PatentDe  


Dokumentenidentifikation EP1113575 18.05.2006
EP-Veröffentlichungsnummer 0001113575
Titel Signalverarbeitungsvorrichtung und Verfahren
Anmelder Matsushita Electric Industrial Co., Ltd., Kadoma, Osaka, JP
Erfinder Norimatsu, Takeshi, Kobe City, Hyougo Pref. 651-1301, JP;
Tsushima, Mineo, Katano City, Osaka Pref. 576-0021, JP;
Ishikawa, Tomokazu, Toyanaka City, Osaka Pref. 561-0828, JP;
Sawada, Yoshiaki, Moriguchi City, Osaka Pref. 570-0046, JP;
Miyasaka, Shuji, Neyagawa City, Osaka Pref. 572-0004, JP
Vertreter derzeit kein Vertreter bestellt
DE-Aktenzeichen 60027122
Vertragsstaaten DE, FR, GB
Sprache des Dokument EN
EP-Anmeldetag 12.12.2000
EP-Aktenzeichen 001271659
EP-Offenlegungsdatum 04.07.2001
EP date of grant 05.04.2006
Veröffentlichungstag im Patentblatt 18.05.2006
IPC-Hauptklasse H03H 17/02(2006.01)A, F, I, 20051017, B, H, EP
IPC-Nebenklasse G06F 9/46(2006.01)A, L, I, 20051017, B, H, EP   G06F 9/38(2006.01)A, L, I, 20051017, B, H, EP   G10L 19/14(2006.01)A, L, I, 20051017, B, H, EP   

Beschreibung[en]

The present invention relates to a signal processing devices defined by claims 1, 10 and a signal processing suitable for methods defined by claims 9, 18 performing a compressing/decoding process for an audio signal.

Recently, for the sake of recording, downloading, and the like of an audio signal, techniques for a compressing/decoding process for a number of audio signals have been developed rapidly. As these types of compressing/decoding process techniques, Layer 3 of MPEG/AUDIO (MP3), Advanced Audio Coding (AAC), and the like are known. Any of these employs technologies, such as subband encoding, MDCT, quantization, Huffman encoding, and the like, as elementary technologies.

The document by V.Z Mesasovic et al. "DTS Multichannel audio decoder on a 24-bit fixed - point dual-DSP architecture", AES 106th convention preprint # 4964, 8 may 1999 discloses an example of tasks division in an audio decorder. It considers underflow and overflow of data but does not practicle a solution for synchronizing tasks of different opérations.

The present invention is to provide a signal processing device in which the power consumption can be reduced effectively even when there is a deviation in the throughputs of plural processes when a low power consumption is sought by making the processing of an audio signal parallel and pipelined.

The present invention comprises a signal processing device having first to Nth signal processing sub sections and a signal processing main section and converting a first digital signal to a second digital signal by performing a second process after performing a first process for each frame signal of the first digital signal which is framed for each predetermined time interval.

Here, the first to Nth signal processing sub sections can be those to which a (N × t + i)th frame signal (i and t are integers, N is a natural number, and 0 ≤ i < N) of the first digital signal is given, wherein each of the sections completes the first process within a time period (N × T) (T is a real number). The signal processing main section can be the one which completes the second process within a time period T for the signal processed in the (i + 1)th signal processing sub section.

The signal processing main section may be the one to which a (N × t + i)th frame signal (i and t are integers, 0 ≤ t, and 0 ≤ i < N) of the first digital signal is given and which completes the first process within a time period T (T is a real number). The first to Nth signal processing sub sections may be those to which the (i + 1)th frame signals are given after processed in the signal processing main section, respectively, and each of which completes the second process within the time period (N × T) (N is a natural number).

The signal processing device and method have the advantage that the power consumption is reduced by making the processing parallel.

Preferred embodiments of the invention will now be described in detail in conjunction with the accompanying drawings in which

  • FIG. 1 is a view showing a flow of signal processing in the case where parallel processing in an audio signal processing device is not performed;
  • FIG. 2 is a view showing a flow of signal processing in the case where parallel processing in an audio signal processing device is performed;
  • FIG. 3 is a view showing a flow of signal processing in the case where parallel processing is not performed when there is a deviation in the throughputs of processes A and B in an audio signal processing device;
  • FIG. 4 is a view showing a problem in the case where parallel processing is performed in an audio signal processing device;
  • FIG. 5 is a block diagram showing a configuration of a signal processing device of Embodiment 1 of the present invention;
  • FIG. 6 is a block diagram showing a signal processing device performing audio signal decoding in accordance with a concrete example of Embodiment 1;
  • FIG. 7 is a block diagram showing a signal processing device performing audio signal decoding in accordance with another concrete example of Embodiment 1;
  • FIG. 8 is a time chart showing a flow of signal processing of Embodiment 1 of the present invention in the order of time;
  • FIG. 9 is a block diagram showing a configuration of a signal processing device of Embodiment 2 of the present invention;
  • FIG. 10 is a time chart showing a flow of signal processing of Embodiment 2 of the present invention in the order of time;
  • FIG. 11 is a block diagram showing a configuration of a signal processing device of Embodiment 3 of the present invention;
  • FIG. 12 is a block diagram showing a signal processing device performing audio signal encoding in accordance with a concrete example of Embodiment 3;
  • FIG. 13 is a block diagram showing a signal processing device performing audio signal encoding in accordance with another concrete example of Embodiment 3;
  • FIG. 14 is a block diagram showing a flow of signal processing of Embodiment 3 of the present invention in the order of time;
  • FIG. 15 is a block diagram showing a configuration of a signal processing device of Embodiment 4 of the present invention;
  • FIG. 16 is a time chart showing a flow of signal processing of Embodiment 4 of the present invention in the order of time; and
  • FIG. 17 is an outlined view showing one example of a portable type apparatus containing a signal processing device in accordance with the present invention.

