Audio encoder device and an audio decoder device having efficient gain coding in dynamic range control

Description

Dynamic range control (DRC) in the context of this document refers to a digi-tal signal processing technique to reduce the dynamic range of audio signals in a controlled way [1]. The desired reduction of the dynamic range is achieved by reducing the level of loud sound components and/or amplifying soft parts of the audio signals.

A typical application for DRC is to adapt the dynamic properties of an audio signal to a listening environment. For example, when listening to music in noisy environment, the dynamic range should be reduced in order to allow for an overall signal amplification without driving the resulting amplified signal into clipping. In this case, high signal peaks should be attenuated, e.g. by means of a limiter. Additionally, soft signal components should be amplified relative to the loud parts in order to improve their intelligibility in a noisy lis-tening environment.

It's an object of the present invention to provide an enhanced concept for dynamic range control in the context of audio transmission.

This object is achieved by an audio encoder device comprising;

an audio encoder configured for producing an encoded audio bitstream from an audio signal comprising consecutive audio frames;

a dynamic range control encoder configured for producing an encoded dy-namic range control bitstream from an dynamic range control sequence cor-responding to the audio signal and comprising consecutive dynamic range control frames, wherein each dynamic range control frame of the dynamic

range control frames comprises one or more nodes, wherein each node of the one or more nodes comprises gain information for the audio signal and time information indicating to which point in time the gain information corre-sponds;

wherein the dynamic range control encoder is configured in such way that the encoded dynamic range control bitstream comprises for each dynamic range control frame of the dynamic range control frames a corresponding bitstream portion;

wherein the dynamic range control encoder is configured for executing a shift procedure, wherein one or more nodes of the nodes of one reference dynam-ic range control frame of the dynamic range control frames are selected as shifted nodes, wherein a bit representation of each of the one or more shifted nodes of the one reference dynamic range control frame is embedded in the bitstream portion corresponding to the dynamic range control frame subse-quent to the one reference dynamic range control frame, wherein a bit repre-sentation of each remaining node of the nodes of the one reference dynamic range control frame of the dynamic range control frames is embedded into the bitstream portion corresponding to the one reference dynamic range con-trol frame.

The invention addresses the situation of audio transmission using coding of the audio signal, wherein the gain information is not directly applied to the audio signal, but also encoded and transmitted together with the encoded audio signal. At the decoder, both, the audio signal and the gain information, may be decoded and the gain information may be applied to the correspond-ing audio signal. As explained more detailed below, the invention achieves an efficient coding of the gain information. More precisely, it avoids bitrate peaks in the encoded dynamic range control bitstream.

The process of applying dynamic range control to an audio signal can be ex-pressed by a simple multiplication of the audio signal x(k) by a time-variant gain value g(k):

y(k) = g(k)x(k) (1)

where k denotes a sample time index. The value of the gain g(k) may be computed, e.g. based on a short-term estimate of the root-mean square of the audio signal x(k). More details about strategies to determine suitable gains values are discussed in [1]. In the following we refer to the time-variant gains g(k) as a gain sequence.

In the following, the coding of dynamic range control gain sequences is ex-plained. First, the dynamic range control gain sequence is divided into so-called dynamic range control frames of gain samples, containing a fixed number of gain samples. Usually, a temporal frame size for the dynamic range control frames is chosen to be equal to the temporal size of an audio frame of the corresponding audio encoder. Within each dynamic range con-trol frame, so-called nodes are selected, preferably on a uniform time grid.

The spacing of this grid defines the highest available time resolution, i.e., the minimum distance in samples between two nodes equals to samples having the highest available time resolution. Each node is represented by the sam-ple position within the dynamic range control frame, the gain information, which may be expressed as a gain value, for that position and optionally in-formation about the slope of the gain values at the node positions. For the following discussion it will be useful to define the maximum number of nodes that can be selected within one frame.

The dynamic range control encoder encodes the gain information from the nodes, e.g.. by using quantized differential values of pairs of consecutive gain nodes. At the decoder, the original gain sequence is reconstructed as

good as possible by using spline interpolation or linear interpolation based on the transmitted information of the nodes (gain value, sample position within the dynamic range control frame, and slope information if applicable).

An efficient approach for encoding the dynamic range control gain sequence is to use a quantized value of the gain difference (typically in dB) of pairs of consecutive nodes, as well as the time difference of the sample positions of these nodes within the considered dynamic range control frame. The slope information is usually not represented as a difference between two nodes. Since there is no preceding node for the first node within a frame, its gain value is not encoded in a differential way, but the values are encoded explic-itly. The time difference of the first node is usually determined as the offset to the beginning of the dynamic range control frame.

The encoder may then assign a fixed code word e.g. of a pre-defined Huff-man table (code book) to each of the gain and time differences of pairs of nodes.

At the dynamic range control decoder, the dynamic range control bitstream is decoded and the relevant information (gain value, sample position within the dynamic range control frame, and slope information if applicable) at the posi-tions of the transmitted nodes is reconstructed. The gain values for the re-maining gain samples within a frame are obtained by interpolation between pairs of transmitted and decoded nodes. The interpolation can be based on splines if the slope information of the gain nodes has been transmitted or, alternatively, using linear interpolation if only the gain differences between pairs of nodes are available and the slope information is discarded.

In principle, dynamic range control encoder/decoder chains can be operated in two modes. The so-called full-frame mode refers to the case where after decoding of a received dynamic range control bitstream, corresponding to a reference dynamic range control frame, the gains at each sample position of

the reference dynamic range control frame can be immediately determined after interpolation based on the decoded nodes. This implies that a node has to be transmitted at each frame border, i.e., at the sample position corre-sponding to the last sample of the reference dynamic range control frame. If the dynamic range control frame length is N this means the last transmitted node has to be located at the sample position N within the reference dynamic range control frame.

The invention avoids this disadvantage as it is based on the second mode, which is referred to as "delay mode". In this case, there is no need for trans-mitting a node for the last sample position within the reference dynamic range control frame. Therefore, the dynamic range control decoder has to wait for decoding the dynamic range control frame subsequent to the refer-ence dynamic range control frame in order to perform the required interpola-tion of all gain values following the last node within reference dynamic range control frame. This is because the information of the first node of the subse-quent dynamic range control frame has to be known to perform the interpola-tion between the last node of the reference dynamic range control frame and the first node of the subsequent dynamic range control frame in order to de-termine the gain value in between via interpolation.

In practice the delay caused by using the delay mode for coding of the dy-namic range control information is not an issue. This is because audio co-decs that commonly accompany the dynamic range control coding scheme also introduce an inherent delay of one audio frame when subsequently ap-plying the encoding and decoding steps. Important examples of such audio codecs are the ISO/IEC 13818-7, Advanced Audio Coding (MPEG-2 AAC), ISO/EC 14496-3, subpart 4 (MPEG-4 AAC), or ISO/IEC 23003-3, part 3, Unified Speech and Audio Coding (USAC). Such audio coding schemes re-quire the reference audio frame and the audio frame subsequent to the refer-ence audio frame in order to compute (using an overlap-add structure) the correct audio samples corresponding to the reference dynamic range control audio frame.

It is important to note that the number of nodes that are required to sufficient-ly approximate the original dynamic range control gain sequence significantly varies from dynamic range control frame to dynamic range control frame. That results from the fact that more nodes are required to represent highly time-variant gains compared to the case where only slowly changing gain values have to be encoded. This observation implies that the required bitrate to transmit gain sequences can vary significantly from frame to frame. Some frames may require a large number of nodes to be encoded, resulting in high bitrate peaks. This is not desirable, especially, when the audio signal and the dynamic range control gain sequence are transmitted in a joint bitstream comprising the encoded dynamic range control bitstream and the encoded audio bitstream, which should have almost constant bitrate. Then, a peak in the dynamic range control related bitrate reduces the available bitrate for the audio encoder, which often result in a degradation of the audio quality after decoding. However, with the current state-of-the-art methods for the coding of dynamic range control gain sequences, a reduction of the dynamic range control related bitrate in a certain frame is only achieved by reducing the number of nodes that are selected to represent the gain sequence within that frame. This again may lead to large errors between the original gain se-quence and the one that is reconstructed after the dynamic range control de-coding process. The invention overcomes these disadvantages by reducing the peak bitrates of encoded dynamic range control bitstream without in-creasing the error between the original and the reconstructed dynamic range control sequence.

In this section, the coding of dynamic range control gain sequences accord-ing to the invention is presented. The invention allows controlling the peak bitrate required for a reference dynamic range control frame without changing the resulting bitstream sequence compared to the case where the proposed

method is not used. The proposed approach exploits the inherent delay of one frame introduced by state-of-the-art audio coders to reduce peaks of number of nodes within one frame by distributing the transmission of some of the nodes to the next subsequent dynamic range control frame. The details of the proposed method are presented in the following.

As explained above, when combined with an audio coding scheme that intro-duces a frame delay relative to the dynamic range control gains, the decoded dynamic range control gains are delayed by one frame before being applied to the audio signal. This means that the nodes of the reference dynamic range control frame are applied to the valid audio decoder output at dynamic range control frame subsequent to the reference dynamic range control frame. This implies that in the default delay mode it is sufficient to transmit the nodes of the reference dynamic range control frame together with the nodes of the dynamic range control frame subsequent to the reference dy-namic range control frame and apply the corresponding dynamic range con-trol gains without a delay directly to the corresponding audio output signal at the decoder.

