EP3544003B1 - Device and method of determining an estimated value - Google Patents

Abstract

The device and method are used for a video or audio signal (100). A first step (102) provides levels for allowable interference (nb(b)) and the signal energy in a given frequency band (e(b)). These signals are processed in a second step (104) which receives a frequency band energy distribution signal (nl(b)) from a third step (106) and calculates an estimated value (pe).

Description

The present invention relates to coders for coding a signal comprising audio and / or video information, and more particularly to estimating a need for information units for coding this signal.

The known encoder is shown below. An audio signal to be coded is fed in at an input 1000. This is first fed to a scaling stage 1002, in which what is known as AAC gain control is carried out in order to determine the level of the audio signal. Page information from the scaling is fed to a bitstream formatter 1004, as shown by the arrow between block 1002 and block 1004. The scaled audio signal is then fed to an MDCT filter bank 1006. In the AAC coder, the filter bank implements a modified discrete cosine transform with 50% overlapping windows, the window length being determined by a block 1008.

Generally speaking, block 1008 is provided for windowing transient signals with shorter windows and windowing more stationary signals with longer windows. The purpose of this is that due to the shorter window for transient signals a higher time resolution is achieved (at the expense of the frequency resolution), while for more stationary signals a higher frequency resolution (at the expense of the time resolution) through longer windows is achieved, with longer windows tending to be preferred because they promise a greater coding gain. At the output of the filter bank 1006 there are temporally successive blocks of spectral values which, depending on the embodiment of the filter bank, can be MDCT coefficients, Fourier coefficients or even sub-band signals, each sub-band signal having a certain limited bandwidth that is determined by the corresponding sub-band channel in the filter bank 1006 is determined, and each sub-band signal has a certain number of sub-band samples.

The following is an example of the case in which the filter bank outputs successive blocks of MDCT spectral coefficients viewed over time, which, generally speaking, represent successive short-term spectra of the audio signal to be encoded at input 1000. A block of MDCT spectral values is then fed into a TNS processing block 1010, in which temporal noise shaping takes place (TNS = temporal noise shaping). The TNS technique is used to shape the temporal shape of the quantization noise within each window of the transformation. This is achieved by applying a filtering process to parts of the spectral data of each channel. The coding is done on a window basis. In particular, the following steps are carried out in order to use the TNS tool on a window of spectral data, i.e. on a block of spectral values.

First, a frequency range is selected for the TNS tool. A suitable choice is to cover a frequency range from 1.5 kHz up to the highest possible scale factor band with a filter. It should be noted that this frequency range depends on the sampling rate as specified in the AAC standard (ISO / IEC 14496-3: 2001 (E)).

An LPC calculation (LPC = linear predictive coding) is then carried out, specifically with the spectral MDCT coefficients which lie in the selected target frequency range. For increased stability, coefficients corresponding to frequencies below 2.5 kHz are excluded from this process. Usual LPC procedures, as they are known from speech processing, can be used for the LPC calculation, for example the known Levinson-Durbin algorithm. The calculation is carried out for the maximum allowable order of the noise shaping filter.

The expected prediction gain PG is obtained as the result of the LPC calculation. Furthermore, the reflection coefficients or Parcor coefficients are obtained.

If the prediction gain does not exceed a certain threshold, the TNS tool is not used. In this case, control information is written into the bit stream so that a decoder knows that TNS processing has not been carried out.

However, if the prediction gain exceeds a threshold, TNS processing is applied.

In a next step, the reflection coefficients are quantized. The order of the noise shaping filter used is determined by removing all reflection coefficients with an absolute value less than a threshold from the "tail" of the reflection coefficient array. The number of remaining reflection coefficients is in the order of magnitude of the noise shaping filter. A suitable threshold is 0.1.

The remaining reflection coefficients are typically converted to linear prediction coefficients, a technique also known as the "step-up" procedure.