FIG. 1 shows a flow of processing in the case where signal processing composed of a first process A and a second process B is performed without being made parallel in a signal processing device. FIG. 2 shows a flow of the process A and the process B in the case where the processing shown in FIG. 1 is made parallel. In these examples, one frame period is T. When the processing is not made parallel, first, the first process A [1] is performed for an input frame signal [1] as shown in FIG. 1, and then the second process B [1] is performed to generate an output frame signal [1]. In the next frame period, the process A [2] is performed for an input frame signal [2], and then the process B [2] is performed to generate an output frame signal [2]. Thus, the processing of the process A and the process B together is completed within the one frame period T.

When processing is made parallel, two processing devices A, B are employed. The process A [1] is performed for the input frame signal [1] in the processing device A, taking the period T, as shown in FIG. 2. In the next frame period, The process A [2] is performed for the input frame signal [2] in the processing device A, taking the period T. In this frame period, the process B [1] is performed for the signal being after the process A [1] is finished in the processing device B, taking the period T. This type of processing is repeated for each frame period so that the processing of the process A and the process B is made parallel.

With this parallel processing, although both the process A and the process B together have to be completed within the period T originally as shown in FIG. 1, it becomes possible that each one of the process A and the process B is completed within the period T, whereby 1/2 of calculation capability becomes sufficient. That is, the operation frequency for the processing can be made 1/2, and thus the power consumption is reduced.

If there is a deviation in the throughputs of the process A and the.process B, it is necessary to complete both the processes A and B within one frame period T unless the process A and the process B are made parallel as shown in FIG. 3. FIG. 4 shows a flow of the processing in the case where parallel processing is employed when the processing amount of the process A is double the processing amount of the process B. In this case, first, the process A [1] is performed for the input frame signal [1] in the processing device A, taking the period T, as shown in FIG. 4. In the next frame period, the process A [2] is performed for the input frame signal [2] in the processing device A, taking the period T. In this frame period, the process B [1] is performed for the signal which exists after the process A [1] is finished in the processing device B, taking the period T/2.

Although the process A and the process B are performed with parallel as described above, in this type of parallel processing, the processing in which the process A is originally completed within a period 2 × T / 3 merely comes to be allowed to be completed within the period T at the most. With respect to the process B, the process is completed in a sufficiently short period, compared with the given period T. Therefore, dead time in which no processing is performed is generated in the processing device B, and thus reduction in power consumption cannot be executed efficiently even by the parallel processing. The inventors are to find out and solve the problem in the case where signal processing is made parallel in audio signal processing.

Embodiment 1

In the present embodiment, explained is a signal processing device which is intended for an audio signal and performs a first process A and then a second process B for a frame input signal that is a first digital signal of an input framed for each time period T (T is a real number) to generate a frame signal that is a second digital signal of an output. Here, the embodiment is explained in the setting that the processing amount of the process A is N times (N is a natural number, here, N=3) the processing amount of the process B.

FIG. 5 shows a configuration of a signal processing device according to the present embodiment. This signal processing device has one signal processing main section 10 and first to third three signal processing sub sections 11 to 13. A distribution section 14 distributes a frame signal of the input to either one of the signal processing sub sections 11 to 13 in accordance with a frame number. A selection section 15 selects either one of the signal processing sub sections 11 to 13 in accordance with the frame number to send it to the signal processing main section 10. A frame number management section 16 updates the frame number each time one frame period T elapses to give the number to the distribution section 14 and the selection section 15. Here, it is supposed that the respective signal processing sub sections 11 to 13 have the capabilities to process the first process A within the time period 3 × T, and the signal processing main section 10 has the capability to process the process B within the time period T.

Here, since the first process A has to start the process of the next frame time before the processing of the past frame time is completed, it is necessary that the process A is a non-chain process, that is, a process in which the information generated in the past frame time is not employed. Conversely, the second process B can be a process in which the information generated in the past frame time is employed, that is, a chain process. This is because the processing of the next frame time is always started after the processing of the past frame time is completed.

The first digital signal may, for example, be a compressed and encoded signal of an audio signal, and the second digital signal may be a PCM signal of an audio signal. The first process may contain a process picking out information from that compressed and encoded signal and converting that information into the information of a frequency spectrum, and second process may contain a process converting the information of that frequency spectrum into a time base PCM signal.

FIG. 6 is a block diagram showing a signal processing device performing an audio signal decoding process that is a concrete example of the present signal processing device. This audio signal decoding processing device has first to third Huffman decoding sections 111, 121, and 131 as the signal processing sub sections 11 to 13 performing the first process A. The Huffman decoding process is a decoding process of a variable length code for decoding encoded information for each frame from an encoded bit stream of the input. It is supposed that the signal processing main section 10 performing the second process B is an inverse MDCT processing section 101. The inverse MDCT process is a process performing an inverse MDCT process for a signal inversely quantized.

Another example of the audio decoding processing device is shown in FIG. 7. It can be set that the signal processing sub sections 11 to 13 are inverse quantizing sections 112, 122, 132 inversely quantizing encoded informing and that the signal processing main section 10 is a sub-band synthesis filter bank processing section 102.

FIG. 8 is a view showing a flow of the processing of the signal processing device of the present embodiment in the order of time. Operations of the signal processing device will be explained below. First, the frame number management section 16 outputs a frame number incremented for each frame period to the distribution section 14 and the selection section 15. The distribution section 14 sends the frame signal to the (i+1)th signal processing sub section when the frame number is (N × t + i) (t and i are integers, and 0 ≤ i < N). In this case, N is 3. As shown in FIG. 8, each frame signal is distributed to a predetermined signal processing sub section one after another.

The 0th frame signal is sent to the first signal processing sub section 11.

The first frame signal is sent to the second signal processing sub section 12.

The second frame signal is sent to the third signal processing sub section 13.

The third frame signal is sent to the first signal processing sub section 11.

