the far end. Accordingly, they are both needed at the far end for use in reconstructing the frequency coefficients. The signal index Ai(F) has only three bits and accordingly is applied directly to the bit stream controller 20 for transmission. The spectral estimates S(j) however are converted to logarithmic spectral estimates Log2 S2(j) and are encoded using a well known differential pulse code modulation encoder (herein "DPCM") 39 to reduce the number of bits to be transmitted. (The DPCM is preferably a first order DPCM with a unity gain predictor coefficient.) The encoded logarithmic spectral estimates Se(j) are further encoded using a Huffman encoder 49 to further reduce the number of bits to be transmitted. The resultant Huffman codes Sh(j) are provided to bit stream controller 20 for transmission.
[0047] Far end audio decoder 28 reconstructs the logarithmic spectral estimates Log2 S2(j) from the Huffman codes Sh(j). However, due to the operation of the DPCM encoder 39, the reconstructed spectral estimates Log2 Sq 2(j) are not identical to the original estimates Log2 S2(j). Accordingly, a decoder 41 in near end encoder 16 decodes the encoded estimates Se(j) in the same manner as performed at the far end, and uses the thus decoded estimates Log2 Sq 2(j) in quantizing and encoding the frequency coefficients F(k). Thus, (as explained more fully below) the audio encoder 16 encodes the frequency coefficients using the identical estimates Log2 Sq 2(j) as used by the far end audio decoder 28 in reconstructing the signal Log2 Sq 2(j) at the far end.
[0048] Based on the values of the decoded spectral estimates Log2 Sq 2(j) and the signal index Ai(F), a bit rate estimator 42 selects groups of frequency coefficients F(k) which collectively have a sufficient amount of voice information to merit being transmitted to the far end. Next, the bit rate estimator selects for each coefficient to be transmitted, a quantization step size which determines the number of bits to be used in quantizing the coefficients for transmission. For each selected group of frequency coefficients, bit rate estimator 42 first computes a group quantization step size Q(j) and a "class" C(j) (to be described below) for each frequency band j. The group quantization step sizes are then interpolated to yield a coefficient quantization step size Q(k) for each frequency coefficient F(k). Q(k) is then applied to a coefficient quantizer 44 which, based on the assigned step size, quantizes each frequency coefficient in the band to provide a corresponding quantization index I(k).
[0049] In response to each quantization index I(k), a limit controller 45 generates a Huffman index Ih(k). The Huffman indices Ih(k) provided by the limit controller are further encoded by Huffman encoder 47 to yield Huffman codes Fh(k). The Huffman codes are applied to bit stream controller 20 for transmission over communication channel 22.
[0050] The limit controller begins the above described encoding process starting with the low frequency quantization indices (i.e., k=0) and continuing until all indices are encoded or until the limit controller 45 concludes that the number of encoded bits exceeds the capacity of channel 22 allocated to the microphone signal. If the limit controller 45 concludes that the capacity has been exceeded, it discards the remaining indices Ih(k) and provides a unique Huffman index Ih(k) indicating that the frame's remaining frequency coefficients will not be coded for transmission. (As explained more fully below, the far end decoder estimates the uncoded coefficients from the transmitted spectral estimates).
Spectrum Estimator for Calculating Spectral Estimates S(j)
[0051] Referring to Figs. 2(a-c) and 3(a-b), the following describes the spectrum estimation module 38 in further detail. Module 38 first separates the frequency coefficients F(k) into bands of L adjacent coefficients, where L is preferably equal to 10. (Step 110). For each band, j, the module prepares a first approximation sample, representative of the entire band, by computing the energy of the spectrum in the band. More specifically, module 38 squares each frequency coefficient F(k) in the band and sums the squares of all coefficients in the band. (Steps 112, 114) (See also Figs. 2(b), 2(c))
[0052] This approximation may provide a poor representation of the spectrum if the spectrum includes a high energy tone at the border between adjacent bands. For example, the spectrum shown Fig. 2(b) includes a tone 50 at the border between band 52 and band 54. Interpolating between approximation sample 58 (representing the sum of the squares of all coefficients in band 54) and approximation sample 56 (representing the sum of the square of all coefficients in band 52) yields a value 59 which does not accurately reflect the presence of tone 50.
[0053] Accordingly, module 38 also employs a second approximation technique wherein the spectral estimates for both bands 52 and 54 will reflect the presence of a tone 50 near the border between the bands.
[0054] More specifically, module 38 derives a second approximation for each band using the ten samples in the band and five neighboring samples from each adjacent band. (See Fig. 2(a)) Module 38 performs the logical inclusive "OR" operation on the binary values of all twenty samples. (Step 116). This operation provides a computationally inexpensive estimate of the magnitude of the largest sample in the set. More specifically, each digit of the binary result of an "OR" operation is set to one if any of the operand binary values have a one in the digit (e.g., 0110 "OR" 0011 = 0111). Accordingly, the result of the "OR" operation is at a minimum equal to the magnitude of the largest sample in the set and at most equal to twice its magnitude. The result of the OR operation is then doubled to form the second approximation. (Step 117).
[0055] As shown in Fig. 2(a), the second approximation in band 52 yields a relatively large approximation value 60 for band 52 since the approximation includes tone 50 from band 54. The second approximation 60 more accurately reflects the presence of tone 50 than the first approximation 56. Accordingly, for each band, module 38 compares the second approximation to the first approximation (Step 118) and selects the larger of the two as the squared spectral estimate S2(j). (Steps 120-122). Finally, module 38 computes the logarithm of the squared spectral estimate Log2 S2(j)). (Step 224).
Speech Detector for Calculating The Signal index Ai
[0056] Referring to Fig. 4, speech detector 40 computes a signal index Ai(F) representative of the relative amount of voice energy in the frame. Toward this end, detector 40 groups the samples of spectral estimates into broad bands of frequencies. The broad bands have varying widths, nonuniformly distributed across the spectrum. For example, the first broad band may be larger than the second broad band which is smaller than the third.
[0057] Referring to Fig. 5, the speech detector estimates the amount of background noise Sn in each broad band. For each broad band, speech detector 40 forms an aggregate estimate Sa(F) as follows:
where F is an integer identifying the current frame, and Y and X are numbers representing the upper and lower frequencies respectively of the broad band. (Step 210). As will be explained in more detail below, the speech detector then compares the aggregate estimate of the current frame against a pair of noise thresholds (derived from prior band aggregate estimates from prior frames) to determine the amount of the noise in the band. (Step 211). In general, a frame having a relatively low aggregate estimate likely includes little voice energy in the broad band. Accordingly, the aggregate estimate in such a frame provides a reasonable estimate of the background noise in the broad band.
[0058] To compare aggregate estimates against the noise thresholds, the speech detector must first unnormalize each aggregate estimate. Otherwise, the estimates from the present frame will be on different scales than those of the prior frames and the comparison will not be meaningful. Accordingly, the speech detector unnormalizes the aggregate estimate by first computing Log2(gq 2) to place the normalization gain in the same logarithmic scale as the aggregate estimates. It then adds the scaled normalization gain Log2(gq 2) to the aggregate estimate to unnormalize the estimate. (Step 212).
[0059] This unnormalized estimate is then compared to an upper threshold Sr(F) and a lower threshold Sf(F) where F identifies the current frame. (Step 214). As explained more fully below, the thresholds are computed for each frame based on the value of a noise estimate from the prior frame. Since the first frame lacks a predecessor, the upper threshold for the first frame, Sr(0), is initialized to a value "r" and the lower threshold Sf(0) is initialized to -f, where f is substantially greater than r (e.g., r=1, and f=10).
[0060] If the aggregate estimate Sa(F) is between the thresholds, the speech detector sets the noise estimate for the frame equal to the aggregate estimate (Step 216):
[0061] If the aggregate estimate is greater than the upper threshold, the noise estimate is set equal to the upper threshold (Step 218):
[0062] Finally, if the aggregate estimate is below the lower threshold, the noise estimate is set equal to the lower threshold (Step 220):
[0063] Before computing the noise estimate for the next frame, the speech detector adjusts the thresholds for the next frame Sf(F+1), Sr(F+1) such that they straddle the noise estimate for the current frame Sn(F). (Step 222) More specifically, for the next frame, F = F + 1, the speech detector calculates an upper noise threshold from the current frame's noise estimate as follows:
Similarly, it calculates a lower threshold Sf(F + 1) = Sn(F) - f.
Thus, for each new frame, the speech detector calculates a noise estimate and adjusts the upper and lower noise thresholds to straddle the noise estimate.
[0064] This technique adaptively adjusts the noise estimates over time such that the noise estimate in a given broad band of a current frame approximately equals the most recent minimum aggregate estimate for the broad band. For example, if a series of frames arrive having no voice component in a broad band, the aggregate estimates will be relatively small since they reflect only the presence of background noise. Thus, if these aggregate values are below the lower threshold, Sf, the above technique will quickly reduce the noise estimate in relatively large increments, f, until the noise estimate equals a value of the relatively low aggregate estimate.