This fact is exploited in the invention in order to reduce the maximum number of nodes transmitted within one dynamic range control frame. According to the invention some of the nodes of the reference dynamic range control frame are shifted to the subsequent dynamic range control frame, which may be done before encoding. As it will be discussed in the following, the shifted nodes may be "preceding" the first node in the subsequent dynamic range control frame only for the encoding of the gain differences and the slope in-formation. For the coding of the time difference information, a different meth-od may be applied.

According to a preferred embodiment of the invention the shift procedure is initiated in case that a number of the nodes of the reference dynamic range control frame is greater than a predefined threshold value.

According to a preferred embodiment of the invention the shift procedure is initiated in case that a sum of a number of the nodes of the reference dynam-ic range control frame and a number of shifted nodes from the dynamic range control frame preceding the reference dynamic range control frame to be embedded in the bitstream portion corresponding to the reference dynamic range control frame is greater than a predefined threshold value.

According to a preferred embodiment of the invention the shift procedure is initiated in case that a sum of a number of the nodes of the reference dynam-ic range control frame and a number of shifted nodes from the dynamic range control frame preceding the reference dynamic range control frame to be embedded in the bitstream portion corresponding to the reference dynamic range control frame is greater than a number of the nodes of the dynamic range control frame subsequent to the reference dynamic range control frame.

Independent from the conditions defined under which the shift procedure is initiated, the first node of the reference dynamic range control frame should not be shifted to the subsequent dynamic range control frame as its value is needed for interpolation of the gain control values at the beginning of the ref-erence dynamic range control frame. Furthermore, a node should be shifted only one time in order to avoid a delay when decoding the bitstream.

According to a preferred embodiment of the invention the time information of the one or more nodes is represented in such way that the one or more shift-ed nodes may be identified by using the time information.

According to a preferred embodiment of the invention the time information of the one or more shifted nodes is represented by a sum of a time difference from a beginning of the dynamic range control frame to which the respective node belongs to the temporal position of the respective node within the dy- namic range control frame to which the respective node belongs and an off-set value being greater than or equal to a temporal size of the dynamic range control frame subsequent to the respective dynamic range control frame.

According to a preferred embodiment of the invention the gain information of the bit representation of the shifted node, which is at a first position of the bitstream portion corresponding to the dynamic range control frame subse-quent to the reference dynamic range control frame, is represented by an absolute gain value and wherein the gain information of each bit representa-tion of the shifted nodes at a position after the bit representation of the node, which is at the first position of the bitstream portion corresponding to the dy-namic range control frame subsequent to the reference dynamic range con-trol frame, is represented by a relative gain value which is equal to a differ-ence of a gain value of the bit representation of the respective shifted node and the gain value of the bit representation of the node, which precedes the bit representation of the respective node.

According to preferred embodiment of the invention, in case that the bit rep-resentations of one or more shifted nodes of the reference dynamic range control frame is embedded in the bitstream portion corresponding to the dy-namic range control frame subsequent to the reference dynamic range con-trol frame, the gain information of the bit representation of the node of the subsequent dynamic range control frame at a first position of the bitstream portion corresponding to the dynamic range control frame subsequent to the reference dynamic range control frame after the one or more positions of the bit representations of the one or more shifted nodes is represented by a rela-tive gain value which is equal to a difference of a gain value of the bit repre-sentation of the respective node and a gain value of the bit representation of the shifted node, which precedes the bit representation of the respective node.

According to a preferred embodiment of the invention a temporal size of the audio frames is equal to a temporal size of the dynamic range control frames

According to a preferred embodiment of the invention the one or more nodes of one of the dynamic range control frame are selected from a uniform time grid.

According to a preferred embodiment of the invention each node of the one or more nodes comprises slope information.

According to a preferred embodiment of the invention the dynamic range control encoder is configured for encoding the nodes using an entropy en- coding technique, such as Huffman coding or arithmetic coding.

The encoder may assign a fixed code word e.g. of a pre-defined Huffman table (code book) to each of the gain and time differences of pairs of nodes. Examples of suitable Huffman tables for encoding the time differences of pairs of consecutive nodes are given in Table 1 and Table 2, respectively.

Table 1: Example of a Huffman table for the coding of time differences of DRC gain nodes.

Table 2: Example of a Huffman table for the coding of time differences of DRC gain nodes, where Z=ceil(log2(2*nNodesMax))

In a further aspect of the invention the objective is achieved by an audio de-coder device comprising:

an audio decoder configured for decoding an encoded audio bitstream in or-der to reproduce an audio signal comprising consecutive audio frames;

a dynamic range control decoder configured for decoding an encoded dy-namic range control bitstream in order to reproduce an dynamic range control sequence corresponding to the audio signal and comprising consecutive dy-namic range control frames;

wherein the encoded dynamic range control bitstream comprises for each dynamic range control frame of the dynamic range control frames a corre-sponding bitstream portion;

wherein the encoded dynamic range control bitstream comprises bit repre-sentations of nodes, wherein each bit representation of one node of the nodes comprises gain information for the audio signal AS and time infor-mation indicating to which point in time the gain information corresponds; wherein the encoded dynamic range control bit stream comprises bit repre-sentations of shifted nodes selected from the nodes of one reference dynam-ic range control frame of the dynamic range control frames, which are em-bedded in a bitstream portion corresponding to the dynamic range control frame subsequent to the one reference dynamic range control frame, wherein the bit representation of each remaining node of the nodes of the one refer-ence dynamic range control frame of the dynamic range control frames is embedded into the bitstream portion corresponding to the one reference dy-namic range control frame; and

wherein the dynamic range control decoder is configured for decoding the bit representation of each remaining node of the remaining nodes of the one reference dynamic range contral frame of the dynamic range control frames in order to reproduce each remaining node of the one reference dynamic range control frame of the dynamic range control frames, for decoding the bit representation of each shifted node of the shifted nodes selected from the nodes of the one reference dynamic range control frame of the dynamic range control frames in order to reproduce each shifted node of the shifted nodes selected from the nodes of the one reference dynamic range control frame of the dynamic range control frames and for combining the reproduced remaining nodes and the reproduced shifted nodes in order to reconstruct the reference dynamic range control frame.

The dynamic range control decoder receives the dynamic range control bit-stream. The dynamic range control bitstream, which corresponds to the node information (sample position, gain value, and slope information if applicable), may be decoded in the following way:

A value for the time difference between two nodes (e.g. as an integer multiple of the minimum distance between two nodes) is determined from the re-ceived code word based e.g. on the rules shown in a Huffman code book. The corresponding sample position of the currently decoded node is obtained by adding the time difference value to the sample position value computed for the previous node.

After decoding the nodes of the reference dynamic range control frame the nodes of the subsequent dynamic range control frame are decoded.

If the determined sample position within the subsequent dynamic range con-trol frame corresponds to a value that is larger than the length of a subse-quent dynamic range control frame, the dynamic range control decoder knows that the current temporal node information refers to a node originally located in the reference dynamic range control frame.

To obtain the correct sample position within the reference dynamic range control frame, an offset is subtracted from the computed sample position. A practical example is to subtract the value that corresponds to the length of a dynamic range control frame (which implies that the encoder has added the same value to the original sample position). A typical example for the offset value is the temporal size of a dynamic range control frame.

After decoding and if applicable correcting the time information of all nodes in the entire subsequent dynamic range control frame, the decoder knows how many nodes have been shifted back to the reference dynamic range control frame (without explicitly providing this information at the encoder) and on which sample position they are located within the reference dynamic range control frame.

The dynamic range control decoder further determines the gain value infor-mation of all nodes of a received frame by decoding the differential gain in-formation in the bitstream.

From the decoding step of the time information, the decoder knows how many of the decoded gain values have to be assigned to the nodes of the

reference dynamic range control frame (and to which sample position) and which gain values are assigned to nodes in the reference dynamic range control frame.

The decoding of the slope information and the assignment to the correct nodes are performed analogously to the decoding process of the gain values.

After decoding all nodes of the subsequent dynamic range control frame, it can be assured that all nodes required for computing the gain values for each sample of the reference dynamic range control frame via interpolation are available. After the interpolation step, the dynamic range control gain values for each sample can be applied to the corresponding correct audio samples.

According to a preferred embodiment of the invention the dynamic range control decoder is configured for identifying the one or more shifted nodes by using the time information.

According to a preferred embodiment of the invention the dynamic range control decoder is configured for decoding the time information of the one or more shifted nodes, which is represented by a sum of a time from a begin-ning of the dynamic range control frame to which the respective node be-longs to the temporal position of the respective node within the dynamic range control frame to which the respective node belongs and an offset value being greater than or equal to a temporal size of the dynamic range control frame subsequent to the respective dynamic range control frame.

According to preferred embodiment of the invention the dynamic range con-trol decoder is configured for decoding the gain information of the bit repre-sentation of the shifted node, which is at a first position of the bitstream por-tion corresponding to the dynamic range control frame subsequent to the ref-erence dynamic range control frame, is represented by an absolute gain val-ue and wherein the gain information of each bit representation of the shifted nodes at a position after the bit representation of the node, which is at the first position of the bitstream portion corresponding to the dynamic range control frame subsequent to the reference dynamic range control frame, is represented by a relative gain value which is equal to a difference of a gain value of the bit representation of the respective shifted node and the gain value of the bit representation of the node, which precedes the bit represen-tation of the respective node

According to a preferred embodiment of the invention the dynamic range control decoder is configured for decoding the gain information of the bit rep-resentation of the node of the subsequent dynamic range control frame at a first position of the bitstream portion corresponding to the dynamic range control frame subsequent to the reference dynamic range control frame after the one or more positions of the bit representations of the one or more shifted nodes is represented by a relative gain value which is equal to a difference of a gain value of the bit representation of the respective node and a gain value of the bit representation of the shifted node, which precedes the bit represen-tation of the respective node.

