RU2802659C1 - Selective bass post-filter - Google Patents

Abstract

FIELD: computer technology for decoding video data.

SUBSTANCE: decoding of the bitstream signal as a preliminary temporal audio signal in one of the plurality of decoding modes, the said plurality of decoding modes including at least one first decoding mode including the post-filtering stage, and at least one second decoding mode not including the post-filtering stage. The post-filtering stage applies a tone boost filter to the preliminary temporal audio signal, thereby obtaining a temporal audio signal, wherein the tone boost filter is controlled by the post-filter gain, wherein in the said at least one first decoding mode the post-filtering step is selectively turned off in response only to the post-filtering information encoded in the bitstream signal by setting the post-filter gain to zero. The post-filtering information indicates the decision on the encoder side as to whether to disable post-filtering.

EFFECT: increased accuracy of reproduction of the sound signal.

7 cl, 11 dwg

Description

Field of technical application

The present invention generally relates to digital audio coding and, more precisely, to methods for encoding audio signals containing components of different natures.

BACKGROUND OF THE INVENTION

A widely used class of coding methods for audio signals containing speech or singing includes code-excited linear prediction (CELP), used in time interleave with other coding methods, including frequency domain coding methods, especially those adapted for music, or in general ways to take into account changes in character between successive periods of time of the sound signal. For example, a simplified decoder according to the Moving Picture Experts Group (MPEG) standard for Unified Voice and Audio Coding (USAC; see ISO/IEC 23003-3) as shown at the top of the accompanying FIG. 2 can operate in at least three decoding modes: advanced audio coding (AAC; see ISO/IEC 13818-7 standard), algebraic CELP (ACELP) mode, and transform coded excitation (TCX) mode.

Various embodiments of CELP are adapted to the properties of the human speech organs and possibly to the human auditory perception. As used herein, the term "CELP" will refer to all possible implementations and variations, including, but not limited to, wideband and narrowband CELP, SB-CELP (subband CELP), CELP with low and high speed, RCELP (Attenuated CELP), LD-CELP (Low Latency CELP), CS-CELP (Coupled CELP), CS-ACELP (Coupled ACELP), PSI-CELP (Synchronous Core Innovative CELP) tone) and VSELP (vector sum excitation linear prediction). The principles of CELP are discussed by R. Schroeder and S. Atal in Proceedings of the IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), vol. 10, pp. 937-940, 1985, and some of its applications are described in references 25-29, cited in Chen and Gersho, IEEE Transactions on Speech and Audio Processing, vol. 3, no. 1, 1995. As described in more detail in the first article, a CELP decoder (or, similarly, a CELP speech synthesizer) may include a pitch predictor that recovers the periodic component of the encoded speech signal, and a pulse codebook from which a generator sequence is added. The pitch predictor may, in turn, include a long-delay predictor designed to recover formants using special spectral envelope shaping. In this context, fundamental tone is understood as the fundamental frequency of the tonal component of sound generated by the vocal cords and further colored by the resonant portions of the vocal tract. This frequency, along with its harmonics, will dominate speech and singing. Generally speaking, CELP methods are best suited for processing solo, or single-voice, singing, for which the pitch frequency is well defined and relatively easy to determine.

To improve the perceived quality of CELP-encoded speech, a common practice is to combine CELP with post-filtering (or, in other words, pitch enhancement). US Patent No. 4,969,192 and Section II of the Chen and Gersho paper disclose the desired properties of such post-filters, namely their ability to suppress noise components located between the harmonics of the detected voice pitch (long-term part, see Section IV). It is assumed that a significant portion of this noise is the result of spectral envelope shaping. The long-term part of a simple post filter can be designed to have the following transfer function:

,

where T is the estimated pitch period in terms of the number of samples, and α is the post-filter gain, as shown in FIG. 1 and 2. In a manner similar to a comb filter, such a filter attenuates the frequencies 1/(2 T ), 3/(2 T ), 5/(2 T ), which are in the middle between the harmonics of the fundamental frequency, and neighboring frequencies. The attenuation depends on the value of the gain α . Slightly more complex post filters apply this gain only to the low frequencies—hence the commonly used term “bass post filter”—where noise is perceived the most. This can be expressed as cascading the transfer function H _E described above and the low pass filter H _LP . The post-processed decoded S _E created by the post-filter will look like this in the transformation area:

, Where

,

and S is the decoded signal that is fed to the post-filter input. Fig. 3 shows an embodiment of a post-filter with such properties, which is further discussed in section 6.1.3 of the ETSI TS 126 290 technical specification, version 6.3.0, publication 6. As this figure shows, pitch information is encoded as a parameter in the bitstream signal and extracted a pitch tracking module communicatively connected to a long-term prediction filter performing operations expressed as P _LT .

The long-term part described in the previous paragraph can be used separately. Alternatively, it is placed in series with a noise conversion filter, which preserves components in the frequency ranges corresponding to the formants, and attenuates noise in other regions of the spectrum (short-term part, see section III), i.e. in the “spectrum valleys” of the formant envelope. As another possible modification, this filter bank is also supplemented with a sequential high-pass type filter to reduce the perceived degradation due to short-term spectral slope.

Sound signals containing a mixture of components of different origins - for example, tonal, non-tonal, vocal, instrumental, non-musical - are not always reproduced satisfactorily by available encoding technologies. It would be more accurate to note that the available technologies are insufficient to process such heterogeneous sound material and, as a rule, give preference to one of the components to the detriment of the others. In particular, music containing single or multi-instrumental singing or choral passages encoded in the manner described above will often be decoded with perceptible artifacts that detract from some of the listening experience.

Brief description of the invention

In order to alleviate at least some of the disadvantages described in the previous section, it is an object of the present invention to provide methods and devices adapted for audio encoding and decoding signals containing a mixture of components of different origins. As a particular object, the invention seeks to provide methods and devices that are suitable in terms of encoding efficiency and (perceived) fidelity, or both.

The invention achieves at least one of these objects by providing an encoder system, a decoder system, an encoding method, a decoding method, and computer program products for implementing each of these methods, which are defined in the independent claims. Dependent claims define embodiments of the invention.