[0065] Once frames having a voice component begin to arrive, the noise estimate remains relatively low, stepping upward in relatively small increments r. By allowing the noise level to increment upward, the speech detector is able to detect increases in the background noise. However, since a relatively small increment "r" is used, the noise estimate tends to remain near the most recent minimum aggregate estimate. This prevents the speech detector from mistaking transient voice energy for background noise.
[0066] After calculating the noise estimates for a frame, the speech detector subtracts the noise estimate Sn(F) from the frame's aggregate estimate Sa(F) to obtain an estimate of the voice signal in the broad band. (Step 224). It also subtracts a threshold constant ST from the aggregate estimate to generate a signal Sout representative of the extent to which the voice signal exceeds the specified threshold ST. (Step 224). Finally, the speech detector computes the index Ai(F) for the entire frame from the collection of Sout signals and the normalization gain gq. (Step 226).
[0067] Referring to Fig. 6, the following describes in further detail the calculation of Ai(F) from the collection of Sout signals. The speech detector first selects the largest Sout from all broad bands. (Step 246). If the selected value, Smax, is less than or equal to zero, all broad bands likely have no voice component. (Step 248). Accordingly, the speech detector sets the index Ai(F) to zero indicating that the broad bands contain only noise. (Step 250).
[0068] If Smax is greater than zero, (Step 248) the speech detector calculates the index Ai(F) from the value of Smax. Toward this end, it scales Smax by a fixed gain Gs i.e., S'max = Smax * Gs where Gs preferably is approximately 0.2734. (Step 252). It next computes a correspondingly attenuated representation go of the normalization gain as follows:
where Tg and Gg are predefined constants e.g., Tg = 4096 and Gg = 0.15625. (Step 254). If go is greater than S'max, the speech detector assumes that Smax is less than the voice energy in the frame. Accordingly, it selects go as the index Ai(F). (Step 256). Otherwise, it selects S'max as the index Ai(F). (Step 256). Finally, the speech detector compares the selected index to a maximum index imax. (Step 258). If the selected index exceeds the maximum index, the speech detector sets the index Ai(F) equal to its maximum value, imax. (Step 262). Otherwise, the selected value is used as the index. (Step 260).
Bit Rate Estimator for Calculating Step Sizes Q(k) and Class Information C(j)
[0069] Referring to Fig. 1(a), the bit rate estimator 42 receives the index Ai(F) and the quantized log spectral estimates Log2Sq 2(j). In response, it computes a step size Q(k) for each frequency coefficient and a class indicator c(j) for each band j of frequency coefficients. The quantization step size Q(k) is used by quantizer 44 to quantize the frequency coefficients. The class indicator c(j) is used by Huffman encoder 47 to select the appropriate coding tables.
[0070] The following describes in further detail the procedure used by bit rate estimator 42 to compute the step size and class information. Referring to Fig. 7, the bit estimator includes a first table 70 which contains a predetermined aggregate allowable distortion value DT for each value of the signal index Ai(F). Each value DT represents an aggregate quantization distortion error for the entire frame. A second table 72 contains, for each value of index Ai(F), a predetermined maximum number of bits rmax initially allocated to the frame. (As explained below, more bits may be allocated to the frame if necessary.) The stored bit rates rmax increase with Ai(F). To the contrary, the stored distortion values DT decrease with each increase in Ai(F). For example, for Ai(F) equals zero, (i.e., 100% noise) the tables provide a small bit rate and a high allowable distortion. For Ai(F) equal seven, (i.e., 100% audio) the tables provide a high bit rate and a low allowable distortion. Based on the value of index Ai(F), the first and second tables select an aggregate allowable quantization distortion DT and an allowable maximum number of bits rmax. This allowable distortion DT is provided to a first distortion approximation module 74. Module 74 then computes a first allowable sample distortion value dl representative of the allowable distortion per estimate. The sample distortion value dl is then used in deriving c(j) and a block quantization step size Q(k).
Calculation of a First Estimate of the Allowable Distortion, d1
[0071] Referring to Figs. 8(a) and 8(b), module 74 first computes an initial value of the sample distortion value d1 by dividing the allowable aggregate distortion DT by the number of spectral estimates P. (Step 310). As will be explained more fully below, only frequency coefficients from a band whose squared spectral estimate is sufficiently greater than a final quantization distortion value dw will be coded for transmission. Thus, the allowable quantization value operates as a noise threshold for determining which coefficients will be transmitted. Accordingly, module 74 performs an inverse logarithm operation on the log spectral estimates Log2 Sq 2(j) to form the square of the quantized spectral estimates Sq 2(j). Module 74 tentatively assumes that if the squared spectral estimate Sq 2(j) is less than or equal to d1, the spectral estimate's constituent frequency coefficients (referred to herein as "noisy samples") will not be coded. (Step 312). Module 74 accordingly increases the sample distortion value d1 to reflect the fact that such constituent coefficients will not be coded. More specifically, it computes the sum DNT of all squared spectral estimates which are less than or equal to d1 (Step 314):
It then subtracts the sum from the aggregate allowable distortion DT (Step 316) and divides the result by the number of remaining spectral estimates N to compute an adjusted sample distortion value d1 (Step 318):
where N is the number of squared spectral estimates above the initial distortion value d1. Since d1 may now be greater than its initial value, module 74 compares each squared spectral estimate to the new d1 to determine if any other coefficients will not be coded. (Step 320). If so, the estimator repeats the process to compute an adjusted sample distortion value d1. (Steps 322, 324). The search terminates when no additional squared spectral coefficients are less than or equal to an adjusted sample distortion value d1 (Step 320). It also terminates after a maximum number of allowable iterations. (Steps 322). The resultant sample distortion value d1 is then provided to a bit rate comparator 76 (Fig. 7). (Step 326).
[0072] Referring again to Fig. 7, based on the first sample distortion value d1 and the log spectral estimate Log2 Sq 2(j), comparator 76 computes a tentative number of bits per frame "r" as follows:
[0073] It then compares the estimated number of bits to the maximum allowable number of bits per frame, rmax. If r is less than the maximum, rmax, comparator 76 signals module 78 to compute step sizes for the frequency coefficients based on the first sample distortion value d1. However, if r exceeds the maximum rmax, comparator 76 assumes that more distortion per estimate must be tolerated to keep the number of bits below rmax.
Accordingly, it signals a second distortion approximation module 80 to begin an iterative search for a new distortion value d2 which will yield a bit rate below rmax.
Second Estimate of the Allowable Distortion, d2
[0074] Referring to Fig. 9, approximation module 80 initially computes a distortion increment value Di which satisfies the relation (Step 410):
The distortion increment value Di is an estimate of a necessary increment in the first sample distortion value to reduce the bit rate below the maximum Rmax. Accordingly, the approximation module computes a new distortion value d2 which satisfies the relation (Step 412):
[0075] In the same manner described above, it then compares each squared spectral estimate Sq 2(j) to d2 to determine which estimates are less than or equal to the distortion, thereby predicting which frequency coefficients will not be coded. (Step 414). Based on this prediction, module 80 again computes the total number of bits r required for the frame according to the following equation (Step 416):
Module 80 again compares the bit rate r to the maximum rmax to determine if the new distortion value d2 yields a bit rate below the maximum. (Step 418). If so, module 80 provides d2 to module 78 and notifies it to calculate the quantization step sizes Q(j) based on both d2 and d1. (Step 422). If not, module 80 performs another iteration of the process in an attempt to find a distortion d2 which will yield a sufficiently low bit rate r. (Steps 418, 420-428). However, if a maximum number of iterations have been tried without finding such a distortion estimate, the search is terminated and the most recent value of d2 is supplied to module 78. (Steps 420, 422).
Calculation of Step Sizes and Class Information From the Distortion Estimates
[0076] Referring to Fig. 10, the module 78 combines the two estimates of allowable sample distortion values d1, d2 to form a weighted distortion dw satisfying the following relation:
where a is a constant weighting factor. (Step 510). (Note: If no value d2 is calculated, the weighted distortion dw is set equal to the first estimate d1). Based on the weighted distortion estimate dw, module 78 computes the class parameter c'(j) for each spectral estimate S(j) as follows (Step 512):
[0077] The class parameter c'(j) is then rounded upward to the nearest integer value in the range of zero to eight to form a class integer c(j). (Step 520). A class value of zero indicates that all coefficients in the class should not be coded. Accordingly, the class value c(j) for each spectral estimate is provided to the quantizer to indicate which coefficients should be coded.
[0078] Class values greater than or equal to one are used to select huffman tables used in encoding the quantized coefficients. Accordingly, the class value c(j) for each spectral estimate is provided to the huffman encoder 47.