According to preferred embodiment of the invention a temporal size of the audio frames is equal to a temporal size of the dynamic range control frames.

According to a preferred embodiment of the invention the one or more nodes of one of the dynamic range control frames are selected from a uniform time grid.

According to preferred embodiment of the invention each node of the one or more nodes comprises slope information.

According to preferred embodiment of the invention the dynamic range con-trol decoder is configured for decoding the bit representations of the nodes using an entropy decoding technique.

The objective is further obtained by a system comprising an audio encoder device according to the invention and an audio decoder device according to the invention.

The invention further provides a method for operating an audio encoder, the method comprises the steps:

producing an encoded audio bitstream from an audio signal comprising con-secutive audio frames;

producing an encoded dynamic range control bitstream from an dynamic range control sequence corresponding to the audio signal and comprising consecutive dynamic range control frames, wherein each dynamic range control frame of the dynamic range control frames comprises one or more nodes, wherein each node of the one or more nodes comprises gain infor-mation for the audio signal and time information indicating to which point in time the gain information corresponds

wherein the encoded dynamic range control bitstream comprises for each dynamic range control frame of the dynamic range control frames a corre-sponding bitstream portion;

executing a shift procedure, wherein one or more nodes of the nodes of one reference dynamic range control frame of the dynamic range control frames are selected as shifted nodes, wherein a bit representation of each of the one or more shifted nodes of the one reference dynamic range control frame is embedded in the bitstream portion corresponding to the dynamic range con-trol frame subsequent to the one reference dynamic range control frame, wherein a bit representation of each remaining node of the nodes of the one reference dynamic range control frame of the dynamic range control frames is embedded into the bitstream portion corresponding to the one reference dynamic range control frame.

The invention further provides a method for operating an audio decoder, the method comprises the steps:

decoding an encoded audio bitstream in order to reproduce an audio signal comprising consecutive audio frames;

decoding an encoded dynamic range control bitstream in order to reproduce an dynamic range control sequence corresponding to the audio signal and comprising consecutive dynamic range control frames;

wherein the encoded dynamic range control bitstream comprises for each dynamic range control frame of the dynamic range control frames a corre-sponding bitstream portion;

wherein the encoded dynamic range control bitstream comprises bit repre-sentations of nodes, wherein each bit representation of one node of the nodes comprises gain information for the audio signal AS and time infor-mation indicating to which point in time the gain information corresponds;

wherein the encoded dynamic range control bit stream comprises bit repre-sentations of shifted nodes selected from the nodes of one reference dynam-ic range control frame of the dynamic range control frames, which are em-bedded in a bitstream portion corresponding to the dynamic range control frame subsequent to the one reference dynamic range control frame, wherein the bit representation of each remaining node of the nodes of the one refer-ence dynamic range control frame of the dynamic range control frames is embedded into the bitstream portion corresponding to the one reference dy-namic range control frame; and

wherein the bit representation of each remaining node of the remaining nodes of the one reference dynamic range control frame of the dynamic range control frames is decoded in order to reproduce each remaining node of the one reference dynamic range control frame of the dynamic range con-trol frames;

wherein the bit representation of each shifted node of the shifted nodes se-lected from the nodes of the one reference dynamic range control frame of the dynamic range control frames is decoded in order to reproduce each shifted node of the shifted nodes selected from the nodes of the one refer-ence dynamic range control frame of the dynamic range control frames; and

wherein the reproduced remaining nodes and the reproduced shifted nodes are combined in order to reconstruct the reference dynamic range control frame.

In another aspect the invention provides a program for, when running on a processor, executing the method according to the invention.

Preferred embodiments of the invention are subsequently discussed with re-spect to the accompanying drawings, in which;

Fig. 1 illustrates an embodiment of an audio encoder device according to the invention in a schematic view;

Fig. 2 illustrates the principle of dynamic range control applied in the context of audio coding in a schematic view;

Fig. 3 illustrates the different modes for the coding of dynamic range control gain sequences in a schematic view;

Fig. 4 illustrates the application of dynamic range control in the con- text of audio coding in a schematic view;

Fig. 5 illustrates a shift procedure for nodes according to the invention in a schematic view;

Fig. 6 illustrates the coding of time information according to the inven- tion in a schematic view;

Fig. 7 illustrates the coding of gain information according to the inven- tion in a schematic view;

Fig. 8 illustrates the coding of slope information according to the in- vention in a schematic view; and

Fig. 9 illustrates an embodiment of an audio decoder device according to the invention in a schematic view.

Fig. 1 illustrates an embodiment of an audio encoder device 1 according to the invention in a schematic view. The audio encoder device 1 comprises:

an audio encoder 2 configured for producing an encoded audio bitstream ABS from an audio signal AS comprising consecutive audio frames AFP, AFR, AFS;

a dynamic range control encoder 3 configured for producing an encoded dy-namic range control bitstream DBS from an dynamic range control sequence DS corresponding to the audio signal AS and comprising consecutive dynam-ic range control frames DFP, DFR, DFS, wherein each dynamic range control frame DFP, DFR, DFS of the dynamic range control frames DFP, DFR, DFS comprises one or more nodes A0 ... A5; B0 ... B2; C0, wherein each node of the one or more nodes A0 ... A5; B0 ... B2; C0 comprises gain information GA0 ... GA5; GB0 ... GB2; GC0 for the audio signal AS and time information TA0 ... TA5; TB0 ... TB2; TC0 indicating to which point in time the gain information GA0 ... GA5; GB0 ... GB2; GC0 corresponds;

wherein the dynamic range control encoder 3 is configured in such way that the encoded dynamic range control bitstream DBS comprises for each dy-namic range control frame DFP, DFR, DFS of the dynamic range control frames DFP, DFR, DFS a corresponding bitstream portion DFP', DFR', DFS';

wherein the dynamic range control encoder 2 is configured for executing a shift procedure, wherein one or more nodes B1, B2 of the nodes B0 ... B2 of one reference dynamic range control frame DFR of the dynamic range con-trol frames DFP, DFR, DFS are selected as shifted nodes B1, B2, wherein a bit representation Β'1, B'2 of each of the one or more shifted nodes B1, B2 of the one reference dynamic range control frame DFR is embedded in the bit-stream portion DFS' corresponding to the dynamic range control frame DFS subsequent to the one reference dynamic range control frame DFR, wherein a bit representation B'0 of each remaining node B0 of the nodes B0 ... B2 of the one reference dynamic range control frame DFR of the dynamic range control frames DFP, DFR, DFS is embedded into the bitstream portion DFR' corresponding to the one reference dynamic range control frame DFR.

The invention allows controlling the peak bitrate required for a reference dy-namic range control frame DFR without changing the resulting bitstream se-quence DBS compared to the case where the proposed method is not used. The proposed approach exploits the inherent delay of one frame introduced by state-of-the-art audio coders to reduce peaks of number of nodes within one frame by distributing the transmission of some of the nodes to the next subsequent dynamic range control frame. The details of the proposed meth-od are presented in the following.

As explained above, when combined with an audio coding scheme that intro-duces a frame delay relative to the dynamic range control gains, the decoded dynamic range control gains are delayed by one frame before being applied to the audio signal. This means that the nodes of the reference dynamic range control frame are applied to the valid audio decoder output at dynamic range control frame subsequent to the reference dynamic range control frame. This implies that in the default delay mode it is sufficient to transmit the nodes of the reference dynamic range control frame together with the nodes of the dynamic range control frame subsequent to the reference dy-namic range control frame and apply the corresponding dynamic range con-trol gains without a delay directly to the corresponding audio output signal at the decoder.

This fact is exploited in the invention in order to reduce the maximum number of nodes transmitted within one dynamic range control frame. According to the invention some of the nodes of the reference dynamic range control frame are shifted to the subsequent dynamic range control frame, which may be done before encoding. As it will be discussed in the following, the shifted nodes may be "preceding" the first node in the subsequent dynamic range control frame only for the encoding of the gain differences and the slope in-formation. For the coding of the time difference information, a different meth-od may be applied.

In the example shown in Fig. 1 the preceding dynamic range control frame DFP contains six nodes A0 ... A5 of which the nodes A4, A5 are shifted into the bitstream portion DFR'. Furthermore, the reference dynamic range con-trol frame DFR contains three nodes B0 ... B2. The sum of the number of the shifted nodes A4, A5 and the nodes B0 ... B2 of the reference dynamic range control frame DFR is equal to five which is bigger than the number of the nodes Co of the subsequent dynamic range control frame DFS so that a shift procedure is initiated in such way that nodes B1, B2 are shifted into the bit-stream portion DFS'. Although the maximum number of nodes within the dy-namic range control frames DFS, DFR, DFP is equal to six, is the maximum

number of nodes within the bitstream portions DFS', DFR', DFP' own equal to four so that bitstream peak is avoided.

According to preferred embodiment of the invention a temporal size of the audio frames AFP, AFR, AFS is equal to a temporal size of the dynamic range control frames DFP, DFR, DFS.

According to preferred embodiment of the invention the one or more nodes A0 ... A5; B0 ... B2; C0 of one of the dynamic range control frame DFP, DFR, DFS are selected from a uniform time grid.

According to a preferred embodiment of the invention the dynamic range control encoder 3 is configured for encoding the nodes A0 ... A5; B0 ... B2 C0 using an entropy encoding technique.

