RU2575809C2 - Encoder using forward aliasing cancellation - Google Patents

Abstract

FIELD: physics, computer engineering.

SUBSTANCE: invention relates to an encoder which supports switching between time-domain aliasing cancellation transform coding mode and time-domain coding mode. The result is achieved by adding a further syntax portion to frames, depending on which the parser of a decoder may select between a first action of expecting the current frame to contain, and thus reading forward aliasing cancellation data from the current frame and a second action of not expecting the current frame to contain, and thus not reading forward aliasing cancellation data from the current frame.

EFFECT: codec is made less liable to frame loss.

20 cl, 27 dwg

Description

The present invention relates to a codec supporting a coding mode with conversion with suppression of sampling interference in the time domain and a coding mode of the time domain, as well as direct suppression of sampling interference to switch between both modes.

It is convenient to combine different coding modes in order to encode common audio signals representing a combination of different types of audio signals, such as speech, music, or the like. Separate encoding modes can be adapted for specific types of audio, and thus a multi-mode audio encoder can take advantage of a time-varying encoding mode corresponding to a change in the type of audio content. In other words, a multi-mode audio encoder may decide, for example, to encode portions of an audio signal having speech content using an encoding mode specifically dedicated to encoding speech, and use a different encoding mode to encode other portions of audio content representing non-speech content, such as music. Time-domain coding modes, such as linear codebook drive predictive coding modes, tend to be more suitable for encoding speech content, while transform coding modes tend to be superior to time-domain coding modes, as it relates to music coding.

There were already solutions to solve the problem of copying with the coexistence of different types of audio within the same audio signal. The current USAC, for example, offers a switch between the frequency-domain coding mode, mainly consistent with the AAC standard, and two additional linear prediction modes, similar to the AMR-WB plus subframe modes, namely based on MDCT (Modified Discrete Cosine Transformation) a variant of the TCX mode (TCX = transform encoded excitation) and the ACELP mode (linear adaptive codebook excitation prediction). To be more precise, in the AMR-WB + standard, TCX is based on the DFT transform, but in the USAC TCX has the basis of the MDCT transform. A specific frame structure is used to switch between the FD coding region, similar to AAC, and the linear prediction region, similar to AMR-WB +. The AMR-WB + standard itself uses its own personnel structure, forming a subframe structure similar to the USAC standard. The AMR-WB + standard allows for a specific unit configuration that subdivides AMR-WB + frames into smaller TCX and / or ACELP frames. Similarly, the AAC standard uses a basic frame structure, but provides the ability to use different window lengths in order to encode the contents of the frame with conversion. For example, either a long window and an associated long transform length can be used, or eight short windows with associated shorter transform lengths.

MDCT causes interference sampling. This, therefore, is true at the boundaries of TCX and FD frames. In other words, just like any frequency domain encoder using MDCT, sampling interference occurs in regions of window overlap, which are suppressed by adjacent frames. That is, for transitions between two FD frames or between two TCX- (MDCT-) frames or transitions between either FD to TCX or TCX to FD, there is an implicit suppression of sampling interference through the overlap / add procedure as part of reconstruction on the decoding side . Then there is no more sampling interference after adding overlap. However, in the case of transitions from ACELP, there is no inherent suppression of sampling interference. Then a new tool should be introduced, which could be called FAC (Direct Suppression of Sample Interference). FAC is used to suppress sampling interference from adjacent frames if they are different from ACELP.

In other words, sampling interference suppression problems always arise when transitions occur between a transform coding mode and a time domain coding mode, such as, for example, ACELP. In order to perform the conversion from the time domain to the spectral region as efficiently as possible, transform-encoding transformations with noise suppression in the time domain, such as MDCT, i.e. an encoding mode using an overlapping transform, where the overlapping window portions of the signal are converted using a transform according to which the number of transform coefficients per section is less than the number of samples per section, so that sampling interference occurs because individual sections are related, with sampling interference, being suppressed by suppressing sampling interference in the time domain, i.e. by adding overlapping sections with interference sampling of adjacent sections of the reconverted signal. MDCT is such a time-domain noise reduction transform. Adverse TDAC (Time Domain Sample Noise Reduction) is not available for transitions between the TC coding mode and the time domain coding mode.

In order to solve this problem, direct sampling interference cancellation (FAC) can be used, according to which the encoder communicates additional FAC data within the current frame inside the data stream whenever there is a change in encoding mode from encoding with conversion to time domain encoding. This, however, forces the decoder to compare the encoding modes of consecutive frames in order to determine if the currently decoded frame contains FAC data within its syntax. This, in turn, means that there may be frames for which the decoder cannot be sure as to whether it should be read or parsed from the current frame. In other words, in the case where one or more frames were lost during transmission, the decoder does not know for the immediately following (received) frames as to whether a change in the encoding mode has occurred, and as to whether the bit stream of encoded data of the current frame contains FAC data. Therefore, the decoder must discard the current frame and wait for the next frame. Alternatively, the decoder can parse the current frame by performing decoding tests, one assuming that FAC data is not present and the other assuming that FAC data is not present, with the subsequent decision that one of the two alternatives is not executed. The decoding process most likely could cause the decoder to crash under one of two conditions. That is, in reality, the possibility of the latter is not an acceptable approach. The decoder must at any time know how to interpret the data and not rely on its own assumptions about how to interpret the data.

Therefore, it is an object of the present invention to provide a codec that is more error tolerant or resistant to frame loss, with, however, support for switching between a transform encoding mode with suppression of time domain sampling interference and a time domain encoding mode.

This goal is achieved by the subject of the invention according to any one of the independent claims, attached by this.

The present invention is based on the discovery that a more error-resistant and frame-lossless codec supporting switching between a time-domain transform coding mode and a time-domain coding mode is achievable if an additional syntax section is added to the frames, depending on which the parser of the decoder can choose between the first wait action that the current frame contains, and thus by reading the data of the direct suppression of sampling interference from the current frame, and the second action of the surprise that the current frame contains, and, thus, not reading the data of the direct suppression of sampling interference from the current frame. In other words, while a small part of the coding efficiency is lost due to the provision of the second syntax section, this is just the second syntax section, which is provided for the ability to use the codec in the case of a communication channel with frame loss. Without the second syntax section, the decoder would not be able to decode any section of the data stream after loss and would crash when trying to resume parsing. Thus, in an error-prone environment, coding efficiency is prevented from tending to zero by introducing a second syntax section.

In addition, preferred embodiments of the present invention are the subject of the dependent claims. In addition, preferred embodiments of the present invention are described in detail below with respect to the drawings. In particular,

figure 1 shows a schematic block diagram of a decoder according to a variant implementation;

figure 2 shows a schematic block diagram of an encoder according to a variant implementation;

figure 3 shows a block diagram of a possible implementation of the reconstructor of figure 2;

figure 4 shows a block diagram of a possible implementation of the FD-decoding module of figure 3;

figure 5 shows a block diagram of a possible implementation of the LPD decoding module of figure 3;

6 is a schematic diagram illustrating an encoding procedure for generating FAC data in accordance with an embodiment;

7 is a schematic diagram of a possible re-mapping for a TDAC transform in accordance with an embodiment;

Figs. 8, 9 are block diagrams for illustrating the linear structure of the FAC data path with the additional processing encoder in the encoder in order to test a change in the encoding mode in the sense of optimization;

figure 10, 11 shows a block diagram of the processing of the decoder in order to reach the FAC data, figures 8 and 9, from the data stream;

12 is a schematic diagram of a reconstruction based on FAC on a decoding side across from frame boundaries of different encoding modes;

on Fig, 14 schematically shows the processing performed by the transition processor of figure 3 in order to perform the reconstruction of figure 12;

15-19 show portions of a syntax structure in accordance with an embodiment; and

20-22 show sections of a syntax structure in accordance with another embodiment.

1 shows a decoder 10 according to an embodiment of the present invention. Decoder 10 is for decoding a data stream containing a sequence of frames 14a, 14b and 14c into which time segments 16a-c of information signal 18 are encoded, respectively. As illustrated in FIG. 1, time segments 16a through 16c are nonoverlapping segments that are directly adjacent to each other in time and are sequentially ordered in time. As illustrated in FIG. 1, the time segments 16a through 16c may be of equal size, but alternative embodiments are also valid. Each of the time segments 16a through 16c is encoded into a corresponding one of the frames 14a through 14c. In other words, each time segment 16a through 16c is uniquely associated with one of the frames 14a through 14c, which, in turn, also has an order specified among them, which follows the order of the segments 16a through 16c, which are encoded into frames 14a by 14c, respectively. Although FIG. 1 suggests that each frame 14a through 14c is of equal length, measured in, for example, coded bits, this is, of course, not mandatory. Rather, the length of frames 14a to 14c may vary according to the complexity of the time segment 16a to 16c with which frame 14a to 14c is associated.

For ease of explanation of the embodiments described below, it is assumed that the information signal 18 is an audio signal. However, it should be noted that the information signal could also be any other signal, such as a signal output by a physical sensor or the like, such as an optical sensor or the like. In particular, the signal 18 can be sampled at a specific sampling rate, and the time segments 16a to 16c can cover directly successive portions of this signal 18, equal in time to the number of samples, respectively. The number of samples per time segment 16a to 16c may, for example, be 1024 samples.

The decoder 10 comprises a parser 20 and a reconstructor 22. The parser 20 is arranged to parse the data stream 12 and, when parsing the data stream 12, read the first syntax section 24 and the second syntax section 26 from the current frame 14b, i.e. e. frame to be decoded at the current moment. 1, for example, it is assumed that frame 14b is a frame that should be decoded at the current time, while frame 14a is a frame that was decoded immediately before. Each frame 14a through 14c has a first syntax section and a second syntax section included therein, with their significance or meaning described below. In FIG. 1, the first syntax section inside frames 14a through 14c is indicated by the rectangle “1” and the second syntax section is indicated by the rectangle “2”.

Naturally, each frame 14a through 14c also has additional information included therein, which is necessary to represent the associated time segment 16a through 16c, as described in more detail below. This information is indicated in FIG. 1 by a hatched block in which reference numeral 28 is used for additional information of the current frame 14b. The parser 20 is configured to, when parsing the data stream 12, also read information 28 from the current frame 14b.

