RU2782182C1 - Audio encoder with signal-dependent precision and number control, audio decoder and related methods and computer programs - Google Patents

Abstract

FIELD: computer technology.

SUBSTANCE: invention relates to the field of computer technology for encoding audio data. The effect is achieved by pre-processing the input audio data to obtain the audio data to be encoded; encoding the audio data to be encoded; and controlling encoding such that, depending on the first signal characteristic of the first frame of audio data to be encoded, the number of audio data elements for the audio data to be encoded for the first frame is reduced compared to the second signal characteristic of the second frame, and the first number of information units used to encode the reduced number of audio data items for the first frame is more significantly improved compared to the second number of information units for the second frame.

EFFECT: increasing the bit rate and quality of audio data playback.

35 cl, 14 dwg, 1 tbl

Description

The present invention relates to audio signal processing, and in particular to audio encoders/decoders employing signal dependent precision and number control.

Modern transform-based audio encoders apply a sequence of psycho-acoustically conditioned processing to the spectral representation of an audio segment (frame) to produce a residual spectrum. This residual spectrum is quantized and the coefficients are encoded using entropy coding.

In this process, the quantization step size, which is usually controlled via the global gain, has a direct impact on the bit consumption of the entropy encoder and must be chosen to match the bit budget, which is usually limited and often fixed. Because the bit consumption of the entropy encoder, and in particular the arithmetic encoder, is not exactly known prior to encoding, the calculation of the optimal global gain can only be performed in a closed-loop quantization and encoding iteration. However, this is impractical under certain complexity constraints because arithmetic coding entails significant computational complexity.

The prior art encoders contained in the 3GPP EVS codec therefore include a bit consumption estimator to extract a first global gain estimate, which typically operates in the power spectrum of the residual signal. Depending on the complexity constraint, it may be followed by a rate loop to refine the first estimate. Using this estimate alone, or in combination with very limited correction bandwidth, reduces complexity but also reduces accuracy, resulting in significant under- or overestimation of bit consumption.

Overestimating bit consumption results in redundant bits after the first encoding stage. Prior art encoders use them to refine the quantization of coded coefficients in a second coding stage called "residual coding". Residual coding is fundamentally different from the first coding stage because it operates at granularity per bit and therefore does not involve entropy coding. Moreover, residual coding is typically only applied at frequencies with non-zero quantized values, leaving dead zones that do not improve further.

On the other hand, underestimating the consumption of bits inevitably leads to a partial loss of the spectral coefficients, usually the highest frequencies. In prior art encoders, this effect is reduced by applying noise replacement in the decoder, which is based on the assumption that high frequency content is typically noisy.

In this configuration, it is quite obvious that it is desirable to encode as much of the signal as possible in the first coding step, which uses entropy coding and is therefore more efficient than the residual coding step. Therefore, one may want to choose a global gain with a bit estimate as close as possible to the available bit budget. Although the power spectrum estimator works well for most audio content, it can cause problems for high-pitched signals, in which the first stage estimation is mainly based on irrelevant filterbank frequency decomposition sidelobes, while important components are lost due to underestimation of bit consumption. .

The object of the present invention is to provide an improved concept for encoding or decoding audio which is nevertheless efficient and achieves good audio quality.

This task is solved by the audio encoder of claim 1, the input audio data encoding method of claim 33 and the audio decoder of claim 35, the encoded audio data decoding method of claim 41, or the computer program of claim 42.

The present invention is based on the conclusion that in order to increase efficiency, in particular in terms of bit rate on the one hand and sound quality on the other hand, a signal-dependent change is required relative to the exemplary situation, which is determined by the considered psychoacoustic factors. Conventional psychoacoustic models or considered psychoacoustic factors lead to good sound quality at low bit rates for all signal classes on average, i.e. for all frames of audio signals, regardless of their signal characteristics, when an average result is provided. However, it has been found that for certain classes of signals or for signals having certain signal characteristics, e.g. quite tonal signals, a simple psychoacoustic model or simple psychoacoustic encoder control only leads to sub-optimal results in terms of audio quality (when the bit rate is kept constant) or relative to the bit rate (when the sound quality is kept constant).

Therefore, in order to overcome this shortcoming of the conventional psychoacoustic factors considered, the present invention provides, in the context of an audio encoder with a preprocessor for preprocessing audio input data to obtain the audio data to be encoded and an encoder processor to encode the audio data to be encoded, a controller for controlling the encoder processor such in such a way that, depending on a certain characteristic of the frame signals, the number of audio data elements for audio data to be encoded by the encoder processor is reduced compared to the usual simple results obtained by the considered psychoacoustic factors of the prior art. In addition, this reduction in the number of audio data elements is performed in a signal-dependent manner such that for a frame with a certain first signal characteristic, the number is more significantly reduced than for another frame with a different signal characteristic that is different from the signal characteristic of the first frame. This reduction in the number of audio data elements may be considered an absolute number reduction or a relative number reduction, although this is not decisive. However, there is a feature that the information units that are "saved" by deliberately reducing the number of audio data items are simply not lost, but are used to more accurately encode the remaining number of data items, i.e. data items that are not excluded by intentionally reducing the number of audio data items.

According to the invention, the controller for controlling the encoder processor operates in such a way that, depending on the first characteristic of the signals of the first frame of audio data to be encoded, the number of audio data elements for the audio data to be encoded by the encoder processor for the first frame is reduced compared to the second characteristic of the signals of the second frame, and at the same time, the first number of information units used to encode the reduced number of audio data elements for the first frame is more significantly improved compared to the second number of information units for the second frame.

In the preferred embodiment, the reduction is performed in such a way that for more tone frames of the signal, a more significant reduction is performed, and at the same time the number of bits for individual lines is more significantly improved compared to a frame that is less tonal, i.e. which is noisier. Here, the number does not decrease to such a high degree, and accordingly, the number of information units used to encode less pitched audio data items does not increase as much.

The present invention provides an infrastructure in which, in a signal-dependent manner, the commonly envisaged psychoacoustic factors in question are violated to some extent. However, on the other hand, this violation is not treated in the same way as in normal encoders, in which the violation of the considered psychoacoustic factors, for example, is carried out in an emergency situation, for example, in a situation where, in order to maintain the required bit rate, higher frequency parts are given zero. Instead, in accordance with the present invention, such disruption of the normal psychoacoustic factors in question occurs regardless of the emergency, and "saved" information units are used to further refine the "surviving" audio data elements.

In preferred embodiments, a two-stage encoder processor is used which has as its initial encoding stage, for example, an entropy encoder such as an arithmetic encoder or a variable length encoder such as a Huffman encoder. The second coding stage serves as the granularity stage, and this second encoder is usually implemented in the preferred embodiments as a residual encoder or bit encoder operating in bit granularity, which for example can be implemented by adding a certain predetermined offset in the case of the first value information unit or offset subtraction in the case of the opposite value of the information unit. In an embodiment, this detail encoder is preferably implemented as a residual encoder adding an offset in case of a first bit value and subtracting an offset in case of a second bit value. In the preferred embodiment, reducing the number of audio data elements results in a situation where the distribution of available bits in a conventional fixed frame rate scenario is changed such that the start coding stage accepts a lower bit budget than the detail coding stage. By now, the paradigm is that the seed coding stage is to receive a bit budget that is as high as possible regardless of the characteristics of the signals, since the seed coding stage, such as the arithmetic coding stage, is considered to have the highest efficiency and therefore encodes much better than the residual coding cascade from an entropy point of view. However, in accordance with the present invention, this paradigm is removed because it is found that for certain signals, such as, for example, signals with higher pitch, the efficiency of an entropy encoder, such as an arithmetic encoder, is not as high as the efficiency obtained by a subsequent connected residual encoder such as a bit encoder. However, while it should be recognized that the entropy coding stage is highly efficient for audio signals on average, the present invention now solves this problem by eliminating the average from consideration, but by reducing the bit budget for the initial coding stage in a signal dependent and preferably , for tonal parts of the signal.

In a preferred embodiment, shifting the bit budget from the initial coding stage to the detail coding stage based on the characteristics of the input data signals is performed such that at least two detail information units are available for at least one and preferably 50% and even more preferably all elements. audio data that remains valid after the number of data items has been reduced. In addition, it has been found that a very efficient procedure for calculating these detail information units at the encoder side and applying these detail information units at the decoder side is an iterative procedure in which, in a certain order, for example, from low frequency to high frequency, the remaining bits from the bit budget for the detail coding stage are consumed one by one. Depending on the number of audio data elements still valid, and depending on the number of information units for the detail coding stage, the number of iterations can be significantly greater than two, and it has been found that for high tone signal frames, the number of iterations can be four, five, or even higher.