The calculated LPC coefficients are then used as encoder noise shaping filter coefficients, that is to say as prediction filter coefficients. This FIR filter is guided over the specified target frequency range. An autoregressive filter is used for decoding, while a so-called moving average filter is used for coding. Finally, the page information for the TNS tool is also fed to the bitstream formatter, as shown by the arrow between the block TNS processing 1010 and the bitstream formatter 1004 in FIG Fig. 3 is shown.

Several in Fig. 3 Run through optional tools, not shown, such as a long-term prediction tool, an intensity / coupling tool, a prediction tool, a noise substitution tool, until finally a middle / side encoder 1012 is reached. The middle / side encoder 1012 is active when the audio signal to be encoded is a multi-channel signal, that is to say a stereo signal with a left channel and a right channel. So far, i.e. in the processing direction before block 1012 in Fig. 3 were the left and right stereo channels processed separately from each other, i.e. scaled, transformed through the filter bank, subjected to TNS processing or not, etc.

In the middle / side coder, it is then first checked whether middle / side coding makes sense, that is, whether there is any coding gain at all. A middle / side coding will bring a coding gain if the left and right channels are more similar, because then the middle channel, i.e. the sum of the left and right channels, is almost the same as the left or right channel, apart from the scaling by the factor 1/2, while the side channel has only very small values, since it is equal to the difference between the left and right channels. It can thus be seen that when the left and right channels are approximately the same, the difference is approximately zero or only includes very small values which - this is the hope - are quantized to zero in a subsequent quantizer 1014 and thus can be transmitted very efficiently, since the quantizer 1014 is followed by an entropy coder 1016.

A psychoacoustic model 1020 supplies the quantizer 1014 with a permitted disturbance per scale factor band. The quantizer works iteratively, ie an outer iteration loop is first called, which then calls an inner iteration loop. Generally speaking, starting from quantizer step size start values, a block of values is first quantized at the input of quantizer 1014. In particular, the inner loop quantizes the MDCT coefficients, consuming a certain number of bits. The outer loop calculates the distortion and modified energy of the coefficients using the scale factor to call an inner loop again. This process is iterated until a certain set of conditions is reached is satisfied. For each iteration in the outer iteration loop, the signal is reconstructed in order to calculate the disturbance introduced by the quantization and to compare it with the permitted disturbance supplied by the psycho-acoustic model 1020. Furthermore, the scale factors are increased by one level from iteration to iteration, specifically for each iteration of the outer iteration loop.

Then, when a situation is reached in which the quantization interference introduced by the quantization is below the permitted interference determined by the psycho-acoustic model, and when bit requirements are met at the same time, namely that a maximum bit rate is not exceeded, the iteration becomes the analysis-through-synthesis method is terminated, and the scale factors obtained are encoded, as is carried out in block 1014, and supplied in encoded form to the bitstream formatter 1004, as indicated by the arrow between block 1014 and the Block 1004 is drawn. The quantized values are then fed to the entropy coder 1016, which typically entropy encodes using several Huffman code tables for different scale factor bands to translate the quantized values into a binary format. As is known, in entropy coding in the form of Huffman coding, code tables are used which are created on the basis of expected signal statistics, and in which frequently occurring values are given shorter code words than less frequently occurring values. The entropy-coded values are then also fed to the bit stream formatter 1004 as the actual main information, which then outputs the coded audio signal on the output side according to a specific bit stream syntax.

The data reduction of audio signals has become a well-known technique that is the subject of a number of international standards (e.g. ISO / MPEG-1, MPEG-2 AAC, MPEG-4).

What the above-mentioned methods have in common is that the input signal is brought into a compact, data-reduced representation by means of a so-called encoder using perception-related effects (psychoacoustics, psychooptics). For this purpose, a spectral analysis of the signal is usually carried out and the corresponding signal components are quantized, taking into account a perception model, and then coded as a so-called bit stream in the most compact way possible.

So-called perceptual entropy (PE) can be used to estimate before the actual quantization how many bits a certain section of the signal to be coded will need. The PE also provides a measure of how difficult it is for the encoder to encode a particular signal or parts of it.