The reconstructor 22 is configured to reconstruct the current time segment 16b of the information signal 18 associated with the current frame 14b, based on the additional information 28, using the selected one of the decoding mode with conversion with suppression of sampling noise in the time domain and the decoding mode of the time domain. The choice depends on the first syntax element 24. Both decoding modes are distinguished from each other by the presence or absence of any transition from the spectral region back to the time domain using repeated transformation. Repeated conversion (along with its corresponding transformation) introduces sampling interference, since separate time segments are considered, however, these sampling interference are compensated by suppressing sampling interference in the time domain, since transitions at the boundaries between consecutive frames encoded in the encoding mode with transformation with suppression of sampling noise in the time domain. The time-domain decoding mode does not make any re-conversion necessary. Rather, decoding remains in the time domain. Thus, generally speaking, the decoding mode with the conversion with suppression of sampling interference in the time domain of the reconstructor 22 involves a re-conversion that is performed by the re-constructor 22. This re-conversion assigns a first number of transform coefficients as obtained from the information 28 of the current frame 14b (being a decoding mode with TDAC conversion) to a segment of the reconverted signal having a sample length of a second number of samples that is longer than the first number, thereby calling sampling noise. The decoding mode of the time domain, in turn, can use the decoding mode with linear prediction, according to which the excitation and linear prediction coefficients are reconstructed from the information 28 of the current frame, which in this case is the encoding mode of the time domain.

Thus, as it became clear from the above, in the decoding mode with conversion with suppression of sampling noise in the time domain, the reconstructor 22 obtains from the information 28 a signal segment for reconstructing the information signal in the corresponding time segment 16b by means of repeated conversion. The segment of the transformed signal is longer than the current time segment 16b, and is involved in the reconstruction of the information signal 18 within the time section, which includes and extends beyond the time segment 16b. Figure 1 illustrates the conversion window 32 used in the conversion of the original signal during both conversion and re-conversion. You can see that the window 32 may contain a zero section 32 ₁ at its beginning and a zero section 32 ₂ at its rear end and sections 32 ₃ and 32 ₄ with sampling interference in the front and rear edges of the current time segment 16b, in which section 32 ₅ without interference sampling, where the window 32 is one, can be located between both sections 32 ₃ and 32 ₄ with interference sampling. Zero sections 32 ₁ and 32 ₂ are optional. It is also possible that only one of the null sections 32 ₁ and 32 _{2 is} present. As shown in FIG. 1, the window function may be monotonically increasing / decreasing within areas with sampling interference. Sampling interference occurs within portions 32 ₁ and 32 ₂ with sampling interference, where window 32 continuously leads from zero to one or vice versa. Sampling interference is not critical as long as the previous and next time segments are also encoded in a coding mode with conversion with suppression of sampling interference in the time domain. This feature is illustrated in FIG. 1 with respect to the time segment 16c. The dashed line illustrates the corresponding conversion window 32 'for the time segment 16c, the portion with sampling interference of which converges with portion 32 ₄ with sampling interference of the current segment 16b. Adding the signals of the converted segments of the time segments 16b and 16c by the reconstructor 22 suppresses the sampling noise of both segments of the converted signals between them.

However, in the case where the previous or next frame 14a to 14c is encoded in the time-domain encoding mode, the transition between different encoding modes results in the leading or trailing edge of the current time segment 16b, and in order to take into account the corresponding sampling interference, the data stream 12 contains data direct suppression of sampling interference within the corresponding frame immediately following the transition, to enable decoder 10 to compensate for sampling interference arising from this appropriate transition. For example, it may happen that the current frame 14b belongs to a time-domain transform transform coding mode, but the decoder 10 does not know whether the previous frame 14a belonged to a time-domain encoding mode. For example, frame 14a may be lost during transmission, and decoder 10, respectively, will not have access to it. However, depending on the encoding mode of frame 14a, the current frame 14b contains data of direct suppression of sampling interference in order to compensate for sampling interference that occurs in section 32 ₃ with sampling interference or not. Similarly, if the current frame 14b belonged to a time-domain coding mode, and the previous frame 14a was not received by the decoder 10, then the current frame 14b has direct suppression of sampling noise, included in it, or independent of the mode of the previous frame 14b. In particular, if the previous frame 14a belonged to a different encoding mode, i.e. mode with conversion with noise suppression of sampling in the time domain, then the data of direct noise suppression of sampling would be present in the current frame 14b to suppress sampling interference that otherwise occurs at the boundary between time segments 16a and 16b. However, if the previous frame 14a belonged to the same encoding mode, i.e. mode encoding the time domain, the parser 20 should not expect the presence of data directly suppressing sampling interference in the current frame 14b.

Therefore, the parser 20 uses the second parsing portion 26 to determine whether or not the direct sampling interference data 34 is present in the current frame 14b. When parsing the data stream 12, the parser 20 can select one of the first wait action that the current frame 14b contains, and thus, read the data directly suppressing sampling interference from the current frame 14b, and the second surprise action that the current frame contains , and thus, without reading the data 34 of the direct suppression of sampling interference from the current frame 14b, the choice being dependent on the second syntax section 26. If present, the reconstructor 22 is configured to perform to explicitly suppress sampling interference at the boundary between the current time segment 16b and the previous time segment 16a of the previous frame 14a using the data of direct suppression of sampling interference.

Thus, compared to a situation where the second syntax portion is not present, the decoder of FIG. 1 should not discard or unsuccessfully parse the current frame 14b, even if the encoding mode of the previous frame 14a is unknown to the decoder 10, for example, due to loss frame. Rather, the decoder 10 is able to use the second syntax portion 26 to determine if the current frame 14b has data for direct suppression of sampling noise 34. In other words, the second syntax section provides a clear criterion as to whether one of the alternatives applies and guarantees, i.e. FAC data for the boundary for a predictable frame, present or not, that any decoder can also behave independently of their implementation, even in the case of frame loss. Thus, the above-described embodiment introduces mechanisms to overcome the problem of frame loss.

Before a detailed description of the embodiments below, an encoder capable of generating the data stream 12 of FIG. 1 is described with reference to FIG. 2. The encoder of FIG. 2 is generally indicated by a reference numeral 40 and is used to encode the information signal into the data stream 12, so that the data stream 12 contains a sequence of frames into which the time segments 16a to 16c of the information signal are encoded, respectively. The encoder 40 includes a designer 42 and an insertion device 44. The designer is configured to encode the current time segment 16b of the information signal into information of the current frame 14b using the first selected one of the encoding mode with conversion with suppression of sampling noise in the time domain and the encoding mode of the time domain. The insertion device 44 is configured to insert information 28 into the current frame 14b, along with the first syntax section 24 and the second syntax section 26, in which the first syntax section reports the first selection, i.e. selection of coding mode. Constructor 42, in turn, is configured to determine direct sampling noise suppression data for directly sampling noise suppression at a boundary between the current time segment 16b and the previous time segment 16a of the previous frame 14a, and inserts direct sampling noise reduction data 34 into the current frame 14b, when the current frame 14b and the previous frame 14a are encoded using different modes from a transform coding mode with suppression of time domain sampling and coding mode time domain, and refrains from inserting any data of direct suppression of sampling interference into the current frame 14b when the current frame 14b and the previous frame 14a are encoded using the same modes from the encoding mode with transformation with the suppression of sampling interference in the time domain and the encoding mode of the time domain . That is, whenever the constructor 42 of the encoder 40 decides that it is preferable, in a certain sense of optimization, to switch from one of the two encoding modes to the other, the constructor 42 and the insertion device 44 are configured to detect and insert direct sampling noise reduction data 34 into the current frame 14b, and if the encoding mode between frames 14a and 14b is maintained, FAC data 34 is not inserted into the current frame 14b. In order to enable the decoder to detect from the current frame 14b, without knowing the contents of the previous frame 14a, whether FAC data 34 is present inside the current frame 14b, a specific syntax section 26 is defined depending on whether the current frame 14b and the previous frame 14a are encoded with using the same or different modes from the encoding mode with conversion with suppression of sampling noise in the time domain and the encoding mode of the time domain. Specific examples for understanding the second syntax section 26 will be indicated below.

An embodiment is further described, according to which the codec to which the decoder and encoder of the above-described embodiments belong, supports a special type of frame structure, according to which frames 14a to 14c themselves are subjects to form subframes, and there are two versions of the coding mode with canceled conversion time domain sampling interference. In particular, according to these embodiments, further described below, the first syntax section 24 associates the corresponding frame from which it was read with the first type of frame, hereinafter referred to as the FD encoding mode (frequency domain), or with the second type of frame, called hereinafter referred to as the LPD coding mode, and if the corresponding frame belongs to the second type of frame, associates subframes of the subunits of the corresponding frame composed of the number of subframes with the corresponding one of the first type of subframes RA and the second type of subframe. As will be described in more detail below, the first type of subframe can use the corresponding subframes that should be encoded by TCX, while the second type of subframe can use these corresponding subframes that should be encoded using ACELP, i.e. linear prediction of adaptive codebook excitation. Or, any other linear prediction codebook coding mode may also be used.

The reconstructor 22 of FIG. 1 is configured to process the capabilities of these different encoding modes. At this point, the reconstructor 22 may be constructed as shown in FIG. According to the embodiment of FIG. 3, the reconstructor 22 comprises two switches 50 and 52 and three decoding modules 54, 56 and 58, each of which is capable of decoding frames and subframes of a particular type, as will be described in detail below.

The switch 50 has an input to which information 28 of the currently decoded frame 14b is supplied, and a control input by which the switch 50 is controllable depending on the first syntax portion 25 of the current frame. The switch 50 has two outputs, one of which is connected to the input of the decoding module 54 responsible for FD decoding (FD = frequency domain), and the other is connected to the input of the sub-switch 52, which also has two outputs, one of which is connected to the input of the module 56 decoding responsible for decoding with linear prediction encoded with conversion of excitation, and the other to the input of module 58, responsible for decoding with linear prediction of excitation codebook. All decoding modules 54 through 58 derive signal segments by reconstructing the corresponding time segments associated with the respective frames and subframes from which these signal segments were obtained by the corresponding decoding mode, and the transition processor 60 receives the signal segments at its respective inputs to perform processing transition and suppression of sampling interference, described above and described in more detail below, to output the reconstructed information signal to its output. Transition handler 60 uses direct sampling interference suppression data 34, as illustrated in FIG.

According to the embodiment of FIG. 3, the reconstructor 22 operates as follows. If the first syntax section 24 associates the current frame with the first type of frame in the FD encoding mode, the switch 50 sends the information 28 to the FD decoding unit 54 to use frequency domain decoding as the first version of the decoding mode with time-domain conversion suppression for reconstructing the time segment 16b associated with the current frame 15b. Otherwise, i.e. if the first syntax section 24 associates the current frame 14b with the second type of frame in the LPD encoding mode, the switch 50 sends the information 28 to the sub-switch 52, which, in turn, works on the subframe structure of the current frame 14b. More specifically, in accordance with the LPD mode, the frame is divided into one or more subframes, a division corresponding to the division of the corresponding time segment 16b into non-overlapping sub-sections of the current time segment 16b, as will be described in more detail below with reference to the following figures. Syntax section 24 reports for each of one or more sub-sections whether it is associated with the first or second type of subframe, respectively. If the corresponding subframe belongs to the first type of subframe, the sub-switch 52 sends the corresponding information 28 belonging to this subframe to the TCX decoding unit 56 to use the linearly predicted decoding of the encoded transform excitation decoding as the second version of the decoding mode with transform suppression of sampling interference into time domain for reconstructing the corresponding sub-section of the current time segment 16b. If, however, the corresponding subframe does not belong to the second type of subframe, the sub-switch 52 sends the information 28 to the module 58 in order to perform linear prediction of the codebook excitation coding as a time-domain decoding mode for reconstructing the corresponding subsection of the current time segment 16b.