The inventors have realized that some artifacts perceived in decoded audio signals of heterogeneous origin are a consequence of inappropriate switching between different encoding modes, at least one of which includes post-filtering in the decoder, and at least one of which does not. More precisely, the available post filters remove not only interharmonic noise (and, where applicable, noise in the valleys of the spectrum), but also signal components representing instrumental or vocal accompaniment and other material of a “desirable” nature. The fact that the audibility threshold in the spectral troughs can be as high as 10 dB (as noted by Ghitza and Goldstein, IEEE Trans. Acoust, Speech, Signal Processing, vol. ASSP-4, pp. 697-708, 1986) can be considered a justification for many developers, which allows you to strictly filter these frequency bands. However, the degradation due to interharmonic attenuation (and valley attenuation) itself may be less important than the degradation due to switching. When the post filter is turned on, the background sounds of the singing voice are suddenly muted, and when the filter is deactivated, the background immediately becomes more sonorous. If switching occurs frequently due to the nature of the audio signal or the configuration of the encoder, switching artifact will occur. As an example, the USAC decoder can operate in either ACELP mode combined with post-filtering, or TCX mode without post-filtering. The ACELP mode is used in fragments where there is a dominant vocal component. Therefore, switching to ACELP mode can be triggered at the beginning of singing, such as at the beginning of a new musical phrase, at the beginning of a new verse, or simply after a passage where the accompaniment is considered to drown out the singing voice in the sense that the vocal component is no longer prominent. Experiments have confirmed that the alternative solution, or workaround, in which TLC coding is used throughout (and ACELP mode is disabled) does not correct the problem, since artifacts like artificial echo appear.

Accordingly, in the first and second aspects, the invention provides an audio encoding method (and an audio encoding system with associated features) which is characterized in that a decision is made as to whether a device that will decode a bit stream output from the encoding method should apply postfiltration, including attenuation of interharmonic noise. The result of this decision is encoded in a bitstream and is available to the decoder.

According to the invention, the decision about whether to use a post filter is made separately from the decision about the most appropriate encoding mode. This makes it possible to maintain one post-filtering state for an entire period of such duration that switching will not annoy the listener. Thus, the encoding method may dictate that the post-filter be maintained in an inactive state even when it switches to an encoding mode where the filter would normally be active.

It should be noted that the decision on whether to apply post-filtering is usually made on a frame-by-frame basis. Therefore, firstly, postfiltering is not applied to less than one frame at a time. Second, the decision about whether to turn off post-filtering is valid only for the duration of the current frame, and for the next frame it can either be saved or revised. In an encoding format that allows a basic frame format and a shortened format that is a fraction of the normal format, for example 1/8 of its length, there may be no need to make post-filtering decisions for individual shortened frames. Instead, a number of reduced frames summed to a normal frame may be considered, and parameters related to the filtering decision may be obtained by calculating the average or median of the reduced frames included therein.

The third and fourth aspects of the invention provide an audio decoding method (and an audio decoding system that includes corresponding features), wherein the decoding step is followed by a post-filtering step that includes inter-harmonic noise attenuation and is characterized in that it includes a post-filter off step in accordance with information about post-filtering encoded in the bitstream signal.

Due to its ability to deactivate the post-filter depending only on the post-filtering information, i.e. Regardless of factors such as the current encoding mode, a decoding method with such characteristics is well suited for encoding mixed-origin audio signals. When using encoding methods in which post-filter activity is traditionally associated with specific encoding modes, the ability to disable post-filtering makes possible a new operating mode, namely, the use without filtering of a traditional filtered decoding mode.

In a further aspect, the invention also provides a computer program product for performing one of the methods described above. The invention further provides a post-filter for attenuating inter-harmonic noise, which can operate in either an active mode or a pass-through mode, as indicated by a post-filtering signal supplied to the post-filter. The post-filter may include a decision section for autonomously controlling the activity of the post-filter.

As is clear to specialists, an encoder adapted to work together with a decoder is equipped with functionally equivalent modules so that it allows reliable reproduction of the encoded signal. Said equivalent modules may be identical or similar modules or modules having identical or similar transmission characteristics. In particular, the encoder and decoder modules, respectively, may be similar or dissimilar processing units executing corresponding computer programs that perform equivalent sets of mathematical operations.

In one embodiment of the invention, the present method includes deciding whether a post-filter includes spectral trough attenuation (relative to the formant envelope, see above). This corresponds to the short-term part of the post-filter. Therefore, it is preferable to adapt the criterion on which the decision is based to the essence of the post-filter.

One embodiment of the invention is directed to an encoder, particularly one adapted for speech encoding. Since some of the problems motivating the invention were observed when a mixture of vocal and other components was encoded, the independent post-filtering decision making provided by the invention is particularly advantageous. In particular, such a decoder may include a code-excited linear prediction encoding module.

In one embodiment of the invention, the encoder bases its decision on the detected simultaneous presence of a signal component with a dominant fundamental frequency (the pitch) and another signal component below the fundamental frequency. Detection may also be aimed at finding the co-occurrence of a component with a dominant fundamental frequency and another component with energy between the harmonics of that fundamental frequency. In such a situation, artifacts of the type in question are often encountered. Then, if the specified simultaneous presence is established, the encoder will decide that post-filtering is not appropriate, which will be indicated accordingly in the post-filtering information contained in the bitstream.

One embodiment of the invention uses as a detection criterion the total signal power content of the temporal audio signal below the pitch frequency, where the pitch frequency is possibly estimated at the encoder using long-term prediction. If the specified power is greater than a predetermined threshold value, other significant components in addition to the pitch component (including harmonics) are considered to be present, which will cause the post-filter to turn off.

An encoder containing a CELP module can benefit from the module estimating the pitch frequency of the temporal audio signal. A further detection criterion, as described in more detail above, is then to check the energy reserve between harmonics at or below that frequency.

As a further development of the prior embodiment of the invention including the CELP module, decision making may involve comparing the estimated energy of the audio signal when it is CELP encoded (ie, encoded and decoded) with the estimated energy of the audio signal when it is CELP encoded and post-filtered. If the energy difference is greater than a threshold value, which may indicate that significant non-noise component of the signal will be lost, the encoder will decide to disable the post-filter.

In an advantageous embodiment of the invention, the encoder comprises a CELP module and a TCX module. As is known in the art, TLC coding is advantageous with respect to certain types of signals, particularly non-vocal signals. It is not common practice to apply post-filtering to a TLC-encoded signal. Therefore, the encoder can choose either TLC encoding, CELP encoding with post-filtering, or CELP encoding without post-filtering, thus covering a significant range of signal types.

As one of the further developments of the previous embodiment of the invention, the decision to choose between the three coding modes is made based on the criterion of distortion depending on the data rate, i.e. using and optimizing a procedure that is itself known in the art.

