RU2575393C2 - Encoding and decoding of slot positions with events in audio signal frame - Google Patents

Abstract

FIELD: physics, acoustics.

SUBSTANCE: invention relates to audio signal processing and audio encoding. Disclosed is an apparatus for decoding, an apparatus for encoding, a method for decoding and a method for encoding positions of slots comprising events in an audio signal frame and respective programs and encoded signals, wherein the apparatus for decoding comprises: an analysing unit for analysing a frame slots number indicating the total of slots of the audio signal frame, an event slots number indicating the number of slots comprising the events of the audio signal frame, and an event state number, and a generating unit for generating an indication of a plurality of positions of slots comprising the events in the audio signal frame using the frame slots number, the event slots number and the event state number.

EFFECT: high accuracy of encoding and decoding.

17 cl, 25 dwg, 6 tbl

Description

The present invention relates to the field of audio processing and audio coding, in particular to coding and decoding of the positions of slots with events in an audio signal frame.

Audio processing and / or audio coding have advanced in many ways. In particular, spatial audio applications are becoming increasingly important. Audio processing is often used for decorrelation or reproduction of signals. Moreover, decorrelation and reproduction of signals is used in the process of up-mixing mono-stereo (from mono to stereo), up-mixing from mono / stereo to multi-channel format, for artificial reverb, stereo expansion or user-mixing mixing / playback.

Some audio processing systems use decorrelators. An important example is the use of decorrelation of signals in parametric spatial audio decoders to restore specific decorrelation properties between two or more signals that are reconstructed from one or more signals obtained by downmixing. The use of decorrelators significantly improves the perceived quality of the output signal, for example, when compared with stereo based on signal intensity. Specifically, the use of decorrelators provides the ability to correctly synthesize spatial sound with a wide phonogram, several simultaneously operating sound objects and / or surroundings. However, it is also known that decorrelators introduce artifacts, such as changes in the temporal structure of the signal, timbre, etc.

Other examples of decorrelators in audio processing are, for example, generating artificial reverb to change the spatial impression, or using decorrelators in multi-channel acoustic echo cancellation systems to improve convergence.

One important spatial audio coding scheme is parametric stereo (PS). Figure 1 illustrates the structure of a mono-stereo decoder. A single decorrelator generates a decorrelated signal D (“processed” signal) from the input mono signal M (“unprocessed” signal). The decorrelated signal D is then supplied to the mixer along with signal M. The mixer then applies the mixing matrix H to the input signals M and D to generate the output signals L and R. The coefficients in the mixing matrix H can be fixed, signal dependent, or controlled by the user.

Alternatively, the mixing matrix is driven by additional information that is transmitted along with the downmix and contains a parametric description of how to upmix the downmix signals to produce the desired multi-channel output. Spatial additional information is usually generated during the down-mix to mono process in the respective signal encoder.

Spatial audio coding, as described above, is widely used, for example, in parametric stereo. A typical structure of a parametric stereo decoder is shown in FIG. 2, decorrelation is performed in the transform domain. Spatial parameters can be modified by the user or by additional tools, for example, through post-processing for binaural playback / presentation. In this case, the upmix parameters are combined with the parameters from the binaural filters to calculate the input parameters for the mixing matrix.

The output signal L / R of the mixing matrix H is calculated from the input mono signal M and the decorrelated signal D.

In the mixing matrix, the amount of decorrelated sound output is controlled based on the transmitted parameters, for example interchannel frequency difference (ILD), interchannel correlation / coherence (ICC) and / or fixed or user-defined settings.

Conceptually, the output signal of the D output of the decorrelator replaces the residual signal, which ideally would provide the ideal decoding of the original L / R signals. Using the output D of the decorrelator instead of the residual signal in the boost mixer saves the bit rate that would otherwise be required to transmit the residual signal. The purpose of the decorrelator, therefore, is to generate a signal D from a mono signal M, which shows similar properties as the residual signal, which is replaced by D. Reference is made to the document:

[1] J. Breebaart, S. van de Par, A. Kohlrausch, E. Schuijers, "High-Quality Parametric Spatial Audio Coding at Low Bitrates" in Proceedings of the AES 116 ^th Convention, Berlin, Preprint 6072, May 2004.

Considering MPEG Surround (MPS), structures similar to PS, called one-to-two conversion blocks (OTT blocks), are used in spatial audio decoding trees. This may look like a generalization of the mono-stereo up-mix concept for multi-channel spatial audio coding / audio decoding. The MPS also has two-to-three boost mixing systems (TTT blocks) that can use decorrelators depending on the TTT mode of operation. Details are described in the document:

[2] J. Herre, K. Kjorling, J. Breebaart, et al., "MPEG surround - the ISO / MPEG standard for efficient and compatible multi-channel audio coding", in Proceedings of the 122 ^th AES Convention, Vienna, Austria, May 2007.

Regarding directional audio coding (DirAC), DirAC relates to a parametric coding scheme for a sound field that is not limited to a fixed number of audio output channels with fixed speaker positions. DirAC applies decorrelators in the DirAC playback unit, i.e. in a spatial audio decoder for the synthesis of incoherent components of sound fields. Directional audio coding is further described in:

[3] Pulkki, Ville: "Spatial Sound Reproduction with Directional Audio Coding," in J. Audio Eng. Soc., Vol. 55, No. 6, 2007.

Regarding prior art decorrelators, reference is made to documents:

[4] ISO / IEC International Standard "Information Technology - MPEG audio technologies - Part1: MPEG Surround", ISO / IEC 23003-1: 2007.

[5] J. Engdegard, H. Purnhagen, J. Roden, L. Liljeryd, "Synthetic Ambience in Parametric Stereo Coding" in Proceedings of the AES 116 ^th Convention, Preprint, May 2004.

Lattice frequency-independent IIR structures are used as decorrelators in spatial audio decoders, such as MPS [2, 4]. Other prior art decorrelators (potentially frequency-dependent) apply delays to decorrelated signals or reduce input signals, for example, with exponentially decaying noise emissions. For a review of prior art decorrelators for spatial upmix audio systems, reference is made to document [5]: "Synthetic Ambience in Parametric Stereo Coding".

In general, it is known that stereo or multi-channel applause-like signals encoded / decoded in parametric spatial audio encoders result in reduced signal quality. Applause-like signals are characterized by the content of fairly frequent transient impurities from different directions. Examples of such signals are applause, the sound of rain, galloping horses, etc. Applause-like signals often also contain sound components from distant sound sources that tangibly merge with a noise-like, smoothed background sound field.

Lattice frequency-independent structures used in spatial audio decoders, such as MPEG Surround, act as artificial reverb generators and are therefore well suited to generating uniform, smooth, noise-like, inverse sounds (such as indoor reverb tails). However, they are examples of sound fields with an inhomogeneous spatio-temporal structure, which nevertheless create an immersion in the listener: one notable example is sound fields similar to applause, which create a listener reach not only through uniform noise-like fields, but also through fairly frequent sequences of single pops from different directions. Therefore, the heterogeneous component of the sound field of applause can be characterized by a spatially distributed admixture of transients. These distinguishable claps are not at all homogeneous, smooth and noise-like.

Due to their similar reverberation behavior, trellised frequency-independent decorrelators are not able to generate immersive sound fields with characteristics such as applause. Instead, when applied to similar applause signals, they tend to temporarily blur the transients in the signal. An undesirable result is a noise-like immersive sound field without a distinctive spatio-temporal structure of applause-like sound fields. In addition, transient events, such as single claps of hands, can cause ringing artifacts of decorrelator filters.

USAC (Unified Speech and Audio Coding) is an audio coding standard for encoding speech and audio and mixing them at different bit rates.

The perceived USAC quality can be further improved by stereo coding of applause and applause-like sounds with bit rates of up to 32 kbit / s when parametric stereo coding methods are applicable. USAC encoded applause tends to show a narrow soundstage and lack of coverage if dedicated applause processing is not used inside the codec. In many ways, USAC stereo coding methods and their limitations have been inherited from MPEG Surround (MPS). However, the USAC offers dedicated adaptation to the requirement for proper applause. The above adaptation is called transient control decorrelator (TSD) and is an embodiment of this invention.