The segments of the reconstructed signal output by the modules 54 through 58 are placed by the transition processor 60 in the correct (presentation) time order by performing the corresponding transition processing and adding overlap and processing to suppress sampling noise in the time domain, as described above and described in more detail below.

In particular, the FD decoding unit 54 may be constructed as shown in FIG. 4 and operates as described below. 4, the FD decoding unit 54 comprises a dequantization device 70 and a re-conversion device 72, connected in series to each other. As described above, if the current frame 14b is an FD frame, that is sent to the module 54, and the dequantization device 70 dequantizes with a varying spectrum of transform coefficient information 74 within the information 28 of the current frame 14b using the scale factor information 76 also contained in the information 28 Scale factors were determined on the encoder side, using, for example, psychoacoustic principles to keep the quantization noise below the human protection threshold.

The re-conversion device 72 then re-converts the information of the dequantized transform coefficients to obtain a segment of the re-transformed signal extending in time within and outside the time segment 16b associated with the current frame 14b. As will be described in more detail below, the re-conversion performed by the re-conversion device 72 may be an IMDCT (Inverse Modified Discrete Cosine Transform) using a DCT IV followed by a sweep operation after which window processing using a window is performed re-conversion, which could be equal to or deviate from the conversion window used in generating information 74 conversion factors by using I have the above steps in reverse order, namely windowing, followed by a convolution operation, followed by DCT IV, followed by quantization, which can be controlled by psycho-acoustic principles to keep the quantization noise below the human protection threshold.

It should be noted that the amount of information 28 conversion coefficients, due to the nature of the TDAC re-conversion of the device 72 re-conversion, less than the number of samples, which corresponds to the length of the segment 78 of the reconstructed signal. In the case of IMDCT, the number of transform coefficients within the information 47 is rather equal to the number of samples of the time segment 16b. That is, the underlying transform can be called a critical sampling transform, making it necessary to suppress sampling interference in the time domain in order to suppress sampling interference arising from the transform at the boundaries, i.e. leading and trailing edges of the current time segment 16b.

As an additional note, it should be noted that, similarly to the subframe structure of the LPD frames, the FD frames could also be a subject of the subframe structure. For example, FD frames could belong to a long window mode, in which a single window is used to process a section of a signal that extends beyond the front and back edges of the current time segment to encode the corresponding time segment, or belong to a short window mode in which the corresponding section the signal extending beyond the boundaries of the current time segment of the FD frame is subdivided into smaller sub-sections, each of which is subjected to corresponding processing by the window method and conversion individually. In this case, the FD coding unit 54 would output a segment of the reconverted signal to subsection of the current time segment 16b.

After describing a possible implementation of the FD encoding module 54, a possible implementation of the TCX LP decoding module and the codebook excitation LP decoding module 56 and 58, respectively, is described with respect to FIG. In other words, FIG. 5 deals with a case where the current frame is an LPD frame. In this case, the current frame 14b is structured into one or more subframes. In the present case, structuring into three subframes 90a, 90b and 90c is illustrated. It might be that structuring, by default, is limited to certain substructuring abilities. Each of the sub-sections is associated with a corresponding one of the sub-sections 92a, 92b and 92c of the current time segment 16b. That is, one or more sub-sections c 92a through 92c cover without gaps, without overlapping, the entire time segment 16b. According to the order of the sub-sections c 92a through 92c within the time segment 16b, the sequential order is set among the subframes c 92a through 92c. As illustrated in FIG. 5, the current frame 14b is not completely subdivided into subframes 90a through 90c. In other words, some portions of the current frame 14b belong to all subframes, such as first and second syntax sections 24 and 26, FAC data 34, and potentially additional data, such as LPC information, as will be described in more detail below, although LPC information may be also subdivided into separate subframes.

In order to deal with TCX subframes, the TCX LP decoding unit 56 includes a spectral weighting differentiator 94, a spectral weighting device 96, and a re-conversion device 98. For purposes of illustration, it is shown that the first subframe 90a should be a TCX subframe, while it is assumed that the second subframe 90b should be an ACELP subframe.

In order to process the TCX subframe 90a, the differentiator 94 obtains a spectral weighting filter from the LPC information 104 within the information 28 of the current frame 14b, and the spectral weighting device 96 spectrally weights the transform coefficient information with respect to the subframe 90a using a spectral weighting filter received from differentiating device 94, as shown by arrow 106.

The re-transform device 98, in turn, re-converts the spectrally weighted information of the transform coefficients to obtain a re-transformed signal segment 108, continuing, at time t, within and outside of the subsection 92a of the current time segment. The re-conversion performed by the re-conversion device 98 may be the same as that performed by the re-conversion device 72. In fact, the remapper 72 and 98 may typically have hardware, standard software, or programmable hardware.

The LPC information 104 contained in the information 28 of the current LPD frame 16b may represent the LPC coefficients of one point in time within the time segment 16b or for several points in time within the time segment 16b, such as, for example, one set of LPC coefficients for each subsection with 92a to 92c. The spectral weighting filter differentiating device 94 converts the LPC coefficients into spectral weighting factors by spectrally weighting the transform coefficients within the information 90a according to the transfer function, which is obtained from the LPC coefficients by the differentiating device 94, so that it essentially approximates the synthesizing LPC filter or some modified one version. Any dequantization performed outside of the spectral weighting by the spectral weighting device 96 may be spectrally unchanged. Thus, different from the FD decoding mode, quantization noise, according to the TCX coding mode, is spectrally generated using LPC analysis.

Due to the use of re-conversion, however, the re-transformed signal segment 108 suffers from sampling interference. Using the same re-conversion, however, the segments 78 and 100 of the re-transformed signal of successive frames and subframes, respectively, can have their sampling interference suppressed by the transition handler 60 simply by adding their overlapping portions.

When processing the ACELP subframes 90b, the excitation signal differentiator 100 obtains the excitation signal from the excitation update information within the corresponding subframe 90b, and the synthesis LPC filter 102 performs LPC synthesis filtering on the excitation signal using the LPC information to obtain the synthesized signal LP segment 110 for subsection 92b of the current time segment 16b.

Differentiators 94 and 100 may be configured to perform some interpolation to adapt the LPC information 104 within the current frame 16b to the changing position of the current subframe corresponding to the current sub-frame within the current time segment 16b.

Referring to FIGS. 3-5, various signal segments 108, 110, and 78 are included in a transition handler 60, which in turn places all signal segments in the correct temporal order. In particular, the transition handler 60 suppresses time domain sampling interference within temporarily overlapping window portions at the boundaries between time segments of directly consecutive one of the FD frames and FCX subframes in order to reconstruct the information signal at these boundaries. Thus, there is no need for data directly suppressing sampling interference for the boundaries between consecutive FD frames, the boundaries between FD frames, followed by TCX frames and TCX subframes, followed by FD frames, respectively.

However, the situation always changes when an FD frame or a TCX subframe (both representing a variant of a transform coding mode) transitions to an ACELP subframe (representing a kind of time domain coding mode). In this case, the transition handler 60 receives the synthesized signal with direct suppression of sampling interference from the data of direct suppression of sampling noise from the current frame and adds the first synthesized signal with direct suppression of sampling interference to the segment 100 or 78 of the reconverted signal of the immediately preceding time segment in order to reconstruct the information signal on the corresponding border. If the boundary falls into the interior of the current time segment 16b, since the TCX subframe and the ACELP subframe within the current frame define the boundary between the associated sub-segments of the time segment, the transition processor can detect the excitation of the corresponding data directly suppressing sampling noise for these transitions from the first syntax section 24 and the subframe structure described there. Syntax section 26 is not needed. The previous frame 14a may or may not be lost.

However, in the case of a boundary coinciding with the boundary between consecutive time segments 16a and 16b, the parser 20 must check the second syntax portion 26 within the current frame to determine whether the current frame 14b has direct noise reduction data 34, FAC data 34 for suppressing sampling interference occurring at the front end of the current time segment 16b, since either the previous frame is an FD frame or the last subframe of a previous LPD frame is a TCX subframe. At least the parser 20 needs to know the syntax portion 26 in the case where the contents of the previous frame are lost.

Similar statements apply to transitions to other directions, i.e. from ACELP subframes to FD frames or TCX frames. As long as the corresponding boundaries between the corresponding segments and subsections of the segments fall into the interior of the current time segment, the parser 20 does not have problems in determining the existence of data 34 of direct suppression of sampling interference for these transitions from the current frame 14b, namely, from the first syntax section 24. The second syntax section is unnecessary and even inappropriate. However, if a boundary occurs or coincides with the boundary between the previous time segment 16a and the current time segment 16b, the parser 20 needs to check the second syntax section 26 to determine if direct sampling noise reduction data 34 is present to transition at the front end of the current time segment 16b - at least in the case when there is no access to the previous frame.

In the case of transitions from ACELP to FD or TCX, the transition processor 60 receives a second synthesized signal with direct suppression of sampling from data 34 of direct suppression of sampling and adds a second synthesized signal with direct suppression of sampling to the segment of the reconstructed signal within the current time segment to reconstruct the information signal on the border.

After describing embodiments with reference to FIGS. 3-5, which relate to an embodiment according to which frames and subframes of different coding modes exist, a specific implementation of these embodiments will be described in more detail below. The description of these embodiments simultaneously includes possible steps when generating an appropriate data stream containing such frames and subframes, respectively. Hereinafter, this particular embodiment is described as a single speech and audio codec (USAC), although the principles described may be transferred to other signals.

Switching windows in USAC has several purposes. It mixes FD frames, i.e. frames encoded using frequency coding, and LPD frames, which, in turn, are structured into ACELP- (sub) frames and TCX- (sub) frames. ACELP frames (time-domain coding) apply window processing with rectangular, non-overlapping windows to the input samples, while TCX frames (frequency-domain coding) apply window processing with non-rectangular, overlapping windows to the input samples and then encode the signal using transform with suppression of time domain sampling (TDAC), namely MDCT, for example. To coordinate windows as a whole, TCX frames can use centered windows with a uniform shape, and explicit information is transmitted to control the transitions at the boundaries of ACELP frames to suppress the effects of temporal discretization and window processing of coordinated TCX windows. This additional information may be considered as FAC. The FAC data is quantized in the following embodiment in an LPC weighted area, so that the FAC quantization noise and the decoded MDCT are of the same nature.