[0079] Module 78 next computes a band step size Q(j) for each spectral estimate, based on the value of the weighted distortion dw and the value of the class parameter c (j) for the estimate. (Steps 514-518). More specifically, for spectral estimates whose class values are less than 7.5, the step size Q(j) is calculated to satisfy the following relation:
where z is a constant offset value, e.g., - 1.47156 (Step 516). For spectral coefficients whose class parameter c'(j) is greater than or equal to 7.5, the step size is chosen to satisfy the following relation (Step 518):
[0080] The band step sizes Q(j) for each band j, are then interpolated to derive a step size Q(k) for each frequency coefficient k within the band j. First, each band step size is scaled downward. In this regard, recall that the spectral estimate S(j) for each band j was computed as the greater of 1) the sum of all squared coefficients in the band and 2) twice the logical OR of all coefficients in the block and of the ten neighboring samples. (See Fig. 2(a)-2(c)). Accordingly, the selected spectral estimate roughly approximates the aggregate energy of the entire band. However, the quantization step size for each coefficient should be chosen based on the average energy per coefficient within each band.
[0081] Accordingly, the band step size Q(j) is scaled downward by dividing it by the number of coefficients used in computing the spectral estimate for the band. (i.e., by either ten or twenty depending on the technique chosen to calculate S(j)). Next, bit rate estimator 42 linearly interpolates between the log band step sizes Log2Q(j) to compute the logarithm of coefficient step sizes log2 Q(k). (See Fig. 11). Finally, the reciprocal of the coefficient step size is derived as follows:
Quantization and Encoding of the Frequency Coefficients
[0082] Referring to Figure 1(a), the class integers c(j) and the quantization steps sizes Q(k) are provided to the coefficient quantizer 44. Coefficient quantizer 44 is a mid-tread quantizer which quantizes each frequency coefficient F(k) using its associated inverse step size l/Q(k) to produce an index I(k). The indices I(k) are encoded for transmission by the combined operation of a limit controller 45 and a Huffman encoder 47.
[0083] Huffman encoder 47 includes, for each class c(j), a Huffman table containing a plurality of Huffman codes. The class integers c(j) are provided to Huffman encoder 47 to select the appropriate Huffman table for the identified class.
[0084] In response to an index I(k), limit controller 45 generates a corresponding Huffman index Ih(k) which identifies an entry in the selected Huffman table. Huffman encoder 47 then provides the selected Huffman code Fh(k) to bit stream controller 20 for transmission to the far end.
[0085] Typically, limit controller 45 simply forwards the index I(k) for use as the Huffman index Ih(k). However, the range of possible indices I(k) may exceed the input range of the corresponding Huffman table. Accordingly, for each class c(j), limit controller 45 includes an index maximum and minimum. The limit controller compares each index I(k) with the index maximum and minimum. If I(k) exceeds either the index maximum or minimum, limit controller 45 clips I(k) to equal the respective maximum or minimum and provides the clipped index to encoder 47 as the corresponding Huffman index.
[0086] For each frame, limit controller 45 also maintains a running tally of the number of bits required to transmit the Huffman codes. More specifically, the limit controller includes, for each Huffman table within Huffman encoder 47, a corresponding bit number table. Each entry in the bit number table indicates the number of bits of a corresponding Huffman code stored in the Huffman table of encoder 47. Thus, for each Huffman index Ih(k) generated by limit controller 45, the limit controller internally supplies the Huffman index to the bit number table to determine the number of bits required to transmit the corresponding Huffman code Fh(k) identified by the Huffman index Ih(k). The number of bits are then added to the running tally. If the running tally exceeds a maximum allowable number of bits, limit controller 45 ignores the remaining indices I(k). Limit controller 45 then prepares a unique Huffman index which identifies a unique Huffman code for notifying the far end receiver that the allowable number of bits for the frame has been reached and that the remaining coefficients will not be coded for transmission.
[0087] To transmit the unique Huffman code, the limit controller must allocate bits for transmission of the unique code. Accordingly, it first discards the most recent Huffman code and recomputes the running tally to determine if enough bits are available to transmit the unique Huffman code. If not, the limit controller repeatedly discards the most recent Huffman code until enough bits are allocated for the transmission of the unique Huffman code.
Reconstruction of the Microphone Signal At the Far End
[0088] Referring again to Fig. 1(b), far end audio decoder 28 reconstructs the microphone signal from the set of encoded signals. More specifically, a Huffman decoder 25 decodes the Huffman codes Sh(j) to reconstruct the encoded log spectral estimates Se(j). Decoder 27 (identical to decoder 41 of the audio encoder 16 (Fig. 1(a)) further decodes the encoded log spectral estimates to reconstruct the quantized spectral estimates Log2 Sq 2(j).
[0089] The log spectral estimates Log2 Sq 2(j) and the received signal index Ai(F) are applied to a bit rate estimator 46 which duplicates the derivation of classes c(j) and step sizes Q(k) performed by bit rate estimator 42 at the near end. The derived class information c(j) is provided to a Huffman decoder 47 to decode the Huffman codes Fh(k). The output of the Huffman decoder 47 is applied to a coefficient reconstruction module 48, which, based on the derived quantization step sizes Q(k), reconstructs the original coefficients Fq(k). The bit rate estimator 46 further supplies class information c(j) to a coefficient fill-in module 50 to notify it of which coefficients were not coded for transmission. Module 50 then estimates the missing coefficients using the reconstructed log spectral estimates Log2 Sq 2(j).
[0090] Finally, the decoded coefficients Fq(k) and the estimated coefficients Fe(k) are supplied to a signal composer 52 which converts the coefficients back to the time domain and unnormalizes the time domain signal using the reconstructed normalization gain gq.
[0091] More specifically, an inverse DCT module 51 merges the decoded and estimated coefficients Fq(k), Fe(k) and transforms the resultant frequency coefficient values to a time domain signal m'(n). An inverse normalization module 53 scales the time domain signal m'(n) back to the original scale. The resultant microphone signal m(n) is applied to an overlap decoder 55 which removes the redundant samples introduced by windowing module 32 of the audio encoder 16. (Fig. 1(a)). A signal conditioner 57 filters the resultant microphone signal and converts it to an analog signal L(t) for driving loudspeaker 29.
Estimation of Uncoded Coefficients
[0092] As explained above, the human ear is less likely to notice distortion of an audio signal in regions of the signal's spectrum having a relatively large energy level. Accordingly, bit rate estimator 42 (Fig. 1) adjusts the quantization step size to finely quantize coefficients having a low energy level and coarsely quantize coefficients having a large energy level. In apparent contradiction to this approach, the bit rate estimator simply discards coefficients having a very low energy level.
[0093] The absence of these coefficients would result in noticeable audio artifacts. Accordingly, a coefficient fill-in module 50 prepares a coefficient estimate of each discarded coefficient from the spectral estimates. It then inserts the coefficient estimate in place of the missing coefficient to prevent such audio artifacts. In doing so, the coefficient module considers the level of the signal index Ai for the frame in which the uncoded coefficient resides. If, for example, the signal index is low, (indicating that the frame largely consists of background noise), the fill-in module assumes the missing coefficient represents background noise. Accordingly, it prepares the coefficient estimate largely from a measure of the background noise in the frame. However, if the signal index is large, (indicating that the frame consists largely of a voice signal), the noise fill-in module assumes that the missing coefficient represents a voice signal. Accordingly, it prepares the coefficient estimate largely from the value of the spectral estimate corresponding to the band of frequencies which includes the frequency of the missing coefficient.
[0094] Referring to Fig. 12, coefficient fill-in module 50 (Fig. 1(b)) includes a coefficient estimator module 82 for each band. Each estimator module 82 includes a noise floor module 84 for approximating, for each Frame F, the amount of background noise in the band of frequencies j. The noise estimate Sn(j,F) is derived from a comparison of the log spectral estimate Log2 Sq 2(j) of the current frame with a noise estimate derived from spectral estimates of previous frames. An adder 91 adds the log spectral estimate to the log gain, log2 gq 2, to unnormalize the log spectral estimate. The unnormalized estimate Su(j, F) is applied to a comparator 99 which compares Su(j, F) to the noise estimate Sn(j,F-1) calculated for the previous frame F-1 (Sn is initialized to zero for the first frame). If Su(j, F) is greater than the previous noise estimate Sn(j, F-1), the noise estimate for the present frame F is computed as follows:
where tr is a rise time constant provided by table 100. More specifically, table 100 provides a unique tr for each value of the signal index Ai(F). (e.g., for Ai(F) = 0, a relatively large time constant tr is chosen to yield a long rise time. As Ai(F) increases, the selected time constant decreases).
[0095] If Su(j, F) is less than the previous noise estimate, Sn(j, F - 1), the noise for the current frame is computed as follows:
where tf is a fall time constant provided by table 102. Table 102, like table 100, provides a unique constant tf for each value of the index Ai(F). Adder 93 normalizes the resultant noise estimate Sn(j, F) by subtracting the log gain, log2 g2 q. The output of 93 is then applied to a log inverter 94 which computes the inverse logarithm as show below to provide a normalized noise estimate Snn(j, F):
[0096] For each frame, the normalized band noise estimates, Snn(j, F), and the spectral estimate Sq(j, F) are applied to a weighting function 86 which prepares weighted sum of the two values. Weighting function 86 includes a first table 88 which contains, for each value of signal index Ai(F), a noise weighting coefficient Cn(F). Similarly, a second table 90 includes for each index Ai(F), a voice weighting index Ca(F). In response to the current value of the audio index Ai(F), table 88 provides a corresponding noise weighting coefficient Cn(F) to multiplier 92. Multiplier 92 computes the product of Cn(F) and Snn(j, F) to produce a weighted noise value Snw(j, F). Similarly, table 90 provides a voice weighting coefficient Ca(F) to multiplier 94 to compute a weighted voice value Svw(j, F) where Svw = Sq(j, F) Ca. The weighted values are provided to a composer 98 which computes a weight estimate as follows:
[0097] The weighting coefficients Ca, Cn stored in tables 90, 88 are related as follows Ca = 1 - Cn, where the values of Cn may range between zero (for Ai=7) and one (for Ai=0). Thus, during silence, the noise estimate carries more weight, while the spectral estimate carries gradually more weight as the audio index increases.