In another further development of the previous embodiment, the encoder also includes an encoder according to the advanced audio coding (AAC) standard, which is also known to be particularly well suited for certain types of signals. Preferably, the decision about whether to use AAC coding (in the frequency domain) is made independently of the decision about which other (linear prediction) modes to use. Therefore, such an encoder can be understood as operating in two super modes: AAC or TCX/CELP, where in the latter mode the encoder will choose between TCX, CELP with post-filtering and CELP without filtering. This embodiment of the invention allows for an even wider range of audio signal types to be processed.

In one embodiment of the invention, the encoder may decide that post-filtering during decoding should be applied gradually, i.e. with gradually increasing gain. Likewise, he may decide that post-filtering will be phased out. This gradual application and elimination makes it less noticeable to switch between modes with and without post-filtering. For example, a vocal fragment for which CELP encoding with post-filtering is found suitable may be preceded by an instrumental fragment where TLC encoding is optimal; the decoder of the invention can then apply post-filtering gradually near or at the beginning of the vocal passage so that the benefits of post-filtering are retained while annoying switching artifacts are eliminated.

In one embodiment of the invention, the decision about whether to apply postfiltering is based on an approximation of the difference signal, which serves as an approximation of the signal component that will be removed by the postfilter from the future decoded signal.

As one possibility, the approximate difference signal is calculated as the difference between the time audio signal and the post-filtered time audio signal. As another possibility, the encoding section extracts the intermediate decoded signal so that an approximate difference signal can be calculated as the difference between the temporary audio signal and the post-filtered intermediate decoded signal. The intermediate decoded signal may be stored in the encoder's long-term prediction buffer. It can then represent the excitation signal under the assumption that further synthesis filtering (vocal tract, resonances) will be required to obtain the final decoded signal. The point of using an intermediate decoded signal is that it captures some of the details, particularly weaknesses, of the encoding method and thus allows a more realistic assessment of the effect of the post-filter. As a third possibility, the decoding section extracts the intermediate decoded signal, whereby an approximate difference signal can be calculated as the difference between the intermediate decoded signal and the post-filtered intermediate decoded signal. This procedure probably gives a less reliable estimate than the first two possibilities, but, on the other hand, it can be carried out autonomously by the decoder.

The approximate difference signal thus obtained is then evaluated with respect to one of the following criteria, which, when established in the affirmative sense, will lead to a decision to disable the post-filter:

a) whether the power of the approximate difference signal exceeds a predetermined threshold value indicating that a significant portion of the signal can be eliminated by the post filter;

b) whether the nature of the approximate difference signal is tonal rather than noise-like;

c) whether the difference between the frequency spectra of the absolute values of the approximate difference signal and the temporal audio signal is unevenly distributed relative to frequency, which shows that it is not so much noise as a signal that could be meaningful to the listener;

d) whether the frequency spectrum of the absolute values of the approximate difference signal is localized to frequency intervals that fall within a predetermined envelope of significance based on what can generally be expected from the type of signal being processed; And

e) whether the frequency spectrum of the absolute values of the approximate difference signal is localized in frequency intervals that are within a predetermined significance envelope obtained by thresholding the frequency spectrum of the absolute values of the temporary audio signal by the absolute value of the largest component of the signal, here scaled down by a predetermined scale factor .

When assessing criterion e), it is preferable to use peak tracking in the spectrum of absolute values, i.e. distinguishing peak-shaped areas usually associated with tonal components rather than noise. Components identified by peak tracking, which may take place through some algorithm that is itself known in the art, can be further sorted using a peak height threshold so that the remaining components represent tonal material with a certain absolute value. These components usually represent the meaningful content of the signal rather than noise, which motivates the decision to disable the post filter.

In one embodiment of the invention as a decoder, the decision to disable the post-filter is made by a switch controlled by the control section and capable of bypassing the post-filter in the circuit. In another embodiment of the invention, the post-filter has a variable gain controlled by a control section, or a gain control therein, where the cut-off decision is made by setting the post-filter gain (see previous section) to zero or by setting it to an absolute value below a predetermined threshold value.

In one embodiment of the invention, decoding according to the present invention includes extracting post-filtering information from the decoded bitstream signal. More specifically, the post-filtering information may be encoded in a data field including at least one bit in a format suitable for transmission. Advantageously, the data field is a pre-existing data field defined by the applicable standard, but is not in use, and thus the post-filtering information does not add to the transmission payload.

It should be noted that the methods and apparatus disclosed in this section can be used, with appropriate modifications within the capabilities of one skilled in the art, including through routine experimentation, to encode signals containing multiple components, possibly belonging to different channels, such as stereo channels. Throughout this application, the terms “pitch enhancement” and “post-filtering” are used interchangeably. It should also be noted that AAC is discussed as a representative example of frequency domain coding techniques. Indeed, application of the invention to a decoder or encoder other than AAC operating in a frequency domain coding mode will require few, if any, modifications within the capabilities of one skilled in the art. Similarly, TLC is mentioned as an example of weighted linear prediction transform coding or transform coding in general.

Features from two or more embodiments of the invention described above may be combined, if not explicitly additional, into further embodiments of the invention. The fact that two characteristic features are cited in different claims does not prevent them from being advantageously combined. Likewise, further embodiments of the invention may also be provided with the omission of certain characteristic features that are not necessary or essential for the desired purpose.

Brief description of graphic materials

Embodiments of the present invention will be described below with reference to the accompanying drawings, in which:

fig. 1 is a block diagram showing a traditional post-filter decoder;

fig. 2 is a schematic block diagram of a traditional decoder operating in AAC, ACELP and TCX modes and including a post filter permanently connected in the downstream direction relative to the ACELP module;

fig. 3 – block diagram illustrating the design of the post-filter.

fig. 4 and 5 are block diagrams of two decoders according to the invention;

fig. 6 and 7 are block diagrams illustrating the differences between a conventional decoder (FIG. 6) and a decoder (FIG. 7) according to the invention;

fig. 8 – block diagram of the encoder according to the invention;

fig. 9 and 10 are block diagrams illustrating the differences between a conventional decoder (FIG. 9) and a decoder (FIG. 10) according to the invention; And

fig. 11 is a block diagram of a stand-alone post-filter that can be selectively activated and deactivated.