In a further aspect the invention provides a method for operating an audio encoder 1, the method comprises the steps:

producing an encoded audio bitstream ABS from an audio signal AS com-prising consecutive audio frames AFP, AFR, AFS;

producing an encoded dynamic range control bitstream DBS from an dynam-ic range control sequence DS corresponding to the audio signal AS and comprising consecutive dynamic range control frames DFP, DFR, DFS, wherein each dynamic range control frame DFP, DFR, DFS of the dynamic range control frames DFP, DFR, DFS comprises one or more nodes A0 ... A5; B0 ... B2, C0, wherein each node of the one or more nodes A0 ... A5; B0 ... B2; C0 comprises gain information GA0 ... GA5; GB0 ... GB2; GC0 for the au-dio signal AS and time information TA0 ... TA5; TB0 ... TB2: TC0 indicating to which point in time the gain information corresponds

wherein the encoded dynamic range control bitstream DBS comprises for

each dynamic range control frame DFP, DFR, DFS of the dynamic range control frames DFP, DFR, DFS a corresponding bitstream portion DFP', DFR', DFS';

executing a shift procedure, wherein one or more nodes B1, B2 of the nodes B0 ... B2 of one reference dynamic range control frame DFR of the dynamic range control frames DFP, DFR, DFS are selected as shifted nodes B1, B2, wherein a bit representation Β'1, B'2 of each of the one or more shifted nodes B1, B2 of the one reference dynamic range control frame DFR is embedded in the bitstream portion DFS' corresponding to the dynamic range control frame DFS subsequent to the one reference dynamic range control frame DFR, wherein a bit representation ΒΌ of each remaining node B0 of the nodes B0 ... B2 of the one reference dynamic range control frame DFR of the dynamic range control frames DFP, DFR, DFS is embedded into the bitstream portion DFR' corresponding to the one reference dynamic range control frame DFR.

Fig. 2 illustrates the principle of dynamic range control applied in the context of audio coding in a schematic view.

The process of applying DRC to a signal can be expressed by a simple mul-tiplication of the audio signal x(k) by a time-variant gain value g(k):

y(k) = g(k)x(k) (1)

where k denotes a sample time index. The value of the gain g(k) is comput-ed, e.g. based on a short-term estimate of the root-mean square of the input signal x(k). More details about strategies to determine suitable gains values are discussed in [1]. In the following we refer to the time-variant gains g(k) as a gain sequence.

The invention refers to an application scenario, where both, the audio signal AS and the dynamic range control sequence DS are coded and transmitted.

In this case, the dynamic range control gains are not directly applied to the audio signal AS, but encoded and transmitted together with the encoded au-dio signal ABS. At the decoder 4, both, the audio signal AS and the dynamic range control sequence DS are decoded and the dynamic range control in-formation is applied to the corresponding audio signal AS.

In one aspect the invention provides a system comprising an audio encoder device 1 according to the invention and an audio decoder device 4 according to the invention.

Fig. 3 illustrates the different modes for the coding of dynamic range control gain sequences in a schematic view, namely the full-frame mode (A) and de-lay mode (B). Gain nodes received in frame n are shown as circles and gain nodes received frame n+1 are shown as squares. The solid line illustrates the interpolated DRC gain up to DRC frame n+1.

In principle, the dynamic range control encoder/decoder chain can be oper-ated in two modes. The so-called full-frame mode refers to the case where after decoding of a received dynamic range control bitstream, corresponding to a specific dynamic range control frame, the gains at each sample position of the dynamic range control frame can be immediately determined after in-terpolation based on the decoded nodes. This implies that a node has to be transmitted at each frame border, i.e., at the sample position corresponding to the last sample of the dynamic range control frame. If the dynamic range control frame length is N this means the last transmitted node has to be lo-cated at the sample position N within that frame. This is illustrated at the top in Fig. 3 denoted by "A". As shown, the dynamic range control gains of the nth frame can immediately be applied to the corresponding audio frame.

The second mode is referred to as "delay mode" and it is illustrated in the lower part "B" of Fig, 3. In this case, there is no node transmitted for the last sample position within frame n. Therefore, the DRC decoder has to wait for

decoding the DRC frame n+1 in order to perform the required interpolation of all gain values following the last node within frame n. This is because the in-formation of the first node of frame n+1 has to be known to perform the inter-polation between the last node of frame n and the first node in frame n+1 in order to determine the gain value in between via interpolation.

Fig. 4 illustrates the application of dynamic range control in the context of audio coding in a schematic view where the audio coder introduces one frame delay relative to the dynamic range coding scheme.

Fig. 5 illustrates a shift procedure for nodes according to the invention in a schematic view. The left-hand side shows the situation when using a state-of-the-art approach, whereas the right-hand side shows the proposed method, where each square corresponds to a node A0 ... A5; B0 ... B2; C0.

According to a preferred embodiment of the invention the shift procedure is initiated in case that a number of the nodes B0 ... B2 of the reference dynamic range control frame DFR is greater than a predefined threshold value.

According to a preferred embodiment of the invention the shift procedure is initiated in case that a sum of a number of the nodes B0 ... B2 of the refer-ence dynamic range control frame DFR and a number of shifted nodes A4, A5 from the dynamic range control frame DFP preceding the reference dynamic range control frame DFR to be embedded in the bitstream portion DFR' cor-responding to the reference dynamic range control frame DFR is greater than a predefined threshold value.

According to preferred embodiment of the invention the shift procedure is initiated in case that a sum of a number of the nodes B0 ... B2 of the refer-ence dynamic range control frame DFR and a number of shifted nodes A4, A5 from the dynamic range control frame DFP preceding the reference dynamic range control frame DFR to be embedded in the bitstream portion DFR' cor- responding to the reference dynamic range control frame DFR is greater than a number of the nodes C0 of the dynamic range control frame DFS subse-quent to the reference dynamic range control frame DFR.

As explained above, when combined with an audio coding scheme that intro-duces a frame delay relative to the dynamic range control frames, the decod-ed dynamic range control gains are delayed by one frame before being ap-plied to the audio signal. Considering the left-hand side in Fig. 5, this means that the nodes Ai of the nth frame are applied to the valid audio decoder out-put at frame n+1 . This implies that in the default delay mode it would be suffi-cient to transmit the nodes A, together with the node Bo in frame n+1 and ap-ply the corresponding DRC gains without a delay directly to the correspond-ing audio output signal at the decoder.

This fact is exploited in the proposed method to reduce the maximum number of nodes transmitted within one frame. This is illustrated on the right-hand side in Figure 4. The nodes A4 and A5 are shifted to frame n+1 before encod-ing, i.e., the maximum number of nodes in frame n is reduced from 6 to 4 in the given example. As it will be discussed in the following, the nodes A4 and A5 are "preceding" the first node in frame n+1 , i.e., B0 only for the encoding of the gain differences and the slope information. For the coding of the time dif-ference information, a different method has to be applied.

Fig. 6 illustrates the coding of time information according to the invention in a schematic view.

According to a preferred embodiment of the invention the time information

TA0 ... TA5: TB0 ... TB2; TC0 of the one or more nodes A0 ... A5; B0 ... B2; C0 is represented in such way that the one or more shifted nodes A4, A5; Β1, B2 may be identified by using the time information TA4. TA5: TB1, TB2.

According to preferred embodiment of the invention the time information TA4, TA5; TB1, TB2 of the one or more shifted nodes A4, A5; B1 B2 is represented by a sum of a time difference t_A4, t_A5; t _B1, t _B2 from a beginning of the dynamic range control frame DFP; DFR to which the respective node A4, A5; B1, B2 belongs to the temporal position of the respective node A4, A5; B1 B2 within the dynamic range control frame DFP; DFR to which the respective node A4, A5; B1, B2 belongs and an offset value drcFrameSize being greater than or equal to a temporal size of the dynamic range control frame DFR; DFS subsequent to the respective dynamic range control frame DFP; DFR.

First we consider the encoding of the time differences between pairs of nodes. In Fig. 6 the situation for determining the time differences for pairs of nodes is depicted for the example according to Figure 4, where t_Ai denotes the sample position of node Ai on the grid of possible node positions within a frame. As discussed earlier nodes can be selected on a uniform time grid, where the spacing of this grid defines the highest available time resolution deltaTmin, Thus, the time information t_Ai is given in samples, where the time differences between two nodes are always integer multiples of del-taTmin.

The temporal position information of a node is encoded in a differential way, i.e., relative to the position of the previous node. If a node is the first node within a frame, the time difference is determined relative to the beginning of a frame. The left-hand side of Fig. 6 depicts the situation if no node shifting is applied. In this case, the differential time information of node A4 tDrcDelta_A4 is computed as tDrcDelta_A4 = t_A4- t_A3. This differential time value is then encoded using the corresponding entry in an appropriate Huffman table, e.g. according to Table 1 or 2. As another example we look at the encoded time difference of node B0. Since it is the first node of frame n+1, the correspond-ing time difference is determined relative to the beginning of the frame, i.e. tDrcDelta__B0 = t_B0.

Let us now consider the encoding of the node position for the proposed node reservoir technique using node shifting. For the example shown on the right-hand side of Fig. 6, the nodes A4 and A5 have been shifted to the next frame for encoding. The representation of nodes A0 to A3 has not changed and the encoded time differences are therefore also not changed. The same is true for the encoded position information of node B0. However, the time infor-mation of node A4 and node A5 is now processed differently. As shown in Fig.

6, the original value t_A4 indicating the sample position of node A4 is modified at the encoder by adding an offset of drcFrameSize. Since the resulting posi-tion information exceeds the maximum value that would be possible in case of regular encoding, the offset indicates the decoder that the corresponding node has to be further processed within the context of the previous frame. Furthermore, the decoder knows that the original sample position t_A4 is ob-tained by subtracting the offset drcFrameSize from the decoded value.