Figure 6 shows the processing in the encoder in frame 120 encoded by transform coding (TC), which is preceded and followed by frame 122, 124 encoded using ACELP. In accordance with the discussion above, the concept of TC includes MDCT over long and short blocks using AAC, as well as TCX based on MDCT. That is, frame 120 can be either an FD frame or a TCX (sub) frame, such as subframe 90a, 92a of FIG. 5, for example. Figure 6 shows time-domain markers and frame boundaries. The boundaries of a frame or time segment are indicated by dashed lines, while time domain markers are short vertical lines along the horizontal axes. It should be noted that in the following description, the terms “time segment” and “frame” are sometimes used synonymously because of the unique relationship between them.

Thus, the vertical dashed lines in FIG. 6 show the beginning and end of frame 120, which may be a subframe / subpart of a time segment or a frame / time segment. LPC1 and LPC2 should indicate the center of the analysis window corresponding to the coefficients of the LPC filter or LPC filters, which are used in order to suppress sampling interference.

These filter coefficients are obtained at the decoder by, for example, a reconstructor 22 or differentiators 90 and 100 by using LPC information 104 (see FIG. 5). LPC filters contain: LPC1, corresponding to its calculation at the beginning of frame 120, and LPC2, corresponding to its calculation at the end of frame 120. It is assumed that frame 122 was encoded using ACELP. The same applies to frame 124.

Fig.6 is structured into 4 lines numbered on the right side of Fig.6. Each line represents a stage in processing at the encoder. It should be understood that each line is time aligned with the line above.

Line 1 in FIG. 6 represents the original audio signal segmented into frames 122, 120, and 124, as described above. From here to the left of the “LPC1” marker, the original signal is encoded using ACELP. Between the “LPC1” and “LPC2” markers, the original signal is encoded using TC. As described above, with TC, the noise restriction is applied directly in the transform domain, and not in the time domain. To the right of the LPC2 marker, the original signal is again encoded using ACELP, i.e. time domain coding mode. This sequence of coding modes (ACELP, then TC, then ACELP) is selected to illustrate FAC processing, since the FAC is related to both transitions (from ACELP to TC and from TC to ACELP).

However, it should be noted that the transitions at LPC1 and LPC2 in FIG. 6 may occur inside the internal part of the current time segment or may coincide with its front end. In the first case, the existence of the associated FAC data can be determined by the parser 20 only on the basis of the first parsing section 24, while in the event of a frame loss, the parsing device 20 may need the parsing section 26 to do so in the latter case.

Line 2 in FIG. 6 corresponds to decoded (synthesized) signals in each of frames 122, 120, and 124. Therefore, reference numeral 110 of FIG. 5 is used in frame 122, respectively, of the possibility that the last sub-section of frame 122 is a sub-section encoded by ACELP as 92b in FIG. 5, while a combination of 108/78 is used to indicate the proportion of the signal for frame 120, similar to FIGS. 5 and 4. Again, to the left of the LPC1 marker, it is assumed that the synthesis of this frame 122 was encoded using ACELP . From here, the synthesized signal 110 to the left of the LPC1 marker is identified as the synthesized ACELP signal. There is, in principle, a high similarity between ACELP synthesis and the original signal in this frame 122, since ACELP performs waveform coding as accurately as possible. Then, the segment between the markers LPC1 and LPC2 on line 2 in FIG. 2 represents the inverse MDCT output of this segment 120, as seen in the decoder. Again, segment 120 may be a temporary segment 16b of the FD frame or a sub-section of a subframe encoded by TCX, such as, for example, 90b in FIG. 5, for example. In this figure, this segment 108/78 is called "TC frame output." In Figs. 4 and 5, this segment was called the segment of the reconverted signal. In the case of frame / segment 120, which is a subpart of the TCX segment, the output of the TC frame is a window-processed synthesized TLP signal, where TLP means “linear prediction transform coding” to indicate that in the case of TCX, the noise limitation of the corresponding segment completed in the transform domain by filtering MDCT coefficients using spectral information from the LPC filters LPC1 and LPC2, respectively, which was also described above with respect to FIG. 5 with respect to the spectral weighted device 96 and I. It should also be noted that the synthesized signal, i.e. a pre-reconstructed signal, including sampling noise, between the markers "LPC1" and "LPC2" on line 2 in Fig.6, i.e. signal 108/78, contains the effects of windowing and sampling interference in the time domain at its beginning and end. In the case of MDCT as TDAC transforms, time domain sampling interference can be symbolically represented as sweeps 126a and 126b, respectively. In other words, the upper curve in line 2 of FIG. 6, which extends from the beginning to the end of this segment 120 and is indicated by 108/78, shows the effect of windowing due to transformative windowing that is flat in the middle to leave the converted signal is unchanged, but not at the beginning and end. The convolution effect is shown by the lower curves 126a and 126b at the beginning and end of segment 120 with a minus sign at the beginning of the segment and a plus sign at the end of the segment. This effect of windowing and time domain sampling (or convolution) interference is common to MDCT, which serves as a clear example for TDAC transforms. Sampling interference can be suppressed when two consecutive frames are encoded using MDCT, as described above. However, in the case where the frame “encoded with MDCT” is not preceded and / or followed by other MDCT frames, its window processing and time domain sampling interference are not suppressed and remain in the time domain signal after the inverse MDCT. Sample Noise Reduction (FAC) can then be used to correct for these effects, as described above. Finally, it is also contemplated that segment 124 after the LPC2 marker in FIG. 6 should be encoded using ACELP. It should be noted that to obtain the synthesized signal in this frame, the filter state of the LPC filter 102 (see Fig. 5), i.e. the memory of the long-term and short-term prediction devices at the beginning of frame 124 should be proper, which suggests that the effects of temporal sampling and windowing effects at the end of the previous frame 120 between the LPC1 and LPC2 markers can be suppressed by applying the FAC in a specific way, as explained below. To summarize, line 2 in FIG. 6 contains a synthesis of pre-reconstructed signals from successive frames 122, 120, and 124, including the effect of the window method for sampling noise in the time domain at the output of the inverse MDCT for the frame between the markers LPC1 and LPC2.

To get line 3 in FIG. 6, the difference between line 1 in FIG. 6 is calculated, i.e. in the original audio signal 18, and line 2 in Fig.6, i.e. synthesized signals 110 and 108/78, respectively, as described above. This gives a first difference signal 128.

Additional processing on the encoder side regarding frame 120 is explained below with respect to line 3 in FIG. 6. At the beginning of frame 120, firstly, two lobes taken from ACELP synthesis 110 to the left of the LPC1 marker on line 2 in FIG. 6 are added to each other, as follows:

The first beat 130 is a windowed and time-reversed (maximized) version of the latest synthesized ACELP samples, i.e. the last samples of the signal segment 110 shown in FIG. 5. The length and shape of the window for this time-reversed signal are the same as the portion with the discretization interference of the transform window to the left of frame 120. This fraction 130 can be seen as a good approximation of the time-domain sampling interference present in frame 120 of the MDCT line 2 in FIG. .6.

The second portion 132 is the window-processed response in the absence of an input signal (ZIR) of the LPC1 synthesizing filter with the initial state taken as the final states of this filter at the end of ACELP synthesis 110, i.e. at the end of frame 122. The length and shape of the window of this second lobe may be the same as for the first lobe 130.

With the new line 3 in FIG. 6, i.e. after adding two shares 130 and 132 above, a new difference is taken by the encoder to obtain line 4 in Fig.6. It should be noted that the difference signal 134 stops at the LPC2 marker. An approximate view of the expected envelope of the error signal in the time domain is shown on line 4 in Fig.6. The error in ACELP frame 122 is expected to be approximately flat in amplitude in the time domain. Then, it is expected that the error in the TC frame 120 will manifest a general shape, i.e. the envelope of the time domain, as shown in this segment 120 of line 4 of FIG. 6. This expected shape of the error amplitude is shown here for illustrative purposes only.

It should be noted that if the decoder used only the synthesized signals of line 3 in FIG. 6 to produce or reconstruct the decoded audio signal, then quantization noise would normally be the expected envelope of error signal 136 on line 4 of FIG. 6. Thus, it should be understood that the correction must be sent to the decoder in order to compensate for this error at the beginning and end of TC-frame 120. The error appears due to window-processing effects and time-domain sampling interference inherent to the MDCT / inverse MDCT pair. Window processing and time domain sampling interference were reduced at the beginning of the TC frame 120 by adding fractions 132 and 130 of the cylindrical region from the previous ACELP frame 122, as mentioned above, but could not be completely suppressed, as in a real TDAC operation sequential MDCT frames. To the right of the TC frame 120 on line 4 of FIG. 6, immediately before the LPC2 marker, all windowing and time domain sampling remain from the MDCT / inverse MDCT pair and should thus be completely suppressed by directly suppressing the sampling interference.

Before proceeding to the description of the encoding process, in order to obtain data of direct suppression of sampling interference, reference is made to FIG. 7 to briefly explain MDCT as one example of TDAC transform processing. Both directions of conversion are depicted and described with reference to Fig.7. The transition from the time domain to the transformation domain is illustrated in the upper half of FIG. 7, while the re-transformation is depicted in the lower half of FIG. 7.

When moving from the time domain to the transformation domain, the TDAC transform employs window processing 150 applied to the interval 152 of the signal to be converted, which extends beyond the time segment 154 for which the last resulting transform coefficients are actually transmitted within the data stream. The window used in the window processing 150 is shown in FIG. 7 as containing a sampling interference part Lk intersecting the front end of the time segment 154, and a sampling part Rk at the rear end of the time segment 154 with a non-sampling part Mk continuing between them. MDCT 156 applies to a windowed signal. That is, a convolution 158 is performed to collapse the first quarter of the interval 152, extending between the front end of the interval 152 and the front end of the time segment 154 back along the left (front) border of the time segment 154. The same is done with respect to the sampling interference portion Rk. Then, DCT IV 160 is performed on the resulting processed window method and a convoluted signal having as many samples as the time signal 154 to obtain transform coefficients of the same number. Then, quantization in 162 is performed. Naturally, quantization 162 can be considered as not contained in the TDAC transform.