Applause signals can be represented as composed of single, distinguishable close pops, temporarily separated by several milliseconds, and an imposed noise-like environment arising from very frequent distant pops. In parametric stereo coding with a reasonable additional information rate, the detail of sets of spatial parameters (inter-channel difference in frequency levels, inter-channel correlation, etc.) is too low to guarantee sufficient spatial redistribution of single pops, leading to a lack of coverage. Additionally, claps are treated with a trellised frequency-independent decorrelator. This inevitably gives rise to a temporary dispersion of transients and further reduces subjective quality.

Using a transient control decorrelator (TSD) inside the USAC decoder modifies the MPS processing. The fundamental idea of this approach is to solve the problem of decorrelation of applause, as follows:

- Separate the transients in the QMF region to a trellised frequency-independent decorrelator, i.e.: divide the input of the decorrelator into stream s2 with transients and stream s1 without transients.

- Feed the transient stream to another parameter-controlled decorrelator that is well suited for transient impurities.

- Feed a stream without transients to a frequency-independent MPS decorrelator.

- Add the output signals of both decorrelators D ₁ and D ₂ to get decorrelated signal D.

Figure 3 illustrates the configuration of the conversion from one channel to two (OTT) inside the USAC decoder. The U-shaped transient processing unit of FIG. 3 comprises a parallel signal path proposed for transient processing.

Two parameters that control the TSD process are transmitted as frequency-independent parameters from the encoder to the decoder (see figure 3):

- The binary solution "transients / without transients" of the transient detector running in the encoder is used to control the separation of transients with details of the QMF time slots in the decoder. An efficient lossless coding scheme is used to transmit position data of QMF slots with transients.

- The actual transient decorrelator parameters that are needed for the transient decorrelator to control the spatial distribution of transients. The transient decorrelator parameters indicate the angle between the downmix and its remainder. These parameters are transmitted only for time slots in which the transient content has been detected on the encoder.

In order to evaluate the quality of the technology described above, two MUSHRA listening tests were conducted in a controlled listening test environment using high-quality STAX electrostatic headphones. The test was performed with a stereo configuration of 32 kbps and 16 kbps. Sixteen expert listeners participated in each of the audits.

Since the USAC verification kit does not contain applause elements, to demonstrate the advantage of the proposed technology, additional applause elements were selected. The elements shown in table 1 were included in the check:

Table 1
Listening Check Elements Element The properties ARL_applause applause from low to medium frequency (MPS check kit element) applause4s very frequent applause containing a few distinguishable claps Applse_2ch frequent multi-channel applause - front channels (MPS check kit element) Applse_st frequent multi-channel applause - downmix to stereo (MPS check kit element) Klatschen signal of rare applause

As for the usual twelve elements of the MPEG USAC listening test, the TSD is always inactive. However, these elements do not remain exactly identical in bit, since the TSD enable bit (indicating that the TSD is disabled) is additionally included in the bitstream, and thus this affects the bit budget for the base encoder a bit. Since these differences are very small, these elements were not included in the listening test. Data is provided on the size of these differences to show that these changes are negligible and inconspicuous.

The codec tool, called inter-TES, is part of the USAC reference model 8 (RM8). Since this method was introduced to improve the perceived quality of transients, including applause-like signals, inter-TES was always on for every test condition. With this installation, the best possible quality is guaranteed and the orthogonality of inter-TES and TSD is demonstrated.

System checks have the following configurations:

- RM8: RM8 USAC system

- CE: RM8 USAC system enhanced by Transient Controlling Decorrelator (TSD)

Figures 4 and 5 show MUSHRA scores along with their 95% confidence intervals for the verification scenario at 32 kbps. Student t-distribution was proposed for the verification data. The absolute scores in Fig. 4 show a higher average score for all elements, for four out of five elements there is a significant improvement in terms of 95% confidence. Not a single element has worsened, unlike RM8. The difference scores for USAC + TSD evaluated in the base experiment (CE) TSD with respect to RMAC USAC are plotted in FIG. 5. A significant improvement can be seen here for all elements.

For the verification scheme at 16 kbps, FIGS. 6 and 7 show MUSHRA scores along with their 95% confidence intervals. Student t-distribution for data was proposed. The absolute scores in FIG. 6 show a higher average score for each element. For one element, significance in the sense of 95% confidence may be visible. No item has a worse score than RM8. Difference points are plotted in Fig.7. Significant improvement was again demonstrated for all elements in relation to different data.

The TSD tool is enabled via the bsTsdEnable flag, transmitted in the bitstream. If the TSD is enabled, the actual transient separation is controlled by the TsdSepData transient detection flags, which are also transmitted in the bitstream and which are encoded in bsTsdCodedPos when the TSD is turned on.

In the encoder, the TSD enable flag bsTsdEnable is generated by a segment classifier. Transient detection flags TsdSepData are set by the transient detector.

As already noted, TDS is not activated for the twelve MPEG USAC validation elements. For five additional applause elements, TSD activation is depicted in FIG. 8, showing the logical state of bsTsdEnable versus time.

If TSD is activated, transients are detected in specific QMF time slots, and they are subsequently fed to a dedicated transient decorrelator. For each additional check item, Table 2 lists the percent of slots inside TSD-activated frames that contain transients.

table 2
Percentage of slots with transients (frequency of slots with transients in% of all time slots of TSD frames) Element Slot frequency with transients (%) ARL_applause 23,4 Applause4s 20.1 applse_2ch 24.7 applse_st 23.8 Klatschen 21.3

The transfer of transient separation decisions and decorrelator parameters from the encoder to the decoder requires a certain amount of additional information. However, this value is abundantly offset by bit rate savings resulting from the transmission of spatial broadband labels within the MPS.

As a result, the average bit rate of the additional information MPS + TSD is even lower than the bit rate of the additional information of simple MPS with a simple USAC, as shown in the first column of Table 3. In the proposed configuration used to evaluate subjective quality, the average bit rates given in the second column of table 3 were measured for TDS:

Table 3
MPS bit rates (+ TSD) in bps as part of a scenario with 32 kbps stereo codec Element Average bit rate (bps) of additional MPS information (+ TSD) simple USAC RM8 USAC with TSD ARL_applause 2966 2345 Applause4s 2754 2278 applse_2ch 3000 2544 applse_st 2735 2253 Klatschen 2950 2495

The difficulty in calculating TSD is due to

- decoding slots with transients,

- complexity of transient decorrelator.

Assuming an MPEG Surround spatial frame length of 32 time slots, decoding slot positions requires (64 divisions + 80 multiplications) for each spatial frame in the worst case, i.e. 64 * 25 + 80 = 1680 operations for each spatial frame.

Ignoring copy operations and conditional operators, the complexity of the transient decorrelator is determined by one complex multiplication per slot and a hybrid QMF range.

This results in the following TSD total numbers shown in comparison with the simple USAC complexity numbers in table 4:

Table 4
The complexity of a TSD decoder in PFM (millions of operations per second) and the complexity of a simple USAC decoder The complexity of simple USAC, in pug TSD: Transient Decorrelator Complexity, in MOPS TSD: table position decoder complexity, in pug ∑ (TSD complexity), in pug ∑ (TSD complexity) relatively simple USAC 16 kbps stereo (f _s = 28.8 kHz) 8.7 0.117 0.024 0.141 1.62% 32 kbps stereo (f _s = 40 kHz) 13,2 0.163 0,033 0.196 1.48%

So, these listening tests clearly show a significant improvement in the subjective quality of the applause signals in the difference scores of all elements at both operating points. In terms of absolute scores, all elements under the TSD condition show a higher average score. For 32 kbps, a significant improvement exists for four of the five elements. For 16 kbps, one element shows a significant improvement. No item has a worse score than RM8. Improvement is achieved at a negligible computational cost, as can be seen from the data on complexity. This further emphasizes the advantage of the TSD tool for USAC.

The transient control decorrelator described above greatly improves USAC audio processing. However, as was also seen above, the transient-controlling decorrelator requires information about the existence or non-existence of transients in a particular slot. With USAC, time slot information can be transmitted frame by frame. A frame contains several, for example 32, time slots. As a consequence, it should be recognized that the encoder also transmits information about which slots contain transients on a frame-by-frame basis. Reducing the number of bits to be transmitted is critical in audio processing. Since even a single audio recording contains a huge number of frames, this means that even if the number of bits to be transmitted for each frame is reduced by only a few bits, the overall bit rate can be significantly reduced.