The fourth frame signal is sent to the second signal processing sub section 12.

The fifth frame signal is sent to the third signal processing sub section 13.

In the respective signal processing sub sections 11 to 13, the first process A is executed for the frame signal distributed as described above within the time period 3T.

The selection section 15 then inputs the signal for which the first process A is performed from either one of the signal processing sub sections 11 to 13 and sends the signal to the signal processing main section 10. Here, when the frame number shown by the frame number management section 16 is (N × t + i), the signal outputted from the (i + 1)th signal processing sub section is sent to the signal processing main section 10. The signal that the signal processing main section 10 receives at this time is the signal which the process A is performed for the signal inputted to the (i+1)th signal processing sub section at the time of the (N × (t - 1) + i)th frame period. The signal processing main section 10 executes the second process B within the period T for the received signal for which the process A has been completed.

As shown in FIG. 8, in the 0th frame period, the 0th frame signal is inputted to the first sub signal processing sub section 11, and for this signal the process A [0] is started so that this process is completed within the period 3T.

In the first frame time, the first frame signal is inputted to the second signal processing sub section 12, and for this signal the process A [1] is started so that this process is completed within the period 3T.

In the second frame time, the second frame signal is inputted to the third signal processing sub section 13, and for this signal the process A [2] is started so that this process is completed within the period 3T.

In the third frame time, the third frame signal is inputted to the first signal processing sub section 11, and for this signal the process A [3] is started. At the same time in the signal processing main section 10, the process B [0] is started for the output signal from the first signal processing sub section 11 so that this process is completed within the period T.

By repeating this type of processing for each frame time one after another, the signal processing composed of the process A and the process B is performed for the frame signals inputted at the time T intervals so that the output frame signals are generated at the time T intervals. At this time, as is obvious from FIG. 8, in the signal processing main section 10 and the signal processing sub sections 11 to 13, the processing can be made parallel without dead time.

Since the processing employing the information generated in the past frame time is excluded in the first process A, the need to deliver a signal between the respective signal processing sub sections is eliminated, whereby making the processing parallel can be efficiently performed.

In the present embodiment, since it is set that the processing amount of the process A is three times the processing amount of the process 8, three signal processing sub sections are provided. When the processing amount of the process A is N times (N is a natural number) the processing amount of the process B, if N signal processing sub sections are provided, efficient parallel processing can be performed.

As described above, in the present embodiment, when the processing composed of the process A and the process B is parallel processed, even if there is a deviation in throughputs of the process A and the process B, parallel processing can be performed efficiently.

Embodiment 2

In the present embodiment, explained is a signal processing device that is intended for an audio signal and performs a first process A and then a second process B for an input frame signal that is framed for each time period T to generate an output frame signal. The present embodiment is explained in the setting that the processing amount of the process A is N times (here, N=2) the processing amount of the process B.

FIG. 9 is a block diagram showing a configuration of the signal processing device according to the present embodiment. The signal processing device has one signal processing main section 30 and first and second two signal processing sub sections 31 and 32. A distribution and selection section 33 distributing and selecting to output each frame signal is provided between the signal processing main section 30 and the signal processing sub sections 31 and 32. A frame number management section 34 updates the frame number each time one frame period T elapses to output it to the distribution and selection section 33. A first memory 35 is a memory storing an input frame signal one after another, and a second memory 36 is a memory storing an output frame signal one after another. Here, it is supposed that the signal processing sub sections 31 and 32 have the capabilities to process the first process A within the time period (2 × T), and the signal processing man section 30 has the capability to process the second process B within the time period T.

Here, similar to Embodiment 1, since the first process A has to start the process of the next frame time before the process of the past frame time is completed, it is necessary that the process A is the non-chain process. Conversely, the second process B can be the chain process. This is because the processing of the next frame time is always started after the process of the past frame time is completed. Therefore, similar to Embodiment 1 described above, it is possible to construct a signal processing device for decoding an audio signal, setting that the signal processing main section 30 performs an inverse MDCT process and the signal processing sub sections 31 and 32 perform a Huffman encoding process. It is possible to select the inverse quantize process as the first process A performed in the signal processing sub sections 31 and 32 and a sub-band synthesis filter bank process as the process B performed in the signal processing main section 30.

FIG. 10 is a view showing a flow of the processing of the signal processing device in the order of time. Operations of the signal processing device will be explained below, employing FIG. 9 and FIG. 10.

The signal processing main section 30 picks out the input frame signal of its frame time from the first memory 35 to output it to the distribution and selection section 33. A frame number is given to the distribution and selection section 33 from the frame number management section 34.

In an (even number)th frame time, its frame signal is outputted to the first signal processing sub section 31. The signal processing sub section 31 executes the process A for the input frame signal transferred as described above to complete it within the period (2 × T).

Parallel to this, the signal processing main section 30 starts the process B for the signal received from the first signal processing sub section 31 to complete that process within the period T. This signal for which the process B is completed is sent to the second memory 36.

In an (odd number)th frame time, the distribution and selection section 33 receives the treated signal from the second signal processing sub section 32 to send it to the signal processing main section 30, and to send the frame signal of the frame time to the second signal processing sub section 32. The signal processing sub section 32 executes the process A for the input frame signal transferred to complete it within the period (2 × T).

Parallel to this, the signal processing main section 30 starts the process B for the signal received from the second signal processing sub section 32 to complete that process within the period T. This signal for which the process B has been completed is sent to the second memory 36.

The signal that the signal processing main section 30 receives from the signal processing sub section 31 or 32 is the signal obtained by the process that the signal processing sub section performs the first process A for the frame signal inputted to that signal processing sub section before two frame time.

By repeating this type of processing for each frame time one after another, the signal processing composed of the process A and the process B is performed for the frame signals inputted at the time T intervals so that the output frame signals are generated at the time T intervals. At this time, as is obvious from FIG. 10, in the signal processing main section 30 and the signal processing sub sections 31 and 32, the processing can be made parallel without dead time.