Repeated conversion does the opposite. That is, following the dequantization 164, the IMDCT 166 is executed, using, firstly, DCT ^-1 IV 168 to obtain time samples, the number of which is equal to the number of samples of the time segment 154, which should be reconstructed. Then, the sweep process 168 is performed on the portion of the inverted signal received from module 168, thereby extending the time interval or the number of time samples of the IMDCT result by doubling the length of the sections with sampling interference. Then, window processing is performed in 170 using the remapping window 172, which may be the same as the window used by window processing 150, but may also be different. The remaining blocks in FIG. 7 illustrate TDAC or overlap / add processing performed on overlapping portions of consecutive segments 154, i.e. adding its expanded sections with sampling interference, as performed by the transition processor in figure 3. As illustrated in FIG. 7, TDAC by blocks 172 and 174 suppresses sampling interference.

The description of FIG. 6 now continues. In order to effectively compensate for the effects of windowing and time-domain sampling at the beginning and end of the TC frame 129 on line 4 of FIG. 6, and assuming that the TC frame 120 uses time domain noise limiting (FDNS), a direct correction is applied Discretization Interference (FAC) following the processing described in FIG. First, note that FIG. 8 describes this processing for both: the left side of the TC frame 120 near the LPC1 marker and for the right side of the TC frame 120 near the LPC2 marker. It should be remembered that the TC frame 120 in FIG. 6 is assumed to be preceded by an ACELP frame 122 at the border of the LPC1 marker and an ACELP frame 124 at the border of the LPC2 marker follows.

In order to compensate for the effects of windowing and sampling interference in the time domain near the LPC1 marker, the processing is described in FIG. First, the weighting filter W (z) is calculated from the filter LPC1. The weighting filter W (z) could be a modified analyzing or whitening filter A (z) for LPC1. For example, W (z) = A (z / λ), where λ is a predefined weighting coefficient. The error signal at the beginning of the TC frame is indicated by a reference designation 138, as on line 4 in FIG. 6. This error is called the FAC target in FIG. The error signal 138 is filtered by the filter W (z) in 140, with the initial state of this filter, i.e. with the initial state, if it is a filter memory, being an ACELP error 141 in an ACELP frame 122 on line 4 in FIG. 6. The output of the filter W (z) then forms the input of the conversion 142 in FIG. 6. This conversion is shown as an example MDCT. The transform coefficients derived by the MDCT are then quantized and encoded in the processing unit 143. These coded coefficients could form at least a portion of the above FAC data 34. These coded coefficients may be transmitted to the coding side. The output of the Q process, namely, quantized MDCT coefficients, is an inverse transform input, such as IMDCT 144, to generate a time-domain signal, which is then filtered by a 1 / W (z) inverse filter at 145, which has zero memory (zero initial state) . Filtering through 1 / W (z) continues beyond the FAC target length, using a zero input for samples that continue after the FAC target. The output of the 1 / W (z) filter is a synthesized FAC signal 146, which is a correction signal that can now be applied at the beginning of the TC frame 120 to compensate for the windowing effect and time domain sampling that occurs there.

Now, processing is described for adjusting window processing and time domain sampling interference at the end of TC frame 120 (up to marker LPC2). For this, reference is made to FIG. 9.

The error signal at the end of the TC frame 120 on line 120 in FIG. 6 is provided by 147 and represents the FAC target in FIG. 9. The FAC target 147 undergoes the same process sequence as the FAC target 138 of FIG. 8, by processing that differs only in the initial state of the weighting filter W (z) 140. The initial state of the filter 140 to filter the FAC target 147 is an error in the TC frame 120 on line 4 of FIG. 6, indicated by a reference numeral 148 in FIG. 6. Additional processing steps 142 to 145 are the same as in FIG. 8, which deal with the processing of the FAC target at the beginning of the TC frame 120.

The processing of FIGS. 8 and 9 is performed completely from left to right, when applied in the encoder to obtain local FAC synthesis and to calculate the resulting reconstruction to determine if changing the encoding mode involved by selecting the TC encoding mode of frame 120 is the best choice. In the decoder, the processing in FIGS. 8 and 9 is applied only from the middle to the right. That is, the encoded and quantized transform coefficients transmitted by the Q 143 processor are decoded to form an IMDCT input. See, for example, FIGS. 10 and 11. FIG. 10 is equal to the right side of FIG. 8, while FIG. 11 is equal to the right side of FIG. 9. The transition handler 60 of FIG. 3 may, in accordance with a particular embodiment, be implemented in accordance with FIGS. 10 and 11. That is, the transition handler 60 may subject the transform coefficient information within the FAC data 34 presented within the current frame 14b to a transform to output the first synthesized FAC signal 146 in case of conversion from the ACELP sub-part of the time segment to the temporary FD segment or FCX sub-part, or the second synthesized FAC signal 149, when switching from the temporary FD segment or TCX sub STI time slot in subpart ACELP-time segment.

It should be noted that the FAC data 34 may relate to such a transition occurring within the current time segment when the existence of the FAC data 34 is obtained for the parsing device 20 exclusively from the parsing section 24, while the parsing device 20 is necessary in case of loss of the previous frame , use syntax portion 26 to determine if FAC data 34 exists for such transitions at the leading edge of the current time segment 16b.

FIG. 12 shows how an entire synthesized signal or reconstructed signal for the current frame 120 can be obtained by using the synthesized FAC signals in FIGS. 8-11 and applying the inverse steps of FIG. 6. It should again be noted that the steps shown in FIG. 12 are also performed by an encoder to determine if the encoding mode for the current frame leads to the best optimization, for example, in terms of speed / distortion or the like. 12, it is assumed that the ACELP frame 122 to the left of the LPC1 marker is already synthesized or reconstructed, for example, by the module 58 of FIG. 3, up to the LPC1 marker, thereby leading to the synthesized ACELP signal on line 2 of FIG. 12 with 110. Since the FAC adjustment is also used at the end of the TC frame, it is also assumed that frame 124 after the LPC2 marker will be an ACELP frame. Then, to produce a synthesized signal or reconstructed signal in the TC frame 120 between the markers LPC1 and LPC2 in FIG. 12, the following steps are performed. These steps are also illustrated in FIGS. 13 and 14, where FIG. 13 illustrates the steps performed by the transition handler 60 in order to cope with the transitions from the TC encoded segment or sub-segment to ACELP encoded sub-portion, and on Fig describes the operation of the transition processor for the reverse transitions.

1. One step is to decode the TC frame encoded by MDCT and place the thus obtained time-domain signal between the markers LPC1 and LPC2, as shown on line 2 in FIG. 12. Decoding is performed by module 54 or module 56 and includes an inverse MDCT, as an example for TDAC re-conversion, so that the decoded TC frame contains the effects of windowing and time-domain sampling. In other words, the sub-part of the segment or time segment that should be currently decoded and indicated by the index k in FIGS. 13 and 14 may be the sub-part 92b of the time segment, which is encoded by ACELP, as illustrated in FIG. 13, or the time segment 16b which is a sub-part 92a encoded with FD or encoded with TCX, as illustrated in FIG. In the case of FIG. 13, a previously processed frame is a sub-part of a segment encoded by TC or a time segment, and in the case of FIG. 14, a previously processed time segment is a sub-part encoded by ACELP. The reconstructed or synthesized signal, as output by modules 54 through 58, partially suffers from the effects of sampling interference. This is also true for 78/108 signal segments.

2. Another step in processing the transition handler 60 is to generate a synthesized FAC signal according to FIG. 10 in the case of FIG. 14 and in accordance with FIG. 11 in the case of FIG. 13. That is, the transition processor 60 may re-convert 191 over the transform coefficients within the FAC data 34 to obtain synthesized FAC signals 146 and 149, respectively. The synthesized FAC signals 146 and 149 are located at the beginning and end of a segment encoded by TC, which in turn suffers from the effects of sampling interference and is recorded on the time segment 78/108. In the case of FIG. 13, for example, transition handler 60 positions the synthesized FAC signal 149 at the end of frame k-1 encoded by TC, as also shown on line 1 of FIG. 12. In the case of FIG. 14, transition handler 60 positions the synthesized FAC signal 146 at the beginning of frame k encoded by TC, as shown on line 1 of FIG. 12. It should again be noted that frame k is the frame to be decoded at the current moment, and that frame k-1 is a previously decoded frame.

3. As long as the situation of FIG. 14 is considered, where a change in the coding mode occurs at the beginning of the current TC frame k, processed by the window method and minimized (reversed), the synthesized ACELP signal 130 from the ACELP frame k-1 preceding TC frame k, and the window-processed response in the absence of an input signal, or ZIR, synthesizing filter LPC1, i.e. signal 132 are arranged to register to a segment 78/108 of the reconverted signal having sampling interference. This fraction is shown on line 3 of FIG. As shown in FIG. 14 and as already described above, the transition handler 60 receives a sampling interference suppression signal 132 by continuing to LPC filter the synthesis of the previous CELP subframe outside the front boundary of the current time segment k and processing the window continuation signal 110 inside the current signal k using both steps indicated by reference numbers 190 and 192 in FIG. In order to obtain a sampling interference suppression signal 130, the transition processor 60 also processes the reconstructed signal segment 110 of the previous CELP frame by the 194 window method and uses this windowed and time-reversed signal as signal 130.

4. The fractions of lines 1, 2 and 3 in Fig. 12 and the fractions 78/108, 132, 130 and 146 in Fig. 14 and the fractions 78/108, 149 and 196 in Fig. 13 are added by the transition processor 60 in the registered positions explained above, for generating a synthesized or reconstructed audio signal for the current frame k in the initial region, as shown on line 4 of FIG. 12. It should be noted that the processing of FIGS. 13 and 14 produces a synthesized or reconstructed signal 198 in a TC frame, where the effects of sampling noise in the time domain and window processing are suppressed at the beginning and end of the frame, and where the potential unevenness of the frame boundary near the LPC1 marker was smoothed and perceptually masked by a 1 / W (z) filter in FIG.

Thus, FIG. 13 relates to the current processing of a frame k encoded by CELP and directly suppresses sampling interference at the end of a previous segment encoded by TC. As illustrated in 196, the finally reconstructed audio signal is reconstructed without sampling interference at the boundary between the k-1 and k segments. The processing of FIG. 14 directly suppresses sampling interference at the beginning of the current segment k encoded by TC, as illustrated in reference numeral 198, showing the reconstructed signal at the boundary between segments k-1 and k. The remaining sampling noise at the trailing end of the current segment k is either suppressed by TDAC if the next segment is a segment encoded by TC, or by the FAC of FIG. 13 if the subsequent segment is a segment encoded by ACELP. 13, this latter possibility is mentioned by reference 198 assigned to the signal segment of the time segment k-1.

Concrete possibilities will be mentioned hereinafter as to how the second syntax section 26 can be implemented.