The deviation of the PE from the number of bits actually required is decisive for the quality of the estimate.

Furthermore, the perceptual entropy or any estimated value for a need for information units for coding a signal can be used to estimate whether the signal is transient or stationary, since transient signals also require more bits for coding than stationary signals. The estimation of a transient property of a signal is used, for example, to make a window length decision as indicated by block 1008 in Fig. 3 is indicated to perform.

In Fig. 6 the perceptual entropy calculated according to ISO / IEC IS 13818-7 (MPEG-2 advanced audio coding (AAC)) is shown. To calculate this perceptual entropy, i.e. a band-wise perceptual entropy, the in Fig. 6 is used. In this equation, the parameter pe stands for the perceptual entropy. Furthermore, width (b) stands for the number of spectral coefficients in the respective band b. Furthermore, e (b) is the energy of the signal in this band. Finally, nb (b) is the matching masking threshold or, in more general terms, the permitted interference that can be introduced into the signal, for example by quantization, so that a human listener still hears no or only a negligible interference.

The bands can be derived from the band division of the psychoacoustic model (block 1020 in Fig. 3 ) or the so-called scale factor bands (scfb) used in the quantization. The psychoacoustic masking threshold is the energy value that the quantization error should not exceed.

In the Fig. 6 The figure shown thus shows how well a perceptual entropy determined in this way works as an estimate for the number of bits required for coding. For this purpose, using the example of an AAC coder at different bit rates, the respective perceptual entropy was plotted for each individual block as a function of the bits used. The test piece used contains a typical mixture of music, language and individual instruments.

Ideally, the points would congregate along a straight line through the zero point. The expansion of the point sequence with the deviations from the ideal line illustrates the imprecise estimate.

The disadvantage of the in Fig. 6 The concept shown is the deviation that expresses itself to the effect that, for example, the value for the perceptual entropy is too large, which in turn means that the quantizer is signaled that more bits are required than actually required. This has the result that the quantizer quantizes too finely, that is to say that it does not exhaust the degree of permitted interference, which results in a reduced coding gain. On the other hand, if the value for the perceptual entropy is determined to be too small, the quantizer is signaled that fewer bits than actually required are required to encode the signal. This in turn has the consequence that the quantizer quantizes too roughly, which would immediately lead to an audible disturbance in the signal, unless countermeasures are taken. The countermeasures can be that the quantizer still needs one or more further iteration loops, which increases the computing time of the encoder.

To improve the calculation of the perceptual entropy one could, as it is in Fig. 7 as shown, introduce a constant term such as 1.5 into the logarithm expression. Then there is already a better result, i.e. a smaller deviation upwards or downwards, although it can still be seen that when taken into account of a constant term in the logarithm expression, the case is reduced that the perceptual entropy signals an overly optimistic need for bits. On the other hand it's over Fig. 7 However, it can be clearly seen that too high a number of bits is being signaled, which means that the quantizer will always quantize too finely, i.e. that the bit requirement is assumed to be greater than it actually is, which in turn results in a reduced coding gain. The constant in the logarithm expression is a rough estimate of the bits needed for the side information.

Inserting a term into the logarithm expression does improve the band-wise perceptual entropy, as shown in Fig. 6 is shown, since the bands with a very small distance between energy and masking threshold are more likely to be taken into account, since a certain number of bits is also necessary for the transmission of spectral coefficients quantized to zero.

Another calculation of the perceptual entropy, which is very time consuming, is in Fig. 8 shown. In Fig. 8 the case is shown in which the perceptual entropy is calculated line by line. The disadvantage, however, is the higher computational effort involved in the line-by-line calculation. Instead of the energy, spectral coefficients X (k) are used, where kOffset (b) denotes the first index of band b. If Fig. 8 With Fig. 7 is compared, a reduction in the "deflections" upwards can clearly be seen in the range between 2000 and 3000 bits. The PE estimate will therefore be more precise, i.e. not estimate too pessimistically, but rather be at the optimum, so that the coding gain compared to the in Fig. 6 and 7th shown Calculation method can increase, or the number of iterations in the quantizer is reduced.