Although the present embodiment is explained in a setting that the processing amount of the process A is two times the processing amount of the process B, when it is N times, N signal processing sub sections can be employed to construct the device. In this case, the distribution and selection section 33 is constructed so as to send, (N × t + i)th frame signal (i and t are integers, and 0 ≤ i < N) obtained from the signal processing main section 30 for each time T to a (i + 1)th signal processing sub section, and to receive the signal for which the first process A is completed for (N × (t - 1) + i)th frame signal from the (i + 1) signal processing sub section to send it to the signal processing main section 30. By this, when the processing composed of the process A and the process B is parallel processed, even if there is a deviation in the throughputs of the process A and the process B, parallel processing can be performed efficiently.

Constructing the process A so that the process A excludes a process employing the information generated in a past frame time eliminates the necessity to deliver a signal between the respective signal processing sub sections, whereby making the processing parallel can be efficiently performed.

Embodiment 3

In the present embodiment, explained is a signal processing device that is intended for an audio signal and performs a first process A and then a second process B for an input frame signal that is framed for each time period T (T is a real number) to generate an output frame signal. Here, the present embodiment is explained in the setting that the processing amount of the process A is N times (here, N=3) the processing amount of the process B.

FIG. 11 shows a configuration of a signal processing device according to the present embodiment. The signal processing device has one signal processing main section 50 and first to third three signal processing sub sections 51 to 53. A distribution section 54 distributes an output signal from the signal processing main section 50 to either one of the signal processing sub sections 51 to 53 in accordance with the frame number. A selection section 55 selects the output signal of either one of the signal processing sub sections 51 to 53 in accordance with the frame number to output it. A frame number management section 56 updates the frame number each time one frame period T elapses to give it to the distribution section 54 and the selection section 55. Here, it is supposed that the signal processing main section 50 has the capability to process the first process A within the time period T, and the respective signal processing sub sections 51 to 53 have the capabilities to process the process B within the time period (3 × T).

Here, since with respect to the second process B, the process of the next frame time has to be started before the processing of the past frame time is completed, it is necessary that the process B is the non-chain process. Conversely, the first process A can be the chain process. This is because the processing of the next frame time is always started after the processing of the past frame time is completed.

The first digital signal may, for example, be a PCM signal of an audio signal, the second digital signal may be a compressed and encoded signal of an audio signal, the first process may contain a process converting the PCM signal into the information of a frequency spectrum, and the second process may contain a process compressing/encoding the information of the frequency spectrum.

FIG. 12 is a block diagram showing a signal processing device performing an audio signal encoding process that is a concrete example of the present signal processing device. This audio signal encoding processing device can employ an MDCT processing section 501 as the signal processing main section 50 performing the first process, and first to third Huffman encoding sections 511, 521, and 531 as the signal processing sub sections 51 to 53 performing the second process as shown in FIG. 12. The MDCT process is a process converting the PCM signal in which the input is framed into a frequency spectrum signal while overlapping with the past PCM signal. The Huffman encoding process is, that is, a variable length encoding process in which the present frame signal can be processed without employing the data generated at the time of the past frame processing.

Fig. 13 shows another example of the audio encoding processing device. It can be set that the signal processing main section 51 is a sub-band analysis filter bank processing section 502 and that the signal processing sub sections 51 to 53 are first to third quantizing sections 512, 522 and 532.

FIG. 14 is a view showing a flow of processing of the signal processing device in the order of time. Operations of this signal processing device will be explained below, employing FIG. 11.

First, in the signal processing main section 50, the process A is performed for the input frame signal within the period T.

Then, the distribution section 54 sends the output signal from the signal processing main section 50 to the (i + 1)th signal processing sub section when the frame number shown by the frame number management section 56 is (N × t + i) (t and i are integers, N is a natural number, and t ≥ 0, and 0 ≤ i < N). In this case, N=3.

That is:

  • When the frame number is 0, the output signal from the signal processing main section 50 is sent to the first signal processing sub section 51.
  • When the frame number is 1, it is sent to the second signal processing sub section 52.
  • When the frame number is 2, it is sent to the third signal processing sub section 53.
  • When the frame number is 3, it is sent to the first signal processing sub section 51.
  • When the frame number is 4, it is sent to the second signal processing sub section 52.
  • When the frame number is 5, it is sent to the third signal processing sub section 53.
  • As this, an output signal from the signal processing main section 50 is distributed to the predetermined signal processing sub section one after another.
  • In the signal processing sub sections 51 to 53, the process B is executed for the after-process A signal that is distributed as described above within the time period 3T.

The selection section 55 then inputs the signal for which the process A and the process B are performed from either one of the signal processing sub sections 51 to 53 to output this processed signal. In general, when the frame number shown by the frame number management section 56 is (N × t + i), the signal outputted from the (i + 1)th signal processing sub section is outputted. The signal outputted at this time becomes the signal which is obtained by performing the process B for the signal inputted to the (i + 1)th signal processing sub section at the (N × (t -1) + i)th frame time. Here, N=3.

FIG. 14 is a view showing a flow of signal processing in the order of time. In the 0th frame time, the 0th frame signal is inputted to the signal processing main section 50, and for this signal the process A [0] is started so that this process is completed within the period T.

In the first frame time, the first frame signal is inputted to the signal processing main section 50, and for this signal the process A [1] is started. At the same time in the first signal processing sub section 51, the process B [0] is started for the output signal for which the process A [0] is completed from the signal processing main section 50 so that this process is completed within the period (3 × T). Of course, the process A is completed within the period T.

In the second frame time, the second frame signal is inputted to the signal processing main section 50, and for this signal the process A [2] is started. At the same time in the second signal processing section 52, the process B [1] is started for the output signal for which the process A [1] is completed from the signal processing main section 50 so that this process is completed within the period (3 × T). Of course, the process A is completed within the period T.