For example, to handle the occurrence of lost frames, the syntax section 26 can be implemented as a 2-bit prev_mode field that explicitly reports the encoding mode that was applied in the previous frame 14a according to the following table inside the current frame 14b:

prev_mode ACELP 0 0 TCX 0 one Fd_long one 0 Fd_short one one

In other words, this 2-bit field may be called prev_mode and may thus indicate the encoding mode of the previous frame 14a. In the case of the example just mentioned, four different states are distinguished, namely:

1) The previous frame 14a is an LPD frame, the last subframe of which is an ACELP subframe;

2) the previous frame 14a is an LPD frame, the last subframe of which is a TCX encoded subframe;

3) the previous frame is an FD frame using a long transform window and

4) the previous frame is an FD frame using short conversion windows.

The potential use of different window lengths of the FD coding mode has already been mentioned above with respect to the description in FIG. Naturally, the syntax section 26 can have only three different states, and the FD encoding mode can only be operated with a constant window length, thereby summing up the last two lengths of the above options 3 and 4.

In any case, based on the above 2-bit field, the parser 20 is able to decide whether FAC data is present to transition between the current time segment and the previous time segment 16a within the current frame 14a. As will be described in more detail below, the parser 20 and the reconstructor 22 are able to determine based on prev_mode whether the previous frame was an FD frame using a long window (FD_long) or whether the previous frame was an FD frame using a short window (FD_short) and whether the current frame 14b (if the current frame is an LPD frame) follows an FD frame or an LPD frame, the differentiation of which is necessary according to the following embodiment, in order to correctly parse the data stream and reconstruct the information system cash respectively.

Thus, in accordance with the mentioned possibility of using a 2-bit identifier, as a syntax section 26, each frame 16a to 16c would be provided with an additional 2-bit identifier in addition to the syntax section 24, which defines the encoding mode of the current frame, which should be FD or LPD encoding mode, and subframe structure in the case of LPD encoding mode.

For all embodiments, it should be mentioned above that other intraframe dependencies should also be avoided. For example, the decoder of FIG. 1 could have SBR capability. In this case, the separation frequency could be parsed by the parser 20 from each frame 16a to 16c within the corresponding SBR extension data, instead of parsing such a separation frequency using the SBR header, which could be transmitted within the data stream 12 less frequently. Other intraframe dependencies could be removed in a similar way.

It is advisable to note for all of the above embodiments that the parser 20 could be configured to buffer at least the currently decoded frame 14b inside the buffer and pass all frames 14a through 14c through this FIFO- buffer (first entered - first out) . In buffering, the parser 20 could clear frames from this buffer in frame units 14a through 14c. That is, filling and flushing the buffer of the parser 20 could be performed in units of frames 14a to 14c, so as to satisfy the constraints imposed by the maximum available buffer space, for example, only one or more than one maximum size frame can be accommodated at a time.

Next, an alternative message capability for syntax portion 26 with a reduced bit fraction will be described. According to this alternative, another construction structure of syntax section 26 is used. In the embodiment described previously, syntax section 26 was a 2-bit field that is transmitted in each frame 14a through 14c of the encoded USAC data stream. Since for the FD part for the decoder, it is only important to know whether it should read the FAC data from the bitstream if the previous frame 14a was lost, these 2 bits can be divided into two 1-bit flags, where one of them is reported in each frame from 14a to 14c as fac_data_present. This bit can be inserted into the single_channel_element and channel_pair_element structures, respectively, as shown in the tables in FIGS. 15 and 16. FIGS. 15 and 16 can be considered as determining the high-level syntax structure of frames 14 in accordance with the present embodiment, where the functions are “function_name (. ..) "call the standard routines, and the bold names of the syntax elements indicate the reading of the corresponding syntax element from the data stream. In other words, the marked areas or shaded areas in FIGS. 15 and 16 show that each frame 14a to 14c, in accordance with this embodiment, is provided with the fac_data_present flag. Reference numbers 199 show these sites.

The other 1-bit flag flag prev_frame_was_lpd is then transmitted in the current frame only if it was encoded using the USAC LPD part, and indicates whether the previous frame was also encoded using the USAC LPD path. This is shown in the table of FIG.

The table of FIG. 17 shows a piece of information 28 in FIG. 1 in the case where the current frame 14b is an LPD frame. As shown in 200, each LPD frame is provided with the prev_frame_was_lpd flag. This information is used to parse the syntax of the current LPD frame. This content and position of the 34 FAC data in LPD frames depends on the transition at the front end of the current LPD frame, which is the transition between the TCX coding mode and the CELP coding mode, or the transition from the FD coding mode to the CELP coding mode Fig. 18. In particular, if the frame currently being decoded 14b is an LPD frame that has just been preceded by an FD frame 14a, and fac_data_present reports that the FAC data is present in the current LPD frame (since the front subframe is an ACELP subframe), then the data FACs are read at the end of the LPD frame syntax at 202 with FAC data 34 including, in this case, a fac_gain gain factor, as shown in 204 in FIG. 18. With this gain factor, the fraction 149 of FIG. 13 is adjustable by gain.

If, however, the current frame is an LPD frame, and the previous frame is also an LPD frame, i.e. if the transition between the TCX and CELP subframes occurs between the current frame and the previous frame, the FAC data is read in 206 without the option of adjustability using gain, i.e. without 34 FAC data, including the FAC fac_gain syntax gain element. In addition, the position of the FAC data read in 206 is different from the position in which the FAC data is read in 202 when the current frame is an LPD frame and the previous frame is an FD frame. As long as the reading position 202 occurs at the end of the current LPD frame, the reading of the FAC data in 206 occurs before reading the characteristic data sub-frame, i.e. ACELP and TCX data, i.e. ACELP and TCX data depend on the subframe modes of the subframe structure at 208 and 210, respectively.

In the example of FIGS. 15-18, the LPC information 104 (FIG. 5) is read after the characteristic subframe data, such as 90a and 90b (compare FIG. 5), in 212.

For completeness only, the syntax structure of the LPD frame of FIG. 17 is further explained with respect to FAC data potentially additionally contained within an LPD frame to provide FAC information regarding transitions between TCX and ACELP subframes in the interior of the current time segment encoded by LPD . In particular, in accordance with the embodiment of FIGS. 15-18, the structure of the LPD subframe is limited to subdivide the current time segment encoded by LPD only in units of quarters, with assignment of these quarters to either TCX or ACELP. The exact LPD structure is defined by the lpd_mode syntax element read in 214. The first, second, third and fourth quarters can form a TCX subframe together, while ACELP frames are limited to only a quarter length. The TCX subframe can also continue throughout the entire time segment encoded using LPD, in which case the number of subframes is only one. The condition checking cycle in FIG. 17 gradually passes through the quarters of the time segment currently encoded using LPD, and transmits whenever the current quarter k is the start of a new subframe inside the interior of the time segment currently encoded using LPD. FAC in 216, provided immediately preceding subframe of the currently starting / currently decoded LPD frame, belongs to another mode, i.e. TCX mode if the current subframe belongs to ACELP mode and vice versa.

For the sake of completeness, FIG. 19 shows a possible syntax structure of an FD frame in accordance with the embodiment of FIGS. 15-18. You can see that the FAC data is read at the end of the FD frame when deciding whether FAC data 34 is present by only activating the fac_data_present flag. In comparison, the parsing of fac_data 34 in the case of LPD frames, as shown in FIG. 17, makes it necessary, for correct parsing, to know the prev_frame_was_lpd flag.

Thus, the 1-bit prev_frame_was_lpd flag is only transmitted if the current frame is encoded using the USAC LPD part and it is reported whether the previous frame was encoded using the USAC codec LPD path (see lpd_channel_stream () syntax in FIG. 17).

Regarding the embodiment of FIGS. 15-19, it should be further noted that the additional syntax element could be transmitted in 220, i.e. in the case where the current frame is an LPD frame and the previous frame is an FD frame (the first frame of the current LPD frame being an ACELP frame), so that the FAC data should be read in 202 to address the transition from the FD frame to ACELP subframe at the front end of the current LPD frame. This additional syntax element, read in 220, could indicate whether the previous FD frame 14a belongs to FD_long or FD_short. Depending on the syntax element, FAC data in 202 could be affected. For example, the length of the synthesized signal 149 could be influenced depending on the length of the window used to convert the previous LPD frame. Summarizing the embodiment of FIGS. 15-19 and transferring the features mentioned therein to the embodiment described with respect to FIGS. 1-4, the following could be applied in the latter embodiments either individually or in combination:

1) The FAC data 34 mentioned in the previous figures was intended to first of all mark the FAC data present in the current frame 14b, in order to allow direct suppression of sampling interference between the previous frame 14a and the current frame 14b, t .e. between the respective time segments 16a and 16b. However, additional FAC data may be present. This additional FAC data, however, deals with transitions between subframes encoded with TCX and subframes encoded with CELP located within the current frame 14b when it belongs to the LPD mode. The presence or absence of this additional data does not depend on syntax portion 26. In FIG. 17, this additional FAC data was read in 216. Their presence or existence only depends on lpd_mode read in 214. The last syntax element, in turn, is part of the syntax section 24, revealing the encoding mode of the current frame, lpd_mode along with core_mode, read in 230 and 232, shown in Fig and 16, correspond to syntax section 24.

2) In addition, the syntax section 26 may be composed of more than one syntax element, as described above. The FAC_data_present flag indicates whether fac_data is present for the boundary between the previous frame and the current frame. This flag is present in the LPD frame, as well as in the FD frames. An additional flag, in the above embodiment, called prev_frame_was_lpd, is transmitted in the LPD frames only to indicate whether the previous frame 14a belonged to the LPD mode. In other words, this second flag included in the syntax portion 26 indicates whether the previous frame 14a was an FD frame. The parser 20 waits and reads this flag only when the current frame is an LPD frame. 17, this flag is read at 200. Depending on this flag, the parser 20 can wait for the presence of FAC data and thus read fac_gain gain value from the current frame. The gain value is used by the reconstructor to set the gain of the synthesized FAC signal for the FAC during the transition between the current and previous time segments. In the embodiment of FIG. 15 to 19, this syntax element is read in 204 depending on the second flag, being exempted from comparing conditions leading to the reading of 206 and 202, respectively. Alternatively or additionally, prev_frame_was_lpd may control the position where the parser 20 waits and reads the FAC data. In the embodiment of FIGS. 15-19, these positions were 206 and 202. In addition, the second syntax section 26 may further comprise an additional flag in the case where the current frame is an LPD frame, wherein the front subframe is an ACELP frame and the previous a frame is an FD frame to indicate whether a previous FD frame is encoded using a long transform window or a short transform window. The last flag could be read at 220 in the case of the previous embodiment of FIGS. 15-19. Knowledge of this FD conversion length can be used to determine the length of the synthesized FAC signals and the data size of 38 FAC, respectively. By this measure, FAC data can be sized to overlap the window length of the previous FD frame in order to achieve a better compromise between encoding quality and encoding speed.