The disadvantage of the line-by-line calculation of the perceptual entropy, however, is the computing time that is required to convert the in Fig. 8 evaluate the equation shown.

Such computation time disadvantages do not necessarily play a role if the encoder runs on a powerful PC or a powerful workstation. On the other hand, the situation is completely different when the encoder is housed in a portable device, such as a UMTS cell phone, which on the one hand has to be small and cheap, on the other hand it has to have a low power requirement, and which also has to work quickly to get the Enabling coding of an audio signal or video signal transmitted via the UMTS connection.

The US 2002/103637 A1 discloses a concept for improving the performance of coding systems employing radio frequency reconstruction techniques. For this purpose, a coding difficulty or a measure for the workload of an encoder is calculated on the encoder side in order to control the crossover frequency as a function of this, which determines the frequency up to which a signal is encoded with a source encoder, the proportion of the signal is encoded above the crossover frequency by a high frequency reconstruction method. As a measure of the difficulty of encoding a signal, the Perceptual Entropy is calculated, which is based on a spectral value being squared and then weighted with a number equal to the number of lines in the current band divided by the psychoacoustic threshold for this Band, and then take the logarithm of the result. Summing up all such logarithms in a band then gives the perceptual entropy in this band. As an alternative to this, a distortion energy can also be calculated at the end of the source coding process by adding up the distortion energy in each band and weighting it with a loudness curve.

The object of the present invention is to create an efficient and yet precise concept for determining an estimated value for a requirement of information units for coding a signal.

This object is achieved by a device according to claim 1, a method according to claim 12 or a computer program according to claim 13.

The present invention is based on the knowledge that a frequency band-wise calculation of the estimated value for a requirement for information units for reasons of computing time It should be noted, however, that in order to obtain an accurate estimate of the estimate, the distribution the energy in the frequency band that is to be calculated band by band must be taken into account.

The entropy coder following the quantizer is thus implicitly “drawn into” the determination of the estimated value for the requirement of information units. The entropy coding enables a smaller number of bits to be required for the transmission of smaller spectral values than for the transmission of larger spectral values. The entropy coder is particularly efficient when spectral values quantized to zero can be transmitted. Since these will typically occur most frequently, the code word for transmitting a spectral line quantized to zero is the shortest code word, and the code word for transmitting an ever larger quantized spectral line is always longer. In addition, for a particularly efficient concept for transmitting a sequence of spectral values quantized to zero, it is even possible to use run-length coding, which means that in the case of a run of zeros per spectral value quantized to zero, on average, not even one only bit is needed.

It was found that the band-by-band perceptual entropy calculation used in the prior art to determine the estimated value for the requirement of information units completely ignores the mode of operation of the downstream entropy coder if the distribution of the energy in the frequency band deviates from a completely uniform distribution .

According to the invention, in order to reduce the inaccuracies of the band-by-band calculation, account is taken of how the energy is distributed within a band.

Depending on the implementation, the measure for the distribution of the energy in the frequency band can be determined on the basis of the actual amplitudes, or by estimating the frequency lines that are not quantized to zero by the quantizer. This dimension, which is also referred to as “nl”, where nl stands for “number of active lines”, that is to say for the number of active lines, is preferred for reasons of computing time efficiency. However, the number of spectral lines quantized to zero or a finer subdivision can also be taken into account, this estimation becoming more and more precise the more information from the downstream entropy coder is taken into account. If the entropy coder is built on the basis of Huffman code tables, properties of these code tables can be integrated particularly well, since the code tables are not calculated on-line based on the signal statistics, but rather because the code tables are fixed anyway, regardless of the actual signal.

Depending on the computing time restrictions, however, in the case of a particularly efficient calculation, the measure for the distribution of the energy in the frequency band is carried out by determining the lines that still survive after the quantization, that is to say the number of active lines.

The present invention is advantageous in that an estimated value for a need for information content is determined which is on the one hand more precise and on the other hand more efficient than in the prior art.