Detailed Description of Embodiments of the Invention

Fig. 4 is a schematic illustration of a decoder system 400 according to one embodiment of the invention having a bitstream signal as an input and an audio signal as an output. As with the traditional decoders shown in FIG. 1, the post filter 440 is located downstream of the decoding module 410, but can be turned on or off from the decoding path by operating switch 442. In the position shown in the figure, the post filter is turned on. This can be disabled by setting the switch to the opposite position, whereby the signal from the decoding module 410 will be routed through the bypass bus 444 instead. As an inventive contribution, the switch 442 is controlled by the post-filtering information contained in the bitstream signal, and thus the post-filtering can be applied and removed regardless of the current state of the decoding module 410. Because the post filter 440 operates with some delay, such as the post filter shown in FIG. 3 introduces a delay of an amount equal to at least the period T of the pitch - in order to maintain the modules in synchronized conditions when switching, a delay compensation module 443 is located on the bypass bus 444. Delay module 443 delays the signal for the same period that post filter 440 would delay it. To minimize switching time, delay compensation module 443 receives the same signal as post filter 440 at all times. In an alternative embodiment, where post filter 440 is replaced by a zero-delay post-filter (eg, a causal filter, such as a two-stage filter that is independent of future signal values), the delay compensation module 443 may be omitted.

Fig. 5 illustrates a further embodiment of the invention in accordance with the teachings of the invention with respect to the tri-mode decoder system 500 of FIG. 2. The ACELP decoding module 511 is located in parallel with the TLC decoding module 512 and the AAC decoding module 513. In series with the ACELP decoding module 511 is a post filter 540 designed to attenuate noise, in particular noise located between the harmonics of the pitch frequency, directly or indirectly obtained from the bitstream signal, for which the decoder system 500 is adapted. The bitstream signal also encodes post-filtering information adjusting the positions of the upper switch 541, which operates to turn off the post-filter 540 from the processing path and replace it with error compensation 543, as in FIG. 4. The bottom switch 542 is used to switch between different decoding modes. In this design, when one of the 512, 513, TLC or AAC modules is used, the position of the top switch 541 does not matter; therefore, the post-filtering information does not necessarily indicate this position except in ACELP mode. Whichever decoding mode is currently in use, the signal is supplied from the connection point of the lower switch 542 in the downstream direction to the spectral band replication (SBR) module 550, which outputs the audio signal. Those skilled in the art will appreciate that this illustration is of a conceptual nature, as is evident particularly from the fact that the switches are shown schematically as separate physical entities with movable contact means. In a possible real-life implementation of a decoder system, the switches, as well as other modules, can be implemented via machine-readable instructions.

Fig. 6 and 7 also provide block diagrams of two tri-mode decoder systems operating in ACELP, TCX and frequency domain decoding modes. With reference to the last-mentioned figure, which shows an embodiment of the invention, a bitstream signal is supplied to an input point 701, which in turn is permanently connected via corresponding branches to three decoding modules 711, 712, 713. Input point 701 also includes a connecting branch 702 (not present in the conventional decoding system of FIG. 6) directed to a pitch enhancer 740, which plays the role of the generalized post-filter described above. As is common practice in the art, located downstream of the ACELP and TCX modules 711, 712 is a first transition framing module 703 for transitioning between decoding modules. The second transition module 704 is located downstream of the frequency domain decoding module 713 and the first transition framing module 703 and is configured to transition between two super modes. In the upstream direction, immediately before the output point 705, an SBR module 750 is located. It is clear that the bitstream signal is supplied directly (or, if necessary, after demultiplexing) to all three decoding modules 711, 712, 713 and to the pitch amplification module 740. The information contained in the bitstream controls which decoding module will be active. However, according to the invention, pitch amplification module 740 performs similar self-activation and, in response to post-filtering information in the bitstream, can act as a post-filter or simply as a pass channel. For example, this may be accomplished by providing the pitch enhancer 740 with a control section (not shown) through which the post-filtering action can be turned on or off. When the system operates in the frequency domain decoding mode or in the TLC decoding mode, where, strictly speaking, any post-filtering information is not necessary, the pitch amplification module 740 is always in the pass-through mode. It should be understood that modules which do not form part of the inventive contribution and whose presence is obvious to those skilled in the art, such as a demultiplexer, have been omitted from FIG. 7 and other similar illustrations.

Alternatively, the decoder system of FIG. 7 may be equipped with a control module (not shown) to decide whether post-filtering should be applied using an analysis-by-synthesis approach. Said control module is in communicative communication with the pitch amplification module 740 and with the ACELP module 711, from which it extracts intermediate decoded signals , representing one of the intermediate stages of the decoding process, preferably a stage corresponding to signal excitation. The detection module contains the necessary information to simulate the action of the pitch amplification module 740, as determined by the transfer functions P _LT ( z ) and H _LP ( z ) (see Background and FIG. 3), or equivalently, their filter impulse response responses p _LT (z) and hLP ₍ n ). As follows from the discussion in the Background section, the component that will be subtracted by post-filtering can _{be estimated using an approximate difference signal sAD} ( n ), which is proportional , where * denotes discrete convolution. This approximation is an approximation of the true difference between the original audio signal and the post-filtered decoded signal, as follows:

,

where α is the post-filter gain. By studying the total energy, low band energy, tone, current absolute value spectrum, or the past absolute value spectrum of this signal, as disclosed in the Summary of the Invention and the claims, the control section can find a basis for deciding whether to activate or deactivate the module 740 pitch gain.

Fig. 8 shows an encoder system 800 according to one embodiment of the invention. The encoder system 800 is adapted to process digital audio signals, which are generally produced by collecting sound wave data with a microphone and converting the wave into an analog electrical signal. The electrical signal is then sampled into a digital signal capable of being transmitted in a suitable format to the encoder system 800. In general, the system consists of an encoding module 810, a decision module 820, and a multiplexer 830. Using switches 814, 815 (represented symbolically), the encoding module can operate in either the CELP mode, the TCX mode, or the AAC mode by selective casting. into operation of modules 811, 812, 813. Decision module 820 applies one or more predefined criteria to decide whether to disable post-filtering during decoding of the bit stream signal generated by encoder system 800 will encode the audio signal. To do this, the decision module 820 may examine the audio signal directly or may receive data from the encoding module 810 via communication line 816. The signal indicative of the decision made by the decision module 820, together with the encoded audio signal from the encoding module 810, is delivered to the multiplexer 830. , which concatenates the signals into a bitstream that constitutes the output of the decoder system 800.

Preferably, decision module 820 bases its decision on an approximate difference signal calculated from the intermediate encoded signal , which can be subtracted from the decoding unit 810. The intermediate decoded signal represents an intermediate step in the decoding process, as discussed in previous paragraphs, but can be extracted from a corresponding step in the encoding process. However, the original audio signal S _ORIG is available to the encoder system 800, so preferably the approximate difference signal is generated as:

.

This approximation is based on the fact that an intermediate decoded signal is used instead of the final decoded signal. This allows the essence of the component that the postfilter might remove during decoding to be assessed and, using one of the criteria discussed in the Summary of the Invention , decision module 820 will be able to decide whether to disable postfiltering.