In the preferred embodiment, the determination of the control value by the controller is performed in an indirect manner, i.e. without explicit definition of signal characteristics. To this end, the control value is computed on the basis of the manipulated input data, where the manipulated input data is, for example, the input data to be quantized or amplitude-related data extracted from the data to be quantized. Although the control value for the encoder processor is determined based on the keyed data, the actual quantization/encoding is performed without this keying. Thus, a signal-dependent procedure is obtained by determining a keying value for signal-dependent manipulation, the manipulation having some effect on the resulting reduction in the number of audio data elements without explicit knowledge of the specific characteristics of the signals.

In another implementation, a direct mode may be used, in which a certain characteristic of the signals is directly evaluated and depends on the result of this signal analysis, a certain reduction in the number of data elements is performed in order to obtain higher accuracy for the remaining data elements.

In a further implementation, a split procedure may be applied for purposes of reducing audio data elements. In the split procedure, a certain number of data items are already obtained by quantization controlled by the usually psychoacoustically conditioned control of the quantizer, and based on the input audio signal, the quantized audio data items are reduced in relation to their number, and preferably this reduction is performed by eliminating the smallest audio data items in relation to their amplitude , their energies or their powers. Control for reduction, again, can be obtained through direct/explicit signal characterization, or through indirect or implicit signal control.

In a further preferred embodiment, an integrated procedure is applied in which the variable quantizer is controlled to perform one quantization, but based on manipulated data, while non-manipulated data is quantized at the same time. The quantizer control value, for example, the global gain is computed using the signal-dependent keyed data, while the data without this keying is quantized and the result of the quantization is encoded using all available information units such that in the case of two-stage coding, usually a large amount of information units for the detail coding stage remains.

Embodiments provide a solution to the quality loss problem for high-pitched content that is based on a modification of the power spectrum that is used to estimate the bit consumption of the entropy encoder. This modification exists for the signal-adaptive noise floor adder, which keeps the estimate for overall flat-residual audio content unchanged in practice, while it increases the bit budget estimate for high-pitched content. The effect of this modification is twofold. First, it causes filterbank noise and irrelevant sidelobes of harmonic components, which are superimposed by a noise floor that can be quantized to zero. Second, it shifts the bits from the first coding stage to the residual coding stage. While this shift is not required for most signals, it is fully effective for high-pitched signals because the bits are used to improve the quantization accuracy of the harmonic components. This means that they are used to encode bits with low significance, which usually adhere to a uniform distribution and are therefore completely efficiently encoded using a binary representation. In addition, the procedure is computationally inexpensive, making it a very efficient tool for solving the above problem.

The following describes the preferred embodiments of the present invention with reference to the accompanying drawings, in which:

Fig. 1 is an embodiment of an audio encoder;

Fig. 2 illustrates a preferred implementation of the encoder processor of FIG. one;

Fig. 3 illustrates a preferred implementation of the detail coding stage;

Fig. 4a illustrates an exemplary frame syntax for the first or second frame with iterative detail bits;

Fig. 4b illustrates a preferred implementation of the audio chip reduction module as a variable quantizer;

Fig. 5 illustrates a preferred implementation of an audio encoder with a spectral preprocessor;

Fig. 6 illustrates a preferred embodiment of an audio decoder with a temporary post processor;

Fig. 7 illustrates an implementation of the encoder processor for the audio decoder of FIG. 6;

Fig. 8 illustrates a preferred implementation of the detail decoding stage of FIG. 7;

Fig. 9 illustrates an indirect mode implementation for calculating control values;

Fig. 10 illustrates a preferred implementation of the keying value calculation module of FIG. 9;

Fig. 11 illustrates the calculation of control values in direct mode;

Fig. 12 illustrates an implementation of split reduction in the number of audio data elements; and

Fig. 13 illustrates an implementation of integrated audio element reduction.

Fig. 1 illustrates an audio encoder for encoding input audio data 11. The audio encoder comprises a preprocessor 10, an encoder processor 15, and a controller 20. The preprocessor 10 preprocesses the input audio data 11 to obtain audio data per frame or audio data to be encoded illustrated in element 12. Audio data that to be encoded are input to the encoder processor 15 to encode the audio data to be encoded, and the encoder processor outputs the encoded audio data. The controller 20 is connected, relative to its input, to the audio data per preprocessor frame, but, alternatively, the controller can also be connected to be able to receive audio input without preprocessing. The controller is configured to decrease the number of audio data elements per frame depending on the signal in the frame, and at the same time the controller increases the number of information units or preferably bits for the reduced number of audio data elements depending on the signal in the frame. The controller is configured to control the encoder processor 15 such that, depending on the first signal characteristic of the first frame of audio data to be encoded, the number of audio data elements for the audio data to be encoded by the encoder processor for the first frame is reduced compared to the second signal characteristic of the second frame. frame, and the number of information units used to encode the reduced number of audio data items for the first frame is more significantly improved compared to the second number of information units for the second frame.

Fig. 2 illustrates a preferred encoder processor implementation. The encoder processor comprises an initial encoding stage 151 and a detailed encoding stage 152 . In an implementation, the initial encoding stage comprises an entropy encoder such as an arithmetic encoder or a Huffman encoder. In another embodiment, the granularity encoding stage 152 comprises a bit encoder or a residual encoder operating at a granularity per bit or information unit. In addition, the functionality for reducing the number of audio data items is implemented in FIG. 2 by means of an audio chip reduction module 150, which, for example, may be implemented as a variable quantizer in the integrated reduction mode illustrated in FIG. 13, or alternatively, as a separate element already operating on quantized audio data elements, as illustrated in the split reduction mode 902, and in a further non-illustrated embodiment, the audio element reduction module can also operate on unquantized elements by setting such unquantized elements to zero. elements or by weighting so as to exclude data elements with a certain weight number such that such audio data elements are quantized to zero and thereby excluded in a subsequent coupled quantizer. The audio element reduction unit 150 of FIG. 2 may operate on unquantized or quantized data elements in a split reduction procedure, or may be implemented by a variable quantizer specifically controlled by a signal-dependent control value, as illustrated in the integrated reduction mode of FIG. 13.

The controller 20 of FIG. 1 is configured to reduce the number of audio data elements encoded by the initial encoding stage 151 for the first frame, and the initial encoding stage 151 is configured to encode the reduced number of audio data elements for the first frame using the initial number of information units of the first frame, and the calculated bits/ones from the initial number of information units are output by block 151, as illustrated in FIG. 2, item 151.

In addition, the detail coding stage 152 is configured to use the remaining number of first frame information units for detail coding for the reduced number of audio data items for the first frame, and the initial number of first frame information units added to the remaining number of first frame information units results in a predetermined number information units for the first frame. Specifically, the detail coding stage 152 outputs the remaining number of bits of the first frame and the remaining number of bits of the second frame, and there are at least two detail bits for at least one, or preferably at least 50%, or even more preferably all non-zero audio data elements. , i.e. for audio data elements that remain in effect after reduction of audio data elements and that are initially encoded by the initial coding stage 151 .

Preferably, the predetermined number of information units for the first frame is equal to the predetermined number of information units for the second frame, or is close enough to the predetermined number of information units for the second frame such that constant or substantially constant bit rate operation is obtained for the audio encoder.

As illustrated in FIG. 2, the audio item reduction module 150 reduces the number of audio data items beyond the psychoacoustically determined number in a signal-dependent manner. Thus, for the first signal characteristic, the number decreases slightly by the psychoacoustically determined number, and in a frame with the second signal characteristic, for example, the number improves significantly beyond the psychoacoustically determined number. Also, preferably, the audio element reduction module excludes the lowest amplitude/power/energy elements, and this operation is preferably performed via indirect selection obtained in integrated mode, whereby the reduction in the number of audio data elements is carried out by quantizing certain audio data elements to zero. In an embodiment, the start encoding stage encodes only the audio data elements that are not quantized to zero, and the detail encoding stage 152 details only the audio data elements already processed by the start encoding stage, i.e. audio data elements that are not quantized to zero by the audio element reduction unit 150 of FIG. 2.