In addition, the present invention can be scaled for various applications, since, depending on the desired accuracy of the estimated value, more and more properties of the entropy coder can be included in the estimation of the bit requirement, but at the cost of increased computing time.

Fig. 1

ein Blockschaltbild der erfindungsgemäßen Vorrichtung zum Ermitteln eines Schätzwerts;

Fig. 2a

eine bevorzugte Ausführungsform der Einrichtung zum Berechnen eines Maßes für die Verteilung der Energie in dem Frequenzband;

Fig. 2b

eine bevorzugte Ausführungsform der Einrichtung zum Berechnen des Schätzwerts für den Bedarf an Bits;

Fig. 3

ein Blockschaltbild eines bekannten AudioCodierers;

Fig. 4

eine Prinzipdarstellung zur Erläuterung des Einflusses der Energieverteilung innerhalb eines Bandes auf die Ermittlung des Schätzwerts;

Fig. 5

ein Diagramm zur Schätzwertberechnung gemäß der vorliegenden Erfindung;

Fig. 6

ein Diagramm zur Schätzwertberechnung gemäß ISO/IEC IS 13818-7(AAC);

Fig. 7

ein Diagramm zur Schätzwertberechnung mit konstantem Term;

Fig. 8

ein Diagramm zur linienweisen Schätzwertberechnung mit konstantem Term.

Preferred embodiments of the present invention are explained in detail below with reference to the enclosed times. Show it:

Fig. 1

a block diagram of the device according to the invention for determining an estimated value;

Fig. 2a

a preferred embodiment of the device for calculating a measure for the distribution of energy in the frequency band;

Figure 2b

a preferred embodiment of the means for calculating the estimate for the demand for bits;

Fig. 3

a block diagram of a known audio encoder;

Fig. 4

a schematic diagram to explain the influence of the energy distribution within a band on the determination of the estimated value;

Fig. 5

a diagram for estimated value calculation according to the present invention;

Fig. 6

a diagram for the estimated value calculation according to ISO / IEC IS 13818-7 (AAC);

Fig. 7

a diagram for estimated value calculation with a constant term;

Fig. 8

a diagram for the line-by-line estimate calculation with constant term.

Reference is made below to Fig. 1 the device according to the invention for determining an estimated value for a requirement of information units for coding a signal is shown. The signal, which can be an audio and / or a video signal, is fed in via an input 100. The signal is preferably already available as a spectral representation with spectral values. However, this is not absolutely necessary, since some calculations can also be carried out with a time signal by means of appropriate bandpass filtering, for example.

The signal is fed to a device 102 for providing a measure of an allowable interference for a frequency band of the signal. The permitted disturbance can, for example, by means of a psycho-acoustic model, as it is based on Fig. 3 (Block 1020) has to be determined. The means 102 is also effective to provide a measure of the energy of the signal in the frequency band. A prerequisite for a band-by-band calculation is that a frequency band for which a permitted interference or signal energy is specified contains at least two or more spectral lines of the spectral representation of the signal. In typical standardized audio encoders, the frequency band will preferably be a scale factor band because the bit requirement estimation is required directly by the quantizer in order to determine whether a quantization that has taken place fulfills a bit criterion or not.

The device 102 is designed to feed both the permitted interference nb (b) and the signal energy e (b) of the signal in the band to a device 104 for calculating the estimated value for the requirement for bits.

According to the invention, the means 104 for calculating the estimated value for the requirement of bits is designed to take into account, in addition to the permitted interference and the signal energy, a measure nl (b) for a distribution of the energy in the frequency band, the distribution of the energy in the frequency band from deviates from a completely even distribution. The measure for the distribution of the energy is calculated in a device 106, the device 106 requiring at least one band, namely the considered frequency band of the audio or video signal either as a bandpass signal or directly as a sequence of spectral lines, e.g. to be able to perform a spectral analysis of the band in order to obtain the measure for the distribution of the energies in the frequency band.