The problem of decoding the positions of slots with events in the audio frame, however, is not limited to the problem of decoding transients. Moreover, it would be useful to also decode the positions of the slots of other events, such as whether the slot of the audio frame is tonal (or not), whether it contains noise (or not) and the like. In fact, a device for effectively encoding and decoding the positions of slots with events in an audio frame would be very useful for a large number of all kinds of events.

When this document refers to slots or slot positions of an audio frame, the slots in this sense can be time slots, frequency slots, time-frequency slots, or any other type of slots. In addition, it should be understood that the present invention is not limited to the processing of audio and audio frames by USAC, but instead relates to any type of audio frames and any type of audio formats, such as MPEG1 / 2 level 3 ("MP3"), advanced audio coding (AAC ) and the like. Effectively encoding and decoding the positions of the event slots in the audio frame would be very useful for any type of audio frame.

Therefore, it is an object of the present invention to provide an apparatus for encoding slot positions with events in an audio frame using a small number of bits. Moreover, it is an object of the present invention to provide an apparatus for decoding slot positions with events in an audio frame encoded by an encoding apparatus according to the present invention. The objectives of the present invention are achieved by means of a decoding device according to claim 1, a device for encoding according to claim 11, a decoding method according to claim 14, an encoding method according to claim 15, a computer program for decoding according to claim 16 of the claims, a computer program for encoding according to claim 17 and the encoded signal according to claim 18.

According to the present invention, it is contemplated that the number of frame slots indicating the total number of audio frame slots and the number of event slots indicating the number of slots containing audio frame events may be available in the decoding apparatus of the present invention. For example, an encoder may transmit the number of frame slots and / or the number of event slots to a device for decoding. According to an embodiment, the encoder may indicate the total number of slots of the audio frame by transmitting a number that is the total number of slots of the audio frame minus 1. The encoder may further indicate the number of slots containing events of the audio frame by transmitting a number which is the number of slots containing frame events audio signal, minus 1. Alternatively, the decoder can determine the total number of slots of the audio frame and the number of slots containing events of the audio frame without informing cations from the encoder.

Based on these assumptions, according to the present invention, a given number of slot positions containing events in an audio signal frame can be encoded and decoded using the following survey results:

Let N be the total number of slots of the audio frame, and P be the number of slots containing events of the audio frame.

It is assumed that both the encoding device and the decoding device are aware of the values of N and P.

разных комбинаций позиций слотов, содержащих события в кадре аудиосигнала.Knowing N and P, we can determine that there is only $$\left({}_p^N\right)$$

different combinations of slot positions containing events in the audio frame.

разных комбинаций.For example, if the slot positions in the frame are numbered from 0 to N-1 and if P = 8, then the first possible combination of slot positions with events will be (0, 1, 2, 3, 4, 5, 6, 7), the second combination will be (0, 1, 2, 3, 4, 5, 6, 8) and so on, until the combination (N-8, N-7, N-6, N-5, N-4, N-3, N- 2, N-1), so in the end there $$\left({}_p^N\right)$$

different combinations.

комбинаций представлена уникальным числом состояний события и если устройство для декодирования осведомлено о том, какое число состояний события представляет какую комбинацию позиций слотов, содержащих события в кадре аудиосигнала (например, посредством применения соответствующего способа декодирования), то устройство для декодирования может декодировать позиции слотов, содержащие события, с использованием N, P и числа состояний события. Для огромного количества типичных значений для N и P такой метод кодирования использует меньше битов для кодирования позиций слотов с событиями по сравнению с другими методами (например, использующими массив битов с одним битом для каждого слота кадра, в котором каждый бит указывает, встречается ли событие в этом слоте или нет).Moreover, the present invention uses additional research results that the number of event states can be encoded by the encoding device and that the number of event states is transmitted to the decoder. If each of the possible $$\left({}_p^N\right)$$

of combinations is represented by a unique number of event states and if the device for decoding is aware of the number of event states representing which combination of slot positions containing events in the audio signal frame (for example, by applying the appropriate decoding method), then the device for decoding can decode the positions of slots containing events using N, P and the number of event states. For a huge number of typical values for N and P, such an encoding method uses fewer bits to encode slot positions with events compared to other methods (for example, using an array of bits with one bit for each slot in a frame in which each bit indicates whether an event occurs in this slot or not).

. Число требуемых битов, таким образом, составляетIn other words, the problem of coding the positions of slots with events in the audio signal frame can be solved by encoding a discrete number P of positions p _k in the range [0 ... N-1], so that the positions do not overlap, p _k ≠ p _h for k ≠ h , with as few bits as possible. Since the order of the positions does not matter, therefore, the number of unique combinations of positions is a binomial coefficient $$\left({}_p^N\right)$$

. The number of bits required is thus

.

.

In an embodiment, a decoding apparatus is provided in which the decoding apparatus is arranged to perform a check comparing the number of event states or the updated number of event states with a threshold value. Such a check can be used to obtain the positions of slots containing events from the number of event states. A comparison check of the number of event states with a threshold value can be performed by comparing whether the number of event states or the updated number of event states is greater, greater or equal, less than or less than or equal to the threshold value. In addition, it is preferable that the device for decoding was configured to update the number of event states or the updated number of event states depending on the result of the check.

According to an embodiment, a decoding apparatus is provided which is arranged to carry out a check comparing the number of event states or the updated number of event states with respect to the particular slot in question, the threshold value depending on the number of frame slots, the number of event slots and the position of the considered slot inside frame. In addition, the positions of the slots containing the events can be determined on a per-slot basis, with a decision for each slot of the frame, one after the other, whether the slot contains an event.

According to a further embodiment, a decoding apparatus is provided which is arranged to split a frame into a first section of a frame containing a first set of frame slots and into a second section of a frame containing a second set of frame slots, and in which this decoding device is further configured defining positions containing events for each of the sections of the frame individually. In addition, the positions of the slots containing the events can be determined by re-dividing the frame or sections of the frame into even smaller sections of the frame.

Below, embodiments of the present invention are described in more detail with respect to the drawings, in which:

1 shows a typical use of a decorrelator in a mono-stereo boost mixer;

FIG. 2 shows an additional typical use of a decorrelator in a mono-stereo boost mixer;

figure 3 shows an overview of the conversion system of one channel into two (OTT), including transient control decorrelator (TSD);

FIG. 4 is a diagram illustrating absolute scores for 32 kbps stereo comparing RM8 USAC and USAC RM8 + TSD in a basic experiment (CE) TSD;

Fig. 5 is a graph showing difference scores for 32 kbps stereo comparing USAC using a transient-controlled decorrelator with a simple USAC system;

6 is a diagram showing absolute scores for 16 kbps stereo comparing RM8 USAC and USAC RM8 + TSD in a basic experiment (CE) TSD;

Fig. 7 is a diagram showing difference scores for 16 kbps stereo comparing USAC using a transient control decorrelator with a simple USAC system;

Fig. 8 shows TSD activity for five additional elements, depicted as the logical state of the bsTsdEnable flag;

Fig. 9A illustrates an apparatus for decoding slot positions containing events in an audio frame according to an embodiment of the present invention;

Fig. 9B illustrates an apparatus for decoding slot positions containing events in an audio signal frame, according to a further embodiment of the present invention;

Fig. 9C illustrates an apparatus for decoding slot positions containing events in an audio signal frame, according to another embodiment of the present invention;

10 is a flowchart illustrating a decoding process conducted by a decoding apparatus according to an embodiment of the present invention;

11 shows a pseudo code implementing decoding of positions of slots containing events according to an embodiment of the present invention;

12 is a flowchart illustrating an encoding process conducted by an encoding apparatus according to an embodiment of the present invention;

13 is a pseudo-code illustrating a process for encoding slot positions containing events in an audio signal frame, according to a further embodiment of the present invention;

Fig. 14 illustrates an apparatus for decoding slot positions containing events in an audio signal frame, according to a further embodiment of the present invention;

on Fig illustrated a device for encoding the positions of slots containing events in the frame of the audio signal, according to a variant implementation of the present invention;

FIG. 16 illustrates data syntax of MPS 212 of the USAC according to an embodiment;

17 illustrates the syntax of TsdData USAC according to an embodiment;

on Fig illustrated table nBitsTrSlots depending on the length of the MPS frame;

on Fig shows a table related to bsTempShapeConfig USAC according to a variant implementation;

20 depicts the syntax of a TempShapeData USAC according to an embodiment;

FIG. 21 illustrates a decorrelator block D in an OTT decoding block according to an embodiment;

22 depicts the syntax of the EcData USAC according to an embodiment;

23 shows a signal flow diagram for generating TSD data.