In a preferred embodiment, the detail coding stage is configured to iteratively assign the remaining number of information units of the first frame to the reduced number of audio data elements of the first frame in at least two successive iterations. In particular, assigned information unit values for at least two successive iterations are computed, and the calculated information unit values for at least two successive iterations are input into the encoded output frame in a predetermined order. In particular, the detail coding stage is configured to sequentially assign an information unit for each audio data item from the reduced number of audio data items for the first frame, in order from low frequency information for the audio data item to high frequency information for the audio data item in the first iteration. In particular, the audio data elements may be individual spectral values obtained by spectral temporal transform. Alternatively, the audio data elements may be tuples of two or more spectral lines, typically adjacent to each other in the spectrum. Bit values are calculated from a certain start value with low frequency information to a certain end value with most high frequency information, and in an additional iteration, the same procedure is performed, i.e. again processing from low spectral information values/tuples to high spectral information values/tuples. Specifically, the detail encoding stage 152 is configured to check whether the number of information units already assigned is less than the predetermined number of information units for the first frame is less than the initial number of information units of the first frame, and the detail encoding stage is also configured to terminate the second iterations in case of a negative test result, or execution, in case of a positive test result, a certain number of additional iterations until a negative test result is obtained, and the number of additional iterations is 1, 2, ...,. Preferably, the maximum number of iterations is limited to a two-digit number, for example, values from 10 to 30, and preferably 20 iterations. In an alternative embodiment, the check for the maximum number of iterations can be eliminated if non-zero spectral lines are counted first and the number of residual bits is adjusted accordingly for each iteration or for the entire procedure. Therefore, when there are, for example, 20 remaining spectral tuples and 50 residual bits, it is possible, without checking during a procedure in the encoder or decoder, to determine that the number of iterations is three, and in the third iteration, the detail bit should be calculated or be available in the bitstream for the first ten spectral lines/tuples. Thus, this alternative does not need to be checked during iterative processing, since the information regarding the number of non-zero or remaining audio elements is known after the processing of the initial stage in the encoder or decoder.

Fig. 3 illustrates a preferred implementation of the iterative procedure performed by the detail coding stage 152 of FIG. 2, which is made possible by the fact that, unlike other procedures, the number of detail bits per frame is significantly increased for certain frames due to the corresponding reduction in audio data elements for such certain frames.

At 300, the remaining audio data elements are determined. This determination may be automatically performed by managing the audio data elements that have already been processed by the initial encoding stage 151 of FIG. 2. In step 302, the procedure is started at a predetermined audio data element, such as the audio data element with the lowest spectral information. At step 304, bit values for each element of audio data in a predetermined sequence are calculated, and this predetermined sequence, for example, is a sequence from low spectral values/tuples to high spectral values/tuples. The calculation at step 304 is performed using the start offset 305 and according to control 314 such that the detail bits are still available. At element 316, the first iterative detail information items are output, i. e. a bit pattern indicating one bit for each remaining audio data element, where the bit indicates whether the offset, i.e. the start offset 305 to be added or subtracted, or, alternatively, whether the start offset should be added or not added.

At step 306, the offset is reduced using the given rule. This predetermined rule could, for example, be that the offset is divided by two, i.e. that the new offset is half of the original offset. However, other bias reduction rules may also apply that differ from the 0.5 weighting.

At step 308, the bit values for each element in the predetermined sequence are calculated again, but now in the second iteration. As input to the second iteration, the detailed elements after the first iteration illustrated at 307 are input. Thus, for the calculation at step 314, the granularity represented by the first iterative detail information items is already applied, and under such a necessary premise that the detail bits are still available, as indicated at step 314, the second iterative detail information items are computed and output. at 318.

At step 310, the offset is again reduced using the predetermined rule to be ready for the third iteration, and the third iteration is again based on the detail elements after the second iteration illustrated at 309, and again under the necessary premise that the detail bits are still available as indicated. at 314, third iterative detail information items are computed and output at 320.

Fig. 4a illustrates an exemplary frame syntax with information units or bits for a first frame or a second frame. The bit data portion for a frame consists of an initial number of bits, i. e. element 400. In addition, first iterative detail bits 316, second iterative detail bits 318, and third iterative detail bits 320 are also included in the frame. Specifically, according to the frame syntax, the decoder is in a state of identifying which bits of the frame are the seed number of bits, which bits are the first, second, or third iterative detail bits 316, 318, 320, and which bits in frame are any other bits 402, such ancillary information which, for example, may also include an encoded representation of the global gain (gg), for example, which, for example, can be calculated by the controller 200 directly or which, for example, can be affected by the controller through the output information of the controller 21. In sections 316, 318, 320, a certain sequence of individual information units is given. This sequence is preferably such that the bits in the bit sequence are applied to the initially decoded audio data elements to be decoded. Since it is impractical, with respect to bit rate requirements, to explicitly signal anything about the first, second, and third iterative detail bits, the order of the individual bits in blocks 316, 318, 320 must be the same as the corresponding order of the remaining bits. audio data elements. With this in mind, it is preferable to use the same iterative procedure on the encoder side as illustrated in FIG. 3 and on the decoder side, as illustrated in FIG. 8. It is not necessary to signal a specific bit allocation or bit association, at least in blocks 316-320.

In addition, the numbers for the initial number of bits on the one hand and the remaining number of bits on the other hand are only exemplary. Typically, the initial number of bits that typically encode the high-bit part of an audio data element, such as spectral values or tuples of spectral values, is greater than the iterative detail bits that represent the low-order part of the "preserved" audio data elements. In addition, the initial number of 400 bits is usually determined by an entropy encoder or an arithmetic encoder, but the iterative detail bits are determined using a residual or bit encoder operating at a granularity per information unit. Although the detail coding stage does not perform entropy coding and the like, the encoding of the low bit part of the audio data items is nevertheless more efficiently performed by the detail coding stage, since it can be assumed that the low bit part of the audio data items, such as spectral values, is the same. is distributed, and because of this, entropy coding with a variable length code or an arithmetic code, together with a certain context, does not introduce an additional advantage, but instead even introduces an additional amount of overhead.

In other words, for the least significant bit part of the audio data elements, using an arithmetic encoder should be less efficient than using a bit encoder, since a bit encoder does not require a bit rate at all for a particular context. Intentional reduction in the number of audio data items called by the controller not only improves the accuracy of the dominant spectral lines or line tuples, but also provides a highly efficient encoding operation for the purpose of detailing the MSB portions of these audio data items represented by an arithmetic or variable length code.

With this in mind, several and for example the following advantages are obtained by implementing the encoder processor 15 of FIG. 1 as illustrated in FIG. 2 with the start encoding stage 151 on the one hand and the detail encoding stage 152 on the other hand.

An efficient two-stage coding scheme is proposed, containing the first entropy coding stage and the second residual coding stage based on one-bit (non-entropy) coding.

The scheme uses a low complexity global gain estimator that includes an energy-based bit consumption estimator for the first coding stage, comprising a signal-adaptive noise floor adder.

The noise floor combiner efficiently transfers bits from the first coding stage to the second coding stage for high tone signals while leaving the estimate for other types of signals unchanged. This bit shifting from the entropy coding stage to the non-entropy coding stage is fully effective for high pitched signals.

Fig. 4b illustrates a preferred implementation of a variable quantizer, which, for example, may be implemented to perform audio chip reduction in a controlled manner, preferably in the integrated reduction mode illustrated with respect to FIG. 13. To this end, the variable quantizer comprises a weighting module 155 which receives the (non-manipulated) audio data to be encoded illustrated in line 12. This data is also input to the controller 20 and the controller is configured to calculate the global gain 21 but based on non-manipulated data input to the weighing module 155 and using signal dependent keying. The global gain 21 is applied in the weighting module 155, and the output of the weighting module is input to the quantizer core 157, which is based on a fixed quantization step size. Variable quantizer 150 is implemented as a controllable weighting module in which control is performed using a global gain 21 (gg) and a subsequent coupled quantizer core 157 with a fixed quantization step size. However, other implementations may also be performed, such as a quantizer core having a variable quantization step size that is controlled by the output value of the controller 20.

Fig. 5 illustrates a preferred implementation of the audio encoder, and in particular a specific implementation of the preprocessor 10 of FIG. 1. Preferably, the preprocessor comprises a weighted coding module 13 which generates, from the input audio data 11, a frame of weighted time domain audio data using a certain weighted analytic coding function, which may, for example, be a cosine weighted coding function. A frame of time domain audio data is input to spectral transform 14, which may be implemented to perform a modified discrete cosine transform (MDCT) or any other transform such as FFT or MDST, or any other time spectral transform. Preferably, the weighted coding module operates with a certain feedforward control such that overlapping frame generation is performed. In the case of 50% overlap, the leading value of the weighting coding unit is half the size of the analytic weighting coding function applied by the weighting coding unit 13 . The (non-quantized) frame of spectral values output by the spectral converter is input to the spectral processor 15, which is implemented to perform some spectral processing, such as performing a temporal noise shaping operation, a spectral noise shaping operation, or any other operation such as a spectral whitening operation, by which the modified spectral values generated by the spectral processor have a spectral envelope that is flatter than the spectral envelope of the spectral values before being processed by the spectral processor 15. The audio data to be encoded (per frame) is forwarded via line 12 to the encoder processor 15 and to the controller 20, while the controller 20 transmits control information via line 21 to the processor 15 of the encoder. The encoder processor outputs its data to a bitstream recorder 30 implemented as a bitstream multiplexer, for example, and the encoded frames are output on line 35.