Next, we consider the computation of the time difference information that is actually encoded for the situation shown on the right-hand side of Fig. 6. For coding efficiency reasons, the differential position information for node A4 is computed relative to node B0. In contrast to the situation previously dis-cussed for the left-hand side of Fig. 6, the differential time information is now computed according to tDrcDelta_A4 = t_A4 + drcFrameSize - t_B0, i.e., by including the offset. Analogously, for node A5 we obtain tDrcDelta_A5 = t_A5 + drcFrameSize - t_A4- drcFrameSize, which obviously is the same as tDrcDelta_A5 = t_A5 - t_A4. These differential time values are encoded using the corresponding code word entry of the correct Huffman table, e.g. accord-ing to Table 1 or 2.

The method for decoding the temporal position information can be summa-rized as follows. The decoder extracts the time difference information of a node based on the corresponding code word from the bitstream. The time information is obtained by adding the time difference information to the time information of the previous node. If the resulting sample position is larger than drcFrameSize the decoder knows that the present node has to be pro-cessed as if it were the last node in the previous frame, i.e., it has to be ap-pended to the nodes decoded in the previous frame. The correct sample po-sition is determined by subtracting the offset value drcFrameSize from the decoded time value. The same processing steps are applied in an analog way if more shifted nodes occur in a decoded frame.

After decoding and correcting the time information of an entire frame, the de-coder knows how many nodes have been shifted back to the previous frame (without explicitly providing this information at the encoder) and on which sample position they are located within the previous frame. The information about the number of shifted nodes will be further exploited in the context of decoding gain and slope information described below.

Fig. 7 illustrates the coding of gain information according to the invention in a schematic view.

According to preferred embodiment of the invention the gain information GBi of the bit representation Β'1 of the shifted node B1, which is at a first position of the bitstream portion DFS' corresponding to the dynamic range control frame DFS subsequent to the reference dynamic range control frame DFR, is represented by an absolute gain value g_B1 and wherein the gain information GB2 of each bit representation B'2 of the shifted nodes B2 at a position after the bit representation Β'1 of the node B1, which is at the first position of the bitstream portion DFS' corresponding to the dynamic range control frame DFS subsequent to the reference dynamic range control frame DFR, is rep-resented by a relative gain value which is equal to a difference of a gain val-ue g_B2 of the bit representation B'2 of the respective shifted node B2 and the gain value g_B1 of the bit representation B'1 of the nodeB1, which precedes the bit representation B'2 of the respective nodeB2.

According to a preferred embodiment of the invention, in case that the bit representations Β'1, B'2 of one or more shifted nodes B1, B2 of the reference dynamic range control frame DFR is embedded in the bitstream portion DFS' corresponding to the dynamic range control frame DFS subsequent to the reference dynamic range control frame DFR, the gain information GC0 of the bit representation C'0 of the node C0 of the subsequent dynamic range con-trol frame DFS at a first position of the bitstream portion DFS' corresponding to the dynamic range control frame DFS subsequent to the reference dynam-ic range control frame DFR after the one or more positions of the bit repre-sentations Β'1, B'2 of the one or more shifted nodes B1, B2 is represented by a relative gain value which is equal to a difference of a gain value g_C0 of the bit representation C'0 of the respective node C0 and a gain value g_B2 of the bit representation B'2 of the shifted nodeB2, which precedes the bit represen-tation C'0 of the respective node C0.

In Fig. 7 the situation for determining the gain differences for pairs of nodes is depicted for the example according to Figure 5, where g_Ai denotes the gain value of node Ai.

First, the differential gain values for the node A4 is considered. For the ap-proach without node reservoir, depicted on the left-hand side of Fig. 7, the differential gain value gain Deita_A4 is computed from the difference of the gain value (in dB) of the preceding node A3 and the node A4. i.e., gainDel-ta_A4 = g_A4 - g_A3. This differential gain value is then encoded using the corresponding entry in an appropriate Huffman table. Furthermore, we con-sider the first node of frame n+1 on the left-hand side of Fig. 7. Since B0 is the first node of that frame, it gain value is not encoded in a differential way, but according to a specific coding of initial gain values gainInitial, i.e., the gain value is encoded as its actual value: gainDelta_B0 = g_B0.

For the situation shown on the right-hand side, where the node A4 has been shifted to the next frame n+1, the values of the encoded gain information is different. As can be seen, after being shifted, the node A4 becomes the first node in frame n+1 with respect to encoding the gain differences. Thus, its gain value is not encoded in a differential way, but the specific coding of ini-tial gain values is applied as described above. The differential gain value of A5 will remain the same for both situations shown on the left- and the right-hand side. Since node B0 now follows node A5 if the node reservoir is used, its gain information will be determined from the difference of the gains of node B0 and A5, i.e., gainDelta_B0 = g _ B0 - g _A5. Note that only the way how the gain differences are determined changes when applying the node reser-voir technique, whereas the reconstructed values of the gains remain the same for each node. Obviously, after decoding the entire gain related infor-mation of the frames n and n+1, the obtained gain values for the nodes A0 to B0 are identical to that obtained in the left-hand side, and the nodes can be computed "in time" for application of the DRC gains to the corresponding au-dio frame.

As discussed in the previous paragraph, the number of shifted nodes and their sample position within the previous frame are known after decoding the time difference information. As illustrated on the right-hand side of Figure 6, the gain values of shifted nodes from frame n start immediately from the be-ginning of the received gain information of frame n+1. Therefore, the infor-mation on the number of shifted nodes is sufficient for the decoder to assign each gain value to the correct sample position within the correct frame. Con-sidering the example shown on the right-hand side in Figure 6, the decoder knows that the first two decoded gain values of frame n+1 have to be ap-pended to the last gain values of the previous frame, whereas the third gain value corresponds to the correct gain value of the first node in the current frame.

Fig. 8 illustrates the coding of slope information according to the invention in a schematic view.

According to a preferred embodiment of the invention each node A0 ... A5; B0

... B2; C0 of the one or more nodes comprises A0 ... A5; B0 ... B2; C0 slope information SA0 ... SA5; SB0 ... SB2: SC0.

Next, the coding of slope information is considered, which is illustrated in Fig.

8. The slope information of the nodes isn't encoded in a differential way be-tween pairs of nodes, but for each node independently. Therefore, the slope related information remains unchanged in both cases with and without usage of the node reservoir. As in case of coding of gain values, the Huffman tables for generating the code words for slope information remain the same for both cases, with and without using the proposed node shifting. The assignment of the slope information to the correct sample position within the correct frame is performed analogously to the case of decoding the gain values.

After all nodes information received for frame n+1 have been decoded and if applicable shifted back to the preceding frame n, the gain interpolation for frame n using splines or linear interpolation can be performed in the common way and the gain values are applied to the corresponding audio frame.

Fig. 9 illustrates an embodiment of an audio decoder device according to the invention in a schematic view. The audio decoder device 4 comprises:

an audio decoder 5 configured for decoding an encoded audio bitstream ABS in order to reproduce an audio signal AS comprising consecutive audio frames AFP, AFR, AFS;

a dynamic range control decoder 6 configured for decoding an encoded dy-namic range control bitstream DBS in order to reproduce an dynamic range control sequence DS corresponding to the audio signal AS and comprising consecutive dynamic range control frames DFP, DFR, DFS:

wherein the encoded dynamic range control bitstream DBS comprises for

each dynamic range control frame DFP, DFR, DFS of the dynamic range control frames a corresponding bitstream portion DFP', DFR', DFS';

wherein the encoded dynamic range control bitstream DBS comprises bit representations A'0 ... A'5; B'0 ... B'2; C'0 of nodes A0 ... A5; B0 ... B2; C0, wherein each bit representation of one node of the nodes comprises gain information GA0 ... GA5; GB0 ... GB2; GC0 for the audio signal AS and time information TA0 ... TA5; TB0 ... TB2; TC0 indicating to which point in time the gain information GA0 ... GA5; GB0 ... GB2; GC0 corresponds;

wherein the encoded dynamic range control bit stream DBS comprises bit representations Β'1, B'2 of shifted nodes B1, B2 selected from the nodes B0 ... B2 of one reference dynamic range control frame DFR of the dynamic range control frames DFP, DFR, DFS, which are embedded in a bitstream portion corresponding to the dynamic range control frame DFS subsequent to the one reference dynamic range control frame DFR, wherein the bit representa-tion B'0 of each remaining node B0 of the nodes B0 ... B2 of the one reference dynamic range control frame DFR of the dynamic range control frames DFP, DFR, DFS is embedded into the bitstream portion DFR' corresponding to the one reference dynamic range control frame DFR; and

wherein the dynamic range control decoder 6 is configured for decoding the bit representation B'0 of each remaining node B0 of the remaining nodes B'0 of the one reference dynamic range control frame DFR of the dynamic range control frames DFP, DFR, DFS in order to reproduce each remaining node B0 of the one reference dynamic range control frame DFR of the dynamic range control frames DFP, DFR, DFS, for decoding the bit representation B'1, B'2 of each shifted node B1, B2 of the shifted nodes B1, B2 selected from the nodes B0 ... B2 of the one reference dynamic range control frame DFR of the dynamic range control frames DFP. DFR, DFS in order to reproduce each shifted node B1, B2 of the shifted nodes B1, B2 selected from the nodes of the one reference dynamic range control frame DFR of the dynamic range con-

trol frames DFP, DFR, DFS and for combining the reproduced remaining nodes B0 and the reproduced shifted nodes B1, B2 in order to reconstruct the reference dynamic range control frame DFR,

According to a preferred embodiment of the invention the dynamic range controi decoder 6 is configured for identifying the one or more shifted nodes A4, A5; B1, B2 by using the time information TA4, TA5; TB1, TB2.