In the third frame time, the third frame signal is inputted to the signal processing main section 50, and for this signal the process A [3] is started. At the same time in the third signal processing sub section 53, the process B [2] is started for the output signal for which the process A [2] is completed from the signal processing main section 50 so that this process is completed within the period (3 × T): Of course, the process A is completed within the period T.

In the forth frame time, the fourth frame signal is inputted to the signal processing main section 50, and for this signal the process A [4] is started. At the same time in the first signal processing sub section 51, the process B [3] is started for the output signal for which the process A [3] is completed from the signal processing main section 50 so that this process is completed within the period (3 × T). Of course, the process A is completed within the period T.

By repeating this type of processing for each frame time one after another, the signal processing composed of the process A and the process B is performed for the frame signals inputted at the time T intervals so that the output frame signals are generated at the time T intervals. At this time, as is obvious from FIG. 14, in the signal processing main section 50 and the signal processing sub sections 51 to 53, the processing can be made parallel without dead time.

Constructing the process B so that the process excludes a process processing employing the information generated in the past frame time eliminates the necessity to deliver a signal between the respective signal processing sub sections, whereby making the processing parallel can be efficiently performed.

In the present embodiment, since it is set that the processing amount of the process B is three times the processing amount of the process A, three signal processing sub sections are provided. When the processing amount of the process B is N times (N is a natural number) the processing amount of the process A, if N signal processing sub sections are provided, efficient parallel processing can be performed.

As described above, in the present embodiment, when the processing composed of the process A and the process B is parallel processed, even if there is a deviation in throughputs of the process A and the process B, parallel processing can be performed efficiently.

Embodiment 4

In the present embodiment, explained is a signal processing device that is intended for an audio signal and performs a first process A and then a second process B for an input frame signal that is framed for each time period T to generate an output frame signal. The present embodiment is explained in the setting that the processing amount of the process B is N times (here, N=2) the processing amount of the process A.

FIG. 15 is a block diagram showing a configuration of the signal processing device according to the present embodiment. The signal processing device has one signal processing main section 70 and first and second two signal processing sub sections 71 and 72. A distribution and selection section 73 distributing and selecting to output each frame signal is provided between the signal processing main section 70 and the signal processing sub sections 71 and 72. A frame number management section 74 updates a frame number each time one frame period T elapses to output it to the distribution and selection section 73. A first memory 75 is a memory storing an input frame signal one after another, and a second memory 76 is a memory storing an output frame signal one after another. Here, it is supposed that the signal processing sub sections 71 and 72 have the capabilities to process the first process B within the time period (2 × T), and the signal processing main section 70 has the capability to process the process A within the time period T.

Here, similar to Embodiment 3, the first process A can be a process in which the information generated in the past frame time is employed, that is, the chain process. This is because the process of the next frame time is always started after the processing of the past frame time is completed. Conversely, since the second process B has to start the process of the next frame time before the process of the past frame time is completed, it is necessary that the process B is the non-chain process. Therefore, similar to Embodiment 3 described above, it is possible to construct an audio coding processing device performing an MDCT process as the signal processing main section 70 and the Huffman encoding process as the signal processing sub sections 31 and 32. It is possible to select a sub-band analysis filter bank process as the process A performed in the signal processing main section 70 and a quantize process as the process B performed in the signal processing sub section 30.

FIG. 16 is a view showing a flow of the processing of the signal processing device in the order of time. Operations of the signal processing device will be explained below, employing FIG. 15 and FIG. 16.

The signal processing main section 70 picks out the input frame signal of its frame time from the first memory 75 and performs the process A for this frame signal. This process A is completed within the period T. This after-process A signal is outputted to the distribution and selection section 73. A frame number has been given to the distribution and selection section 73 from the frame number management section 74, and the frame signal is outputted to the first signal processing sub section 71 in an (even number)th frame time.

Parallel to this, in the second signal processing sub section 72, the process B is started for the after-process A signal which exists before one frame time, and this process is completed within the period 2 × T. This after-process B signal is sent to the second memory 76 via the distribution and selection section 73 and the signal processing main section 70.

in an (odd number)th frame time, a frame signal of its time is inputted from the first memory 75 to the signal processing main section 70, and for this signal the process A is started so that this process is completed within the period T. This after-process A signal is sent to the second signal processing sub section 72 via the distribution and selection section 73.

Parallel to this, in the first signal processing sub section 71, the process B is started for the after-process A signal which exists before one frame time so that this process is completed within the period 2 × T. This after-process B signal is sent to the second memory 76 via the distribution and selection section 73 and the signal processing main section 70.

At this time, the signal that the signal processing main section 70 receives from the signal processing sub section 71 or 72 is the signal obtained by the process that the signal signal processing sub section performs the process B for the frame signal inputted to this signal processing sub section before two frame time.

By repeating this type of processing for each frame time one after another, the signal processing composed of the process A and the process B is performed for the frame signals inputted at the time T intervals so that the output frame signals are generated at the time T intervals. At this time, as is obvious from FIG. 16, in the signal processing main section 70 and the signal processing sub sections 71 and 72, the processing can be made parallel without dead time.

As described above, in the present embodiment, when the processing composed of the process A and the process B is parallel processed, even if there is a deviation in throughputs of the process A and the process B, parallel processing can be performed efficiently.

Constructing the process B so that the process B does not contain a process employing the information generated in a past frame time eliminates the necessity to deliver a signal between the respective signal processing sub sections, whereby making the processing parallel can be efficiently performed.

In the respective embodiments described above, it is premised that the operations of the signal processing sub sections are defined by a program stored in a commend memory and are operated by the same program. That is, in a signal processing device, a plurality of signal processing sub sections operate exactly the same. Accordingly, it is desired that the memory needed in the processing performed in the signal processing sub sections is smaller than the memory needed in the processing performed in the signal processing main section.