Regarding processing at the decoder side, refer to FIG. 6. The bitstream output by the block 30, for example, can be directly input to the bitstream reader 40 after some storage or transmission. Naturally, any other processing may be performed between the encoder and decoder, such as transmission processing according to a wireless transmission protocol such as the DECT protocol or the Bluetooth protocol, or any other wireless transmission protocol. Data input to the audio decoder shown in FIG. 6 are input to the bit stream reading unit 40 . The bitstream reader 40 reads the data and forwards the data to the encoder processor 50, which is controlled by the controller 60. Specifically, the bitstream reader receives encoded data, the encoded audio data containing, for a frame, the initial number of frame information units and the remaining number of information units. frame. The encoder processor 50 processes the encoded audio data, and the encoder processor 50 comprises an initial decoding stage and a detailed decoding stage, as illustrated in FIG. 7 in the element 51 for the initial decoding stage and in the element 52 for the detailed decoding stage, which are controlled by the controller 60. The controller 60 is configured to control the detailed decoding stage 52 so as to use the initial decoded data elements output by the initial stage 51 in detailing. decoding according to FIG. 7, at least two information units out of the remaining number of information units for detailing the same initially decoded data element. In addition, the controller 60 is configured to control the encoder processor such that the initial decoding stage uses the initial number of frame information units to obtain the initial decoded data elements on the line connecting block 51 and 52 in FIG. 7, wherein, preferably, the controller 60 receives an indicator of the initial number of frame information units on the one hand and the initial remaining number of frame information units from the bitstream reader 40 as indicated by an input line to block 60 of FIG. 6 or FIG. 7. The post processor 70 processes the detailed audio data elements to obtain the decoded audio data 80 in the output of the post processor 70.

In a preferred implementation for an audio decoder that corresponds to the audio encoder of FIG. 5, the post processor 70 includes, as an input stage, a spectrum processor 71 that performs a temporal noise de-shaping operation, or a spectral noise de-shaping operation, or an inverse spectral whitening operation, or any other operation that reduces some of the processing applied by the spectral processor 15. according to fig. 5. The output of the spectral processor is input to a time transform 72 which is operable to perform a spectral-to-time domain transform, and preferably the time transform 72 is the same as the spectral transform 14 of FIG. 5. The output of the time converter 72 is input to the overlap adder stage 73, which performs an overlap adder operation on a certain number of overlapping frames, such as at least two overlapping frames, to obtain decoded audio data 80. Preferably, the overlap adder stage 73 applies a synthesis function. weighting coding to the output of the time converter 72, this synthesis weighting coding function being the same as the analytic weighting coding function applied by the analytic weighting coding module 13 . In addition, the overlap operation performed by the block 73 is the same as the forward block operation performed by the weighted coding unit 13 of FIG. 5.

As illustrated in FIG. 4a, the remaining number of frame information units contains calculated values of information units 316, 318, 320 for at least two consecutive iterations in a given order, with even three iterations illustrated in Embodiment 4a. In addition, the controller 60 is configured to control the detail decoding stage 52 so as to use for the first iteration the calculated values, for example, from block 316 for the first iteration in accordance with a predetermined order, and to use for the second iteration the calculated values from block 318 for the second iteration. in the given order.

Next, a preferred implementation of the detail decoding stage under the control of controller 60 is illustrated with respect to FIG. 8. In step 800, the controller or detail decoding stage 52 of FIG. 7 defines the audio data elements to be refined. These audio data elements are typically all audio data elements that are output by the block 51 of FIG. 7. As indicated in step 802, a start at a predetermined audio data element, such as the lowest spectral information, is performed. Using start offset 805, the first iterative detail information items received from the bit stream or from controller 16, such as the data in block 316 of FIG. 4a are applied 804 to each element in a predetermined sequence, with the predetermined sequence stretching from low to high spectral value/spectral tuple/spectral information. The results are the detailed audio data items after the first iteration, as illustrated by line 807. In step 808, the bit values for each item in a predetermined sequence are applied, with the bit values coming from the second iterated detail information units, as illustrated at 818, and these bits are received from the bitstream reader or controller 60, depending on the particular implementation. The result of step 808 is the detailed items after the second iteration. On the other hand, at step 810, the offset is reduced in accordance with the predetermined offset reduction rule already applied in block 806. With the reduced offset, the bit values for each element in the predetermined sequence are applied as illustrated at 812 using the received third iteratives. detail information items, such as from the bit stream or controller 60. The third iterative detail information items are written to the bit stream at element 320 of FIG. 4a. The result of the procedure at block 812 is the detailed items after the third iteration, as indicated at 821.

This procedure continues until all of the iterative detail bits included in the bit stream for the frame have been processed. This is checked by controller 60 via control line 814, which controls the remaining availability of detail bits, preferably for each iteration, but at least for the second and third iterations processed in blocks 808, 812. At each iteration, controller 60 controls the detail decoding stage in this way to check if the number of information units already read is less than the number of information units in the remaining frame information units for the frame, to terminate the second iteration in case of a negative test result, or, in the case of a positive test result, perform a certain number of additional iterations until until a negative test result is obtained. The number of additional iterations is at least one. Due to the application of similar procedures at the encoder side, explained in the context of FIG. 3 and on the decoder side as shown in FIG. 8, specific signaling is not required at all. Instead, multiple iterative detail processing is performed in a highly efficient manner without a particular amount of overhead. In an alternative embodiment, the check for the maximum number of iterations can be eliminated if non-zero spectral lines are counted first and the number of residual bits is adjusted accordingly for each iteration.

In a preferred implementation, the detail decoding stage 52 is configured to add an offset to the initially decoded data item when the read information data unit of the remaining number of frame information units has a first value, and subtract the offset from the initially decoded item when the read information data unit of the remaining number frame information units has a second meaning. This offset, for the first iteration, is the detail elements of FIG. 8. In the second iteration, as illustrated at 808 in FIG. 8, the reduced offset generated by block 806 is used to add the reduced or second offset to the result of the first iteration when the read information data unit of the remaining number of frame information units has a first value, and to subtract the second offset from the result of the first iteration when the read information unit the data unit of the remaining number of information units of the frame has the second value. Typically, the second offset is lower than the first offset, and preferably the second offset is 0.4-0.6 times the first offset and most preferably 0.5 times the first offset.

In a preferred implementation of the present invention using the indirect mode illustrated in FIG. 9, explicit signal characterization is not required. Instead, the keying value is calculated preferably using the embodiment illustrated in FIG. 9. For the indirect mode, the controller 20 is implemented as shown in FIG. 9. Specifically, the controller includes a control preprocessor 22, a keying value calculation module 23, a combining module 24, and a global gain calculation module 25 that ultimately calculates the global gain for the audio chip reduction module 150 of FIG. 2, which is implemented as the variable quantizer illustrated in FIG. 4b. In particular, the controller 20 is configured to parse the audio data of the first frame to determine the first control value for the variable quantizer for the first frame, and parse the audio data of the second frame to determine the second control value for the variable quantizer for the second frame, the second control value being different from the first control value. . Analysis of the frame audio data is performed by the keying value calculation unit 23 . The controller 20 is configured to perform manipulation of the audio data of the first frame. In this operation, the control preprocessor 20 illustrated in FIG. 9 is not present, and therefore the bypass line for block 22 is active.

However, when the manipulation is not performed on the audio data of the first frame or the second frame, but is applied to the amplitude related values extracted from the audio data of the first frame or the second frame, the control preprocessor 22 is present and the bypass line does not exist. The actual keying is performed by a combining unit 24 which combines the keying value output from the block 23 with amplitude related values extracted from the audio data of a certain frame. In the output of combining unit 24, there is manipulated (preferably energy) data, and based on this manipulated data, global gain calculation unit 25 calculates a global gain or at least a control value for the global gain indicated at 404. Global gain calculation unit 25 must input restrictions on the allowed bit budget for the spectrum such that a certain data rate or a certain number of information units are allowed per frame.

In the direct mode illustrated in FIG. 11, the controller 20 includes an analyzer 201 for characterizing signals per frame, and the analyzer 208 outputs, for example, signal quantity information such as tone information, and controls the control value calculation unit 202 using this preferably quantitative data. One procedure for calculating the tone of a frame is to calculate the Spectral Smoothness Score (SFM) of the frame. Any other tone determination procedures or any other signal characterization procedures may be performed by block 201, and translation from the determined signal characteristic value to the determined control value must be performed to obtain the intended reduction in the number of audio data elements for the frame. The output of the direct mode control value calculation module 202 of FIG. 11 may be a control value to an encoder processor, such as a variable quantizer, or alternatively an initial encoding stage. When the control value is passed to the variable quantizer, the integrated reduction mode is performed, while when the control value is provided for the initial coding stage, the divided reduction is performed. Another implementation of split reduction is to remove or affect specifically selected non-quantized audio data items present prior to actual quantization such that, by means of a particular quantizer, such affected audio data items are quantized to zero and therefore must be excluded for the purposes of entropy coding and subsequent detail coding. .