According to a preferred embodiment of the invention the dynamic range control decoder 6 is configured for decoding the time information TA4, TA5; TB1, TB2 of the one or more shifted nodes A4, A5; B1 B2, which is represented by a sum of a time difference t_A4, t_A5; t_B1, t_B2 from a beginning of the dynamic range control frame DFP; DFR to which the respective node A4, A5; B1, B2 belongs to the temporal position of the respective node A4, A5; B1, B2 within the dynamic range control frame DFP; DFR to which the respective node A4, A5; B1, B2 belongs and an offset value drcFrameSize being greater than or equal to a temporal size of the dynamic range control frame DFR; DFS subsequent to the respective dynamic range control frame DFP; DFR.

According to a preferred embodiment of the invention the dynamic range control decoder 6 is configured for decoding the gain information GB1 of the bit representation Β'1 of the shifted node B1 , which is at a first position of the bitstream portion DFS' corresponding to the dynamic range control frame DFS subsequent to the reference dynamic range control frame DFR, is rep-resented by an absolute gain value g B1 and wherein the gain information GB2 of each bit representation B'2 of the shifted nodes B2 at a position after the bit representation Β'1 of the node B1 , which is at the first position of the bitstream portion DFS' corresponding to the dynamic range control frame DFS subsequent to the reference dynamic range control frame DFR, is rep-resented by a relative gain value which is equal to a difference of a gain val-ue g_B2 of the bit representation B'2 of the respective shifted node B2 and the gain value g_B1 of the bit representation Β'1 of the nodeB1, which precedes the bit representation B'2 of the respective nodeB2

According to a preferred embodiment of the invention the dynamic range control decoder 6 is configured for decoding the gain information GC0 of the bit representation C'0 of the node C0 of the subsequent dynamic range con-trol frame DFS at a first position of the bitstream portion DFS' corresponding to the dynamic range control frame DFS subsequent to the reference dynam-ic range control frame DFR after the one or more positions of the bit repre-sentations Β'1, B'2 of the one or more shifted nodes B1, B2 is represented by a relative gain value which is equal to a difference of a gain value g_C0 the bit representation C'0 of the respective node C0 and the gain value g_B2 of the bit representation B'2 of the shifted nodeB2, which precedes the bit repre-sentation C'0 of the respective node C0.

According to preferred embodiment of the invention a temporal size of the audio frames AFP, AFR, AFS is equal to a temporal size of the dynamic range control frames AFP, AFR, AFS.

According to a preferred embodiment of the invention the one or more nodes A0 ... A5; B0 ... B2; C0 of one of the dynamic range control frames DFP, DFR, DFS are selected from a uniform time grid.

According to preferred embodiment of the invention each node A0 ... A5; B0 ... B2: C0 of the one or more nodes A0 ... A5; B0 ... B2; C0 comprises slope information SA0 ... SA5; SB0 ... SB2: SC0.

According to a preferred embodiment of the invention the dynamic range control decoder 6 is configured for decoding the bit representations of the nodes A'0 ... A'5; B'0 ... B'2; C'0 using an entropy decoding technique.

In another aspect the invention provides a method for operating an audio de-coder, the method comprises the steps:

decoding an encoded audio bitstream ABS in order to reproduce an audio signal AS comprising consecutive audio frames AFP, AFR, AFS;

decoding an encoded dynamic range control bitstream DBS in order to re-produce an dynamic range control sequence DS corresponding to the audio signal AS and comprising consecutive dynamic range control frames DFP, DFR, DFS;

wherein the encoded dynamic range control bitstream DBS comprises for each dynamic range control frame DFP, DFR, DFS of the dynamic range control frames a corresponding bitstream portion DFP', DFR', DFS';

wherein the encoded dynamic range control bitstream DBS comprises bit representations A'0 ... A'5; B'0 ... B'2; C'0 of nodes A0 ... A5; B0 ... B2; C0, wherein each bit representation of one node of the nodes comprises gain information GA0 ... GA5; GB0 ... GB2; GC0 for the audio signal AS and time information TA0 ... TA5; TB0 ... TB2; TC0 indicating to which point in time the gain information GA0 ... GA5; GB0 ... GB2; GC0 corresponds;

wherein the encoded dynamic range control bit stream DBS comprises bit representations B'1, B'2 of shifted nodes B1, B2 selected from the nodes B0 ... B2 of one reference dynamic range control frame DFR of the dynamic range control frames DFP, DFR, DFS, which are embedded in a bitstream portion corresponding to the dynamic range control frame DFS subsequent to the one reference dynamic range control frame DFR, wherein the bit representa-tion B'0 of each remaining node B0 of the nodes B0 ... B2 of the one reference dynamic range control frame DFR of the dynamic range control frames DFP, DFR, DFS is embedded into the bitstream portion DFR' corresponding to the one reference dynamic range control frame DFR; and

wherein the bit representation Β'0 of each remaining node B0 of the remaining nodes B'0 of the one reference dynamic range control frame DFR of the dy-namic range control frames DFP, DFR, DFS is decoded in order to reproduce each remaining node B0 of the one reference dynamic range control frame DFR of the dynamic range control frames DFP, DFR, DFS;

wherein the bit representation Β'1, B'2 of each shifted node B1, B2 of the shift-ed nodes B1, B2 selected from the nodes B0 ... B2 of the one reference dy-namic range control frame DFR of the dynamic range control frames DFP, DFR, DFS is decoded in order to reproduce each shifted node B1, B2 of the shifted nodes B1, B2 selected from the nodes of the one reference dynamic range control frame DFR of the dynamic range control frames DFP, DFR, DFS; and

wherein the reproduced remaining nodes B0 and the reproduced shifted nodes B1, B2 are combined in order to reconstruct the reference dynamic range control frame DFR.

With respect to the decoder, the encoder and the methods of the described embodiments the following shall be mentioned:

Although some aspects have been described in the context of an apparatus, it is clear that these aspects also represent a description of the correspond-ing method, where a block or device corresponds to a method step or a fea-ture of a method step. Analogously, aspects described in the context of a method step also represent a description of a corresponding block or item or feature of a corresponding apparatus.

Depending on certain implementation requirements, embodiments of the in-vention can be implemented in hardware or in software. The implementation can be performed using a digital storage medium, for example a floppy disk, a DVD, a CD, a ROM, a PROM, an EPROM, an EEPROM or a FLASH memory, having electronically readable control signals stored thereon, which cooperate (or are capable of cooperating) with a programmable computer system such that the respective method is performed.

Some embodiments according to the invention comprise a data carrier hav-ing electronically readable control signals, which are capable of cooperating with a programmable computer system such that one of the methods de-scribed herein is performed.

Generally, embodiments of the present invention can be implemented as a computer program product with a program code, the program code being operative for performing one of the methods when the computer program product runs on a computer. The program code may for example be stored on a machine readable carrier.

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

In other words, an embodiment of the inventive method is, therefore, a com-puter program having a program code for performing one of the methods de-scribed herein, when the computer program runs on a computer.

A further embodiment of the inventive methods is, therefore, a data carrier (or a digital storage medium, or a computer-readable medium) comprising, rec-orded thereon, the computer program for performing one of the methods de-scribed herein.

A further embodiment of the inventive method is, therefore, a data stream or a sequence of signals representing the computer program for performing one of the methods described herein. The data stream or the sequence of signals may be configured, for example, to be transferred via a data communication connection, for example via the Internet.

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

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

In some embodiments, a programmable logic device (for example a field pro-grammable gate array) may be used to perform some or all of the functionali-ties of the methods described herein. In some embodiments, a field pro-grammable gate array may cooperate with a microprocessor in order to per-form one of the methods described herein. Generally, the methods are ad-vantageously performed by any hardware apparatus.

While this invention has been described in terms of several embodiments, there are alterations, permutations, and equivalents which fall within the scope of this invention. It should also be noted that there are many alterna-tive ways of implementing the methods and compositions of the present in-vention. It is therefore intended that the following appended claims be inter-preted as including all such alterations, permutations and equivalents as fall within the true spirit and scope of the present invention.

Reference signs:

1 audio encoder device

2 audio encoder

3 dynamic range control encoder

4 audio decoder device

5 audio decoder

6 dynamic range control decoder

ABS encoded audio bitstream

AS audio signal

AFP preceding audio frame

AFR reference audio frame

AFS subsequent audio frame

DBS encoded dynamic range control bitstream

DS dynamic range control sequence

DFP preceding dynamic range control frame

DFR reference dynamic range control frame

DFS subsequent dynamic range control frame

A0 ... A5 nodes of the preceding dynamic range control frame

B0 ... B2 nodes of the reference dynamic range controi frame

C0 node of the subsequent dynamic range control frame

DFP' bit stream portion corresponding to the preceding dynamic range control frame

DFR' bit stream portion corresponding to the reference dynamic range control frame

DFS' bit stream portion corresponding to the subsequent dynamic range control frame

TA0 ... TA5 time information of the nodes of the preceding dynamic range control frame

TB0 ... TB2 time information of nodes of the reference dynamic range control rame

TC0 time information of node of the subsequent dynamic range control frame

t_A0 ... t_A5 time difference of the nodes of the preceding dynamic range control frame

t_B0 ... t_B2 time difference of nodes of the reference dynamic range control frame

t_C0 time difference of node of the subsequent dynamic range control frame

GA0 ... GA5 gain information of the nodes of the preceding dynamic range control frame

GB0 ... GB2 gain information of nodes of the reference dynamic range control frame