Alternatively, decision module 820 may use the original signal instead of the intermediate encoded signal, and thus the approximate difference signal will be the signal . This approximation is likely to be less reliable, but on the other hand, it makes the presence of a communication link 816 between the decision module 820 and the encoding module 810 unnecessary.

In those other variations of this embodiment where decision module 820 examines the audio signal directly, one or more of the following criteria may be applied:

• Does the audio signal contain both a component at a dominant fundamental frequency and a component below the fundamental frequency? (The fundamental frequency may be supplied as a by-product of the encoding unit 810).

• Does the audio signal contain both a component with a dominant fundamental frequency and a component located between the harmonics of the fundamental frequency?

• Does the audio signal contain significant signal energy below the fundamental frequency?

• Is decoding with post-filtering (as a possible outcome) preferable to decoding without filtering in terms of optimal distortion versus data rate?

In all of the described changes to the encoder design shown in FIG. 8, – i.e. Regardless of the basis of the detection criterion, decision section 820 may be connected to decide whether to gradually start or fade out post-filtering in order to achieve smooth transitions. Gradual onset and release can be controlled by adjusting the post-filter gain.

Fig. 9 shows a conventional decoder operating in a frequency decoding mode and a CELP decoding mode depending on a bit stream signal supplied to the decoder. Post-filtering is applied whenever CELP decoding mode is selected. An improvement to this decoder is illustrated in FIG. 10, which shows a decoder 1000 according to one embodiment of the invention. This decoder operates not only in the frequency domain decoding mode, where the frequency domain decoding unit 1013 is active, and in the CELP decoding mode with filtering, where the CELP decoding unit 1011 and the post filter 1040 are active, but also in the CELP mode without filtering, in which the CELP module 1011 provides a signal to the delay compensation module 1043 via bypass bus 1044. Switch 1042 controls which decoding mode is currently active in response to post-filtering information contained in the bitstream signal delivered to decoder 1000. In this decoder and the decoder of FIG. 9, the final processing step is performed by the 1050 SBR module, whose output is the final audio signal.

Fig. 11 shows a post filter 1100 suitable for placement downstream of a decoder 1199. The filter 1100 includes a post filtering module 1140 that is enabled or disabled by a control module (not shown), especially a binary or non-binary gain control, in response to a post filtering signal received from decision module 1120 in post-filter 1100. The decision module performs one or more checks on the signal received from the decoder, reaching a decision as to whether the post-filter module is active or inactive. The decision may be made in the direction of the functionality of the decision module 820 of FIG. 8, which uses the original signal and/or the intermediate encoded signal to predict the action of the post-filter. The decision of the decision module 1120 may also be based on information similar to that which the decision modules use in those embodiments of the invention where an intermediate decoded signal is generated. For example, decision module 1120 may estimate the pitch frequency (unless it is easily extracted from the bitstream signal) and calculate the amount of energy in the signal below the pitch frequency and between its harmonics. If this energy reserve is significant, it likely represents a significant component of the signal rather than noise, which motivates the decision to turn off the post-filtering module 1140.

Six subjects were subjected to a listening test in which pieces of music encoded and decoded in accordance with the invention were compared with reference to pieces containing the same music post-filtered in the traditional manner, but leaving all other parameters unchanged. The results confirm the improvement in perceived quality.

Further embodiments of the present invention will become apparent to those skilled in the art after reading the above description. Although the present description and drawings disclose embodiments and examples, the invention is not limited to these specific examples. Without departing from the scope of the present invention, as defined by the accompanying claims, numerous modifications and changes may be made.

The systems and methods disclosed above may be implemented as software, firmware, hardware, or a combination thereof. Some or all of the components may be implemented as software executed by a digital signal processor or microprocessor, or they may be implemented as firmware or a special purpose integrated circuit. The software may be distributed on a computer-readable storage medium, which may include a computer storage medium (or non-transmitting medium) or a communications medium (or transient medium). As is well known to those skilled in the art, computer storage media includes volatile and non-volatile, removable and non-removable storage media implemented in any method or technology for storing information, such as computer-readable instructions, data structures, program modules or other data. Computer storage media includes, but is not limited to, RAM, ROM, EE-PROM, flash memory or other memory technologies, CD-ROM, digital versatile disk (DVD) or other optical disk storage media, magnetic tape cassettes, magnetic tape, magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to store the necessary information and that can be accessed by a computer. Moreover, as is well known to those skilled in the art, a communication medium typically embodies machine-readable instructions, data structures, program modules, or other data in a modulated information signal, such as a carrier wave or other data transmission mechanism, and includes any means of delivering information.

List of embodiments of the invention

In a first embodiment, a decoder system (400; 500; 700; 1000) is provided for decoding a bitstream signal as a temporary audio signal, which includes:

a decoding section (410; 511, 512, 513; 711, 712, 713; 1011, 1013) for decoding a bit stream signal as a preliminary time audio signal; And

post-filter (440; 540; 740; 1040) interharmonic noise attenuation, designed to filter the preliminary temporal audio signal to obtain a time-dependent audio signal,

characterized by having a control section adapted to disable the post-filter in response to post-filter information encoded in the bitstream signal, wherein the pre-temporal audio signal is output as a temporal audio signal.

In the second embodiment, the post-filter of the decoder system of the first embodiment is also adapted to attenuate noise located in the valleys of the spectrum.

In the third embodiment, the control section of the decoder system of the first embodiment includes a switch (442; 541; 1042) for selectively excluding the post-filter from the signal processing path of the decoder system, whereby the post-filter is disabled.

In the fourth embodiment, the post filter of the decoder system of the first embodiment has a variable gain determining the interharmonic attenuation, and the control section includes a gain control operative to set the absolute value of the gain below a predetermined threshold value, whereby the post filter is turned off.

In the fifth embodiment, said decoding section of the decoder system of the first embodiment includes a speech decoding module.

In the sixth embodiment, said decoding section of the decoder system of the first embodiment includes a code-excited linear prediction (CELP) decoding unit (511; 711; 1011).

In the seventh embodiment, in the decoder system of the fifth embodiment, the pitch frequency estimated in the long-term prediction section in the encoder is encoded into a bit stream signal.

In the eighth embodiment, the post-filter of the decoder system is adapted to attenuate spectral components located between the harmonics of the fundamental frequency.

In the ninth embodiment, in the decoder system of the first embodiment, the bit stream signal contains a pitch frequency representation, and a post filter is adapted to attenuate spectral components located between harmonics of the pitch frequency.

In the tenth embodiment, the post-filter of the decoder system of the eighth or ninth embodiment is adapted to attenuate only those spectral components that are below a predetermined cutoff frequency.