Although the indirect mode of FIG. 9 is shown together with integrated reduction, i.e. that the global gain calculator 25 is configured to calculate a variable global gain, the manipulated data output by the combiner 24 can also be used to directly control the initial coding stage to remove any specific quantized audio data units, such as the smallest quantized data units, or, in alternatively, the control value can also be sent to the non-illustrated audio data, influencing the cascade, which affects the audio data before the actual quantization using a variable quantization control value, which is determined without data manipulation and therefore usually obeys psychoacoustic rules, which, however, however, are intentionally violated by the procedures of the present invention.

As illustrated in FIG. 11 for direct mode, the controller is configured to determine the first tone characteristic as the first signal characteristic and determine the second tone characteristic as the second signal characteristic such that the bit budget for the detail coding stage is increased in the case of the first tone characteristic compared to the bit budget for of the detail coding stage in the case of the second tone characteristic, the first key characteristic indicating a greater key than the second key characteristic.

The present invention does not result in more approximate quantization, which is usually obtained by applying a larger global gain. Instead, this calculation of the global gain based on the signal-dependent manipulated data results only in a shift of the bit budget from the initial coding stage, which accepts a lower bit budget, to the fine-grained decoding stage, which accepts a higher bit budget, but this bit budget shift is performed by the dependent away from the signal in a way and amounts to more for the higher pitched portion of the signal.

Preferably, the control preprocessor 22 of FIG. 9 calculates amplitude related values as a set of power values extracted from one or more audio data values. In particular, it is these power values that are manipulated using the addition of the same keying value by the combining module 24, and this same keying value, which is determined by the keying value calculation module 23, is combined with all the power values of the plurality of power values for the frame.

Alternatively, as indicated by a bypass line, values obtained by the same absolute value of the keying value calculated by block 23, but preferably with randomized signs, and/or values obtained by subtracting slightly different terms from the same absolute value (but preferably with randomized signs), or a complex keying value, or, more generally, values obtained as samples from a certain normalized probability distribution scaled using a calculated complex or real absolute value of the keying value, are added to all audio values from the set of audio values included in the frame. The procedure performed by the control preprocessor 22, such as power spectrum calculation and subsampling, may be included in the global gain calculation module 25 . Therefore, preferably, the noise floor is added either directly to the spectral audio values or, alternatively, to the amplitude-related values extracted from the audio data per frame, i. e. with the output of the control preprocessor 22. Preferably, the controller preprocessor calculates a downsampled power spectrum that corresponds to using exponentiation with an exponent value of 2. However, alternatively, another exponent value greater than 1 may be used. As an example, the exponent value 3 should represent loudness, not degree. But other exponent values can also be used, such as smaller or larger exponent values.

In the preferred implementation illustrated in FIG. 10, the keying value calculation unit 23 includes a search unit 26 for performing a search for the maximum spectral value in a frame and at least one of the calculation of the signal-independent fraction indicated by the element 27 of FIG. 10, or a calculation module for calculating one or more moments per frame, as illustrated by block 28 of FIG. 10. As such, block 26 or block 28 serves here to provide a signal-dependent influence on the keying value for a frame. In particular, the search module 26 is configured to search for the maximum value of the plurality of audio data items or amplitude related values, or to search for the maximum value of the plurality of subsampled audio data or the plurality of subsampled amplitude associated values for a corresponding frame. The actual calculation is performed by block 29 using the output of blocks 26, 27 and 28, blocks 26, 28 actually representing signal analysis.

Preferably, the signal-independent fraction is determined by the bit rate for the actual encoder session, the frame duration, or the sampling rate for the actual encoder session. In addition, the calculation unit 28 for calculating one or more moments per frame is configured to calculate a signal-dependent weight value derived from at least the first sum of absolute values of audio data or subsampled audio data in a frame, the second sum of absolute values of audio data or subsampled audio data in a frame multiplied by the index associated with each absolute value and the quotient of the second sum and the first sum.

In the preferred implementation performed by the global gain calculation module 25 of FIG. 9, the required bit estimate is calculated for each energy value depending on the energy value and a possible value for the actual control value. The required bit scores for the energy values and the candidate value for the control value are accumulated, and it is checked whether the accumulated bit score for the candidate value for the control value meets the allowed bit consumption criterion, for example, as illustrated in FIG. 9 as a bit budget for the spectrum input to the global gain calculation unit 25 . If the allowed bit consumption criterion is not met, the candidate value for the control value is modified, and the calculation of the required bit estimate, the accumulation of the required bit rate, and the check for the allowed bit consumption criterion for the modified candidate value for the control value is repeated. Once such an optimal control value is found, that value is output on line 404 of FIG. 9.

The following illustrates preferred embodiments.

Detailed description of the encoder (for example, Fig. 5)

Designation

Let f be the base sampling rate in Hz, N _ms the base frame duration in milliseconds, and br the base bit rate in bits per second.

Residual spectrum extraction (e.g. preprocessor 10)

, который обычно извлекается посредством временно-частотного преобразования, такого как MDCT, с дальнейшими психоакустически обусловленными модификациями, такими как формирование временного шума (TNS) для удаления временной структуры, и формирование спектрального шума (SNS) для удаления спектральной структуры. Для аудиосодержимого с медленно варьирующейся спектральной огибающей, огибающая остаточного спектра X_f(k) в силу этого является плоской.Embodiment operates with actual residual spectrum

, which is typically extracted by a time-frequency transform such as MDCT, with further psychoacoustically driven modifications such as temporal noise shaping (TNS) to remove temporal structure, and spectral noise shaping (SNS) to remove spectral structure. For audio content with a slowly varying spectral envelope, the residual spectrum envelope X _f (k) is therefore flat.

Global Gain Estimation (eg, Fig. 9)

Spectrum quantization is controlled by the global gain g _glob via the following:

The initial global gain estimate (element 22 of FIG. 9) is extracted from the power spectrum X _f (k) ² after downsampling by a factor of 4:

and a signal-adaptive noise floor N(X _f ), which is given as follows:

(например, элемент 23 по фиг. 9)

(for example, element 23 in Fig. 9)

The regBits parameter depends on the bit rate, frame duration and sample rate and is calculated as follows:

(например, элемент 27 по фиг. 10)

(for example, element 27 in Fig. 10)

where C(Nm _s , f _s ) is as specified in the table below.

N _ms \f _s 48000 96000 2.5 -6 -6 5 0 0 ten 2 5

The lowBits parameter depends on the center of mass of the absolute values of the residual spectrum and is calculated as follows:

(например, элемент 28, фиг. 10)

(for example, element 28, Fig. 10)

where:

and:

are moments of the absolute spectrum.

The global gain is estimated in the form:

from values:

(например, вывода модуля 24 комбинирования по фиг. 9),

(for example, output module 24 combination of Fig. 9),

where gg _off is a bit rate and sample rate dependent offset.

для каждой спектральной линии, перед вычислением спектра мощности.It should be noted that adding the noise floor term N(X _f ) to PX _lp (k) gives the expected result of adding the corresponding noise floor to the residual spectrum X _f (k), such as a randomized addition or subtraction of the term

for each spectral line, before calculating the power spectrum.

Pure estimates based on the power spectrum may already be found, for example, in the 3GPP EVS codec (3GPP TS 26.445, section 5.3.3.2.8.1). In embodiments, the addition of a noise floor N(X _f ) is performed. The noise floor is signal-adaptive in two ways.

First, it scales with the maximum amplitude X _f . Therefore, the effect on energy of a uniform spectrum in which all amplitudes are close to the maximum amplitude is very small. But for high-pitched signals, in which the spectrum and, in extension, also the residual spectrum show a certain number of strong peaks, the total energy increases significantly, which increases the bit estimate in the global gain calculation, as indicated below.

Second, the noise floor is lowered via the lowBits parameter if the spectrum exhibits a low center of mass. In this case, the low frequency content is dominant, whereby the loss of the high frequency components is not likely to be as critical as the high frequency content.