GC0 gain information of node of the subsequent dynamic range contral frame

g_A0 ... g_A5 gain value of the nodes of the preceding dynamic range control frame

9_B0... g_B2 gain value of nodes of the reference dynamic range control frame

g_C0 gain value of node of the subsequent dynamic range control frame

SA0 ... SA5 slope information of the nodes of the preceding dynamic range control frame

SB0 ... SB2 slope information of nodes of the reference dynamic range control frame

SC0 slope information of node of the subsequent dynamic range control frame

References:

[1] D. Giannoulis, M. Massberg, J. D. Reiss, "Digital Dynamic Range Compressor Design - A Tutorial and Analysis" J. Audio Engineering Society, Vol. 60, No. 6, June 2012. in

Claims

1. Audio encoder device comprising;

an audio encoder (2) configured for producing an encoded audio bit- stream (ABS) from an audio signal (AS) comprising consecutive audio frames (AFP, AFR, AFS);

a dynamic range control encoder (3) configured for producing an encoded dynamic range control bitstream (DBS) from an dynamic range control sequence (DS) corresponding to the audio signal (AS) and comprising consecutive dynamic range control frames (DFP, DFR, DFS), wherein each dynamic range control frame (DFP, DFR, DFS) of the dynamic range control frames (DFP, DFR, DFS) comprises one or more nodes (A0 ... A5; B0 , .. B2; C0), wherein each node of the one or more nodes (A0 ... A5; B0 ... B2; C0) comprises gain information (GA0 ... GA5; GB0 ... GB2; GC0) for the audio signal (AS) and time information (TA0 ... TA5; TB0 ... TB2; TC0) indicating to which point in time the gain information (GA0 ... GA5; GB0 ... GB2; GC0) corresponds;

wherein the dynamic range control encoder (3) is configured in such way that the encoded dynamic range control bitstream (DBS) comprises for each dynamic range control frame (DFP, DFR, DFS) of the dynamic range control frames (DFP, DFR, DFS) a corresponding bitstream portion (DFP', DFR', DFS');

wherein the dynamic range control encoder (2) is configured for executing a shift procedure, wherein one or more nodes (B1, B2) of the nodes (B0 ... B2) of one reference dynamic range control frame (DFR) of the dynamic range control frames (DFP, DFR, DFS) are selected as shifted nodes (Bi. B2), wherein a bit representation (B'1, B'2) of each of the one or more shifted nodes (B1, B2) of the one reference dynamic range control frame (DFR) is embedded in the bitstream portion (DFS') corresponding to the dynamic range control frame (DFS) subsequent to the one reference dy- namic range control frame (DFR), wherein a bit representation (B'0) of each remaining node (B0) of the nodes (B0 ... B2) of the one reference dy- namic range control frame (DFR) of the dynamic range control frames (DFP, DFR, DFS) is embedded into the bitstream portion (DFR') corre- sponding to the one reference dynamic range control frame (DFR),

2. Audio encoder device according to claim 1, wherein the shift procedure is initiated in case that a number of the nodes of the reference dynamic range control frame is greater than a predefined threshold value.

3. Audio encoder device according to claim 1, wherein the shift procedure is initiated in case that a sum of a number of the nodes of the reference dy- namic range control frame and a number of shifted nodes from the dy- namic range control frame preceding the reference dynamic range control frame to be embedded in the bitstream portion corresponding to the ref- erence dynamic range control frame is greater than a predefined thresh- old value.

4. Audio encoder device according to claim 1, wherein the shift procedure is initiated in case that a sum of a number of the nodes (B0 ... B2) of the ref- erence dynamic range control frame (DFR) and a number of shifted nodes (A4, A5) from the dynamic range control frame (DFP) preceding the reference dynamic range control frame (DFR) to be embedded in the bit- stream portion (DFR') corresponding to the reference dynamic range con- trol frame (DFR) is greater than a number of the nodes (C0) of the dynam- ic range control frame (DFS) subsequent to the reference dynamic range control frame (DFR).

5. Audio encoder device according to one of the preceding claims, wherein the time information (TA0 ... TA5; TB0 ... TB2; TC0) of the one or more nodes (A0 ... A5; B0 ... B2; C0) is represented in such way that the one or more shifted nodes (A4, A5; B1 B2) may be identified by using the time in- formation (TA4. TA5; TB1 TB2).

6. Audio encoder device according to the preceding claim, wherein the time information (TA4, TA5; TB1, TB2) of the one or more shifted nodes (A4, A5; Β1, B2) is represented by a sum of a time difference (t_A4, t_A5; t_B1, t_B2) from a beginning of the dynamic range control frame (DFP; DFR) to which the respective node (A4, A5; B1 B2) belongs to the temporal position of the respective node (A4, A5; B1, B2) within the dynamic range control frame (DFP; DFR) to which the respective node (A4, A5; B1, B2) belongs and an offset value (drcFrameSize) being greater than or equal to a temporal size of the dynamic range control frame (DFR; DFS) subsequent to the re- spective dynamic range control frame (DFP; DFR).

7. Audio encoder device according to one of the preceding claims, wherein the gain information (GB1) of the bit representation (Β'1) of the shifted node (B1), which is at a first position of the bitstream portion (DFS') corre- sponding to the dynamic range control frame (DFS) subsequent to the reference dynamic range control frame (DFR), is represented by an abso- lute gain value (g_B1) and wherein the gain information (GB2) of each bit representation (B'2) of the shifted nodes (B2) at a position after the bit rep- resentation (B'1) of the node (B1), which is at the first position of the bit- stream portion (DFS') corresponding to the dynamic range control frame (DFS) subsequent to the reference dynamic range control frame (DFR), is represented by a relative gain value which is equal to a difference of a gain value (g_B2) of the bit representation (B'2) of the respective shifted node (B2) and a gain value (g_B1) of the bit representation (Β'1) of the node (B1), which precedes the bit representation (B'2) of the respective node (B2).

8. Audio encoder device according to one of the preceding claims, wherein, in case that the bit representations (Β'1, B'2) of one or more shifted nodes (B1, B2) of the reference dynamic range control frame (DFR) is embedded in the bitstream portion (DFS') corresponding to the dynamic range con- trol frame (DFS) subsequent to the reference dynamic range control frame (DFR), the gain information (GC0) of the bit representation (C'0) of the node (C0) of the subsequent dynamic range control frame (DFS) at a first position of the bitstream portion (DFS') corresponding to the dynamic range control frame (DFS) subsequent to the reference dynamic range control frame (DFR) after the one or more positions of the bit representa- tions (B'1, B'2) of the one or more shifted nodes (B1, B2) is represented by a relative gain value which is equal to a difference of a gain value (g_C0) of the bit representation (C'0) of the respective node (C0) and a gain value (g_B2) of the bit representation (B'2) of the shifted node (B2), which pre- cedes the bit representation (C'0) of the respective node (C0).

9. Audio encoder device according to one of the preceding claims, wherein a temporal size of the audio frames (AFP, AFR, AFS) is equal to a temporal size of the dynamic range control frames (DFP, DFR, DFS).

10. Audio encoder device according to one of the preceding claims, wherein the one or more nodes (A0 ... A5; B0 ... B2; C0) of one of the dynamic range control frame (DFP, DFR, DFS) are selected from a uniform time grid.

11. Audio encoder device according to one of the preceding claims, wherein each node (A0 ... A5; B0 ... B2; C0) of the one or more nodes comprises (A0 ... A5; B0 ... B2; C0) slope information (SA0 ... SA5; SB0 ... SB2 SC0).

12. Audio encoder device according to one of the preceding claims, wherein the dynamic range control encoder (3) is configured for encoding the nodes (A0 ... A5; B0 ... B2; C0) using an entropy encoding technique.

13. Audio decoder device comprising:

an audio decoder (5) configured for decoding an encoded audio bitstream (ABS) in order to reproduce an audio signal (AS) comprising consecutive audio frames (AFP, AFR, AFS);

a dynamic range control decoder (6) configured for decoding an encoded dynamic range control bitstream (DBS) in order to reproduce an dynamic range control sequence (DS) corresponding to the audio signal (AS) and comprising consecutive dynamic range control frames (DFP, DFR, DFS);

wherein the encoded dynamic range control bitstream (DBS) comprises for each dynamic range control frame (DFP, DFR, DFS) of the dynamic range control frames a corresponding bitstream portion (DFP', DFR', DFS');

wherein the encoded dynamic range control bitstream (DBS) comprises bit representations (A'0 ... A'5; B'0 ... B'2; C'0) of nodes (A0 ... A5; B0 ... B2; C0), wherein each bit representation of one node of the nodes comprises gain information (GA0 ... GA5; GB0 ... GB2; GC0) for the audio signal (AS) and time information (TA0 ... TA5; TB0 ... TB2; TC0) indicating to which point in time the gain information (GA0 ... GA5; GB0 ... GB2; GC0) corre- sponds;

wherein the encoded dynamic range control bit stream (DBS) comprises bit representations (Β'1, B'2) of shifted nodes (B1, B2) selected from the nodes (B0 ... B2) of one reference dynamic range control frame (DFR) of the dynamic range control frames (DFP, DFR, DFS), which are embed- ded in a bitstream portion corresponding to the dynamic range control frame (DFS) subsequent to the one reference dynamic range control frame (DFR), wherein the bit representation (ΒΌ) of each remaining node

(B0) of the nodes (B0 ... B2) of the one reference dynamic range control frame (DFR) of the dynamic range control frames (DFP, DFR, DFS) is embedded into the bitstream portion (DFR') corresponding to the one ref- erence dynamic range control frame (DFR); and

wherein the dynamic range control decoder (6) is configured for decoding the bit representation (B'0) of each remaining node (B0) of the remaining nodes (B'0) of the one reference dynamic range control frame (DFR) of the dynamic range control frames (DFP, DFR, DFS) in order to reproduce each remaining node (B0) of the one reference dynamic range control frame (DFR) of the dynamic range control frames (DFP, DFR, DFS), for decoding the bit representation (Β'1, B'2) of each shifted node (B1, B2) of the shifted nodes (B1, B2) selected from the nodes (B0 ... B2) of the one reference dynamic range control frame (DFR) of the dynamic range con- trol frames (DFP, DFR, DFS) in order to reproduce each shifted node (B1, B2) of the shifted nodes (B1, B2) selected from the nodes of the one refer- ence dynamic range control frame (DFR) of the dynamic range control frames (DFP, DFR, DFS) and for combining the reproduced remaining nodes (B0) and the reproduced shifted nodes (B1, B2) in order to recon- struct the reference dynamic range control frame (DFR).