3) By dividing the second syntax section 26 into the three flags just mentioned, it is possible to transmit only one flag or bit for reporting the second syntax section 26 in the case where the current frame is an FD frame, only two flags or bits in the case where the current frame is An LPD frame and the previous frame is also an LPD frame. Only in the case of a transition from the FD frame to the current LPD frame, the third flag should be transmitted in the current frame. Alternatively, as mentioned above, the second syntax portion 26 may be a 2-bit pointer transmitted for each frame and indicating the mode when the frame preceding this frame is needed by the parser to decide if 38 FAC data should be read. from the current frame, and if so, where and how long is the synthesized FAC signal. That is, the specific embodiment of FIGS. 15-19 can be easily transferred to the embodiment using the above 2-bit pointer to implement the second syntax portion 26. Instead of the FAC_data_present in FIGS. 15 and 16, a 2-bit identifier would be transmitted. Flags 200 and 220 should not be transmitted. Instead, the contents of fac_data_present in the conditional statement leading to 206 and 218 could be obtained by the parser 20 from the 2-bit identifier. The following table can be accessed at the decoder to use a 2-bit pointer.

prev_mode core _mode first_lpd_flag current frame (superframe) ACELP one 0 TCX one 0 Fd_long one one Fd_short one one

Syntax section 26 may also have only three different possible values, in this case FD frames will use only one possible length.

A slightly different, but very similar syntax structure to that described above with respect to FIGS. 15-19 is shown in FIGS. 20-22 using the same reference designators as in FIGS. 15-19, so that reference is made to this embodiment to explain the embodiment of FIGS. 20-22.

Regarding the embodiments described with respect to FIG. 3 and the following, it should be noted that any transform coding scheme, if appropriate, sampling interference can be used in connection with TCX frames, unlike MDCT. In addition, a conversion coding scheme, such as, for example, FFT, could also be used, then without sampling interference in the LPD mode, i.e. without FAC for transitions of subframes within LPD frames and, thus, without the need for transmitting FAC data for the boundaries of subframes between LPD boundaries. FAC data would then only be included in the composition for each transition from FD to LPD and vice versa.

Regarding the embodiments described with respect to FIG. 1 and the following, it should be noted that these were directed to the case where the additional syntax section 26 was defined together, i.e. uniquely depending on the comparison between the encoding mode of the current frame and the encoding mode of the previous frame, as specified in the first syntax section of the previous frame, so that in all the above embodiments, the decoder or parser was able to unambiguously anticipate the contents of the second syntax section of the current frame by using or comparing the first syntactic section of these frames, namely the previous and current frame. That is, in the case without frame loss, it was possible for a decoder or parser to obtain from transitions between frames whether FAC data is present in the current frame if the frame is lost; a second syntax section, such as the fac_data_present bit, explicitly provides this information. However, according to another embodiment, the encoder could use this explicit signaling capability proposed by the second syntax section 26 in order to apply reverse coding, according to which the syntax section 26 is adaptive, when solving after execution on a frame-by-frame basis, for example, set so that although the transition between the current frame and the previous frame is of the type that usually goes with the FAC data (such as FD / TCX, i.e. any TC-coding mode, to ACELP, i.e. any coding mode anija time domain or vice versa), the syntax portion of the current frame indicates no FAC. The decoder could then be implemented to strictly act according to syntax section 26, thereby effectively eliminating the possibility, or suppressing, the transmission of FAC data in the encoder that reports this suppression, only setting, for example, fac_data_present = 0. A scenario where this could be a favorable option is when encoding at very low bit rates, where additional FAC data could cost too many bits, while artifacts of the resulting sampling interference could be tolerant compared to the overall sound quality.

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 method step or feature of a method step. Similarly, the aspects described in the context of a method step also provide a description of the corresponding block or element or feature of the corresponding device. Some or all of the steps of the method may be performed by (or using) a hardware device, such as a microprocessor, programmable computer, or electronic circuit. In some embodiments, some one or more of the most important steps of the method may be performed by such a device.

A patented encoded audio signal may be stored in a digital 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.

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, for example, a floppy disk, DVD, Blue-Ray, CD, ROM, PROM, EPROM, EEPROM or flash memory having electronically readable control signals stored on it that communicate (or able to interact) with a programmable computer system, so that the appropriate method is performed. As a result, the digital storage medium may be a computer readable.

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

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

Other embodiments comprise a computer program for performing one of the methods described herein stored on a machine readable medium.

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

An additional embodiment of patentable methods, therefore, is a storage medium (either a digital storage medium or a computer readable medium) comprising a computer program recorded thereon for performing one of the methods described herein. A storage medium, digital storage medium or recording medium is usually tangible and / or permanent.

An additional embodiment of the inventive method, therefore, is 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, over the Internet.

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

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

A further embodiment of the invention comprises a device or system configured to transfer (for example, electronically or optically) a computer program for executing one of the methods described herein into a receiver. The receiver may, for example, be a computer, mobile device, storage device, or the like. The device or system may, for example, comprise a file server for transferring a computer program to a receiver.

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. This is an intention, therefore, to be limited only by the scope of the forthcoming claims, and not by the specific details presented for the purpose of describing and explaining the embodiments herein.

Claims (20)