The need to deliver a signal between the respective signal processing sub sections is eliminated by not performing parallel processing in the process in which a process employing the information generated in a past frame time is contained and performing parallel processing in the process in which a process employing the information generated in a past frame time is not contained, whereby parallel processing and pipeline processing can be efficiently performed, and the power consumption can be reduced.

FIG. 17 is an outlined view showing one example of a portable type apparatus employing a signal processing device in accordance with either one of the embodiments described above. This portable type apparatus is, for example, an apparatus for encoding and decoding to reproduce audio data, such as MP3, AAC, or the like. This apparatus is provided with a microphone 81 as an input device, an input section 82, an output section 83 amplifying a decoded signal, a speaker 84, a memory 85, and a battery 86 in addition to that signal processing device 87 as shown in the drawing. The signal processing device 87 is constructed, including an encoding section 801 performing an encoding process of an audio signal in accordance with Embodiment 3 or 4 described above and a decoding section 802 performing a decoding process in accordance with Embodiment 1 or 2. The memory 85 is a memory holding audio data and is constructed so as to encode the data inputted thereto to write and decode the data written to output. The memory 85 can be constructed as a memory card that is small in size and is easy to attach or detach. With this, efficiently paralleling to encode and decode is possible, whereby advantageous effects can be produced wherein drastic low power consumption is achieved, and possible operating time can be prolonged by one time charge of the portable type apparatus.


Anspruch[de]
Signalverarbeitungsvorrichtung zum Umsetzen eines ersten digitalen Signals in ein zweites digitales Signal durch Durchführen eines ersten Verfahrens und dann eines zweiten Verfahrens für jedes Rahmensignal des ersten digitalen Signals, das für ein vorgegebenes Zeitintervall mit einem Rahmen versehen ist, umfassend: erste bis N-te Signalverarbeitungs- Unterabschnitte (10, 12, 13, 31, 32), wobei jeder (i + 1)-te Signalverarbeitungs- Unterabschnitt mit dem (N * t + i)-ten Rahmen des Signals, wobei i und t ganze Zahlen sind, N eine natürliche Zahl >= 2 ist und 0 ≤ i < N gilt, des ersten digitalen Signals versehen ist und jeder der selben das erste Verfahren innerhalb einer Dauer (N * T) vollständig ausführt, wobei T eine reale Zahl ist; und eine Signalverarbeitungs- Hauptstation (10, 30), die ein Signal, das in dem (i + 1)-ten Signalverarbeitungs- Unterabschnitt verarbeitet wurde, in das zweite digitale Signal umsetzt, indem das zweite Verfahren innerhalb einer Periode T vollständig ausgeführt wird. Signalverarbeitungsvorrichtung nach Anspruch 1, worin die Signalverarbeitungsvorrichtung ferner umfasst: einen Verteilerabschnitt, der das erste digitale Signal an einen der ersten bis N-ten Signalverarbeitungs- Unterabschnitte für jedes Rahmenintervall eines nach dem anderen eingibt; einen Auswahlabschnitt, der wahlweise eines der nach dem Verfahren vorhandenen Signale, die von den ersten bis N-ten Signalverarbeitungs- Unterabschnitten ausgegeben werden, für jedes Rahmenintervall Eines nach dem Anderen auszugeben, um das Signal an den Signalverarbeitungs- Hauptabschnitt einzugeben. Signalverarbeitungsvorrichtung nach Anspruch 1, worin die Signalverarbeitungsvorrichtung ferner umfasst: einen ersten Speicher, der das Rahmensignal für das erste digitale Signal Eines nach dem Anderen speichert; einen zweiten Speicher, der das Rahmensignal des zweiten digitalen Signals Eines nach dem Anderen speichert; und einen Verteiler- und Auswahlabschnitt, der das (N * t + i)-te Rahmensignal, wobei i und t ganze Zahlen sind und 0 ≤ i < N gilt, das von dem Signalverarbeitungs- Hauptabschnitt erhalten wird, an den (i + 1)-ten Signalverarbeitungs- Unterabschnitt sendet, das Signal empfängt, dass das erste Verfahren für das (N * (t -1) + i)-te Rahmensignal von dem (i + 1)-ten Signal Verarbeitungs- Unterabschnitt verarbeitet hat, und der das Signal an den Signalverarbeitungs- Hauptabschnitt ausgibt; und wobei jeder der ersten bis N-ten Signalverarbeitungs- Unterabschnitte mit dem Verteiler- und Auswahlabschnitt verbunden ist, das erste Verfahren für das von dem Verteiler- und Auswahlabschnitt empfangene Signal durchführt und das nach dem Verfahren vorhandene Signal an den Verteiler- und Auswahlabschnitt sendet; und

wobei der Signalverarbeitungs- Hauptabschnitt mit dem ersten und dem zweiten Speicher verbunden ist, das Rahmensignal von dem ersten Speicher für jedes Zeitintervall T Eins nach dem Anderen herausnimmt, um das Rahmensignal an den Verteiler- und Auswahlabschnitt auszugeben, und das zweite Verfahren für das von dem Verteiler- und Auswahlabschnitt empfangene Signal durchführt, um dieses nach dem Verfahren vorhandene Signal in dem zweiten Speicher abzuspeichern.
Signalverarbeitungsvorrichtung nach Anspruch 1, worin

das zweite Verfahren ein Verfahren enthält, das Information verwendet, die in der vergangenen Rahmenzeit erzeugt worden ist; und

das erste Verfahren ein Verfahren ausschließt, das Information verwendet, die in der vergangenen Rahmenzeit erzeugt worden ist.
Signalverarbeitungsvorrichtung nach Anspruch 4, worin

das erste digitale Signal ein komprimiertes und kodiertes Signal eines Audiosignals ist; das zweite digitale Signal ein PCM Signal eines Audiosignals ist;