[0098] This weighted estimate is then supplied to signal composer 52 for use in computing the estimated frequency coefficient Fe(k) for each of the ten uncoded frequency coefficients corresponding to the spectral estimate.
[0099] More specifically, the weighted estimate, W, is used to control the level of "fill-in" for each missing frequency coefficient (i.e. those with a class of zero). The fill-in consists of scaling the output from a random number generator (with a uniform distribution) and inserting the result, Fe in place of the missing frequency coefficients. The following equation is used to generate the fill-in for each missing frequency coefficient.
Where "noise" is the output from the random number generator at given instant (or coef/sample), the range of the random number generator being twice the value n; and wherein e is a constant, e.g., 3. Note that a new value of noise is generated for each of the missing frequency coefficients.
selecting an aggregate allowable quantization distortion value representing a collective allowable quantization distortion error for a frame of samples of said digital signal,
selecting from said frame of samples, a set of samples wherein each of a plurality of samples of said set is greater than a noise threshold,
computing, for each sample of said set, a sample quantization distortion value representing an allowable quantization distortion error for said sample, wherein the sum of all sample quantization distortion values for all samples of said set is approximately equal to said aggregate allowable quantization distortion value,
for each sample of said set, selecting a quantization step size which yields a quantization distortion error approximately equal to said sample's corresponding quantization distortion value, and
quantizing said sample using said quantization step size.
preparing a signal index representing, for at least one sample of said frame, a signal magnitude of said signal component relative to a noise magnitude of said noise component, and
based on said signal index, selecting said aggregate allowable quantization distortion.
selecting a tentative set of samples of said digital signal wherein each sample of said tentative set is greater than a noise threshold determined at least in part by the value of said first sample distortion value, and wherein calculating said sample quantization distortion value comprises the step of:
adjusting said first sample distortion value by an amount determined by the difference between said first sample distortion value and at least one sample excluded from said tentative set.
a) identifying any noisy samples of said tentative set, each said noisy sample having a magnitude below a current value of said first sample distortion value,
b) removing said noisy samples, if any, from said tentative set,
c) if any noisy samples are removed, increasing said first sample distortion value by an amount determined by the difference between said first sample distortion value and at least one said noisy sample, and
repeating steps a, b and c until an adjusted first sample distortion value is reached for which step a identifies no additional noisy samples of said tentative set which are greater than said adjusted first sample distortion.after said adjustment is terminated, estimating a bit number for a quantity of bits required to transmit all samples of said tentative set,
comparing said estimated bit number to a maximum bit number, and
if said estimated bit number is less than or equal to said maximum bit number, selecting a final noise threshold based on said adjusted first sample distortion value.
preparing a second sample distortion value,
a) selecting a second tentative set of samples of said digital signal, each of a plurality of said samples of said second tentative set having a magnitude above said second sample distortion value,
b) estimating the number of bits required to transmit all samples of said second tentative set,
c) comparing said estimated bit number to said maximum bit number, and
d) if said estimated bit number is greater than said maximum bit number, increasing said second sample distortion value by an amount determined by said selection.
calculating said sample distortion value from said adjusted first sample distortion value and said second sample distortion value and,
selecting a final set of samples of said digital signal, each of a plurality of said samples of said final set having a magnitude above a final threshold determined by the sample's corresponding sample distortion value.
preparing an estimation signal representative of said digital signal, yet having fewer samples than said digital signal,
preparing a signal index representing, for at least one sample of said estimation signal, a signal magnitude of said signal component relative to a noise magnitude of said noise component,
based on said signal index, selecting samples of said digital signal having a sufficiently large signal component,
transmitting said selected samples of said digital signal,
transmitting said samples of said digital estimation signal, and
reconstructing said digital signal from said transmitted selected samples and estimation samples.
generating a random number for each nonselected sample of said frequency domain speech signal,
preparing, from at least one said spectral estimate, a noise estimate of the magnitude of a noise component of said nonselected sample,
based on said noise estimate, generating a scaling factor, and
scaling said random number according to said scaling factor to produce a reconstructed sample representative of said nonselected sample.
for a prior frame of said estimation signal, preparing an initial noise estimate from at least one spectral estimate representative of a band of frequencies of said prior frame,
based on the magnitude of said signal index, selecting a rise time constant tr and a fall time constant tf,
adding said rise time constant to said initial noise estimate to form an upper threshold,
subtracting said fall time constant from said initial noise estimate to form a lower threshold,
comparing a current spectral estimate of said current frame, representative of said band of frequencies, to said upper and lower thresholds,
if said current spectral estimate is between said thresholds, setting said current noise estimate equal to said current spectral estimate,
if said current spectral estimate is below said lower threshold, decrementing said initial noise estimate by said fall time constant to form said current noise estimate, and
if said current spectral estimate is above said upper threshold, incrementing said initial noise estimate by said rise time constant to form said current noise estimate.
selecting a noise coefficient based on the value of said signal index,
selecting a voice coefficient based on the value of said signal index,
multiplying said current noise estimate by said noise coefficient,
multiplying said current spectral estimate by said voice coefficient,
adding the products of said multiplication steps, and
forming said scaling factor from the result of said adding step.
means for selecting an aggregate allowable quantization distortion value representing a collective allowable quantization distortion error for a frame of samples of said digital signal,
means for selecting from said frame of samples, a set of samples wherein each of a plurality of samples of said set is greater than a noise threshold,
means for computing, for each sample of said set, a sample quantization distortion value representing an allowable quantization distortion error for said sample, wherein the sum of all sample quantization distortion values for all samples of said set is approximately equal to said aggregate allowable quantization distortion value,
means for selecting, for each sample of said set, a quantization step size which yields a quantization distortion error approximately equal to said sample's corresponding quantization distortion value, and
means for quantizing said sample using said quantization step size.
and wherein said means for selecting an aggregate allowable quantization distortion comprises:
means for preparing a signal index representing, for at least one sample of said frame, a signal magnitude of said signal component relative to a noise magnitude of said noise component, and
means for selecting said aggregate allowable quantization distortion, based on said signal index.
means for selecting a tentative set of samples of said digital signal wherein each sample of said tentative set is greater than a noise threshold determined at least in part by the value of said first sample distortion value, and wherein said means for computing said sample quantization distortion value comprises:
means for adjusting said first sample distortion value by an amount determined by the difference between said first sample distortion value and at least one sample excluded from said tentative set.
means for identifying any noisy samples of said tentative set, each said noisy sample having a magnitude below a current value of said first sample distortion value,
means for removing said noisy samples, if any, from said tentative set,
means for increasing said first sample distortion value by an amount determined by the difference between said first sample distortion value and at least one said noisy sample, if any noisy samples are removed, and
means for repeatedly removing noisy samples and adjusting said first sample distortion value until an adjusted first sample distortion value is reached for which no additional noisy samples of said tentative set are greater than said adjusted first sample distortion.
means for estimating after said adjustment is terminated, a bit number for a quantity of bits required to transmit all samples of said tentative set,
means for comparing said estimated bit number to a maximum bit number, and
means for selecting a final noise threshold based on said adjusted first sample distortion value, if said estimated bit number is less than or equal to said maximum bit number.
means for preparing a second sample distortion value, if said estimated bit number exceeds said maximum bit number,
means for selecting a second tentative set of samples of said digital signal, each of a plurality of said samples of said second tentative set having a magnitude above said second sample distortion value,
means for estimating the number of bits required to transmit all samples of said second tentative set,
means for comparing said estimated bit number to said maximum bit number, and
means for increasing said second sample distortion value by an amount determined by said selection, if said estimated bit number is greater than said maximum bit number.
means for calculating said sample distortion value from said adjusted first sample distortion value and said second sample distortion value and,
means for selecting a final set of samples of said digital signal, each of a plurality of said samples of said final set having a magnitude above a final threshold determined by the sample's corresponding sample distortion value.
means for preparing an estimation signal representative of said digital signal, yet having fewer samples than said digital signal,
means for preparing a signal index representing, for at least one sample of said estimation signal, the magnitude of said signal component relative to the magnitude of said noise component,
means for selecting, based on said signal index, samples of said digital signal having a sufficiently large signal component,
means for transmitting said selected samples of said digital signal,
means for transmitting said samples of said digital estimation signal, and
means for reconstructing said digital signal from said transmitted selected samples and estimation samples.