In the eleventh embodiment, in the decoder system of the sixth embodiment 6:

The decoding section also includes a coded excitation conversion mode decoding module (512; 712) for decoding the bit stream signal as a temporary audio signal,

the control section is adapted to operate the decoder system in at least three of the following modes:

a) the TLC module is connected and the post filter is disabled;

b) CELP module and post filter connected; And

c) The CELP module is connected and the post filter is disabled where the pre-time signal and the audio time signal coincide.

In the twelfth embodiment, in the decoder system of the tenth embodiment:

The decoding section also includes an advanced audio coding (AAC) decoding module (513; 713) for decoding the bit stream signal as a temporary audio signal,

The control section is adapted to operate the decoder also in the following mode:

d) the AAC module is connected and the post filter is disabled.

In the thirteenth embodiment, in the decoder system of the first embodiment, the bitstream signal is segmented into time frames, and the control section is adapted to turn off an entire time frame or a sequence of entire time frames.

In the fourteenth embodiment, in the decoder system of the thirteenth embodiment, the control section is also adapted to receive, for each time frame in the MPEG Moving Picture Experts Group standard bitstream, a data field associated with that time frame and acting in response to the value of this data field to disable the post-filter.

In the fifteenth embodiment, in the decoder system of the fourth embodiment, the control section is adapted to gradually decrease and/or increase the gain of the post filter.

In a sixteenth embodiment, a (400; 500; 700; 1000) decoder system is provided, which includes:

a decoding section (410; 511, 512, 513; 711, 712, 713; 1011, 1013) for decoding a bit stream signal as a preliminary time audio signal; And

post-filter (440; 540; 740; 1040) attenuation of interharmonic noise, designed to filter the preliminary temporary audio signal in order to obtain a temporary audio signal,

characterized in that

a decoding section is adapted to generate an intermediate decoded signal representing the excitation and to supply it to the control section; And

the control section is adapted to calculate an approximate difference signal that serves as an approximation of the signal component to be post-filtered from the decoded signal, as the difference between the intermediate decoded signal and the intermediate post-filtered decoded signal, and to evaluate at least one of the following criteria:

a) whether the power of the approximate difference signal exceeds a predetermined threshold;

b) whether the approximate difference signal is tonal in nature;

c) whether the difference between the frequency spectra of the absolute values of the approximate difference signal and the temporal audio signal is unevenly distributed relative to frequency;

d) whether the frequency spectrum of the absolute values of the approximate difference signal is localized in frequency intervals that are within a predetermined significance envelope; And

e) whether the frequency spectrum of the absolute values of the approximate difference signal is localized in frequency intervals that are within the significance envelope obtained by thresholding the frequency spectrum of the absolute values of the temporary audio signal by the absolute value of the largest component of the signal, here reduced in scale with a predetermined scale factor;

and, in response to the affirmative determination, to turn off the post-filter, whereby the pre-temporal audio signal is output as a temporary audio signal.

In a seventeenth embodiment, an interharmonic noise reduction post filter (440; 550; 740; 1040; 1140) is provided, adapted to receive an input signal that includes a pre-audio signal, and to provide an output audio signal,

characterized by the presence of a control section designed to selectively, in accordance with the value of the post-filtering signal, activate the post-filter in one of the following modes:

i) Filtering mode, where it filters the pre-audio signal to produce a filtered signal and provides it as an output audio signal; And

ii) Passthrough mode, where it provides a pre-audio signal as an output audio signal.

In the eighteenth embodiment, in the post-filter of the seventeenth embodiment, the post-filtering signal is contained in the input signal.

In the nineteenth embodiment, the post-filter of the seventeenth embodiment also includes a decision module (1120) adapted to estimate the pitch frequency of the pre-audio signal and to evaluate at least one of the following criteria:

a) whether the power of spectral components below the fundamental frequency exceeds a predetermined threshold value;

b) whether the spectral components below the fundamental frequency are tonal;

c) whether the power of spectral components between harmonics of the fundamental frequency exceeds a predetermined threshold value; And

d) whether the spectral components between the harmonics of the fundamental frequency are tonal;

and, in response to a positive determination, to decide to generate a negative post-filtering signal, disabling the post-filter.

In the twentieth embodiment, there is provided a method for decoding a bit stream signal as a temporary audio signal, including the steps of:

decoding the bit stream signal as a preliminary time audio signal; And

post-filtering the preliminary temporal audio signal by attenuating the interharmonic noise, whereby a temporal audio signal is obtained,

characterized in that the post-filtering step is selectively skipped in response to post-filtering information encoded in the bitstream signal.

In the twenty-first embodiment, in the method of the twentieth embodiment, the post-filtering step also includes attenuation of noise located in the valleys of the spectrum.

In the twenty-second embodiment, in the method of the twenty-first embodiment, the decoding step includes applying an encoding method adapted for speech encoding.

In the twenty-third embodiment, in the method of the twentieth embodiment, the decoding step includes applying code-excited linear prediction, CELP, decoding.

In the twenty-fourth embodiment, in the method of the twenty-second or twenty-third embodiment, the post-filtering step includes attenuating spectral components located between harmonics of a pitch frequency, where the pitch frequency is extracted from the bitstream signal or estimated in a decoding step.

In the twenty-fifth embodiment, in the method of the twentieth embodiment, the post-filtering step includes attenuating only those spectral components that are below a predetermined cutoff frequency.

In the twenty-sixth embodiment, in the method of the twenty-third embodiment, the decoding and post-filtering steps selectively perform one of the following:

a) TCX decoding;

b) CELP decoding with post-filtering; And

c) CELP decoding without post-filtering.

In the twenty-seventh embodiment, in the method of the twenty-sixth embodiment, the decoding and post-filtering steps selectively perform one of modes a), b), c) and

d) decoding in advanced audio coding mode, AAC.

In the twenty-eighth embodiment, in the method of the twentieth embodiment, the bitstream signal is segmented into time frames, and the post-filtering step is skipped for the entire time frame or for a sequence of entire time frames.

In the twenty-ninth embodiment, in the method of the twenty-eighth embodiment:

the bitstream signal is a bitstream according to the Moving Picture Experts Group, MPEG standard, and includes, for each time frame, an associated data field; And

the post-filtering step is skipped for the time frame in response to the associated data field.

In the thirtieth embodiment, in the method of the twentieth embodiment, said post-filtering pass includes one of the following:

complete attenuation pass,

partial attenuation skip,

gradual increase in attenuation, and

gradual decrease in attenuation.