FIG. 9A illustrates an apparatus 10 for decoding slot positions containing events in an audio frame according to an embodiment of the present invention. The decoding device 10 comprises an analysis unit 20 and a generation unit 30. The number of FSN frame slots indicating the total number of audio frame slots, the number of ESON event slots indicating the number of slots containing audio frame events, and the number of ESTN event states are supplied to decoding apparatus 10. The decoding apparatus 10 then decodes the positions of the slots containing the events by using the number of slots of the FSN frame, the number of slots with ESON events, and the number of states of the ESTN event. Decoding is performed by an analyzing unit 20 and a generating unit 30, which interact in the decoding process. While the analysis unit 20 is responsible for performing checks, for example, comparing the number of states of an ESTN event with a threshold value, the generating unit 30 generates and updates intermediate results of the decoding process, for example, an updated number of states of the event.

In addition, the generating unit 30 generates an indication of a plurality of slot positions containing events in the audio frame. A specific indication of a plurality of slot positions containing events of an audio signal frame may be referred to as an “indication state”.

According to an embodiment, an indication of a plurality of slot positions containing events in an audio signal frame can be generated so that at the first time, the generating unit 30 indicates for the first slot whether the slot contains an event or not, at the second time, the generating unit 30 indicates for the second slots, whether the slot contains an event or not, and so on.

According to a further embodiment, the indication of the multiple positions of the slots containing the events may, for example, be an array of bits indicating for each slot of the frame whether it contains an event.

The analyzing unit 20 and the generating unit 30 can interact so that both units call each other one or more times in the decoding process to produce intermediate results.

9B, a decoding apparatus 40 according to an embodiment of the present invention is illustrated. The decoding apparatus 40, among other things, differs from the decoding apparatus 10 of FIG. 9A in that it further comprises an audio signal processor 50. An audio signal processor 50 receives an audio input signal and an indication of a plurality of slot positions containing events in an audio signal frame that has been generated by the generating unit 45. Depending on the indication, the audio signal processor 50 generates an audio output signal. The audio processor 50 may generate an audio output, for example, by de-correlating an audio input. In addition, the audio signal processor 50 may include a trellis IIR decorrelator 54, a transient decorrelator 56, and a transient separation unit 52 for generating an audio output signal, as illustrated in FIG. If the indication of the multiple positions of the slots containing events in the audio signal frame indicates that the slot contains transients, the audio signal processor 50 will decorrelate the input audio signal related to this slot by the transient decorrelator 56. If, however, an indication of the plurality of slot positions containing events in the audio signal frame indicates that the slot does not contain transients, then the audio signal processor will decorrelate the input audio signal S related to this slot by using the trellis IIR decorrelator 54. The audio signal processor uses block 52 a transient compartment, which decides based on the indication whether the portion of the input audio signal related to the slot is fed to the transient decorrelator 56 or to the trellis IIR decorrelator 54, depending on whether AET specify whether that particular slot contains transients (decorrelation by transient decorrelator 56) or slot contains no transients (decorrelation by trellis decorrelator IIR-54).

9C, a decoding apparatus 60 according to an embodiment of the present invention is illustrated. The decoding apparatus 60 differs from the decoding apparatus 10 of FIG. 9A in that it further comprises a slot selector 90. Decoding is done on a per-slot basis, with a decision for each slot in the frame, one after the other, whether the slot contains an event. The slot selector 90 decides which frame slot to consider. A preferred approach would be one where the 90 slot selector selects frame slots one by one.

The slot-by-bit decoding of the decoding apparatus 60 of this embodiment is based on the following research findings that can be applied to embodiments of the decoding apparatus, the encoding apparatus, the decoding method, and the encoding method of slot positions that contain events in an audio signal frame. The following survey results are also applicable for the corresponding computer programs and encoded signals.

Suppose that N is the (total) number of audio frame slots and P is the number of slots containing frame events (this means that N may be the number of FSN frame slots and P may be the number of ESON event slots). The first slot of the frame is considered. Two cases can be distinguished.

разных возможных комбинаций из P позиций слотов, содержащих событие, относительно оставшихся N-1 слотов кадра.If the first slot is a slot that does not contain an event, then, relative to the remaining N-1 frame slots, there is only $$\left({}_P^{N-\mathrm{one}}\right)$$

different possible combinations of P positions of the slots containing the event relative to the remaining N-1 frame slots.

разных возможных комбинаций из оставшихся P-1 слотов, содержащих событие, относительно оставшихся N-1 слотов кадра.However, if the first slot is the slot that contains the event, then, relative to the remaining N-1 frame slots, there is only $$\left({}_{P-\mathrm{one}}^{N-\mathrm{one}}\right)=\left({}_P^N\right)-\left({}_P^{N-\mathrm{one}}\right)$$

different possible combinations of the remaining P-1 slots containing the event, relative to the remaining N-1 frame slots.

.Based on these survey results, the embodiments are further based on the survey results, that all combinations with the first slot where the event did not occur should be encoded by numbers of event states that are less than or equal to the threshold value. In addition, all combinations with the first slot where the event occurred must be encoded by the numbers of event states that are greater than the threshold value. In an embodiment, all event state numbers may be positive integers or 0, and a suitable threshold value for the first slot may be $$\left({}_P^{N-\mathrm{one}}\right)$$

.

In an embodiment, the decoding apparatus is configured to determine whether the first slot of the frame contains an event by checking whether the number of event states is greater than a threshold value. (Alternatively, the encoding / decoding process of the embodiments may also be implemented, so that the decoding apparatus checks whether the number of event states is greater or equal, less than or equal to or less than a threshold value.) After analyzing the first slot, decoding continues for the second slot of the frame using the adjusted values. In addition to adjusting the number of slots in question (which is reduced by one), the number of slots containing events also ultimately decreases by one (if the first slot contains an event), and the number of event states is adjusted when the number of event states was more than the threshold value in order to remove the portion related to the first slot, based on the number of event states. The decoding process can be continued for further frame slots in a similar manner.

. Напротив, если состояние больше, чем $$\left({}_P^{N-1}\right)$$

, то можно сделать вывод, что первый слот имеет назначенную ему позицию. Следующий алгоритм декодирования может быть результатом этого:In an embodiment, a discrete number P is encoded from positions p _k on the range [0 ... N-1], so that the positions do not overlap, p _k ≠ p _h for k ≠ h. Here, each unique combination of positions in a given range is called a state, and each possible position in this range is called a slot. According to an embodiment of a decoding apparatus, a first slot in a range is considered. If the slot does not have a position assigned to it, then the range can be reduced to N-1, and the number of possible states is reduced to $$\left({}_P^{N-\mathrm{one}}\right)$$

. On the contrary, if the condition is greater than $$\left({}_P^{N-\mathrm{one}}\right)$$

, then we can conclude that the first slot has a position assigned to it. The following decoding algorithm may be the result of this:

For each slot h

, тоIf condition> $$\left({}_P^{N-h-\mathrm{one}}\right)$$

then

Assign position to slot h

Update remaining state as state: = state- $$\left({}_P^{N-h-\mathrm{one}}\right)$$

Decrease the number of positions to the left as P: = P-1

the end

the end

Calculating the binomial coefficient at each iteration would be costly. Because of this, according to embodiments, the following rules can be used to update the binomial coefficient using the value from the previous iteration:

Using these formulas, each binomial coefficient update costs only one multiplication and one division, while an exact estimate would cost P multiplications and divisions at each iteration.