14. Audio decoder device according to claim 13, wherein the dynamic range control decoder (6) is configured for identifying the one or more shifted nodes (A4, A5; B1, B2) by using the time information (TA4, TA5; TB1, TB2).

15. Audio decoder device according to claims 13 or 14, wherein the dynamic range control decoder (6) is configured for decoding the time information (TA4, TA5; TB1 TB2) of the one or more shifted nodes (A4, A5; B1 B2), which is represented by a sum of a time difference (t_A4, t_A5; t_B1, t_ B2) from a beginning of the dynamic range control frame (DFP; DFR) to which the respective node (A4, A5; B1, B2) belongs to the temporal position of the respective node (A4, A5; B1, B2) within the dynamic range control frame (DFP; DFR) to which the respective node (A4, A5; B1, B2) belongs and an offset value (drcFrameSize) being greater than or equal to a temporal size of the dynamic range control frame (DFR; DFS) subsequent to the re- spective dynamic range control frame (DFP; DFR).

16. Audio decoder device according to one of the claims 13 to 15, wherein the dynamic range control decoder (8) is configured for decoding the gain information (GB1) of the bit representation (Β'1) of the shifted node (B1), which is at a first position of the bitstream portion (DFS') corresponding to the dynamic range control frame (DFS) subsequent to the reference dy- namic range control frame (DFR), is represented by an absolute gain val- ue (g_B1) and wherein the gain information (GB2) of each bit representa- tion (B'2) of the shifted nodes (B2) at a position after the bit representation (B'1) of the node (B1), which is at the first position of the bitstream portion (DFS') corresponding to the dynamic range control frame (DFS) subse- quent to the reference dynamic range control frame (DFR), is represented by a relative gain value which is equal to a difference of a gain value (g_B2) of the bit representation B'2 of the respective shifted node B2 and a gain value (g_B1) of the bit representation (Β'1) of the node (B1), which precedes the bit representation (B'2) of the respective node (B2)

17. Audio decoder device according to one of the claims 13 to 16, wherein the dynamic range control decoder (6) is configured for decoding the gain information (GC0) of the bit representation (C'0) of the node (C0) of the subsequent dynamic range control frame (DFS) at a first position of the bitstream portion (DFS') corresponding to the dynamic range control frame (DFS) subsequent to the reference dynamic range control frame (DFR) after the one or more positions of the bit representations (Β'1, B'2) of the one or more shifted nodes (B1, B2) is represented by a relative gain value which is equal to a difference of a gain value (g_C0) of the bit repre- sentation (C'0) of the respective node (C0) and a gain value (g_B2) of the bit representation (B'2) of the shifted node (B2), which precedes the bit representation (C'0) of the respective node (C0).

18. Audio decoder device according to one of the claims 13 to 17, wherein a temporal size of the audio frames (AFP, AFR, AFS) is equal to a temporal size of the dynamic range control frames (AFP, AFR, AFS).

19. Audio decoder device according to one of the claims 13 to 18, wherein the one or more nodes (A0 ... A5; B0 ... B2, C0) of one of the dynamic range control frames (DFP, DFR, DFS) are selected from a uniform time grid.

20. Audio decoder device according to one of the claims 13 to 19, wherein each node (A0 ... A5; B0 ... B2; C0) of the one or more nodes (A0 ... A5; B0 ... B2; C0) comprises slope information (SA0 ... SA5; SB0 ... SB2; SC0).

21. Audio decoder device according to one of the claims 13 to 20, wherein the dynamic range control decoder (6) is configured for decoding the bit representations of the nodes (A'0 ... A'5; B'0 ... B'2: C'0) using an entropy decoding technique.

22. System comprising an audio encoder device (1) according to one of the claims 1 to 12 and an audio decoder device (4) according to one of the claims 13 to 21.

23. Method for operating an audio encoder, the method comprises the steps.

producing an encoded audio bitstream (ABS) from an audio signal (AS) comprising consecutive audio frames (AFP, AFR, AFS);

producing an encoded dynamic range control bitstream (DBS) from an dynamic range control sequence (DS) corresponding to the audio signal

(AS) and comprising consecutive dynamic range control frames (DFP.

DFR, DFS), wherein each dynamic range control frame (DFP, DFR, DFS) of the dynamic range control frames (DFP, DFR, DFS) comprises one or more nodes (A0 ... A5; B0 ... B2; C0), wherein each node of the one or more nodes (A0 ... A5; B0 ... B2; C0) comprises gain information (GA0 ... GA5; GB0 ... GB2; GC0) for the audio signal (AS) and time information (TA0 ... TA5; TB0 ... TB2; TC0) indicating to which point in time the gain in- formation corresponds

wherein the encoded dynamic range control bitstream (DBS) comprises for each dynamic range control frame (DFP, DFR, DFS) of the dynamic range control frames (DFP, DFR, DFS) a corresponding bitstream portion

(DFP', DFR', DFS );

executing a shift procedure, wherein one or more nodes (B1, B2) of the nodes (B0 ... B2) of one reference dynamic range control frame (DFR) of the dynamic range control frames (DFP, DFR, DFS) are selected as shift- ed nodes (B1, B2). wherein a bit representation (B'1, B'2) of each of the one or more shifted nodes (B1, B2) of the one reference dynamic range control frame (DFR) is embedded in the bitstream portion (DFS') corresponding to the dynamic range control frame (DFS) subsequent to the one refer- ence dynamic range control frame (DFR), wherein a bit representation (B'0) of each remaining node (B0) of the nodes (B0 ... B2) of the one refer- ence dynamic range control frame (DFR) of the dynamic range control frames (DFP, DFR, DFS) is embedded into the bitstream portion (DFR') corresponding to the one reference dynamic range control frame (DFR).

24. Method for operating an audio decoder, the method comprises the steps:

decoding an encoded audio bitstream (ABS) in order to reproduce an au- dio signal (AS) comprising consecutive audio frames (AFP. AFR, AFS);

decoding an encoded dynamic range control bitstream (DBS) in order to reproduce an dynamic range control sequence (DS) corresponding to the audio signal (AS) and comprising consecutive dynamic range control frames (DFP, DFR, DFS);

wherein the encoded dynamic range control bitstream (DBS) comprises for each dynamic range control frame (DFP, DFR, DFS) of the dynamic range control frames a corresponding bitstream portion (DFP', DFR', DFS');

wherein the encoded dynamic range control bitstream (DBS) comprises bit representations (A'0 ... A'5; B'0 ... B'2; C'0) of nodes (A0 ... A5; B0 ... B2; C0), wherein each bit representation of one node of the nodes comprises gain information (GA0 ... GA5; GB0 ... GB2; GC0) for the audio signal (AS) and time information (TA0 ... TA5; TB0 ... TB2; TC0) indicating to which point in time the gain information (GA0 ... GA5; GB0 ... GB2; GC0) corre-sponds;

wherein the encoded dynamic range control bit stream (DBS) comprises bit representations (B'1, B'2) of shifted nodes (B1, B2) selected from the nodes (B0 ... B2) of one reference dynamic range control frame (DFR) of the dynamic range control frames (DFP, DFR, DFS), which are embed-ded in a bitstream portion corresponding to the dynamic range control frame (DFS) subsequent to the one reference dynamic range control frame (DFR), wherein the bit representation (ΒΌ) of each remaining node (Bo) of the nodes (B0 ... B2) of the one reference dynamic range control frame (DFR) of the dynamic range control frames (DFP, DFR, DFS) is embedded into the bitstream portion (DFR') corresponding to the one ref-erence dynamic range control frame (DFR); and

wherein the bit representation (B'0) of each remaining node (B0) of the remaining nodes (B'0) of the one reference dynamic range control frame

(DFR) of the dynamic range control frames (DFP, DFR, DFS) is decoded in order to reproduce each remaining node (B0) of the one reference dy- namic range control frame (DFR) of the dynamic range control frames (DFP, DFR, DFS):

wherein the bit representation (Β'1, B'2) of each shifted node (B1, B2) of the shifted nodes (B1 B2) selected from the nodes (B0 ... B2) of the one refer- ence dynamic range control frame (DFR) of the dynamic range control frames (DFP, DFR, DFS) is decoded in order to reproduce each shifted node (B1, B2) of the shifted nodes (B1, B2) selected from the nodes of the one reference dynamic range control frame (DFR) of the dynamic range control frames (DFP, DFR, DFS); and

wherein the reproduced remaining nodes (B0) and the reproduced shifted nodes (B1, B2) are combined in order to reconstruct the reference dynamic range control frame (DFR).

25. Computer program for, when running on a processor, executing the method according to the one of the claims 23 or 24.