means for generating a random number for each nonselected sample of said frequency domain speech signal,
means for preparing, from at least one said spectral estimate, a noise estimate of the magnitude of a noise component of said nonselected sample,
means for generating a scaling factor based on said noise estimate, and
means for scaling said random number according to said scaling factor to produce a reconstructed sample representative of said nonselected sample.
means for preparing, for a prior frame of said estimation signal, an initial noise estimate from at least one spectral estimate representative of a band of frequencies of said prior frame,
means for selecting, based on the magnitude of said signal index, a rise time constant tr and a fall time constant tf,
means for adding said rise time constant to said initial noise estimate to form an upper threshold,
means for subtracting said fall time constant from said initial noise estimate to form a lower threshold,
means for comparing a current spectral estimate representative of said band of frequencies in said current frame, to said upper and lower thresholds,
means for setting said current noise estimate equal to said current spectral estimate, if said current spectral estimate is between said thresholds,
means for decrementing said initial noise estimate by said fall time constant to form said current noise estimate, if said current spectral estimate is below said lower threshold, and
means for incrementing said initial noise estimate by said rise time constant to form said current noise estimate, if said current spectral estimate is above said upper threshold.
means for selecting a noise coefficient based on the value of said signal index,
means for selecting a voice coefficient based on the value of said signal index,
means for multiplying said current noise estimate by said noise coefficient,
means for multiplying said current spectral estimate by said voice coefficient,
means for adding the products of said multiplications, and
means for forming said scaling factor from the result of said adding.
Wählen eines summierten, zulässigen Quantisierungsverzerrungswertes, der einen kollektiven, zulässigen Quantisierungsverzerrungsfehler für einen Block von Abtastwerten des genannten digitalen Signales darstellt,
Auswählen einer Gruppe von Abtastwerten aus dem genannten Block von Abtastwerten, wobei jeder Abtastwert aus einer Mehrzahl von Abtastwerten der genannten Gruppe größer ist als eine Geräuschschwelle,
für jeden Abtastwert der genannten Gruppe Berechnen eines Abastwert-Quantisierungsverzerrungswertes, welcher einen zulässigen Quantisierungsverzerrungsfehler für den genannten Abtastwert darstellt, wobei die Summe sämtlicher Abtastwert-Quantisierungsverzerrungswerte für sämtliche Abtastwerte der genannten Gruppe etwa gleich dem genannten summierten, zulässigen Quantisierungsverzerrungswert ist,
für jeden Abtastwert der genannten Gruppe Wählen einer Quantisierungsschrittgröße, die zu einem Quantisierungsverzerrungsfehler führt, der etwa gleich dem entsprechenden Quantisierungsverzerrungswert des genannten Abtastwertes ist, und
Quantisieren des genannten Abtastwertes unter Verwendung der genannten Quantisierungsschrittgröße.
Bereiten eines Signalindexes, der für zumindest einen Abtastwert des genannten Blockes eine Signalgröße der genannten Signalkomponente relativ zu einer Rauschgröße der genannten Geräuschkomponente darstellt, und
Wählen der genannten summierten, zulässigen Quantisierungsverzerrung auf Grundlage des genannten Signal indexes.
Auswählen einer provisorischen Gruppe von Abtastwerten des genannten digitalen Signales, wobei jeder Abtastwert der genannten provisorischen Gruppe größer ist als eine Geräuschschwelle, die zumindest teilweise durch den Wert des genannten ersten Abtastwert-Verzerrungswertes bestimmt ist, und wobei das Berechnen des genannten Abtastwert-Quantisierungsverzerrungswertes den Schritt aufweist:
Verstellen des genannten ersten Abtastwert-Verzerrungswertes um einen Betrag, der durch die Differenz zwischen dem genannten ersten Abtastwert-Verzerrungswert und zumindest einem Abtastwert bestimmt ist, der aus der genannten provisorischen Gruppe ausgeschlossen ist.
a) Identifizieren jedweder rauschigen Abtastwerte der genannten provisorischen Gruppe, wobei jeder genannte rauschige Abtastwert eine geringere Größe als ein laufender Wert des genannten ersten Abtastwert-Verzerrungswertes besitzt,
b) Entfernen der genannten rauschigen Abtastwerte, falls vorhanden, aus der genannten provisorischen Gruppe,
c) Vergrößern, wenn irgendwelche rauschigen Abtastwerte entfernt werden, des genannten ersten Abtastwert-Verzerrungswertes um einen Betrag, der durch die Differenz zwischen dem genannten ersten Abtastwert-Verzerrungswert und zumindest einem genannten rauschigen Abtastwert bestimmt ist, und
Wiederholen der Schritte a, b und c bis ein eingestellter erster Abtastwert-Verzerrungswert erreicht ist, für den der Schritt a keine zusätzlichen rauschigen Abtastwerte der genannten provisorischen Gruppe identifiziert, welche größer sind als die genannte eingestellte erste Abtastwertverzerrung.Abschätzen, nachdem die genannte Verstellung abgebrochen ist, einer Bitanzahl für eine Menge von Bits, die erforderlich sind, um sämtliche Abtastwerte der genannten provisorischen Gruppe zu übertragen,
Vergleichen der genannten abgeschätzten Bitanzahl mit einer maximalen Bitanzahl und
Wählen, wenn die genannte abgeschätzte Bitanzahl kleiner oder gleich der genannten maximalen Bitanzahl ist, einer endgültigen Geräuschschwelle auf Grundlage des genannten eingestellten ersten Abtastwert-Verzerru ngswertes.
Bereiten eines zweiten Abtastwert-Verzerrungswertes,
a) Auswählen einer zweiten provisorischen Gruppe von Abtastwerten des genannten digitalen Signales, wobei jeder Abtastwert einer Mehrzahl der genannten Abtastwerte der genannten zweiten provisorischen Gruppe eine Größe besitzt, die oberhalb des genannten zweiten Abtastwert-Verzerrungswertes liegt,
b) Abschätzen der Anzahl von Bits, die erforderlich sind, um sämtliche Abtastwerte der genannten zweiten provisorischen Gruppe zu übertragen,
c) Vergleichen der genannten abgeschätzten Bitanzahl mit der genannten maximalen Bitanzahl und
d) Vergrößern, wenn die genannte abgeschätzte Bitanzahl größer ist als die genannte maximale Bitanzahl, des genannten zweiten Abtastwert-Verzerrungswertes um einen Betrag, der durch die genannte Auswahl bestimmt ist.
Berechnen des genannten Abtastwert-Verzerrungswertes aus dem genannten eingestellten, ersten Abtastwert-Verzerrungswert und dem genannten zweiten Abtastwert-Verzerrungswert und
Auswählen einer endgültigen Gruppe von Abtastwerten des genannten digitalen Signales, wobei jeder Abtastwert einer Mehrzahl der genannten Abtastwerte der genannten endgültigen Gruppe eine Größe besitzt, die oberhalb einer endgültigen Schwelle liegt, die durch den entsprechenden Abtastwert-Verzerrungswert des Abtastwertes bestimmt ist.
Bereiten eines abgeschätzten Signales, das für das genannte digitale Signal repräsentativ ist, jedoch weniger Abtastwerte aufweist als das genannte digitale Signal,
Bereiten eines Signalindexes, der, für zumindest einen Abtastwert des genannten abgeschätzten Signales, eine Signalgröße der genannten Signalkomponente relativ zu einer Rauschgröße der genannten Geräuschkomponente darstellt,
Auswählen, basierend auf dem genannten Signalindex, von Abtastwerten des genannten digitalen Signales, die eine ausreichend große Signalkomponente besitzen,
Übertragen der genannten ausgewählten Abtastwerte des genannten digitalen Signales,
Übertragen der genannten Abtastwerte des genannten digitalen, abgeschätzten Signales und
Rekonstruieren des genannten digitalen Signales aus den genannten übertragenen ausgewählten Abtastwerten und den abgeschätzten Abtastwerten.
Generieren einer Zufallszahl für jeden nicht ausgewählten Abtastwert des genannten Frequenzbereich-Sprachsignales,
Bereiten, aus zumindest einem genannten spektralen Schätzwert, einen Geräuschschätzwert für die Größe einer Geräuschkomponente des nicht gewählten Abtastwertes,
Erzeugen, auf Grundlage des genannten Geräuschschätzwertes, eines Skalierungsfaktors und
Skalieren der genannten Zufallszahl gemäß dem genannten Skalierungsfaktor, um einen rekonstruierten Abtastwert zu erzeugen, der für den genannten nicht ausgewählten Abtastwert repräsentativ ist.