In this embodiment, the total decoder complexity is P multiplications and divisions to initialize the binomial coefficient, for each iteration 1 is multiplication, division, and the if operator, and for each coded position 1 is multiplication, addition, and division. It should be noted that in theory it would be possible to reduce the number of divisions needed for initialization to one. In practice, however, this approach would lead to very large integers that are difficult to process. The worst case decoder complexity is then N + 2P divisions and N + 2P multiplications, P additions (can be ignored if MAC operations are used), and N if-statements.

In an embodiment, the encoding algorithm used by the encoding device does not have to iterate over all slots, but only at those that have their assigned position. Thereby

For each position P _h , h = 1 ... P

.Update state as state: = state + $$\left(\stackrel{Ph-\mathrm{one}}{h}\right)$$

.

The worst case coder complexity is P ∙ (P-1) multiplications and P ∙ (P-1) divisions, as well as P-1 divisions.

10 illustrates a decoding process conducted by a decoding apparatus according to an embodiment of the present invention. In this embodiment, decoding is performed on a per-chip basis.

At 110, values are initialized. The device for decoding stores the number of states of the event that it took as an input value in the variable s. In addition, the number of slots containing frame events, as indicated by the number of event slots, is stored in the variable p. In addition, the total number of slots contained in the frame, as indicated by the number of frame slots, is stored in variable N.

At 120, a TsdSepData [t] value is initialized with a value of 0 for all frame slots. The TsdSepData bitmap is the output to be generated. It indicates for each slot position t whether the event contains a slot with the corresponding slot position (TsdSepData [t] = 1) or not (TsdSepData [t] = 0). At step 120, the corresponding values of all frame slots are initialized with a value of 0.

At step 130, the variable k is initialized with a value of N-1. In this embodiment, the slots of a frame containing N elements are numbered 0, 1, 2, ..., N-1. Setting k = N-1 means that the slot with the highest number of slots is considered the first.

At step 140, it is considered whether k≥0 is true. If k <0, the decoding of the positions of the slots has been completed, and the process stops, otherwise the process continues at step 150.

At step 150, it is checked whether p> k is true. If p is greater than k, this means that all remaining slots contain the event. The process continues to step 230, in which all the TsdSepData field values of the remaining slots 0, 1, ..., k are set to 1, indicating that each of the remaining slots contains an event. In this case, the process then stops. However, if at step 150 it is found that p is not greater than k, the decoding process continues at step 160.

, c используется как пороговое значение.At step 160, the value c = $$\left({}_p^k\right)$$

, c is used as a threshold value.

At step 170, it is checked whether the (ultimately updated) number of event states s is greater than c, or equal to it, where c is the threshold value just calculated at step 160.

If s is less than c, this means that the slot in question (with slot position k) does not contain an event. In this case, it is not necessary to take an additional action, since TsdSepData [k] has already been set to 0 for this slot in step 140. The process then continues to step 220. At step 220, k is set so that k: = k-1 , and the next slot is considered.

However, if the check in step 170 shows that s is greater than or equal to c, this means that the slot k in question contains an event. In this case, the number of event states s is updated and set to s: = s-c in step 180. In addition, TsdSepData [k] is set to 1 in step 190 to indicate that slot k contains the event. Moreover, at step 200, p is set to p-1, indicating that the remaining slots to be examined now contain only p-1 event slots.

At step 210, it is checked whether p is 0. If p is 0, the remaining slots do not contain events, and the decoding process ends. Otherwise, at least one of the remaining slots contains an event, and the process continues to step 220, where the decoding process continues with the next slot (k-1).

The decoding process of the embodiment illustrated in FIG. 10 generates an TsdSepData array as an output value, indicating for each slot k of the frame whether the slot contains an event (TsdSepData [k] = 1) or not (TsdSepData [k] = 0).

Returning to FIG. 9c, the decoding apparatus 60 of the embodiment in which the apparatus implements the decoding process illustrated in FIG. 10 comprises a slot selector 90 that decides which slots to consider. Regarding FIG. 10, such a slot selector would be capable of executing steps 130 and 220 of the process of FIG. 10. A suitable analysis unit 70 of this embodiment would be configured to execute processing steps 140, 150, 170 and 210 of FIG. 10. The generating unit 80 according to such an embodiment would be able to carry out all the other processing steps of FIG. 10.

11 shows pseudo code implementing decoding of positions of slots containing events according to an embodiment of the present invention.

12 illustrates an encoding process conducted by an encoding apparatus according to an embodiment of the present invention. In this embodiment, encoding is performed on a per-chip basis. The aim of the encoding process according to the embodiment illustrated in FIG. 12 is to generate the number of event states.

At 310, values are initialized. p_s is initialized with a value of 0. The number of event states is generated by successfully updating the variable p_s. When the encoding process is completed, p_s will carry the number of event states. Step 310 also initializes the variable k by setting k to a value of k: = the number of slots containing events in the frame, -1.

At 320, the variable "slots" is set to slots: = tsdPos [k], where tsdPos is an array holding the positions of the slots containing the events. The positions of the slots in the array are stored in ascending order.

At 330, a check is made to check if k≥slots is true. If so, the process is interrupted. Otherwise, the process continues at block 340.

.At step 340, the value c = $$\left({}_{k+\text{one}}^{slots}\right)$$

.

At step 350, the variable p_s is updated and set to p_s: = p_s + c.

At step 360, k is set to k: = k-1.

Then, at step 370, a check is performed to check if k≥0 is true. In this case, the next slot k-1 is considered. Otherwise, the process is interrupted.

13 is a pseudo code implementing encoding the positions of slots containing events according to an embodiment of the present invention.

FIG. 14 illustrates an apparatus 410 for decoding slot positions containing events in an audio frame according to a further embodiment of the present invention. Again, as in FIG. 9a, the number of FSN frame slots indicating the total number of audio frame slots, the number of ESON event slots indicating the number of slots containing audio frame events, and the number of ESTN event states are supplied to decoding apparatus 410. The decoding apparatus 410 differs from the apparatus of FIG. 9a in that it further comprises a frame splitting unit 440. The frame splitting unit 440 is configured to split the frame into a first frame section containing a first set of frame slots, and into a second frame section containing a second set of frame slots, and in which slot positions containing events are determined individually for each frame section. In addition, the positions of the slots containing the events can be determined by re-dividing the frame or sections of the frame into even smaller sections of the frame.

The “split-based” decoding of the decoding apparatus 410 of this embodiment is based on the following concepts that can be applied to embodiments of the decoding apparatus, the encoding apparatus, the decoding method, and the encoding method of slot positions that contain events in an audio signal frame. The following concepts also apply to related computer programs and encoded signals:

Split-based decoding is based on the idea that a frame is divided into two sections A and B of the frame, with each section of the frame containing a set of slots in which section A of the frame contains N _a slots and in which section B of the frame contains N _b slots, and that N _a + N _b = N. The frame can be arbitrarily divided into two sections, preferably so that sections A and B have approximately the same total number of slots (for example, so that N _a = N _b or N _a = N _b -1). By splitting the frame into two sections, the task of determining the positions of the slots where the events met is also divided into two subtasks, namely, determining the positions of the slots where the events occurred in section A of the frame, and determining the positions of the slots where the events met in section B of the frame.

In this embodiment, it is again assumed that the decoding apparatus is aware of the number of frame slots, the number of slots containing the frame event, and the number of event states. To solve both subproblems, the decoding device must be aware of the number of slots of each section of the frame, the number of slots where events occurred for each section of the frame, and the number of event states of each section of the frame (this number of states of the event of the frame section is now called the "number of substates of the event" )

Since the decoding device itself divides the frame into two sections of the frame, it alone knows that section A of the frame contains N _a slots, and section B of the frame contains N _b slots. The determination of the number of slots containing events for each of the two sections of the frame is based on the following survey results.