In the thirty-first embodiment, there is provided a method for decoding a bit stream signal as a temporary audio signal, including the steps of:

decoding the bit stream signal as a preliminary time audio signal; And

post-filtering the preliminary temporal audio signal by attenuating the interharmonic noise, whereby a temporal audio signal is obtained,

characterized in that the decoding step includes:

extracting an intermediate decoded signal representing the excitation;

calculating an approximate difference signal, which serves as an approximation of the signal component that will be removed by the post-filter from the decoded signal, as the difference between the intermediate decoded signal and the intermediate decoded signal subjected to post-filtering;

assessment of at least one of the following criteria:

a) whether the power of the approximate difference signal exceeds a predetermined threshold;

b) whether the approximate difference signal is tonal in nature;

c) whether the difference between the frequency spectra of the absolute values of the approximate difference signal and the temporal audio signal is unevenly distributed relative to frequency;

d) whether the frequency spectrum of the absolute values of the approximate difference signal is localized in frequency intervals that are within a predetermined significance envelope;

e) whether the frequency spectrum of the absolute values of the approximate difference signal is localized in frequency intervals that are within the significance envelope obtained by thresholding the frequency spectrum of the absolute values of the temporary audio signal by the absolute value of the largest component of the signal, here reduced in scale with a predetermined scale factor;

and, in response to the affirmative determination, disabling the post-filter, whereby the pre-temporal audio signal is output as a temporal audio signal.

In a thirty-second embodiment, an encoder system (800) for encoding a temporary audio signal as a bit stream signal is provided, which includes an encoding section (810) for encoding the temporary audio signal as a bit stream signal,

characterized by having a decision section (820) adapted to decide whether, when decoding a bitstream signal, post-filtering, which includes attenuation of inter-harmonic noise, should be turned off, and to encode this decision as post-filtering information in the bitstream signal.

In the thirty-third embodiment, in the encoder system of the thirty-second embodiment, the decision section is adapted to decide whether to disable post-filtering, which also includes attenuation of noise located in the valleys of the spectrum.

In the thirty-fourth embodiment, in the encoder system of the thirty-second embodiment, the encoding section includes a speech encoding unit.

In the thirty-fifth embodiment, in the encoder system of the thirty-second embodiment, the encoding section includes a code excitation linear prediction (CELP) encoding unit.

In the thirty-sixth embodiment, in the encoder system of the thirty-second embodiment, the decision section is adapted to:

detecting the co-presence of a signal component with a dominant fundamental frequency and a signal component located below the fundamental frequency and, optionally, between its harmonics; And

in response to this, to make a decision to disconnect.

In the thirty-seventh embodiment, in the encoder system of the thirty-fifth embodiment:

the CELP coding module is adapted to estimate the pitch frequency in a temporal audio signal; And

The decision section is adapted to detect spectral components below the estimated pitch frequency and, in response, to make a shutdown decision.

In the thirty-eighth embodiment, in the encoder system of the thirty-fifth embodiment, the decision section is adapted to

calculating the difference between the predicted power of the temporal audio signal under CELP encoding, and the predicted power of the temporal audio signal under CELP encoding and post-filtering, and,

in response to the fact that this difference exceeds a predetermined threshold value - to make a shutdown decision.

In the thirty-ninth embodiment, in the encoder system of the thirty-fifth embodiment:

said encoding section also includes a coded excitation conversion mode encoder, TCX,

where the decision section is adapted to select one of the following encoding modes:

a) TCX coding;

b) CELP encoding with post-filtering; And

c) CELP coding without post-filtering.

In the fortieth embodiment, the encoder system of the thirty-ninth embodiment also includes an encoding selector 814 adapted to select one of the following super modes:

i) advanced audio coding, AAC, where the decision section is disabled; And

ii) TCX/CELP encoding, where the decision section is connected to select one of the encoding modes a), b) or c).

In the forty-first embodiment, in the encoder system of the thirty-ninth embodiment, the decision section is adapted to decide which mode to use based on optimizing the distortion versus data rate relationship.

In the forty-second embodiment, the encoder system of the thirty-second embodiment is also adapted to segment the bit stream signal into time frames, where the decision section is adapted to decide to disable the post filter in time segments consisting of complete frames.

In the forty-third embodiment, in the encoder system of the thirty-second embodiment, the decision section is adapted to decide whether to gradually decrease and/or increase the attenuation of the post filter.

In the forty-fourth embodiment, in the encoder system of the thirty-second embodiment, the decision section is adapted to:

calculating the power of a temporary audio signal below the estimated fundamental frequency; And

in response to the specified power exceeding a predetermined threshold - to make a shutdown decision.

In the forty-fifth embodiment, in the encoder system of the thirty-second embodiment, the decision section is adapted to:

obtaining from the temporary audio signal an approximate difference signal, which serves as an approximation of the signal component that will be eliminated by the post-filter from the future decoded signal;

assessment of at least one of the following criteria:

a) whether the power of the approximate difference signal exceeds a predetermined threshold;

b) whether the approximate difference signal is tonal in nature;

c) whether the difference between the frequency spectra of the absolute values of the approximate difference signal and the temporal audio signal is unevenly distributed relative to frequency;

d) whether the frequency spectrum of the absolute values of the approximate difference signal is localized in frequency intervals that are within a predetermined significance envelope; And

e) whether the frequency spectrum of the absolute values of the approximate difference signal is localized in frequency intervals that are within the significance envelope obtained by thresholding the frequency spectrum of the absolute values of the temporary audio signal by the absolute value of the largest component of the signal, here reduced in scale with a predetermined scale factor;

and, in response to a positive determination, decide to disable the post-filter.

In the forty-sixth embodiment, in the encoder system of the forty-fifth embodiment, the decision section is adapted to calculate an approximate difference signal as the difference between the temporal audio signal and the post-filtered temporal audio signal.

In the forty-seventh embodiment, in the encoder system of the forty-fifth embodiment:

the encoding section is adapted to extract the intermediate encoded signal representing the excitation and transmit it to the decision section; And

the decision section is adapted to calculate an approximate difference signal as the difference between the temporary audio signal and the intermediate post-filtered decoded signal.

In a forty-eighth embodiment, there is provided a method for encoding a temporary audio signal as a bit stream signal, where the method includes the step of encoding a temporary audio signal as a bit stream signal,

different by also having a step of deciding whether post-filtering, which includes attenuation of inter-harmonic noise, should be turned off when decoding the bit stream, and encoding this decision as post-filtering information in the bit stream signal.

In the forty-ninth embodiment, in the method of the forty-eighth embodiment, the decision step relates to post-filtering, which also includes attenuation of noise located in the valleys of the spectrum.