Bereiten, für einen voraufgehenden Block des genannten abgeschätzten Signales, einen anfänglichen Geräuschschätzwert aus zumindest einem spektralen Schätzwert, der für ein Frequenzband des genannten voraufgehenden Blockes repräsentativ ist,
Wählen, basierend auf der Größe des genannten Signalindexes, eine Anstiegszeitkonstante t, und eine Abfallzeitkonstante tf,
Addieren der genannten Anstiegszeitkonstante zu dem genannten anfänglichen Geräuschschätzwert, um eine obere Schwelle zu bilden,
Subtrahieren der genannten Abfallzeitkonstante von dem genannten anfänglichen Geräuschschätzwert, um eine untere Schwelle zu bilden,
Vergleichen eines laufenden spektralen Schätzwertes des genannten laufenden Blockes, der für das genannte Frequenzband repräsentativ ist, mit der genannten oberen und der genannten unteren Schwelle,
Setzen, wenn der genannte laufende spektrale Schätzwert sich zwischen den genannten Schwellen befindet, des genannten laufenden Schätzwertes gleich dem genannten laufenden spektralen Schätzwert,
Dekrementieren, wenn der genannte spektrale Schätzwert unterhalb der genannten unteren Schwelle liegt, des genannten anfänglichen Geräuschschätzwertes mit der genannten Abfallzeitkonstante, um den genannten laufenden Geräuschschätzwert zu bilden, und
Inkrementieren, wenn der genannte laufende spektrale Schätzwert oberhalb der genannten oberen Schwelle liegt, des genannten anfänglichen Geräuschschätzwertes mit der Anstiegszeitkonstante, um den genannten laufenden Geräuschschätzwert zu bilden.
Wählen eines Geräuschkoeffizienten auf der Basis des Werte des genannten Signalindexes,
Wählen eines Tonkoeffizienten auf Grundlage des Wertes des genannten Signalindexes,
Multiplizieren des genannten laufenden Geräuschschätzwertes mit dem genannten Geräuschkoeffizienten,
Multiplizieren des genannten laufenden spektralen Schätzwertes mit dem genannten Tonkoeffizienten,
Addieren der Produkte der genannten Multiplikationsschritte und
Bilden des genannten Skalierungsfaktors aus dem Ergebnis des genannten Additionsschrittes.
ein Mittel zum Wählen eines summierten, zulässigen Quantisierungsverzerrungswertes, der einen kollektiven, zulässigen Quantisierungsverzerrungsfehler für einen Block von Abtastwerten des genannten digitalen Signales darstellt,
ein Mittel zum Auswählen einer Gruppe von Abtastwerten aus dem genannten Block von Abtastwerten, wobei jeder Abtastwert aus einer Mehrzahl von Abtastwerten der genannten Gruppe größer ist als eine Geräuschschwelle,
ein Mittel, um für jeden Abtastwert der genannten Gruppe einen Abtastwert-Quantisierungsverzerrungswert zu berechnen, der einen zulässigen Quantisierungsverzerrungsfehler für den genannten Abtastwert darstellt, wobei die Summe sämtlicher Abtastwert-Quantisierungsverzerrungswerte für sämtliche Abtastwerte der genannten Gruppe etwa gleich dem genannten summierten, zulässigen Quantisierungsverzerrungswert ist,
ein Mittel, um für jeden Abtastwert der genannten Gruppe eine Quantisierungsschrittgröße zu wählen, die zu einem Quantisierungsverzerrungsfehler führt, der etwa gleich dem entsprechenden Quantisierungsverzerrungswert des genannten Abtastwertes ist, und
ein Mittel zum Quantisieren des genannten Abtastwertes unter Verwendung der genannten Quantisierungsschrittgröße.
und bei dem das genannte Mittel zum Wählen einer summierten, zulässigen Quantisierungsverzerrung beinhaltet:
ein Mittel zum Bereiten eines Signalindexes, der für zumindest einen Abtastwert des genannten Blockes eine Signalgröße der genannten Signalkomponente relativ zu einer Rauschgröße der genannten Geräuschkomponente darstellt, und
ein Mittel zum Wählen der genannten summierten, zulässigen Quantisierungsverzerrung auf Grundlage des genannten Singalindexes.
ein Mittel zum Auswählen einer provisorischen Gruppe von Abtastwerten des genannten digitalen Signales, wobei jeder Abtastwert der genannten provisorischen Gruppe größer ist als eine Geräuschschwelle, die zumindest teilweise durch den Wert des genannten ersten Abtastwert-Verzerrungswertes bestimmt ist, und wobei das genannte Mittel zum Berechnen des genannten Abtastwert-Quantisierungsverzerrungswertes aufweist:
ein Mittel zum Verstellen des genannten ersten Abtastwert-Verzerrungswertes um einen Betrag, der durch die Differenz zwischen dem genannten ersten Abtastwert-Verzerrungswert und zumindest einem Abtastwert bestimmt ist, der aus der genannten provisorischen Gruppe ausgeschlossen ist.
ein Mittel zum Identifizieren jedweder rauschiger Abtastwerte der genannten provisorischen Gruppe, wobei jeder genannte rauschige Abtastwert eine geringere Größe als ein laufender Wert des genannten ersten Abtastwert-Verzerrungswertes besitzt,
ein Mittel zum Entfernen der genannten rauschigen Abtastwerte, falls vorhanden, aus der genannten provisorischen Gruppe,
ein Mittel zum Vergrößern des genanten ersten Abtastwert-Verzerrungswertes um einen Betrag, der durch die Differenz zwischen dem genannten ersten Abtastwert-Verzerrungswert und zumindest einem genannten rauschigen Abtastwert bestimmt ist, wenn irgendwelche rauschigen Abtastwerte entfernt werden, und
ein Mittel zum wiederholten Entfernen rauschiger Abtastwerte und Verstellen des genannten ersten Abtastwert-Verzerrungswertes, bis ein eingestellter erster Abtastwert-Verzerrungswert erreicht ist, bei dem keine zusätzlichen rauschigen Abtastwerte der genannten provisorischen Gruppe größer sind als die genannte, eingestellte erste Abtastwert-Verzerrung.
ein Mittel, um, nachdem die genannte Verstellung abgebrochen ist, eine Bitanzahl für eine Menge von Bits abzuschätzen, die erforderlich sind, um sämtliche Abtastwerte der genannten provisorischen Gruppe zu übertragen,
ein Mittel zum Vergleichen der genannten abgeschätzten Bitanzahl mit einer maximalen Bitanzahl und
ein Mittel, um eine endgültige Geräuschschwelle auf Grundlage des genannten eingestellten ersten Abtastwert-Verzerrungswertes zu wählen, wenn die genannte abgeschätzte Bitanzahl kleiner oder gleich der genannten maximalen Bitanzahl ist.
ein Mittel zum Bereiten eines zweiten Abtastwert-Verzerrungswertes, wenn die genannte abgeschätzte Bitanzahl die genannte maximale Bitanzahl übersteigt,
ein Mittel zum Auswählen einer zweiten provisorischen Gruppe von Abtastwerten des genannten digitalen Signales, wobei jeder Abtastwert einer Mehrzahl der genannten Abtastwerte der genannten zweiten provisorischen Gruppe eine Größe besitzt, die oberhalb des genannten zweiten Abtastwert-Verzerrungswertes liegt,
ein Mittel zum Abschätzen der Anzahl von Bits, die erforderlich sind, um sämtliche Abtastwerte der genannten zweiten provisorischen Gruppe zu übertragen,
ein Mittel zum Vergleichen der genannten abgeschätzten Bitanzahl mit der genannten maximalen Bitanzahl und
ein Mittel, um den genannten zweiten Abtastwert-Verzerrungswert um einen Betrag zu vergrößern, der durch die genannte Auswahl bestimmt ist, wenn die genannte abgeschätzte Bitanzahl größer ist als die genannte maximale Bitanzahl.
ein Mittel zum Berechnen des genannten Abtastwert-Verzerrungswertes aus dem genannten eingestellten, ersten Abtastwert-Verzerrungswert und dem genannten zweiten Abtastwert-Verzerrungswert und
ein Mittel zum Auswählen einer endgültigen Gruppe von Abtastwerten des genannten digitalen Signales, wobei jeder Abtastwert einer Mehrzahl der genannten Abtastwerte der genannten endgültigen Gruppe eine Größe besitzt, die oberhalb einer endgültigen Schwelle liegt, die durch den entsprechenden Abtastwert-Verzerrungswert des Abtastwertes bestimmt ist.
ein Mittel zum Bereiten eines abgeschätzten Signales, das für das genannte digitale Signal repräsentativ ist, jedoch weniger Abtastwerte aufweist als das genannte digitale Signal,
ein Mittel zum Bereiten eines Signalindexes, der für zumindest einen Abtastwert des genannten abgeschätzten Signales die Größe der genannten Signalkomponente relativ zu der Größe der genannten Geräuschkomponente darstellt,
ein Mittel, um, basierend auf dem genannten Signalindex, Abtastwerte des genannten digitalen Signales auszuwählen, die eine ausreichend große Signalkomponente besitzen,
ein Mittel zum Übertragen der genannten ausgewählten Abtastwerte des genannten digitalen Signales,
ein Mittel zum Übertragen der genannten Abtastwerte des genannten digitalen, abgeschätzten Signales und
ein Mittel zum Rekonstruieren des genannten digitalen Signales aus den genannten übertragenen ausgewählten Abtastwerten und den abgeschätzten Abtastwerten.