Of course, the audio or video signal can be fed to the device 106 as a time signal, the device 106 then performing band filtering and an analysis in the band. Alternatively, the audio or video signal that is fed to device 106 can already be present in the frequency range, such as, for example, as MDCT coefficients, or as a bandpass signal in the filter bank with a smaller number of bandpasses compared to an MDCT filter bank -Filter.

In a preferred exemplary embodiment, the means 106 for calculating is designed to take into account current amounts of spectral values in the frequency band for calculating the estimated value.

Furthermore, the device for calculating the measure for the distribution of the energy can be designed to determine a number of spectral values as a measure for the distribution of the energy, the amount of which is greater than or equal to a predetermined amount threshold, or the amount of which is less than or equal to the amount threshold , the absolute value threshold preferably being an estimated quantizer stage which, in a quantizer, causes values less than or equal to the quantizer stage to be quantized to zero. In this case, the measure for the energy is the number of active lines, i.e. the number of lines that survive or are not equal to zero after quantization.

Fig. 2a shows a preferred embodiment for the means 106 for calculating the measure for the distribution of the energy in the frequency band. The measure of the distribution of energy in the frequency band is in Fig. 2a denoted by nl (b). The form factor ffac (b) is already a measure of the distribution of energy in the frequency band. As can be seen from block 106, the measure for the spectral distribution nl is obtained from the form factor ffac (b) by weighting with the 4th root of the signal energy e (b) divided by the bandwidth width (b) or number of lines determined in the scale factor band b. In this context, it should be pointed out that the form factor is also an example of a quantity that specifies a measure for the distribution of energies, while nl (b), in contrast, is an example of is a quantity that represents an estimate of the number of lines relevant for quantization.

The form factor ffac (b) is calculated by forming the absolute value of a spectral line and then taking the root of this spectral line and then adding up the "rooted" amounts of the spectral lines in the band.

Figure 2b shows a preferred embodiment of the means 104 for calculating the estimated value pe, where in Figure 2b Another case distinction is introduced, namely when the logarithm to base 2 of the ratio of the energy to the permitted disturbance is greater than a constant factor c1 or equal to the constant factor. In this case, the above alternative in block 104 is used, that is to say the measure for the spectral distribution nl is multiplied by the logarithm expression.

If, on the other hand, it is found that the logarithm to base 2 from the ratio of the signal energy to the permitted interference is less than the value c1, then the lower alternative in block 104 of Figure 2b is used, which also has an additive constant c2 and a multiplicative constant c3, which is calculated from the constants c2 and c1.

The following is based on Figures 4a and 4b illustrated the inventive concept. So shows Figure 4a a band in which there are four spectral lines, all of the same size. The energy in this band is thus evenly distributed over the band. Against it shows Figure 4b a situation where the energy in the band resides in one spectral line while the other three spectral lines are the same are zero. This in Figure 4b For example, the band shown could be before quantization or could be obtained after quantization if the in Figure 4b Spectral lines set to zero before the quantization are smaller than the first quantizer stage and are thus set to zero by the quantizer, ie they do not "survive".

The number of active lines in Figure 4b is therefore equal to 1, with the parameter nl in Figure 4b to the square root of 2 is calculated. In contrast, the value nl, i.e. the measure for the spectral distribution of the energy in Figure 4a calculated to 4. This means that the spectral distribution of the energy is more uniform when the measure for the distribution of the spectral energy is larger.

It should be noted that the band-by-band calculation of the perceptual entropy according to the prior art does not find any difference between the two cases. In particular, no difference is found when in the two bands ending in Figures 4a and 4b shown that the same energy is present.

Obviously, however, the in Figure 4b The case shown can be coded with only one relevant line with fewer bits, since the three spectral lines set to zero can be transmitted very efficiently. Generally speaking, the simpler quantizability of the in Figure 4b The case shown is based on the fact that after the quantization and lossless coding, smaller values and, in particular, values quantized to zero require fewer bits for transmission.