das erste Verfahren ein Verfahren enthält, das Information aus dem komprimierten und kodierten Signal herausnimmt, um die Information in eine Information eines Frequenzspektrums umzusetzen; und

das zweite Verfahren ein Verfahren enthält, das die Information des Frequenzspektrums in das PCM Signal auf Zeitbasis umsetzt.
Signalverarbeitungsvorrichtung nach Anspruch 5, worin

das erste Verfahren ein Dekodierungsverfahren für einen Code variabler Länge enthält; und

das zweite Verfahren ein inverses MDCT Verfahren enthält.
Signalverarbeitungsvorrichtung nach Anspruch 5, worin

das erste Verfahren ein inverses, quantifizierendes Verfahren enthält, dass das komprimierte und kodierte Signal invers quantifiziert, und

das zweite Verfahren ein Unterbandsynthese- Filterbankverfahren enthält.
Signalverarbeitungsvorrichtung nach Anspruch 1, worin

eine Unterteilung für das erste Verfahren und das zweite Verfahren so gemacht wird, dass die Rechendauer, die für das erste Verfahren erforderlich ist, das N-fache der Rechendauer ist, die für das zweite Verfahren erforderlich ist.
Signalverarbeitungsverfahren zum Umsetzen eines ersten digitalen Signals in ein zweites digitales Signal durch Verwendung von ersten bis N-ten Signalverarbeitungs- Unterabschnitten und eines Signalverarbeitungs- Hauptabschnittes umfassend die folgenden Schritte:

Vollenden des ersten Verfahrens mit einer Periode (N * T) in jedem der (i + 1)-ten Signalverarbeitungs- Unterabschnitt für (N * t + i)-te Rahmensignale, wobei i und t ganze Zahlen sind, N eine natürliche Zahl >= 2 ist und, T eine reale Zahl ist und O ≤ i < N gilt, von dem ersten digitalen Signal Eines nach dem Anderen; und

Umsetzen eines Signals, das in dem (i + 1)-ten Signalverarbeitungs- Unterabschnitt bearbeitet wurde, in ein zweites digitales Signal durch Vollenden eines zweiten Verfahrens innerhalb einer Zeit T in dem Signalverarbeitungs- Hauptabschnitt.
Signalverarbeitungsvorrichtung zum Umsetzen eines ersten digitalen Signals in ein zweites digitales Signal durch Durchführung eines ersten Verfahrens und dann eines zweiten Verfahrens für jedes Rahmensignal des ersten digitalen Signals, das für jedes vorgegebene Zeitintervall mit einem Rahmen versehen ist, umfassend: einen Signalverarbeitungs- Hauptabschnitt (50, 70), der mit Rahmensignalen des ersten digitalen Signals beliefert wird und der das erste Verfahren innerhalb einer Periode T vollendet, wobei T eine reale Zahl ist; und erste bis N-te Signalverarbeitungs- Unterabschnitte (51, 52, 53, 71, 72), wobei der (i + 1)-te Signalverarbeitungs- Unterabschnitt mit dem (N * t + i)-ten Rahmensignal beliefert wird, das in dem Signalverarbeitungs- Hauptabschnitt verarbeitet wurde, wobei i und t ganze Zahlen sind, N eine natürliche Zahl >=2 ist und 0 <= i < N gilt, und das Rahmensignal in das zweite digitale Signal durch Vollendung des zweiten Verfahrens in einer Zeitdauer (N * T) umsetzt. Signalverarbeitungsvorrichtung nach Anspruch 10, worin die Signalverarbeitungsvorrichtung ferner umfasst: einen Verteilerabschnitt, der das digitale Rahmensignal, das von dem Signalverarbeitungs-Hauptabschnitt ausgegeben wird, in einen der ersten bis N-ten Signalverarbeitungs- Unterabschnitte für jedes Rahmenintervall Eines nach dem Anderen eingibt; und einen Auswahlabschnitt, der wahlweise eines der nach dem Verfahren vorhandenen Signale, die von einem der ersten bis N-ten Signalverarbeitungs- Unterabschnitte ausgegeben werden, für jedes Rahmenintervall eines nach dem anderen ausgibt. Signalverarbeitungsvorrichtung nach Anspruch 10, wobei die Signalverarbeitungsvorrichtung ferner umfasst: einen ersten Speicher, der das Rahmensignal des ersten digitalen Signals Eines nach dem Anderen speichert; einen zweiten Speicher, der das Rahmensignal des zweiten digitalen Signals Eines nach dem Anderen speichert; und einen Verteiler- und Auswahlabschnitt, der ein Signal, für dass das erste Verfahren durch den Signalverarbeitungs- Hauptabschnitt durchgeführt wird, an den (i + 1)-ten Signalverarbeitungs- Unterabschnitt sendet, ein Signal, für dass das zweite Verfahren für das (N * (t-1) + i)-te Rahmensignal vor dem (i + 1)-ten Signalverarbeitungs- Unterabschnitt durchgeführt worden ist, empfängt und das Signal an den Signalverarbeitungs- Hauptabschnitt ausgibt; und wobei der Signalverarbeitungs- Hauptabschnitt mit dem ersten und dem zweiten Speicher verbunden ist, ein Rahmensignal von dem ersten Speicher für jedes Zeitintervall T Eines nach dem Anderen herausnimmt, das erste Verfahren für das (N *t + 1)-te Rahmensignal ausführt, um das Rahmensignal an den Verteiler- und Auswahlabschnitt abzugeben, und das nach dem Verfahren vorhandene Signal, das von dem Verteiler- und Auswahlabschnitt empfangen wurde, in dem zweiten Speicher abspeichert; und wobei die ersten bis N-ten Signalverarbeitungs- Unterabschnitte, die das zweite Verfahren für das Rahmensignal, das von dem Verteiler- und Auswahlabschnitt empfangen wurde, durchführt und das nach dem Verfahren vorhandene Signal an den Verteiler- und Auswahlabschnitt sendet. Signalverarbeitungsvorrichtung nach Anspruch 10, worin