ein Mittel zum Generieren einer Zufallszahl für jeden nicht ausgewählten Abtastwert des genannten Frequenzbereich-Sprachsignales,
ein Mittel, um aus zumindest einem genannten spektralen Schätzwert einen Geräuschschätzwert für die Größe einer Geräuschkomponente des nicht ausgewählten Abtastwertes zu bereiten,
ein Mittel, um auf der Grundlage des genannten Geräuschschätzwertes einen Skalierungsfaktor zu erzeugen, und
ein Mittel zum Skalieren der genannten Zufallszahl gemäß dem genannten Skalierungsfaktor, um einen rekonstruierten Abtastwert zu erzeugen, der für den genannten nicht ausgewählten Abtastwert repräsentativ ist.
ein Mittel, um für einen voraufgehenden Block des genannten abgeschätzten Signales einen anfänglichen Geräuschschätzwert aus zumindest einem spektralen Schätzwert zu bereiten, der für ein Frequenzband des genannten voraufgehenden Blockes repräsentativ ist,
ein Mittel, um auf Grundlage der Größe des genannten Signalindexes eine Anstiegszeitkonstante tr und eine Abfallzeitkonstante tf zu wählen,
ein Mittel, um die genannte Anstiegszeitkonstante mit dem genannten anfänglichen Geräuschschätzwert zu addieren, um eine obere Schwelle zu bilden,
ein Mittel, um die genannte Abfallzeitkonstante von dem genannten anfänglichen Geräuschschätzwert zu subtrahieren, um eine untere Schwelle zu bilden,
ein Mittel zum Vergleichen eines laufenden spektralen Schätzwertes, der für das Frequenzband in dem genannten laufenden Block repräsentativ ist, mit der genannten oberen und genannten unteren Schwelle,
ein Mittel, um den genannten laufenden Geräuschschätzwert gleich dem genannten laufenden spektralen Schätzwert zu setzen, wenn der genannte laufende spektrale Schätzwert zwischen den genannten Schwellen liegt,
ein Mittel, um den genannten anfänglichen Geräuschschätzwert mit der genannten Abfallzeitkonstante zu dekrementieren, um den genannten laufenden Geräuschschätzwert zu bilden, wenn der genannte laufende spektrale Schätzwert unterhalb der genannten unteren Schwelle liegt, und
ein Mittel, um den genannten anfänglichen Geräuschschätzwert mit der genannten Anstiegszeitkonstante zu inkrementieren, um den genannten laufenden Geräuschschätzwert zu bilden, wenn der genannte laufende spektrale Schätzwert oberhalb der genannten oberen Schwelle liegt.
ein Mittel, um einen Geräuschkoeffizienten auf Grundlage des Wertes des genannten Signalindexes zu wählen,
ein Mittel, um einen Tonkoeffizienten auf Grundlage des Wertes des genannten Signalindexes zu wählen,
ein Mittel, um den genannten laufenden Geräuschschätzwert mit dem genannten Geräuschkoeffizienten zu multiplizieren,
ein Mittel, um den genannten laufenden spektralen Schätzwert mit dem genannten Tonkoeffizienten zu multiplizieren,
ein Mittel, um die Produkte der genannten Multiplikationen zu addieren, und
ein Mittel, um den genannten Skalierungsfaktor aus dem Ergebnis der genannten Addition zu bilden.
sélectionner une valeur de distorsion de quantification acceptable globale représentant une erreur de distorsion de quantification acceptable collective pour une structure d'échantillons dudit signal numérique,
sélectionner à partir de ladite structure d'échantillons, un jeu d'échantillons dans lequel chacune d'une pluralité d'échantillons dudit jeu est supérieure à un seuil de bruit,
calculer, pour chaque échantillon dudit jeu, une valeur de distorsion de quantification d'échantillon représentant une erreur de distorsion de quantification acceptable pour ledit échantillon, et où la somme de toutes les valeurs de distorsion de quantification d'échantillon pour tous les échantillons dudit jeu est approximativement égale à ladite valeur de distorsion de quantification acceptable globale,
sélectionner, pour chaque échantillon dudit jeu, une étape de taille de quantification qui fournit une erreur de distorsion de quantification approximativement égale à ladite valeur de distorsion de quantification correspondante des échantillons, et
quantifier ledit échantillon en utilisant ladite étape de taille de quantification.
préparer un indice de signal représentant, pour au moins un échantillon de ladite structure, une grandeur de signal de ladite composante de signal relative à une grandeur de bruit de ladite composante de bruit, et
en se fondant sur ledit indice de signal, sélectionner ladite distorsion de quantification acceptable globale.
sélectionner un jeu expérimental d'échantillons dudit signal numérique dans lequel chaque échantillon dudit jeu expérimental est supérieur à un seuil de bruit déterminé au moins en partie par la valeur de ladite première valeur de distorsion d'échantillon, et dans lequel le calcul de ladite valeur de distorsion de quantification d'échantillon comprend l'étape consistant à :
ajuster ladite première valeur de distorsion d'échantillon d'une quantité déterminée par la différence entre ladite première valeur de distorsion d'échantillon et au moins un échantillon exclu dudit jeu expérimental.
a) identifier tous les échantillons bruyants dudit jeu expérimental, chacun desdits échantillons bruyants ayant une grandeur située en dessous d'une valeur courante de ladite première valeur de distorsion d'échantillon,
b) retirer lesdits échantillons bruyants, s'il en existe, dudit jeu expérimental,
c) si tous les échantillons bruyants sont retirés, augmenter ladite première valeur de distorsion d'échantillon d'une quantité déterminée par la différence entre ladite première valeur de distorsion d'échantillon et au moins un dit échantillon bruyant, et
répéter les étapes a), b) et c) jusqu'à ce que soit atteinte une première valeur de distorsion d'échantillon ajustée dans laquelle étape aucun échantillon bruyant additionnel dudit jeu expérimental n'est identifié qui soit supérieur à ladite première distorsion d'échantillon ajustée.après terminaison dudit ajustement, estimer un nombre binaire pour une quantité d'éléments binaires requise de façon à transmettre tous les échantillons dudit jeu expérimental,
comparer ledit nombre binaire estimé à un nombre binaire maximum, et
si ledit nombre binaire estimé est inférieur ou égal audit nombre binaire maximum, sélectionner un seuil de bruit final en se fondant sur ladite première valeur de distorsion d'échantillon ajustée.
préparer une seconde valeur de distorsion d'échantillon,
a) sélectionner un second jeu expérimental d'échantillons dudit signal numérique, chaque pluralité desdits échantillons dudit second jeu expérimental ayant une grandeur supérieure à ladite seconde valeur de distorsion d'échantillon,
b) estimer le nombre d'éléments binaires requis pour transmettre tous les échantillons dudit second jeu expérimental,
c) comparer ledit nombre binaire estimé audit nombre binaire maximum, et
d) si ledit nombre binaire estimé est supérieur audit nombre binaire maximum, augmenter ladite seconde valeur de distorsion d'échantillon d'une quantité déterminée par ladite sélection.
calculer ladite valeur de distorsion d'échantillon à partir de ladite première valeur de distorsion ajustée d'échantillon et ladite seconde valeur de distorsion d'échantillon et,
sélectionner un jeu final d'échantillons dudit signal numérique, chaque pluralité desdits échantillons dudit jeu final ayant une grandeur supérieure à un seuil final déterminé par la valeur de distorsion d'échantillon correspondante des échantillons.
préparer un signal d'estimation représentatif dudit signal numérique en ayant déjà moins d'échantillons que ledit signal numérique,
préparer un indice de signal représentant, pour au moins un échantillon dudit signal d'estimation, une grandeur de signal de ladite composante de signal relative à une grandeur de bruit de ladite composante de bruit,
en se fondant sur ledit indice de signal, sélectionner des échantillons dudit signal numérique ayant une composante de signal suffisamment grande,
transmettre lesdits échantillons sélectionnés dudit signal numérique,
transmettre lesdits échantillons dudit signal d'estimation numérique, et
reconstruire ledit signal numérique à partir desdits échantillons sélectionnés transmis et des échantillons d'estimation.
engendrer un nombre aléatoire pour chaque échantillon non sélectionné dudit signal à fréquence du domaine de la parole,
préparer, à partir d'au moins une dite estimation spectrale, une estimation de bruit de la grandeur d'une composante de bruit dudit échantillon non sélectionné,
en se fondant sur ladite estimation de bruit, engendrer un facteur d'échelle, et
proportionner ledit nombre aléatoire en conformité avec ledit facteur d'échelle de façon à produire un échantillon reconstruit représentatif dudit échantillon non sélectionné.