In the fiftieth embodiment, in the method of the forty-eighth embodiment, the encoding step includes applying an encoding method adapted for speech encoding.

In the fifty-first embodiment, in the method of the forty-eighth embodiment, the encoding step includes applying code-excited linear prediction, CELP encoding.

In the fifty-second embodiment, the method of the forty-eighth embodiment also includes the step of detecting the co-presence of a signal component with a dominant fundamental frequency and a signal component below the fundamental frequency and optionally between its harmonics,

where the decision to disable post-filtering is made in case of a positive detection result.

In the fifty-third embodiment, in the method of the fifty-first embodiment

said CELP encoding step includes estimating a pitch frequency in a temporal audio signal; And

The decision stage involves detecting spectral components below the estimated pitch frequency, and a decision to disable post-filtering is made if the detection is positive.

In the fifty-fourth embodiment, the method of the fifty-first embodiment also includes the step of calculating a difference between the predicted power of the temporal audio signal under CELP encoding and the predicted power of the temporal audio signal under CELP encoding and post-filtering,

where the decision to disable post-filtering is made if this difference exceeds a predefined threshold value.

In the fifty-fifth embodiment, in the method of the fifty-first embodiment:

the encoding step includes selectively applying either CELP encoding or coded excitation transform, TLC encoding; And

the step of detecting whether post-filtering should be disabled is performed only when CELP encoding is applied.

In the fifty-sixth embodiment, in the method of the fifty-fifth embodiment, the decision step includes selecting, based on optimization of the distortion versus data rate relationship, one of the following operating modes:

a) TCX coding;

b) CELP encoding with post-filtering; And

c) CELP coding without post-filtering.

57. The method according to embodiment 55, wherein the decision step includes selecting, based on optimization of the distortion versus data rate relationship, one of the following operating modes

a) TCX coding;

b) CELP encoding with post-filtering;

c) CELP coding without post-filtering; And

d) coding in advanced audio coding mode, AAC.

In the fifty-eighth embodiment, in the method of the forty-eighth embodiment:

the encoding step includes segmenting the temporal audio signal into time frames and generating a bit stream signal containing the corresponding time frames; And

The decision step that post-filtering should be disabled occurs once every time frame.

In the fifty-ninth embodiment, in the method of the forty-eighth embodiment, the result of the decision step that post-filtering should be disabled is selected from the following options:

no weakening,

complete weakening

partial weakening,

gradually increasing weakening, and

gradually decreasing weakening.

In the sixtieth embodiment, in the method of the forty-eighth embodiment, the decision step includes calculating the power of a temporary audio signal below the estimated pitch frequency and, in response to that power exceeding a predetermined threshold, turning off the post-filter.

In the sixty-first embodiment, in the method of the forty-eighth embodiment:

the encoding step includes obtaining from the temporary audio signal an approximate difference signal, which serves as an approximation of the signal component that will be eliminated by the post-filter from the future decoded signal; And

The decision-making phase involves evaluating at least one of the following criteria:

a) whether the power of the approximate difference signal exceeds a predetermined threshold;

b) whether the approximate difference signal is tonal in nature;

c) whether the difference between the frequency spectra of the absolute values of the approximate difference signal and the temporal audio signal is unevenly distributed relative to frequency;

d) whether the frequency spectrum of the absolute values of the approximate difference signal is localized in frequency intervals that are within a predetermined significance envelope; And

e) whether the frequency spectrum of the absolute values of the approximate difference signal is localized in frequency intervals that are within the significance envelope obtained by thresholding the frequency spectrum of the absolute values of the temporary audio signal by the absolute value of the largest component of the signal, here reduced in scale with a predetermined scale factor;

and, in response to at least one positive determination, disabling the post-filter.

In the sixty-second embodiment, in the method of the sixty-first embodiment, the approximate difference signal is calculated as the difference between the temporary audio signal and the post-filtered temporary audio signal.

In the sixty-third embodiment, in the method of the sixty-first embodiment:

the encoding step includes extracting an intermediate decoded signal representing the excitation; And

the decision step involves calculating an approximate difference signal as the difference between the temporal audio signal and the intermediate post-filtered decoded signal.

In a sixty-fourth embodiment, a computer-readable storage medium is provided containing a computer program product storing instructions for executing a method according to one of the above embodiments of the invention.

Claims (13)

1. A decoder system (400; 500; 700; 1000) for decoding a bitstream signal as a temporary audio signal, comprising: a decoding section (410; 511, 512, 513; 711, 712, 713; 1011, 1013) for decoding the bitstream signal as a preliminary temporal audio signal, the decoding section being configured to operate in at least one first decoding mode including post-filtering , and at least one second decoding mode not including post-filtering; And a tone gain filter (440; 540; 740; 1040) for post-filtering a preliminary temporary audio signal to obtain a temporary audio signal, wherein the tone gain is controlled by a post-filter gain, characterized in that it comprises a control section configured to selectively disable the tone gain filter by setting the post-filter gain to zero in said at least one first decoding mode in response only to post-filtering information encoded in the bitstream signal, wherein the post-filtering information indicates a decision on the encoder side as to whether to disable post-filtering, whereby the pre-temporal audio signal is output as a temporary audio signal. 2. The decoder system according to claim 1, wherein said decoding section comprises a code-excited linear prediction (CELP) decoding module (511; 711; 1011). 3. The decoder system according to claim 2, characterized in that the decoding section further comprises a frequency domain decoding module (713; 1013). 4. The decoder system of claim 1, wherein the tone gain filter includes a gain controller for setting the post filter gain to zero when the tone gain filter is disabled. 5. The decoder system according to claim 1, characterized in that the tone enhancement filter is a bass post-filter. 6. A method for decoding a bitstream signal as a temporary audio signal, comprising the steps of: decoding the bitstream signal as a preliminary temporal audio signal in one of a plurality of decoding modes, the plurality of decoding modes including at least one first decoding mode including a post-filtering step, and at least one second decoding mode not including a post-filtering step; wherein in the post-filtering step, a tone enhancement filter is applied to the preliminary temporary audio signal, thereby obtaining a temporary audio signal, wherein the tone enhancement filter is controlled by the post-filter gain, characterized in that, in said at least one first decoding mode, the post-filtering step is selectively disabled in response only to post-filtering information encoded in the bitstream signal by setting the post-filter gain to zero, wherein the post-filtering information indicates an encoder-side decision regarding whether to disable postfiltering. 7. A storage medium on which instructions are stored that, when executed by a digital signal processor, cause the digital signal processor to execute the method of claim 6.