Since the frame was divided into two sections, each of the slots containing the events is now located either in section A or in section B. In addition, it is assumed that P is the number of slots containing frame section events, and N is the total number of slots section of the frame, and that f (P, N) is a function that returns the number of different combinations of slot positions with events of the frame section, then the number of different combinations of slot positions with events of the entire frame (which was divided into section A and section B) is:

The number of slots containing events in section A The number of slots containing events in section B The number of different combinations in the whole audio frame with this configuration 0 P f (0, N _a ) ∙ f (P, N _b ) one P-1 f (1, N _a ) ∙ f (P-1, N _b ) 2 P-2 f (2, N _a ) ∙ f (P-2, N _b ) .... .... .... P 0 f (P, N _a ) ∙ f (0, N _b )

Based on the foregoing considerations, according to an embodiment, all combinations with the first configuration, where section A has 0 slots containing events, and where section B has P slots containing events, should be encoded with the number of event states less than the first threshold value . The number of event states can be encoded as an integer value that is positive or 0. Since there are only f (0, N _a ) ∙ f (P, N _b ) combinations with the first configuration, a suitable first threshold value can be f (0, N _a ) ∙ f (P, N _b ).

All combinations with the second configuration, where section A has 1 slot containing events, and where section B has P-1 slots containing events, must be encoded with the number of event states greater than or equal to but less than the first threshold value than the second value, or equal to it. Since there are only f (1, N _a ) ∙ f (P-1, N _b ) combinations with the second configuration, a suitable second value might be f (0, N _a ) ∙ f (P, N _b ) + f (1 , N _a ) ∙ f (P-1, N _b ). The number of event states for combinations with other configurations is defined similarly.

According to an embodiment, decoding is performed by dividing the frame into two sections A and B of the frame. Then, it is checked whether the number of event states is less than the first threshold value. In a preferred embodiment, the first threshold value may be f (0, N _a ) ∙ f (P, N _b ).

If the number of event states is less than the first threshold value, it can be concluded that section A contains 0 slots containing events, and section B contains all P slots of the frame where the events occurred. Decoding is then performed for both partitions with an appropriately defined number representing the number of slots containing events of the corresponding partition. In addition, the first number of event states is determined for section A, and the second number of event states is determined for section B, which are respectively used as the new number of event states. For the purposes of this document, the number of event states of a frame section event is referred to as the “number of event sub-states”.

However, if the number of event states is greater than or equal to the first threshold value, the number of event states can be updated. In a preferred embodiment, the number of event states can be updated by subtracting the value from the number of event states, preferably by subtracting the first threshold value, for example f (0, N _a ) ∙ f (P, N _b ). In the next step, it is checked whether the updated number of event states is less than the second threshold value. In a preferred embodiment, the second threshold may be f _{(1, N a) ∙ f (P} - 1, N _b). If the number of event states is less than the second threshold value, then it can be obtained that section A contains 1 slot containing events, and section B contains P-1 slots containing events. Decoding is then carried out for both sections with an appropriately defined number of slots containing the events of each section. The first event substate value is used to decode section A, and the second event substate value is used to decode section B. However, if the number of event states is greater than or equal to the second threshold value, the number of event states can be updated. In a preferred embodiment, the number of event states can be updated by subtracting the value from the number of event states, preferably f (1, N _a ) ∙ f (P-1, N _b ). The decoding process is similarly applied to the remaining distribution possibilities of slots containing events regarding two sections of the frame.

In an embodiment, the event substate value for section A and the event substate value for section B can be used to decode section A and section B, where both event substate values are determined by dividing:

value of event states / f (number of slots containing events of section B, N _b ).

Preferably, the number of substates of the event of section A is an integer part of the above division, and the number of substates of the event of section B is the remainder of this division. The number of event states used in this division may be the initial number of frame event states or the updated number of event states, for example, updated by subtracting one or more threshold values, as described above.

To illustrate the above concept of separation-based decoding, a situation is considered where a frame has two slots containing events. In addition, if f (p, N) is again a function that returns the number of different combinations of slot positions with frame section events, in which p is the number of slots containing frame section events, and N is the total number of slots of this frame section, then, for each of the possible position distributions, the result is the following number of possible combinations:

Items in Section A Position in section B The number of combinations in this configuration 0 2 f (0, N _a ) ∙ f (2, N _b ) one one f (1, N _a ) ∙ f (1, Nb) 2 0 f (2, N _a ) ∙ f (0, N _b )

Thus, the conclusion can be made that if the number of states of the encoded frame events is smaller than f (0, N _a) ∙ f (2, N _b), the slots comprising events shall be distributed as 0 and 2. Otherwise, f (0, N _a ) ∙ f (2, N _b ) is subtracted from the number of event states, and the result is compared with f (1, N _a ) ∙ f (1, N _b ). If it is smaller, then the positions are distributed as 1 and 1. Otherwise, only the distribution of 2 and 0 remains, and the positions are distributed as 2 and 0.

Below is provided a pseudo code according to an embodiment for decoding the positions of slots containing certain events (here: "pulses") in an audio signal frame. In this pseudo-code, "pulses_a" is the (estimated) number of slots containing events in section A, and "pulses_b" is the (estimated) number of slots containing events in section B. In this pseudo-code, the (ultimately updated) number of event states is called "state". The numbers of event substates of sections A and B are still co-encoded in the state variable. According to the co-coding scheme of the embodiment, the number of substates of an event from A (herein referred to as “state_a”) is an integer part of the division state / f (pulses_b, N _b ), and the number of substates of an event from B (herein referred to as “state_b" ) is the remainder of this division. In addition, the length (total number of section slots) and the number of coded positions (number of slots containing events in a section) of both sections can be decoded using the same approach:

Function x = decodestate (state, pulses, N)

1. Break the vector into two sections of length Na and Nb.

2. For pulses_a from 0 to pulses

a. pulses_b = pulses-pulses_a

b. if state <f (pulses_a, Na) * f (pulses_b, Nb), then

stop the "for" loop.

c. state: = state-f (pulses_a, Na) * f (pulses_b, Nb)

3. The number of possible states for section B is

no_states_b = f (pulses_b, Nb)

4. States state_a and state_b of sections A and B,

respectively, are the integer part and the remainder

from the division state / no_states_b.

5. If Na> 1, then the decoded partition vector A

it turns out recursively as

xa = decodestate (state_a, pulses_a, Na)

Otherwise (Na == 1), and the vector xa is a scalar, and

we can set xa = state_a.

6. If Nb> 1, then the decoded section vector B

it turns out recursively as

xb = decodestate (state_b, pulses_b, Nb)

Otherwise (Nb == 1), and the vector xb is a scalar, and

we can set xb = state_b.

7. The final output signal x is obtained by

unions xa and xb as x = [xa xb].

The output of this algorithm is a vector that has one (1) at each coded position (i.e., the slot position for the slot containing the event) and zero (0) elsewhere (i.e. at slot positions that are not contain events).

Below is provided a pseudo code according to an embodiment for encoding slot positions containing events in an audio signal frame that uses the same variable names with the same value as above:

State = encodestate (x, n) function

1. Break the vector into two sections xa and xb

the length of Na and Nb.

2. Count pulses in sections A and B in pulses_a and

pulses_b and set pulses = pulses_a + pulses_b.

3. Set the state to 0.

4. For k from 0 to pulses_a-1

a. state: = state + f (k, Na) * f (pulses-k, Nb).

5. If Na> 1, encode section A as

state_a = encodestate (xa, Na);

Otherwise (Na == 1), set state_a = xa.

6. If Nb> 1, encode section B as

state_b = encodestate (xb, Nb);

Otherwise (Nb == 1), set state_b = xb.

7. Encoding states together

state: = state + state_a * f (pulses_b, Nb) + state_b.

Here, it is assumed that, similarly to the decoder algorithm, each coded position (i.e., the position of the slot for the slot containing the event) is identified by one (1) in the vector x, and all other elements are zero (0) (i.e., in the positions slots that do not contain events.)

The above recursive methods formulated in pseudo-code can easily be implemented in a non-recursive manner using standard methods.

According to an embodiment of the present invention, the function f (p, N) can be implemented as a correspondence table. When the positions do not overlap, as, for example, in the current context, the function of the number of states f (p, N) is simply a binomial function that can be calculated over the network. Is available

According to an embodiment of the present invention, both the encoder and the decoder have a loop with a counter, where the product f (pk, N _a ) * f (k, N _b ) is calculated for consecutive values from k. For efficient calculation, this can be written as

In other words, the sequential terms for subtraction / addition (in step 2b and 2c in the decoder and in step 4a in the encoder) can be calculated by three multiplications and one division per iteration.