According to the invention, it is thus taken into account how the energy is distributed within the band. This is done as it is carried out by replacing the number of lines per band in the familiar equation ( Fig. 6 ) by estimating the number of lines that are non-zero after quantization. This estimate is in Fig. 2a shown.

It should also be noted that the in Fig. 2a The form factor shown is also required elsewhere in the encoder, for example within the quantization block 1014 for determining the quantization step size. If the form factor is already calculated elsewhere, it does not have to be recalculated for bit estimation, so that the inventive concept for improved estimation of the measure for the required bits manages with a minimum of additional computing effort.

As has already been explained, X (k) is the spectral coefficient to be quantized later, while the variable kOffset (b) designates the first index in band b.

How it looks Figures 4a and 4b can be seen, gives the spectrum in Figure 4a a value nl = 4, while the spectrum in Figure 4b gives a value of 1.41. With the help of the form factor, a measure for the characterization of the spectral field structure within the band is available.

The new formula for calculating an improved band-wise perceptual entropy is based on the multiplication of the measure for the spectral distribution of the energy and the logarithm expression in which the signal energy e (b) occurs in the numerator and the permitted disturbance in the denominator, depending on requirements a term inserted within the logarithm can be, as it is already in Fig. 7 is shown. This term can, for example, also be 1.5, but can also be equal to zero, as in the in Figure 2b case shown, this being e.g. B. can be determined empirically.

At this point, open again Fig. 5 indicated, from which the perceptual entropy calculated according to the invention can be seen, plotted over the required bits. A higher accuracy of the estimation compared to the comparative examples in the Fig. 6 , 7th and 8th can be clearly seen. The modified band-by-band calculation according to the invention also performs at least equally as compared to the line-by-line calculation.

Depending on the circumstances, the method according to the invention can be implemented in hardware or in software. The implementation can take place on a digital storage medium, in particular a floppy disk or CD with electronically readable control signals which can interact with a programmable computer system in such a way that the method is carried out. In general, the invention thus also consists in a computer program product with a program code stored on a machine-readable carrier for carrying out the method according to the invention when the computer program product runs on a computer. In other words, the invention can thus be implemented as a computer program with a program code for carrying out the method when the computer program runs on a computer.

Claims (13)

1. Apparatus for determining an estimate of a need for information units for encoding a signal having audio or video information, the signal having several frequency bands, comprising:

   means (102) for providing a measure for an admissible interference for a frequency band of the signal, wherein the frequency band includes at least two spectral values of a spectral representation of the signal, and a measure for an energy of the signal in the frequency band;

   characterized by

   means (106) for calculating a measure for a distribution of the energy in the frequency band, wherein the distribution of the energy in the frequency band deviates from a completely uniform distribution; and

   means (104) for calculating the estimate using the measure for the interference, the measure for the energy, and the measure for the distribution of the energy.
2. Apparatus of claim 1, wherein the means (106) for calculating is configured to take magnitudes of spectral values in the frequency band into account for calculating the measure for the distribution of the energy.
3. Apparatus of claim 1 or 2, wherein the means (106) for calculating the measure for the distribution of the energy is configured to determine, as a measure for the distribution of the energy, a number of spectral values the magnitudes of which are greater than or equal to a predetermined magnitude threshold, or the magnitudes of which are smaller than or equal to the magnitude threshold.
4. Apparatus of claim 3, wherein the magnitude threshold is an exact or estimated quantizer stage causing, in a quantizer, values smaller than or equal to the quantizer stage to be quantized to zero.
5. Apparatus of one of the preceding claims, wherein the means (106) for calculating is configured to calculate a form factor according to the following equation: $$\mathit{ffac}\left(b\right)={\displaystyle {\sum }_{k=\mathit{kOffset}\left(b\right)}^{\mathit{kOffset}\left(b+1\right)-1}\sqrt{\left|X\left(k\right)\right|}},$$