обозначает битовый бюджет для кодирования спектра. Оценка потребления битов (накопленная в переменной tmp) основана на значениях E(k) энергии с учетом контекстной зависимости в арифметическом кодере, используемом для кодирования в каскаде 1.The actual global gain estimation is performed (eg, block 25 of FIG. 9) by a low complexity bisectional search as indicated in the C code below, where

denotes the bit budget for spectrum coding. The bit consumption estimate (accumulated in the tmp variable) is based on the context sensitive energy E(k) values in the arithmetic encoder used to encode in stage 1.

fac=0.3;

gg _ind =255;

for (iter=0; iter<8; iter++)

{

fac>>=1;

gg _ind -=fac;

tmp=0;

iszero=1;

for (i=N/4-1; i>=0; i--)

{

if (E[i]*28/20<(gg _ind +gg _off ))

{

if (iszero==0)

{

tmp+=2.7*28/20;

}

}

else

{

if ((gg _ind +gg _off )<E[i]*28/20-43*28/20)

{

tmp+=2*E[i]*28/20-2*(gg _ind +gg _off )-36*28/20;

}

else

{

tmp+=E[i]*28/20-(gg _ind +gg _off )+7*28/20;

}

iszero=0;

}

}

*1,4*28/20 andand iszero==0)if (tmp>

*1.4*28/20 andand iszero==0)

{

gg _ind +=fac;

}

}

Residual coding (eg, Fig. 3)

Residual coding uses redundant bits that are available after arithmetic coding of the quantized spectrum X _q (k). Let B denote the number of redundant bits, and let K denote the number of encoded non-zero coefficients X _q (k). In addition, let k _i , i=1, ... K, denote the enumeration of these non-zero coefficients from the smallest to the largest frequency. The residual bits b _i (j) (taking the values 0 and 1) for the coefficient k _i are calculated in such a way as to minimize the error:

This can be done in an iterative way, checking if the following is true:

If (1) is true, then the nth residual bit bi(n) for the coefficient k _i is set to 0, and otherwise it is set to 1. The calculation of the residual bits is performed by calculating the first residual bit for each k _i and then the second bit and etc. until all residual bits are consumed, or the maximum number nmax of iterations is not executed. This leaves:

residual bits for coefficient X _q (k _i ). This residual coding scheme improves upon the residual coding scheme applied in the 3GPP EVS codec, which consumes at most one bit per non-zero coefficient.

The calculation of the residual bits with nmax=20 is illustrated with the following pseudocode, with gg standing for the global gain:

iter=0;

nbits_residual=0;

offset=0.25;

while (nbits_residual<nbits_residual_max andand iter<20)

{

k=0;

while (k<N _E andand nbits_residual<nbits_residual_max)

{

if (Xq[k] ! ₌ 0)

{

if ( _Xf [k]> _=Xq [k]*gg)

{

res_bits[nbits_residual]=1;

X _f [k] -=offset*gg;

}

else

{

res_bits[nbits_residual]=0;

_Xf [k]+=offset*gg;

}

nbits_residual++;

}

k++;

}

iter++;

offset/=2;

}

Description of the decoder (for example, Fig. 6)

получается посредством энтропийного декодирования. Остаточные биты используются для детализации этого спектра, как продемонстрировано следующим псевдокодом (см. также, например, фиг. 8).At the decoder, entropy coded spectrum

obtained by entropy decoding. The residual bits are used to refine this spectrum, as demonstrated by the following pseudocode (see also, eg, FIG. 8).

iter=n=0;

offset=0.25;

while (iter<20 andand n<nResBits)

{

k=0;

while (k<N _E andand n<nResBits)

{

[k] !=0)if (

[k] !=0)

{

if (resBits[n++]==0)

{

[k] -=offset;

[k]-=offset;

}

else

{

[k]+=offset;

[k]+=offset;

}

}

k++;

}

iter++;

offset/=2;

}

The decoded residual spectrum is given as follows:

Conclusions

An efficient two-stage coding scheme is proposed, containing the first entropy coding stage and the second residual coding stage based on one-bit (non-entropy) coding.

The scheme uses a low complexity global gain estimator that includes an energy-based bit consumption estimator for the first coding stage, comprising a signal-adaptive noise floor adder.

The noise floor combiner efficiently transfers bits from the first coding stage to the second coding stage for high tone signals while leaving the estimate for other types of signals unchanged. This bit shifting from the entropy coding stage to the non-entropy coding stage is claimed to be fully effective for high-pitched signals.

Fig. 12 illustrates a procedure for reducing the number of audio data elements in a signal-dependent manner using split reduction. In step 901, quantization is performed using non-manipulated information such as a global gain calculated from the signal data without being manipulated. To this end, the (full) bit budget for the audio data items is required, and in the output of block 901, quantized data items are obtained. In block 902, the number of audio data elements is reduced by eliminating a (controlled) number of preferably the smallest audio data elements based on a signal-dependent control value. At the output of block 902, the reduced number of data elements is obtained, and at block 903, the initial encoding stage is applied, and with the bit budget for the residual bits that remain due to the controlled reduction, the detail encoding stage is applied, as illustrated at 904.

In addition, in the procedure of FIG. 12, reduction block 902 may also be performed prior to actual quantization using a global gain value or, in general, a specific quantizer step size that is determined using non-keyed audio data. This reduction in the number of audio data elements can therefore also be carried out in the non-quantized region by setting certain preferably small values to zero, or by weighting certain values with weighting factors, which ultimately leads to values quantized to zero. In the implementation for divided reduction, an explicit quantization step on the one hand and an explicit reduction step on the other hand are performed, while control for specific quantization is performed without data manipulation.

In contrast, FIG. 13 illustrates an integrated reduction mode according to an embodiment of the present invention. In block 911, manipulated information is determined by the controller 20, such as, for example, the global gain illustrated in the output of block 25 of FIG. 9. In block 912, quantization of the non-keyed audio data is performed using the keyed global gain, or in general, the keyed information computed in block 911. In the output of the quantization procedure of block 912, the reduced number of audio data items that are initially encoded in block 903 is obtained, and the granularity , encoded in block 904. Due to signal-dependent audio chip reduction, residual bits for at least one full iteration and at least part of the second iteration, and preferably even more than two iterations, remain. The shift of the bit budget from the initial coding stage to the detail coding stage is performed in accordance with the present invention and in a signal dependent manner.

The present invention can be implemented in at least four different modes. The determination of the control value may be performed in direct mode with explicit signal characterization, or in indirect mode without explicit signal characterization, but adding the signal-dependent noise floor to the audio data or extracted audio data as an example for manipulation. At the same time, the reduction in the number of audio data items is performed in an integrated manner or in a divided manner. Indirect determination and integrated reduction or indirect control value generation and divided reduction may also be performed. In addition, direct determination along with integrated reduction and direct determination of the control value along with divided reduction can also be performed. For low efficiency purposes, it is preferable to indirectly determine the control value along with the integrated reduction of audio data elements.

It should be noted here that all alternatives or aspects explained above and all aspects defined in the independent claims in the following claims may be used individually, i. without alternatives or objectives other than the intended alternative, objective or independent claim. However, in other embodiments, two or more of the alternatives or aspects or independent claims may be combined with each other, and in other embodiments, all aspects or alternatives and all independent claims may be combined with each other.

The audio signal encoded according to the invention may be stored on a digital storage medium or a permanent storage medium, or may be transmitted over a transmission medium such as a wireless transmission medium or a wired transmission medium such as the Internet.

Although some aspects are described in the context of a device, it is obvious that these aspects also represent a description of the corresponding method, with the block or device corresponding to a method step or a feature of a method step. Likewise, the aspects described in the context of a method step also provide a description of the corresponding block or element, or feature of the corresponding device.

Depending on certain implementation requirements, embodiments of the invention may be implemented in hardware or software. The implementation may be performed using a digital storage medium, such as a floppy disk, DVD, CD, ROM, PROM, EPROM, EEPROM, or flash memory, having stored electronically readable control signals that interact (or are capable of interacting) with a programmable computer system. so that the corresponding method is carried out.

Some embodiments of the invention comprise a storage medium having electronically readable control signals that are capable of interacting with a programmable computer system such that one of the methods described herein is carried out.

In general, embodiments of the present invention may be implemented as a computer program product with program code, wherein the program code is configured to perform one of the methods when the computer program product is running on the computer. The program code may, for example, be stored on a computer-readable medium.

Other embodiments comprise a computer program for carrying out one of the methods described herein stored on a computer-readable medium or on a permanent storage medium.

In other words, therefore, an embodiment of the method according to the invention is a computer program having program code for carrying out one of the methods described herein when the computer program is running on a computer.

Therefore, a further embodiment of the methods of the invention is a storage medium (digital storage medium or computer-readable medium) containing a recorded computer program for carrying out one of the methods described herein.

Therefore, a further embodiment of the method according to the invention is a data stream or sequence of signals representing a computer program for implementing one of the methods described herein. The data stream or signal sequence, for example, may be configured to be transmitted over a data connection, such as the Internet.

An additional embodiment includes processing means, such as a computer or programmable logic device, configured to perform one of the methods described herein.

An additional embodiment comprises a computer having a computer program installed to implement one of the methods described herein.

In some embodiments, a programmable logic device (eg, a field programmable gate array) may be used to perform some or all of the functionality of the methods described herein. In some embodiments, a user-programmable gate array may interface with a microprocessor to implement one of the methods described herein. In general, the methods are preferably carried out by any hardware device.