Similarly, as in the method described previously, the state of a long vector (multi-slot frame) can be a very large integer that easily extends the presentation length in standard processors. As a result, it will be necessary to use arithmetic functions that can handle very large integers.

Regarding complexity, the method discussed here, in contrast to the above post-bit processes, is a divide-and-conquer algorithm. Assuming that the length of the input vector has a power of two, then recursion has a depth of log2 (N).

Since the number of pulses remains constant at each depth of the recursion, the number of iterations of the loop with the counter is the same for each recursion. It follows that the number of cycles is pulses ∙ log2 (N).

As explained above, each update _{f (pk, N a) ∙} f (k, N b) can be done using three multiplications and one division.

It should be noted that subtractions and comparisons in the decoder can be assumed as one operation.

It can be easily seen that partitions combine log2 (N) -1 times. When co-encoding states in an encoder, it is therefore necessary to multiply and add log2 (N) -1 times. Similarly, when decoding states together in a decoder, it is necessary to divide log2 (N) -1 times.

It should be noted that of divisions, only when co-coding states in a decoder, divisions are needed where the denominator is a long integer. Other divisions always have relatively short integers in the denominator. Since divisions with long denominators are the most difficult operations, they should be avoided when possible.

So, this number of arithmetic operations with long integers occurs in the decoder

Multiplications (3 ∙ pulses + 1) ∙ log2 (N) -1 Divisions (pulses + 1) ∙ log2 (N) -1 of which divisions with a long denominator log2 (N) -1 Addition and Subtraction pulses ∙ log2 (N) Similarly, in the encoder occur Multiplications (3 ∙ pulses + 1) ∙ log2 (N) -1 Divisions (pulses + 1) ∙ log2 (N) -1 of which divisions with a long denominator 0 Addition and Subtraction (pulses + 2) ∙ log2 (N)

Only log2 (N) -1 divisions with a long denominator are required.

In further embodiments, the above-described embodiments that comprise or are configured to use recursive processing steps are modified so that some or all of the recursive processing steps are implemented in a non-recursive manner using standard methods.

FIG. 15 illustrates a device (510) for encoding slot positions containing events in an audio frame according to an embodiment. The encoding device (510) comprises an event state number generator (530), which is configured to encode slot positions by encoding the number of event states. In addition, the device comprises a slot information unit (520) configured to provide the number of frame slots and the number of event slots to an event state generator (530). The event state number generator may implement one of the above encoding methods.

In a further embodiment, an encoded audio signal is provided. The encoded audio signal contains the number of event states. In another embodiment, the encoded audio signal further comprises a number of event slots. Moreover, the encoded audio signal frame may also contain the number of frame slots. In the audio frame, the positions of slots containing events in the audio frame can be decoded according to one of the decoding methods described above. In an embodiment, the number of event states, the number of event slots, and the number of frame slots are transmitted, so that the positions of the slots containing the events in the audio frame can be decoded using one of the above methods.

The inventive encoded audio signal may be stored in a digital storage medium or non-transitory storage medium or may be transmitted in a transmission medium such as a wireless transmission medium or a wired transmission medium such as the Internet.

The following explains USAC syntax definitions configured to support a transient control decorrelator (TSD) according to an embodiment.

16, MPS data (MPEG Surround) 212 is illustrated. MPS data 212 is a data block containing useful data for the stereo module MPS 212. MPS data 212 contains TSD data.

17 depicts the syntax of TSD data. They contain the number of transient slots (bsTsdNumTrSlots) and the TSD transient phase data (bsTsdTrPhaseData) for the slots in the MPS 212 data frame. If the slot contains transient data (TsdSepData [ts] is set to 1), bsTsdTrPhaseDataTata tase phase data ] is set to 0.

nBitsTrSlots sets the number of bits used to carry the number of transient slots (bsTsdNumTrSlots). nBitsTrSlots depends on the number of slots in the MPS 212 data frame (numSlots). On Fig illustrated the relationship of the number of slots in the data frame of the MPS 212 and the number of bits used to transfer the number of slots with transients.

In Fig.19 sets the value of tempShapeConfig. tempShapeConfig indicates the mode of operation of the temporary formation (STP or GES) or the activation of transient-controlled decorrelation in the decoder. If tempShapeConfig is set to 0, temporary shaping does not apply at all; if tempShapeConfig is set to 1, temporary subband processing (STP) is applied; if tempShapeConfig is set to 2, controlled envelope shaping (GES) is applied; and if tempShapeConfig is set to 3, transient control decorrelation (TSD) is applied.

20 illustrates the syntax of TempShapeData data. If bsTempShapeConfig is set to 3, TempShapeData contains bsTsdEnable, indicating that TSD is included in the frame.

21, a decorrelator block D according to an embodiment is illustrated. The decorrelator unit D in the OTT decoding unit comprises a signal separation unit, two decorrelator structures and a signal combining unit.

D _AP means: frequency-independent decorrelator, as specified in subsection 7.11.2.5 (frequency-independent decorrelator).

D _TR means: decorrelator of transients.

с транзиентами и поток $${\nu }_{X,nonTr}^{n,k}$$

без транзиентов согласно:If the TSD tool is active in the current frame, i.e. if (bsTsdEnable == 1), the input signal is split into a stream $${\nu }_{X,Tr}^{n,k}$$

with transients and flow $${\nu }_{X,nonTr}^{n,k}$$

without transients according to:

The flag of the post-stranded transients separation TsdSepData (n) is decoded from the variable-length codeword bsTsdCodedPos by TsdTrPos_dec (), as described below. The length of the codeword from bsTsdCodedPos, i.e. nBitsTsdCW, calculated according to:

Returning to FIG. 11, FIG. 11 illustrates decoding of bsTsdCodedPos transient slot separation data in TsdSepData [n] according to an embodiment. An array of length numSlots, consisting of '1' for coded positions with transients and '0' for others, is specified as illustrated in FIG. 11.

If the TSD tool is disabled in the current frame, i.e. if (bsTsdEnable == 0), the input signal is processed as if TsdSepData (n) = 0 for all n.

The components of the signal with transients are processed in the structure of the decorrelator of transients D _TR , as follows:

, $$d_{X,Tr}^{n,k}=\{\begin{array}{ll}{e}^{j{\varphi }_{TSD}^n}\cdot v_{X,Tr}^{n,k}\hfill & \text{if bsTsdEnable}=\mathrm{one}\hfill \\ 0\hfill & \text{, otherwise}\hfill \end{array}$$

,

Where

The signal components without transients are processed in the frequency-independent decorrelator D _AP , as specified in the next subsection, which outputs the output signal of the decorrelator for signal components without transients

The output signals of the decorrelator are added to form a decorrelated signal containing components with both transients and without transients

On Fig illustrates the syntax of EcData containing bsFrequencyResStrideXXX. The bsFreqResStride syntax element provides the ability to use broadband tags in MPS. XXX should be replaced with a data type value (CLD, ICC, IPD).

A decorrelator controlling transients in the structure of an OTT decoder provides the possibility of applying a specialized decorrelator to components that are similar to applause with transients. The activation of this TSD function is controlled by the encoder-generated flag bsTsdEnable, which is transmitted once per frame.

The TSD data in the module for converting two channels into one channel (R-OTT) of the encoder is generated as follows:

- Launch a semantic signal classifier that detects applause-like signals. The classification result is transmitted once per frame: the bsTsdEnable flag is set to 1 for applause-like signals, otherwise it is set to 0.

- If bsTsdEnable is set to 0 for the current frame, no additional TSD data is generated / transmitted for this frame.

- If bsTsdEnable is set to 1 for the current frame, do the following:

○ Enable broadband calculation of spatial OTT parameters.

○ Detect transients in the current frame (binary solution for each MPS time slot).

○ Encode the positions of the slots with transients tsdPosLen in the tsdPos vector according to the following pseudo-code, where the positions of the slots in tsdPos are assumed in ascending order. FIG. 13 illustrates pseudo-code for encoding slot positions with transients in tsdPosLen.

○ Send the number of slots with transients (bsTsdNumTrSlots = (the number of detected slots with transients) -1).

○ Transfer coded positions with transients (bsTsdCodedPos).

○ For each transient slot, calculate the phase value, which represents the phase difference in broadband transmission between the signal obtained by downmix and the residual signal.