   

   wherein X(k) is a spectral value at a frequency index k, wherein kOffset is a first spectral value in a band b, and wherein ffac(b) is the form factor.
6. Apparatus of one of the preceding claims,
   wherein the means (106) for calculating is configured to take a fourth root of a ratio between the energy in the frequency band and a width of the frequency band or number of spectral values within the frequency band into account.
7. Apparatus of one of the preceding claims,
   wherein the means (106) for calculating is configured to calculate the measure for the distribution of the energy according to the following equations: $$\mathit{nl}\left(b\right)=\frac{\mathit{ffac}\left(b\right)}{{\left(\frac{e\left(b\right)}{\mathit{width}\left(b\right)}\right)}^{0.25}}$$

   

   $$\mathit{ffac}\left(b\right)={\displaystyle {\sum }_{k=\mathit{kOffset}\left(b\right)}^{\mathit{kOffset}\left(b+1\right)-1}\sqrt{\left|X\left(k\right)\right|}},$$

   

   wherein X(k) is a spectral value at a frequency index k, wherein kOffset is a first spectral value in a band b, wherein ffac(b) is a form factor, wherein nl(b) represents the measure for the distribution of the energy in the band b, wherein e(b) is a signal energy in the band b, and wherein width(b) is a width of the band.
8. Apparatus of one of the preceding claims,
   wherein the means (104) for calculating the estimate is configured to use a quotient of the energy in the frequency band and the interference in the frequency band.
9. Apparatus of one of the preceding claims,
   wherein the means (104) for calculating the estimate is configured to calculate the estimate using the following expression: $$\mathit{pe}={\displaystyle \sum _b\mathit{nl}\left(b\right)\cdot {\log }_2\left(\frac{e\left(b\right)}{\mathit{nb}\left(b\right)}+s\right)}$$

   

   wherein pe is the estimate, wherein nl(b) represents the measure for the distribution of the energy in the band b, wherein e(b) is an energy of the signal in the band b, wherein nb(b) is the admissible interference in the band b, and wherein s is an additive term preferably equal to 1.5.
10. Apparatus of one of the preceding claims,
    wherein the means (104) for calculating the estimate is configured to calculate the estimate according to the following equation: $$\mathit{pe}={\displaystyle \sum _b\mathit{nl}\left(b\right)\cdot {\log }_2\left(\frac{e\left(b\right)}{\mathit{nb}\left(b\right)}+s\right)}$$

    

    wherein: $$\mathit{nl}\left(b\right)=\frac{\mathit{ffac}\left(b\right)}{{\left(\frac{e\left(b\right)}{\mathit{width}\left(b\right)}\right)}^{0.25}},$$

    

    and
    wherein: $$\mathit{ffac}\left(b\right)={\displaystyle {\sum }_{k=\mathit{kOffset}\left(b\right)}^{\mathit{kOffset}\left(b+1\right)-1}\sqrt{\left|X\left(k\right)\right|}},$$

    

    wherein pe is the estimate, wherein nl(b) represents the measure for the distribution of the energy in the band b, wherein e(b) is an energy of the signal in the band b, wherein nb(b) is the admissible interference in the band b, wherein s is an additive term preferably equal to 1.5, wherein X(k) is a spectral value at a frequency index k, wherein kOffset is a first spectral value in a band b, wherein ffac(b) is a form factor, and wherein width(b) is a width of the band.
11. Apparatus of one of the preceding claims,
    wherein the signal is given as a spectral representation with spectral values.
12. Method of determining an estimate of a need for information units for encoding a signal having audio or video information, the signal having several frequency bands, comprising the steps of:

    providing (102) a measure for an admissible interference for a frequency band of the signal, wherein the frequency band includes at least two spectral values of a spectral representation of the signal, and a measure for an energy of the signal in the frequency band;

    characterized by

    calculating (106) a measure for a distribution of the energy in the frequency band, wherein the distribution of the energy in the frequency band deviates from a completely uniform distribution; and

    calculating (104) the estimate using the measure for the interference, the measure for the energy, and the measure for the distribution of the energy.
13. Computer program with program code for performing the method of determining an estimate of a need for information units for encoding a signal of claim 12, when the program is executed on a computer.

Images