The above described embodiments are merely illustrative of the principles of the present invention. It should be understood that modifications and changes to the configurations and details described herein should be apparent to those skilled in the art. Therefore, it is intended to be limited only by the scope of the following claims, and not by the specific details provided herein as a description and explanation of the embodiments.

Claims (113)

1. Audio encoder for encoding input audio data (11), containing: - a preprocessor (10) for pre-processing the input audio data (11 for obtaining the audio data to be encoded; - an encoder processor (15) for encoding the audio data to be encoded; and - a controller (20) for controlling the encoder processor (15) in such a way that, depending on the first characteristic of the signals of the first frame of audio data to be encoded, the number of audio data elements for the audio data to be encoded by the encoder processor (15) for the first frame is reduced by compared to the second signal characteristic of the second frame, and the first number of information units used to encode the reduced number of audio data elements for the first frame is more significantly improved compared to the second number of information units for the second frame. 2. Audio encoder according to claim 1, - in which the encoder processor (15) comprises an initial encoding stage (151) and a detail encoding stage (152), - wherein the controller (20) is configured to reduce the number of audio data elements encoded by the initial coding stage (151) for the first frame, wherein the initial encoding stage (151) is configured to encode the reduced number of audio data elements for the first frame using the initial number of information units of the first frame, and wherein the detail coding stage (152) is configured to use the remaining number of information units of the first frame for detail coding for the reduced number of audio data elements for the first frame, wherein the result of adding the initial number of information units of the first frame with the remaining number of information units of the first frame is a given the number of information units for the first frame. 3. Audio encoder according to claim 2, - in which the controller (20) is configured to reduce the number of audio data elements encoded by the initial coding stage (151) for the second frame to a higher number of audio data elements compared to the first frame, wherein the initial encoding stage (151) is configured to encode the reduced number of audio data elements for the second frame using the initial number of information units of the second frame, wherein the initial number of information units of the second frame is greater than the initial number of information units of the first frame, and wherein the detail coding stage (152) is configured to use the remaining number of information units of the second frame for detail coding for the reduced number of audio data elements for the second frame, wherein the result of adding the initial number of information units of the second frame with the remaining number of information units of the second frame is a given the number of information units for the first frame. 4. An audio encoder according to any one of the preceding claims, - in which the encoder processor (15) comprises an initial encoding stage (151) and a detail encoding stage (152), wherein the initial encoding stage (151) is configured to encode a reduced number of audio data elements for the first frame using the initial number of information units of the first frame, wherein the detail coding stage (152) is configured to use the remaining number of information units of the first frame for detail coding for the reduced number of audio data elements for the first frame, wherein the result of adding the initial number of information units of the first frame with the remaining number of information units of the first frame is a given the number of information units for the first frame, and wherein the controller (20) is configured to control the encoder processor (15) such that the detail coding stage (152) performs detail coding of at least one of the reduced number of audio data elements of the first frame using at least two information units, or such that the detail coding stage (152) performs detail coding on more than 50 percent of the reduced number of audio data items using at least two information units for each audio data item, or wherein the controller (20) is configured to control the encoder processor (15) such that the detail coding stage (152) performs detail coding of all audio data elements of the second frame using less than two information units, or such that the detail coding stage performs detail encoding less than 50 percent of the reduced number of audio data items using at least two information units for each audio data item. 5. An audio encoder according to any one of the preceding claims, - in which the encoder processor (15) comprises an initial encoding stage (151) and a detail encoding stage (152), wherein the initial encoding stage (151) is configured to encode a reduced number of audio data elements for the first frame using the initial number of information units of the first frame, wherein the detail coding stage (152) is configured to use the remaining number of information units of the first frame for detail coding for the reduced number of audio data elements for the first frame, - at the same time, the detail coding stage (152) is configured to iteratively assign (300, 302) the remaining number of information units of the first frame to a reduced number of audio data elements on at least two consecutive iterations, calculating (304, 308, 312) the value of the assigned information units for at least two sequentially performed iterations and input (316, 318, 320) the calculated values of information units for at least two sequentially performed iterations into the encoded output frame in a given order. 6. The audio encoder of claim 5, wherein the detail encoding stage (152) is configured to sequentially compute (304) an information unit for each audio data element from the reduced number of audio data elements for the first frame, in order from low frequency information for the audio data element to high frequency information for the audio data element in the first iteration, wherein the detail coding stage (152) is configured to sequentially compute (308) an information unit for each audio data element from the reduced number of audio data elements for the first frame, in order from low frequency information for the audio data element to high frequency information for the audio data element in the second iteration, and wherein the detail coding stage (152) is configured to check (314) whether the number of information units already assigned is less than the predetermined number of information units for the first frame, which is less than the initial number of information units of the first frame, and terminate the second iteration at in the case of a negative test result, or, in the case of a positive test result, performing (312) a certain number of additional iterations until a negative test result is obtained, and the number of additional iterations is at least one, or - at the same time, the detail coding stage (152) is configured to count the number of non-zero audio elements and determine the number of iterations from the number of non-zero audio elements and a given number of information units for the first frame, which is less than the initial number of information units of the first frame. 7. An audio encoder according to any one of the preceding claims, - in which the encoder processor (15) comprises an initial encoding stage (151) and a detail encoding stage (152), wherein the initial coding stage (151) is configured to encode the number of MSIs for each audio data element from the reduced number of audio data elements for the first frame using the initial number of information units of the first frame, said number being greater than one, and wherein the detail coding stage (152) is configured to use the remaining number of information units of the first frame to encode the number of lower information units for each audio data element from the reduced number of audio data elements for the first frame, said number being greater than one for at least one audio data element of the reduced number of audio data items for the first frame. 8. An audio encoder according to any one of the preceding claims, wherein the first signal characteristic constitutes a first tone value, wherein the second signal characteristic constitutes a second tone value, wherein the first tone value indicates a higher tone than the second tone value, and - while the controller (20) is configured to reduce the number of audio data elements for the first frame to a first number less than the number of audio data elements for the second frame, and increase the average number of information units used to encode each audio data element from the reduced number of audio data elements of the first frame such that it exceeds the average number of information units used to encode each audio data element from the reduced number of audio data elements of the second frame. 9. An audio encoder according to any one of the preceding claims, wherein the encoder processor (15) comprises: - a variable quantizer (150) for quantizing the audio data of the first frame to obtain quantized audio data for the first frame, and for quantizing the audio data of the second frame to obtain quantized audio data for the second frame; - an initial coding stage (151) for encoding the quantized audio data of the first frame or the second frame; - a detail coding stage (152) for encoding the residual data of the first frame and the second frame; - while the controller (20) is configured to analyze (26, 28) the audio data of the first frame to determine the first control value (21) for the variable quantizer (150) for the first frame, and analyze (26, 28) the audio data of the second frame to determine the second a control value for the variable quantizer (150) for the second frame, the second control value being different from the first control value (21), and - while the controller (20) is configured to perform (23, 24) manipulation of the audio data of the first frame or the second frame or amplitude-related values extracted from the audio data of the first frame or the second frame, depending on the audio data to determine the first control value (21) or a second control value (21), wherein the variable quantizer (150) is configured to quantize the audio data of the first frame or the second frame without said manipulation. 10. Audio encoder according to any one of paragraphs. 1-9, wherein the encoder processor (15) comprises: - a variable quantizer (150) for quantizing the audio data of the first frame to obtain quantized audio data for the first frame, and for quantizing the audio data of the second frame to obtain quantized audio data for the second frame; - an initial coding stage (151) for encoding the quantized audio data of the first frame or the second frame; - a detail coding stage (152) for encoding the residual data of the first frame and the second frame; - while the controller (20) is configured to analyze the audio data of the first frame to determine the first control value (21) for the variable quantizer (150), for the initial coding stage (151) or for the module (150) for reducing the number of audio data elements for the first frame, and with the possibility of analyzing the audio data of the second frame to determine the second control value for the variable quantizer (150), for the initial coding stage (151), or for the module (150) for reducing the number of audio data elements for the second frame, the second control value being different from the first control value, and wherein the controller (20) is configured (201) to determine the first tone characteristic as the first signal characteristic for determining the first control value, and the second tone characteristic as the second signal characteristic for determining the second control value, so that the bit budget for of the detail coding stage (152) is increased in the case of the first keynote compared to the bit budget for the detail coding stage (152) in the case of the second keynote, with the first keynote indicating a greater key than the second keynote. 11. The audio encoder according to claim 9 or 10, wherein the initial coding stage (151) is an entropy coding stage for entropy coding, or the detail coding stage (152) is a residual or binary coding stage for encoding the residual data of the first frame and the second frame . 12. Audio encoder according to any one of paragraphs. 9-11, - in which the controller (20) is configured to determine the first or second control value so that the first information unit budget for the initial encoding stage (151) is less than or equal to a predetermined value, and the controller (20) is configured to extract a second information unit budget for the detail coding cascade (152) using the first information unit budget and the maximum number of information units for the first or second frame, or a predetermined value. 13. Audio encoder according to any one of paragraphs. 9-12, in which the controller (20) is configured to calculate (22) amplitude related values as a plurality of power values extracted from one or more audio data audio values, and manipulate (24) the power values using the addition of the same manipulation value with all power values from a plurality of power values, or - while the controller (20) is configured to: - randomized addition or subtraction (24) of the same manipulation value on all audio values from or from a plurality of audio values included in the frame, or - addition or subtraction of values obtained by means of the same absolute value of the manipulation value, but preferably with randomized signs, or - adding or subtracting values obtained by subtracting slightly different terms from the same absolute value, - adding or subtracting values obtained as samples from a normalized probability distribution scaled using the computed complex or real absolute value of the keying value, or - while the controller (20) is configured to calculate (22) amplitude-related values using the exponentiation of the audio data of the first or second frame or subsampled audio data of the first or second frame with an exponent value, and the exponent value is greater than 1. 14. Audio encoder according to any one of paragraphs. 9-13, in which the controller (20) is configured to calculate (23) a keying value to be manipulated using the maximum value (26) of the set of audio data or amplitude related values, or using the maximum value of the set of subsampled audio data or the set of subsampled amplitude related values. for the first or second frame. 15. Audio encoder according to any one of paragraphs. 9-14, wherein the controller (20) is configured to calculate (23) a keying value for keying further using a signal-independent weight value (27), wherein the signal-independent weight value depends on at least one of the bit rate for the first or second frame, frame duration and sample rate. 16. Audio encoder according to any one of paragraphs. 9-15, wherein the controller (20) is configured to calculate (23, 29) a keying value to be manipulated using a signal dependent weight value derived from at least one of a first sum of absolute values of audio data or subsampled audio data in a frame, a second the sum of the absolute values of the audio data or subsampled audio data in the frame, multiplied by the index associated with each absolute value, and the quotient of the second sum and the first sum. 17. Audio encoder according to any one of paragraphs. 9-16, - in which the controller (20) is configured to calculate (29) a manipulation value for manipulation based on the following equation:

, where k is a frequency index, where X _f (k) is the audio data value for frequency index k before quantization, where max is a function of the maximum, where regBits is the first signal-independent weight value, and where lowBits is the second signal-dependent weighting value. 18. An audio encoder according to any one of the preceding claims, wherein the preprocessor (10) further comprises: - frequency-time converter (14) for converting time domain audio data into frame spectral values; and - a spectral processor (15) for calculating modified spectral values having a flatter spectral envelope than the spectral envelope of the spectral values, wherein the modified spectral values represent the audio data of the first or second frame to be encoded by the encoder processor (15). 19. The audio encoder of claim 18, wherein the spectral processor (15) is configured to perform at least one of a temporal noise shaping operation, a spectral noise shaping operation, and a spectral whitening operation. 20. Audio encoder according to any one of paragraphs. 9-19, wherein the controller (20) is configured to calculate a control value using a plurality of energy values as amplitude-related values for a frame, each energy value being extracted (22, 23, 24) from the power value as associated with an amplitude value and a signal-dependent keying value for said keying. 21. Audio encoder according to claim 20, wherein the controller (20) is configured to: - calculating the required bit estimate of each energy value depending on the energy value and the possible value for the control value, - accumulation of the required bit estimates for energy values and a possible value for the control value, - checking if the accumulated bit score for the possible value for the control value meets the allowed bit consumption criterion, and - modifying the candidate value for the control value in case the allowed bit consumption criterion is not met, and repeating the computation of the required bit estimate, accumulating the required bit rate, and checking until the allowed bit consumption criterion for the modified candidate value is met for control value. 22. Audio encoder according to claim 20 or 21, - in which the controller (20) is configured to calculate a plurality of energy values based on the following equation:

where E(k) is the energy value for index k, where PX _lp (k) is the power value for index k as an amplitude related value, and N(X _f ) is the signal-dependent keying value. 23. Audio encoder according to any one of paragraphs. 9-22, wherein the controller (20) is configured to calculate a first or second control value based on an estimate of the accumulated information units required for each manipulated audio data value or manipulated amplitude-related value. 24. Audio decoder according to any one of paragraphs. 9-23, - in which the controller (20) is configured to manipulate in such a way that, due to manipulation, the bit budget for the initial encoding stage (151) is increased or the bit budget for the detailed encoding stage (152) is reduced. 25. Audio decoder according to any one of paragraphs. 9-24, - in which the controller (20) is configured to be manipulated such that the manipulation results in a larger residual coding stage bit budget for the first key signal compared to the second key signal, with the second key lower than the first key. 26. Audio decoder according to any one of paragraphs. 9-25, - in which the controller (20) is configured to manipulate such that the energy of the audio data from which the bit budget for the initial coding stage (151) is calculated is increased relative to the energy of the audio data to be quantized by the variable quantizer (150). 27. An audio encoder according to any one of the preceding claims, wherein the encoder processor (15) comprises a variable quantizer (150) for quantizing the audio data of the first frame to obtain quantized audio data for the first frame and for quantizing the audio data of the second frame to obtain quantized audio data for the second frame, - while the controller (20) is configured to calculate the global gain for the first or second frame, and - while the variable quantizer (150) contains: module (155) weighting for weighting with global gain; and a quantizer kernel (157) having a fixed quantization step size. 28. An audio encoder according to any one of the preceding claims, wherein the encoder processor (15) comprises an initial encoding stage (151) and a detail encoding stage (152), wherein the detail coding stage (152) is configured to calculate the detail bits for the quantized audio values in a plurality of iterations, with the detail bit indicating a different number at each iteration, or where the detail bit at the lower iteration indicates a larger number than the detail bit at the higher iteration, or wherein the amount is a fractional amount that is a fraction of the quantizer step size indicated by the control value. 29. An audio encoder according to any one of the preceding claims, wherein the encoder processor (15) comprises a detail coding stage (152), wherein the detail coding stage (152) is configured (304, 308, 312): - performing iterative processing having at least two iterations, - checking whether the quantized audio value, or the quantized audio value, together with the potential first quantity associated with the detail bit for the quantized audio value at the first iteration, is added to or subtracted from the second quantity for the second iteration, when weighted by global gain, is greater or less than the unquantized audio value, and - setting the detail bit for the second iteration, depending on the result of the check. 30. An audio encoder according to any one of the preceding claims, wherein the encoder processor (15) comprises a variable quantizer (150) and a detail coding stage (152), wherein the detail coding stage (152) is configured to compute a detail bit only for audio values that are not quantized to zero by a variable quantizer (150). 31. An audio encoder according to any one of the preceding claims, - in which the controller (20) is configured to reduce the effect of manipulation for audio data having a center of mass at a lower frequency, and wherein the initial encoding stage (151) of the encoder processor (15) is configured to remove high-frequency spectral values from the audio data if it is determined that the bit budget for the first or second frame is not sufficient to encode the quantized audio data of the frame. 32. Audio encoder according to any one of the preceding claims, - in which the controller (20) is configured to perform a bisectional search for each frame separately using the manipulated spectral energy values for the first or second frame as the manipulated amplitude-related values for the first or second frame. 33. A method for encoding input audio data, comprising the steps of: - pre-processing the input audio data (11) to obtain the audio data to be encoded; encode the audio data to be encoded; and controlling the encoding such that, depending on the first signal characteristic of the first frame of audio data to be encoded, the number of audio data elements for the audio data to be encoded for the first frame is reduced compared to the second signal characteristic of the second frame, and the first number of information units used to encode the reduced number of audio data items for the first frame is more significantly improved compared to the second number of information units for the second frame. 34. The method of claim 33, wherein the encoding comprises the steps of: - variably quantizing the audio data of the frame to obtain quantized audio data; - perform entropy encoding of the quantized audio data of the frame; and - encode the residual data of the frame; - while the control contains the step at which determine the control value for the variable quantization, and the definition contains the steps at which: analyze the audio data of the first or second frame; and manipulating the audio data of the first or second frame or the amplitude related values extracted from the audio data of the first or second frame depending on the audio data to determine the control value, wherein the variable quantization quantizes the audio data of the frame without manipulation, or wherein the control comprises the step of determining the first or second audio data sentiment characteristic and determining the control value such that the bit budget for the residual coding is increased in the case of the first sentiment characteristic compared to the bit budget for the residual coding stage in the case of the second sentiment characteristic, wherein the first key characteristic indicates a greater key than the second key characteristic. 35. A physical storage medium on which a computer program is stored for implementation when executed on a computer or in a processor of the method according to claim 33 or 34.

Images