○ For each transient slot, encode and transmit the phase difference value for broadband transmission (bsTsdTrPhaseData).

Finally, FIG. 23 illustrates a signal flow diagram for generating TSD data in a two channel to one channel conversion (R-OTT) module.

Although some aspects have been described in the context of the device, it is clear that these aspects also represent a description of the corresponding method, where the unit or device corresponds to a step of the method, or an attribute of a step of the method. Likewise, aspects described in the context of a method step also provide a description of a corresponding unit, or element, or feature of a corresponding device.

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

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

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

Other embodiments comprise a computer program for performing one of the methods described herein stored on a computer-readable medium or in a non-transitory storage medium.

In other words, an embodiment of the invented method is therefore a computer program comprising program code for executing one of the methods described herein when a computer program is executed on a computer.

An additional embodiment of the invented methods is therefore a data medium (or digital storage medium or computer-readable medium) comprising a computer program recorded thereon for performing one of the methods described herein.

An additional embodiment of the invented method is therefore a data stream or a sequence of signals representing a computer program for performing one of the methods described herein. The data stream or sequence of signals may, for example, be configured to be carried over a data connection, for example via the Internet.

A further embodiment comprises processing means, for example, a computer or programmable logic device, configured to or configured to perform one of the methods described herein.

A further embodiment comprises a computer comprising a computer program installed thereon for executing one of the methods described herein.

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

The above embodiments are merely illustrative of the principles of the present invention. It should be understood that modifications and variations of these arrangements and the details described herein will be apparent to those skilled in the art. Accordingly, it is intended that the invention be limited only by the scope of the appended claims, and not by the specific details presented for the purpose of describing and explaining the embodiments herein.

Bibliography

[1] J. Breebaart, S. van de Par, A. Kohlrausch, E. Schuijers, "High-Quality Parametric Spatial Audio Coding at Low Bitrates" in Proceedings of the AES 116 ^th Convention, Berlin, Preprint 6072, May 2004.

[2] J. Herre, K. Kjorling, J. Breebaart et al., "MPEG surround - the ISO / MPEG standard for efficient and compatible multi-channel audio coding", in Proceedings of the 122 ^th AES Convention, Vienna, Austria May 2007.

[3] Pulkki, Ville: "Spatial Sound Reproduction with Directional Audio Coding," in J. Audio Eng. Soc., Vol. 55, No. 6, 2007.

[4] ISO / IEC International Standard "Information Technology - MPEG audio technologies - Part1: MPEG Surround", ISO / IEC 23003-1: 2007.

[5] J. Engdegard. H. Purnhagen, J. Roden, L. Liljeryd , "Synthetic Ambience in Parametric Stereo Coding" in Proceedings of the AES 116 ^th Convention, Berlin, Preprint, May 2004.

Claims (17)

1. Device (10; 40; 60; 410) for decoding an encoded audio signal containing an audio signal frame containing slots and events associated with these slots, comprising:
an analysis unit (20; 42; 70; 420) for analyzing the number of frame slots indicating the total number of audio frame slots, the number of event slots indicating the number of slots containing audio frame events, and the number of event states; and
a generating unit (30; 45; 80; 430) for generating an indication of a plurality of slot positions containing events in an audio signal frame using the number of frame slots, the number of event slots and the number of event states. 2. The device (10; 40; 60; 410) for decoding according to claim 1,
wherein the device (10; 40; 60; 410) for decoding is configured to decode the positions of the slots with transients in the frame of the audio signal. 3. The device (10; 40; 60; 410) for decoding according to claim 1,
in which the analyzing unit (20; 42; 70; 420) is configured to perform a check comparing the number of event states or the updated number of event states with a threshold value. 4. The device (10; 40; 60; 410) for decoding according to claim 3,
in which the analyzing unit (20; 42; 70; 420) is configured to perform verification by comparison, is
whether the number of event states or the updated number of event states is greater, greater, or equal, less than, less than, or equal to a threshold value, and
in which the generating unit (30; 45; 80; 430) is further configured to update the number of event states or the updated number of event states depending on the result of the check. 5. The device (10; 40; 60) for decoding according to claim 3,
wherein the device (10; 40; 60) for decoding further comprises a selector (90) slots,
wherein the slot selector (90) is configured to select a slot as the slot in question,
while the analyzing unit (20; 42; 70) is configured to perform checks in relation to the slot in question,
and the threshold value depends on the number of slots of the frame, the number of slots with events and the position of the considered slot inside the frame. 6. The device (10; 40) for decoding according to claim 5,
in which the analyzing unit (20; 42; 70) is configured to perform a check comparing the number of event states or the updated number of event states with a threshold value, where the threshold value is

,
where N is the total number of slots of the audio frame, where P is the number of slots containing events of the audio frame or the considered portion of the audio frame, and where h
is the position of the slot in question within the frame. 7. The device (10; 40; 410) for decoding according to claim 1,
wherein the device for decoding (10; 40; 410) further comprises a frame splitting unit (440),
the block (440) for dividing the frame is configured to split the frame into a first section of a frame containing a first set of frame slots, and into a second section of a frame containing a second set of frame slots, and the device (10; 40; 410) for decoding additionally configured to determine the positions of slots containing events for each of the sections of the frame individually. 8. The device (10; 40; 60; 410) for decoding according to claim 1, further comprising:
an audio signal processor (50) for generating an output audio signal using indicating a plurality of slot positions containing events in an audio signal frame using the number of frame slots, the number of event slots and the number of event states. 9. The device (10; 60; 410) for decoding according to claim 8,
in which the audio signal processor (50) is configured to generate an output audio signal according to the first method, if the indication of the plurality of positions of slots containing the events is in a first indicating state, and in which the audio signal processor (50) is configured to generate an audio output according to a different second method if the indication of the multiple positions of the slots containing the events is in the second indication state,
which differs from the first indication state. 10. The device (10; 40; 60; 410) for decoding according to claim 9,
in which the audio signal processor (50) is configured such that the first method comprises the step of using a transformer decorrelator (56) to decode the slot, if the first indication state indicates that the slot contains a transient, and wherein the second method comprises the step of the decoding of the slot uses a second decorrelator (54) if the second indication state indicates that the slot does not contain a transient. 11. A device (510) for encoding slot positions containing events in an audio signal frame, comprising:
an event state number generator (530) for encoding slot positions by encoding an event state number; and
a slot information unit (520) configured to provide the number of frame slots indicating the total number of audio frame slots and the number of event slots indicating the number of slots containing audio frame events to the event state generator (530),
wherein the number of event states, the number of frame slots and the number of event slots together indicate a plurality of slot positions containing events in the audio signal frame. 12. The device (510) for encoding according to claim 11,
wherein the event state number generator (530) is configured to generate an event state number by adding a positive integer value for each slot containing the event. 13. The device (510) for encoding according to claim 11,
wherein the event state number generator (530) is configured to generate the number of event states by determining the first number of event substates for the first section of the frame, by determining the second number of event substates for the second section of the frame, and by combining the first and second number of event states to generate a number event states. 14. A method for decoding positions of slots containing events in an audio signal frame, comprising the steps of:
analyzing the number of frame slots indicating the total number of audio frame slots, the number of event slots indicating the number of slots containing audio frame events, and the number of event states; and
generating an indication of the plurality of slot positions containing events in the audio signal frame using the number of frame slots, the number of event slots, and the number of event states. 15. A method of encoding slot positions containing events in an audio signal frame, comprising the steps of:
receive or determine the number of frame slots indicating the total number of audio frame slots,
receive or determine the number of event slots indicating the number of slots containing events of the audio frame,
encode the number of event states based on the number of event states, the number of frame slots and the number of event slots, so that the indication of the multiple positions of the slots containing events in the audio signal frame can be encoded using the number of frame slots, the number of event slots and the number of event states. 16. A computer-readable medium containing a computer program recorded on it, which, when executed on a computer, instructs the computer to perform a method of decoding the positions of slots with events in an audio signal frame according to claim 14. 17. A computer-readable medium containing a computer program recorded on it, which, when executed on a computer, instructs the computer to perform a method of encoding slot positions with events in an audio signal frame according to claim 15.

Images