pour une structure préalable dudit signal d'estimation, préparer une estimation initiale de bruit à partir d'au moins une estimation spectrale représentative d'une bande des fréquences de ladite structure préalable,
en se fondant sur la grandeur dudit indice de signal, sélectionner une constante de temps de montée tr et une constante de temps de chute tf,
additionner ladite constante de temps de montée à ladite estimation initiale de bruit pour former un seuil supérieur,
soustraire ladite constante de temps de chute de ladite estimation initiale de bruit pour former un seuil inférieur,
comparer une estimation spectrale courante de ladite structure courante, représentative de ladite bande de fréquences, auxdits seuils supérieur et inférieur,
si ladite estimation spectrale courante est comprise entre lesdits seuils, établir ladite estimation courante de bruit pour qu'elle soit égale à ladite estimation spectrale courante,
si ladite estimation spectrale courante est inférieure audit seuil inférieur, décrémenter ladite estimation initiale de bruit de ladite constante de temps de chute pour former ladite estimation courante de bruit, et
si ladite estimation spectrale courante est supérieure audit seuil supérieur, incrémenter ladite estimation initiale de bruit de ladite constante de temps de montée de façon à former ladite estimation courante de bruit.
sélectionner un coefficient de bruit en se fondant sur la valeur dudit indice de signal,
sélectionner un coefficient vocal en se fondant sur la valeur dudit indice de signal,
multiplier ladite estimation courante de bruit par ledit coefficient de bruit,
multiplier ladite estimation spectrale courante par ledit coefficient vocal,
additionner les produits desdites étapes de multiplication, et
former ledit facteur d'échelle à partir du résultat de ladite étape d'addition.
des moyens pour sélectionner une valeur de distorsion de quantification acceptable globale représentant une erreur de distorsion de quantification acceptable collective pour une structure des échantillons dudit signal numérique,
des moyens pour sélectionner à partir de ladite structure des échantillons, un jeu d'échantillons dans lequel chaque pluralité d'échantillons dudit jeu est supérieure à un seuil de bruit,
des moyens pour calculer, pour chaque échantillon dudit jeu, une valeur de distorsion de quantification d'échantillon représentant une erreur de distorsion de quantification acceptable pour ledit échantillon, et où la somme de toutes les valeurs de distorsion de quantification d'échantillon pour tous les échantillons dudit jeu est approximativement égale à ladite valeur de distorsion de quantification acceptable globale,
des moyens pour sélectionner, pour chaque échantillon dudit jeu une étape de taille de quantification qui fournit une erreur de distorsion de quantification approximativement égale à ladite valeur de distorsion de quantification correspondante des échantillons, et
des moyens pour quantifier ledit échantillon en utilisant ladite étape de taille de quantification.
et dans lequel lesdits moyens pour sélectionner une distorsion de quantification acceptable globale comprennent :
des moyens pour préparer un indice de signal représentant, pour au moins un échantillon de ladite structure, une grandeur de signal de ladite composante de signal relative à une grandeur de bruit de ladite composante de bruit, et
des moyens pour sélectionner ladite distorsion de quantification acceptable globale en se fondant sur ledit indice de signal.
des moyens pour sélectionner un jeu expérimental des échantillons dudit signal numérique dans lequel chaque échantillon dudit jeu expérimental est supérieur à un seuil de bruit déterminé au moins en partie par la valeur de ladite première valeur de distorsion d'échantillon, et où lesdits moyens pour calculer ladite valeur de distorsion de quantification d'échantillon comprennent :
des moyens pour ajuster ladite première valeur de distorsion d'échantillon au moyen d'une quantité déterminée par la différence entre ladite première valeur de distorsion d'échantillon et au moins un échantillon exclu dudit jeu expérimental.
des moyens pour identifier tous les échantillons bruyants dudit jeu expérimental, chaque dit échantillon bruyant ayant une grandeur inférieure à une valeur courante de ladite première valeur de distorsion d'échantillon,
des moyens pour retirer lesdits échantillons bruyants, s'il en existe, dudit jeu expérimental,
des moyens pour augmenter ladite première valeur de distorsion d'échantillon d'une quantité déterminée par la différence entre ladite première valeur de distorsion d'échantillon et au moins un dit échantillon bruyant, si des échantillons bruyants sont retirés, et
des moyens pour retirer de façon répétitive des échantillons bruyants et ajuster ladite première valeur de distorsion d'échantillon jusqu'à ce qu'une première valeur de distorsion d'échantillon ajustée soit atteinte pour laquelle aucun échantillon bruyant additionnel dudit jeu expérimental ne soit supérieur à ladite première distorsion d'échantillon ajustée.
des moyens pour estimer après terminaison dudit ajustement, un nombre binaire pour une quantité d'éléments binaires requis de façon à transmettre tous les échantillons dudit jeu expérimental,
des moyens pour comparer ledit nombre binaire estimé à un nombre binaire maximum, et
des moyens pour sélectionner un seuil de bruit final en se fondant sur ladite première valeur de distorsion d'échantillon ajustée, si ledit nombre binaire estimé est inférieur ou égal audit nombre binaire maximum.
des moyens pour préparer une seconde valeur de distorsion d'échantillon, si ledit nombre binaire estimé excède ledit nombre binaire maximum,
des moyens pour sélectionner un second jeu expérimental d'échantillons dudit signal numérique, chaque pluralité desdits échantillons dudit second jeu expérimental ayant une grandeur supérieure à ladite seconde valeur de distorsion d'échantillon,
des moyens pour estimer le nombre d'éléments binaires requis pour transmettre tous les échantillons dudit second jeu expérimental,
des moyens pour comparer ledit nombre binaire estimé audit nombre binaire maximum, et
des moyens pour augmenter ladite seconde valeur de distorsion d'échantillon d'une quantité déterminée par ladite sélection, si ledit nombre binaire estimé est supérieur audit nombre binaire maximum.
des moyens pour calculer ladite valeur de distorsion d'échantillon à partir de ladite première valeur de distorsion d'échantillon ajustée et de ladite seconde valeur de distorsion d'échantillon et,
des moyens pour sélectionner un jeu final des échantillons dudit signal numérique, chaque pluralité desdits échantillons dudit jeu final ayant une grandeur supérieure à un seuil final déterminé par ladite valeur de distorsion d'échantillon correspondante des échantillons.
des moyens pour préparer un signal d'estimation représentatif dudit signal numérique ayant déjà moins d'échantillons que ledit signal numérique,
des moyens pour préparer un indice de signal représentant, pour au moins un échantillon dudit signal d'estimation, la grandeur de ladite composante de signal relative à la grandeur de ladite composante de bruit,
des moyens pour sélectionner, en se fondant sur ledit indice de signal, des échantillons dudit signal numérique ayant une composante de signal suffisamment grande,
des moyens pour transmettre lesdits échantillons sélectionnés dudit signal numérique,
des moyens pour transmettre lesdits échantillons du signal d'estimation numérique, et
des moyens pour reconstruire ledit signal numérique à partir desdits échantillons sélectionnés transmis et des échantillons d'estimation.
des moyens pour engendrer un nombre aléatoire pour chaque échantillon non sélectionné dudit signal à fréquence du domaine de la parole,
des moyens pour préparer, à partir d'au moins une dite estimation spectrale, une estimation de bruit de la grandeur d'une composante de bruit dudit échantillon non sélectionné,
des moyens pour engendrer un facteur d'échelle en se fondant sur ladite estimation de bruit, et
des moyens pour proportionner ledit nombre aléatoire en conformité audit facteur d'échelle de façon à produire un échantillon reconstruit représentatif dudit échantillon non sélectionné.
des moyens pour préparer, pour une structure préalable dudit signal d'estimation, une estimation initiale de bruit à partir d'au moins une estimation spectrale représentative d'une bande de fréquences de ladite structure préalable,
des moyens pour sélectionner, en se fondant sur la grandeur dudit indice de signal, une constante de temps de montée tr, et une constante de temps de chute tf,
des moyens pour additionner ladite constante de temps de montée à ladite estimation initiale de bruit de manière à former un seuil supérieur,
des moyens pour soustraire ladite constante de temps de chute de ladite estimation initiale de bruit pour former un seuil inférieur,
des moyens pour comparer une estimation spectrale courante représentative de ladite bande de fréquences dans ladite structure courante, auxdits seuils supérieur et inférieur,
des moyens pour établir ladite estimation courante de bruit pour qu'elle soit égale à ladite estimation spectrale courante, si ladite estimation spectrale courante est comprise entre lesdits seuils,
des moyens pour décrémenter ladite estimation initiale de bruit de ladite constante de temps de chute pour former ladite estimation courante de bruit, si ladite estimation spectrale courante est inférieure audit seuil inférieur, et
des moyens pour incrémenter ladite estimation initiale de bruit au moyen de ladite constante de temps de montée de manière à former ladite estimation courante de bruit, si ladite estimation spectrale courante est supérieure audit seuil supérieur.
des moyens pour sélectionner un coefficient de bruit en se fondant sur la valeur dudit indice de signal,
des moyens pour sélectionner un coefficient vocal en se fondant sur la valeur dudit indice de signal,
des moyens pour multiplier ladite estimation courante de bruit par ledit coefficient de bruit,
des moyens pour mulitplier ladite estimation spectrale courante par ledit coefficient vocal,
des moyens pour additionner les produits desdites multiplications, et
des moyens pour former le facteur d'échelle à partir du résultat de ladite addition.