1. A decoder (10) for decoding a data stream (12), comprising a sequence of frames into which time segments of the information signal (18) are encoded, respectively, containing:
a parsing device (20) configured to parse the data stream (12), the parsing device being configured to read the first syntax section (24) and the second syntax section from the current frame when parsing the data stream (12) (14b); and
reconstructor (22) configured to reconstruct the current time segment (16b) of the information signal (18) associated with the current frame (14b) based on information (28) obtained from the current frame by parsing using the first selected one of the mode decoding with conversion with suppression of sampling interference in the time domain and the decoding mode of the time domain, and the first choice depends on the first syntax section (24),
moreover, the parsing device (20) is configured to, when parsing the data stream (12), perform the first wait action that the current frame (14b) contains, and thus reading the data (34) of the direct suppression of sampling interference from the current frame (14b), or a second act of surprise that the current frame (14b) contains, and thus, without reading data (34) directly suppressing sampling interference from the current frame (14b), the parser makes a second choice, which from the first of action and a second action is performed, and depends on the syntactic second portion,
moreover, the reconstructor (22) is configured to perform direct suppression of sampling interference at the boundary between the current time segment (16b) and the previous time segment (16a) of the previous frame (14a) using data (34) of direct suppression of sampling interference. 2. The decoder (10) according to claim 1, wherein the first and second syntax sections are contained in each frame, in which the first syntax section (24) associates the corresponding frame from which it was read, with the first frame type or the second frame type, and if the corresponding frame belongs to the second type of frame, associates the subframes of the subdivision of the corresponding frame composed of a number of subframes with the corresponding one of the first type of subframe and the second type of subframe in which the reconstructor (22) is executed if the first syntactic section (24) associates the corresponding frame with the first type of frame, with the possibility of using frequency domain decoding as the first version of the decoding mode with transforming the suppression of sampling noise in the time domain to reconstruct the time segment associated with the corresponding frame, and if the first syntax section ( 24) associates the corresponding frame with the second type of frame, use, for each subframe of the corresponding frame, linear prediction decoding encoded with excitation conversion as the second version of the decoding mode with conversion with suppression of sampling interference in the time domain for reconstructing the corresponding subsection of the current time segment of the corresponding frame that is associated with the corresponding subframe if the first syntax section (24) associates the corresponding subframe of the corresponding frame with the first type subframe, and linear predictive codebook excitation decoding as mode decoded I the time domain to reconstruct subplot time slot of a respective frame, which is associated with a corresponding subframe, if the first parsing section (24) associates the corresponding subframe with the second subframe type. 3. The decoder (10) according to claim 1, in which the second syntax section has a set of possible values, each of which is uniquely associated with one of the set of possibilities, containing:
a previous frame (14a) belonging to the first type of frame, a previous frame (14a) belonging to the second type of frame, with its last subframe belonging to the first type of subframe, and
a previous frame (14a) belonging to the second type, with its last subframe belonging to the second type of subframe, and
the parser (20) is arranged to make a second selection based on a comparison between the second syntax section of the current frame (14b) and the first syntax section (24) of the previous frame (14a). 4. The decoder according to claim 3, in which the parser (20) is arranged to read data (34) directly suppressing sampling interference from the current frame (14b), if the current frame (14b) belongs to the second type of frame, depending on the previous frame (14a) belonging to the second type of frame with its last subframe belonging to the first type of subframe, or the previous frame (14a) belonging to the first type of frame, since the gain of direct suppression of sampling interference is obtained by parsing and from the data (34) of the direct suppression of sampling interference in the case where the previous frame (14a) belonging to the first type of frame is not there, if the previous frame belonging to the second type of frame with its last subframe belonging to the first type of subframe, the reconstructor ( 22) is configured to perform direct suppression of sampling interference with an intensity that depends on the gain of direct suppression of sampling interference in the case in which the previous frame (14a) belongs to the first type of frame. 5. The decoder (10) according to claim 4, wherein the parser (20) is arranged to read if the current frame (14b) belongs to the first type of frame, amplification of the direct suppression of sampling interference from the data (34) of the direct suppression of sampling in which the reconstructor is configured to perform direct suppression of sampling interference with an intensity that depends on the gain of direct suppression of sampling interference. 6. The decoder (10) according to claim 1, in which the second syntax section has a set of possible values, each of which is uniquely associated with one of the set of possibilities, containing:
the previous frame (14a) belonging to the first type of frame, using a long conversion window,
the previous frame (14a) belonging to the first type of frame, using short conversion windows,
a previous frame (14a) belonging to the second type of frame, with its last subframe belonging to the first type of subframe, and
a previous frame (14a) belonging to the second type of frame, with its last subframe belonging to the second type of subframe, and
the parser is configured to make a second selection based on a comparison between the second syntax portion of the current frame (14b) and the first syntax portion (24) of the previous frame (14a) and reading data (34) to directly suppress sampling interference from the current frame (14b) if the previous frame (14a) belongs to the first type of frame, depending on the previous frame (14a) involving a long conversion window or short conversion windows, so that the amount of data (34) is directly suppressed by ex sampling greater if the previous frame (14a) uses a long transform window and smaller if the previous frame (14a) uses short transform window. 7. The decoder (20) according to claim 2, wherein the reconstructor is capable, for each frame of the first type of frame, of performing dequantization (70) with a changing spectrum of information of transform coefficients inside the corresponding frame of the first type of frame based on the information of the scale factor inside the corresponding frame the first type of frame and re-converting the information of the dequantized transform coefficients to obtain a segment (78) of the re-transformed signal, continuing in time within and outside the time segment associated with the corresponding frame of the first type of frame, and
for each frame of the second type of frame,
for each subframe of a first type of subframe of a corresponding frame of a second type of frame,
obtaining (94) a spectral weighting filter from the LPC information inside the corresponding frame of the second type of frame,
spectral weighting (96) of transform coefficient information within a corresponding subframe of a first type of subframe using a spectral weighting filter, and
re-transforming (98) the spectrally-weighted information of the transform coefficients to obtain a segment of the transformed signal lasting in time within and outside the portion of the time segment associated with the corresponding subframe of the first type of subframe, and
for each subframe of the second type of subframe of the corresponding frame of the second type of frame,
obtaining (100) the excitation signal from the excitation update information within the corresponding subframe of the second type of subframe,
performing LPC filtering (102) of the synthesis on the excitation signal, using LPC information inside the corresponding frame of the second type of frame in order to obtain a segment (110) of the synthesized LP signal for subsection of the time segment associated with the corresponding subframe of the second type of subframe, and
performing suppression of sampling interference in the time domain inside temporarily overlapping window sections at the boundaries between the time segments of immediately following one of the frames of the first type of frames and sub-segments of time segments that are associated with the subframes of the first type of subframe to reconstruct the information signal (18) between them,
if the previous frame belongs to the first frame type or the second frame type with its last subframe belonging to the first subframe type, and the current frame (14b) belongs to the second frame type with its first subframe belonging to the second subframe type, obtaining the first synthesized signal with direct suppression of sampling interference from the data (34) of direct suppression of sampling interference and adding the first synthesized signal with direct suppression of sampling interference to the segment (78) of the transformed signal inside edyduschego time slice for reconstruction of the information signal (18) at the border between the previous and current frames (14a, 14b), and
if the previous frame (14a) belongs to the second frame type with its first subframe belonging to the second subframe type, and the current frame (14b) belongs to the first frame type or the second frame type with its last subframe belonging to the first subframe type, obtain the second synthesized signal with direct suppression of sampling interference from the data (34) of the direct suppression of sampling interference and adding a second synthesized signal with direct suppression of sampling interference to the segment of the reconverted signal and the current time slot (16b) for reconstruction of the information signal (18) at the border between the previous and current time segments (16a, 16b). 8. The decoder (10) according to claim 7, in which the reconstructor is configured to:
obtaining a first synthesized signal with direct suppression of sampling interference from the data (34) of the direct suppression of sampling interference by performing the conversion of the conversion coefficient information contained in the data (34) of the direct suppression of sampling interference, and / or
obtaining a second synthesized signal with direct suppression of sampling interference from the data (34) of the direct suppression of sampling interference by performing the conversion of the conversion coefficient information contained in the data (34) of the direct suppression of sampling interference. 9. The decoder according to claim 1, in which the second syntax section comprises a first flag indicating whether data (34) of direct suppression of sampling interference is present in the corresponding frame, and the parser is configured to make a second selection depending on the first flag, and in which the second syntax section further comprises a second flag only inside frames of the second type of frames, the second flag indicating whether the previous frame belongs to the first type of frame or to the second type of frame with its last a subframe belonging to the first type of subframe. 10. The decoder according to claim 9, in which the parser is arranged to read data (34) to directly suppress sampling interference from the current frame (14b) if the current frame (14b) belongs to the second type of frame, depending on the second flag wherein, the gain of direct suppression of sampling interference is obtained by parsing from data (34) of the direct suppression of sampling interference in the case where the previous frame belongs to the first type of frame and not if the previous frame belongs to the second the frame ip with its last subframe belonging to the first type of subframe, in which the reconstructor is configured to perform direct suppression of sampling interference with an intensity that depends on the gain of the direct suppression of sampling interference in the case where the previous frame belongs to the first type of frame. 11. The decoder according to claim 10, in which the second syntax section further comprises a third flag, indicating whether the previous frame uses a long conversion window or short conversion windows, only within frames of the second type of frame, if the second flag indicates that the previous frame belongs to the first type a frame in which the parser (20) is arranged to read data (34) to directly suppress sampling interference from the current frame (14b), depending on the third flag, so that the amount of data (34) direct suppression of sampling interference is greater if the previous frame uses a long conversion window, and less if the previous frame uses short conversion windows. 12. The decoder according to claim 1, wherein the reconstructor is configured to if the previous frame belongs to the second type of frame with its last subframe belonging to the second type of subframe, and the current frame (14b) belongs to the first type of frame or second type of frame with its last subframe, belonging to the first type of subframe, of performing window processing of a segment of the synthesized LP signal of the last subframe of the previous frame to obtain a first segment of a noise reduction signal and add a first drove the suppression of sampling interference into the segment of the converted signal within the current time segment. 13. The decoder according to claim 7, in which the reconstructor is configured to if the previous frame belongs to the second type of frame with its last subframe belonging to the second type of subframe, and the current frame (14b) belongs to the first type of frame or second type of frame with its last subframe, belonging to the first type of subframe, continuing the LPC synthesis filtering performed on the excitation signal from the previous frame to the current frame, performing window processing of the thus obtained continuation of the synthesized LP segment the signal of the previous frame (14b) within the current frame to obtain a second segment of the signal to suppress sampling interference and add a second segment of the signal to suppress sampling noise in the segment of the transformed signal inside the current time segment. 14. The decoder according to claim 1, in which the parsing device (20) is configured to, when parsing the data stream (12), make a second choice depending on the second syntax section and regardless of whether the current frame (14b) and the previous frame (14a) encoded using the same or different modes from the encoding mode with conversion with suppression of sampling noise in the time domain and the encoding mode of the time domain. 15. An encoder for encoding an information signal (18) into a data stream (12), so that the data stream (12) contains a sequence of frames into which time segments of the information signal (18) are encoded, respectively containing:
a designer (42) configured to encode the current time segment (16b) of the information signal (18) into information of the current frame (14b) using the first selected one of the encoding mode with conversion with suppression of sampling noise in the time domain and the time domain encoding mode; and
an insertion device (44) configured to insert information (28) into the current frame (14b) along with a first syntax section (24) and a second syntax section in which the first syntax section (24) reports the first selection,
moreover, the constructor (42) and the insertion device (44) are configured to:
determining direct sampling noise suppression data (34) for directly sampling noise suppression at the boundary between the current time segment (16b) and the previous time segment of the previous frame and inserting direct sampling noise suppression data (34) into the current frame (14b) in the case where the current a frame (14b) and a previous frame (14a) are encoded using different modes from a transform coding mode with suppression of sampling interference in a time domain and a time domain coding mode,
and refraining from inserting any data (34) of direct suppression of sampling interference into the current frame (14b) in the case when the current frame (14b) and the previous frame (14a) are encoded using the same modes from the encoding mode with conversion with suppression of sampling interference in the time domain and the encoding mode of the time domain,
moreover, the second syntax section (26) is set depending on whether the current frame (14b) and the previous frame (14a) are encoded using the same or different modes from the encoding mode with conversion with suppression of sampling noise in the time domain and the encoding mode of the time domain . 16. The encoder according to claim 15, in which the encoder is made,
if the current frame (14b) and the previous frame (14a) are encoded using the same modes from the encoding mode with conversion with suppression of sampling noise in the time domain and the encoding mode of the time domain, with the possibility of setting the second syntax section to the first state, reporting lack of data (34 ) direct suppression of sampling interference in the current frame, and
if the current frame (14b) and the previous frame (14a) are encoded using different modes from the encoding mode with conversion with suppression of sampling noise in the time domain and the encoding mode of the time domain, with the possibility of solving in the sense of optimizing speed / distortion so that
refrain from inserting data (34) for direct suppression of sampling interference into the current frame (14b), although the current frame (14b) and the previous frame (14a) are encoded using different modes from the encoding mode with conversion with suppression of sampling interference in the time domain and the encoding mode time domain, while setting the second syntax section so that it reports the absence of data (34) for direct suppression of sampling interference in the current frame (14b), or
insert the data (34) of the direct suppression of sampling interference into the current frame (14b), while setting the second syntax section so that it reports the insertion of data (34) of the direct suppression of sampling interference in the current frame (14b). 17. A method for decoding a data stream (12), comprising a sequence of frames into which time segments of the information signal (18) are encoded, respectively comprising:
parsing the data stream (12), wherein the parsing the data stream comprises reading a first syntax section (24) and a second syntax section from the current frame (14b); and
reconstruction of the current time segment of the information signal (18) associated with the current frame (14b) based on additional information obtained from the current frame (14b) by parsing using the first selected one of the decoding mode with conversion with suppression of sampling noise in the time domain and the decoding mode of the time domain, and the first choice depends on the first syntax section (24),
moreover, when parsing the data stream (12), the first wait action is performed that the current frame (14b) contains, and thus, reading data (34) directly suppresses sampling interference from the current frame (14b), or the second surprise action, that the current frame contains, and thus, without reading the data (34) of the direct suppression of sampling interference from the current frame (14b), the second choice as to which of the first action and second action is performed, depending on the second syntax section ,
moreover, the reconstruction includes performing direct suppression of sampling interference at the boundary between the current time segment and the previous time segment (16a) of the previous frame, using data (34) of direct suppression of sampling interference. 18. A method for encoding an information signal (18) into a data stream (12) so that the data stream (12) contains a sequence of frames into which time segments of the information signal (18) are encoded, respectively containing
encoding the current time segment of the information signal (18) into the information of the current frame (14b) using the first selected one of the encoding mode with conversion with suppression of sampling noise in the time domain and the encoding mode of the time domain; and
inserting information into the current frame (14b) along with the first syntax section (24) and the second syntax section, in which the first syntax section (24) reports the first selection,
determining data (34) for direct suppression of sampling interference at the boundary between the current time segment and the previous time segment of the previous frame and inserting data (34) for direct suppression of sampling interference in the current frame (14b) in the case where the current frame (14b) and the previous frame ( 14a) are encoded using different modes from the encoding mode with conversion with suppression of sampling noise in the time domain and the encoding mode of the time domain and refraining from inserting any data (34) of direct noise suppression data into the current frame (14b) in the case where the current frame (14b) and the previous frame (14a) are encoded using the same modes from the encoding mode with conversion with suppression of sampling noise in the time domain and the encoding mode of the time domain,
moreover, the second syntax section is set depending on whether the current frame (14b) and the previous frame are encoded using the same or different modes from the encoding mode with conversion with suppression of sampling noise in the time domain and the encoding mode of the time domain. 19. Machine-readable medium storing a computer program having program code for executing, when executed on a computer, the method of claim 17. 20. Machine-readable medium storing a computer program having program code for executing, when executed on a computer, the method of claim 18.

Images