das erste Verfahren ein Verfahren enthält, das Information verwendet, die in der ersten Rahmenzeit erzeugt wurde; und

das zweite Verfahren ein Verfahren ausschließt, das Information verwendet, die in der vergangenen Rahmenzeit erzeugt wurde.
Signalverarbeitungsvorrichtung nach Anspruch 13, worin

das erste digitale Signa! ein PCM Signal eines Audiosignals ist;

das zweite digitale Signal ein komprimiertes und kodiertes Signal eines Audiosignals ist;

das erste Verfahren ein Verfahren enthält, dass das PCM Signal in Information eines Frequenzspektrums umsetzt; und

das zweite Verfahren ein Verfahren enthält, das die Information des Frequensspektrums kodiert und komprimiert.
Signalverarbeitungsvorrichtung nach Anspruch 14, worin

das erste Verfahren ein MDCT Verfahren enthält; und

das zweite Verfahren ein Huffman- Kodierungsverfahren enthält.
Signalverarbeitungsvorrichtung nach Anspruch 14, worin

das erste Verfahren ein Unterbandanalyse- Filterbankverfahren enthält; und

das zweite Verfahren ein Quantisierunsgsverfahren enthält.
Signalverarbeitungsvorrichtung nach Anspruch 10, worin

eine Unterteilung für das erste Verfahren und das zweite Verfahren so gemacht wird, dass die Rechenzeit, die für das zweite Verfahren erforderlich ist, N mal der Rechenzeit ist, die für das erste Verfahren erforderlich ist.
Signalverarbeitungsverfahren, dass ein erstes digitales Signal in ein zweites digitales Signal umsetzt, in dem erste bis N-te Signalverarbeitungsunterabschnitte und ein Signalverarbeitungs- Hauptabschnitt verwendet werden, dass die folgenden Schritte umfasst: Vollenden des ersten Verfahrens innerhalb einer Periode T in dem Hauptverarbeitungsabschnitt für Rahmensignale des ersten digitalen Signals, dass für jedes vorgegebene Zeitintervall mit einem Rahmen versehen ist; und Umsetzen des (N + t + i)-ten Rahmensignals, das von dem Signalverarbeitungs- Hauptabschnitt verarbeitet und an den (1 + i)-ten Signalverarbeitungs- Unterabschnitt Eins nach dem Anderen gegeben wurde, wobei i und t gerade Zahlen sind, N eine natürliche Zahl >=2 ist, T eine reale Zeit ist und 0 <= i < N gilt, in ein zweites digitales Signal durch Vollenden des zweiten Verfahrens innerhalb einer Periode (N + T) in den ersten bis N-ten Signalverarbeitungs- Unterabschnitten. Tragbare Vorrichtung umfassend: einen Audiosignal- Eingangsabschnitt, der ein kodiertes Audiosignal eingibt; eine Signalverarbeitungsvorrichtung nach Anspruch 5, die das kodierte Audiosignal dekodiert; und einen Audiosignal- Ausgangsabschnitt, der das dekodierte Audiosignal ausgibt. Tragbare Vorrichtung umfassend: einen Audiosignal- Eingangsabschnitt, der ein Audiosignal eingibt; eine Signalverarbeitungsvorrichtung nach Anspruch 14, die das Audiosignal kodiert; und einen Speicher, der das kodierte Audiosignal hält.
Anspruch[en]
A signal processing device for converting a first digital signal into a second digital signal by performing a first process and then a second process for each frame signal of said first digital signal framed for each predetermined time interval, comprising: first to Nth signal processing sub sections (11,12,13,31,32) wherein each (i + 1)th signal processing sub sections is provided with the (N * t + i)th frame of the signal, wherein i and t are integers, N is a natural number >= 2, and 0 ≤ i <N, of said first digital signal and each of which completes said first process within a period (N * T), wherein T is a real number; and a signal processing main section (10,30) which converts a signal processed in said (i + 1)th signal processing sub section into said second digital signal by completing said second process within a period T. The signal processing device of claim 1, wherein said signal processing device further comprising: a distribution section inputting said first digital signal to one of said first to Nth signal processing sub sections for each frame interval one after another; and a selection section selectively outputting one of the after-process signal outputted from said first to Nth signal processing sub sections for each frame interval one after another to input the signal to said signal processing main section. The signal processing device of claim 1, wherein said signal processing device further comprising: a first memory storing said frame signal of said first digital signal one after another; a second memory storing said frame signal of said second digital signal one after another; and a distribution and selection section which sends the (N* t + i)th frame signal, wherein i and t are integers, and 0 ≤ i < N, obtained from said signal processing main section to said (i + 1)th signal processing sub section, receives the signal which said first process has been performed for the (N *(t - 1) + i)th frame signal from the (i + 1)th signal processing sub section, and outputs said signal to said signal processing main section; and wherein each of said first to Nth signal processing sub section is connected to said distribution and selection section, performs the first process for said frame signal received from said distribution and selection section, and sends the after-process signal to said distribution and selection section; and

wherein said signal processing main section is connected to said first and second memories, picks out said frame signal from said first memory for each time interval T one after another to output said frame signal to said distribution and selection section, and performs said second process for said signal received from said distribution and selection section to store this after-process signal in said second memory.
The signal processing device of claim 1, wherein

said second process contains a process employing information generated in the past frame time; and

said first process excludes a process employing information generated in the past frame time.
The signal processing device of claim 4, wherein

said first digital signal is a compressed and encoded signal of an audio signal;

said second digital signal is a PCM signal of an audio signal;

said first process contains a process picking out information from the compressed and encoded signal to convert the information into information of a frequency spectrum; and

said second process contains a process converting said information of said frequency spectrum into said PCM signal of time base.
The signal processing device of claim 5, wherein

said first process contains a decoding process of a variable length code; and

said second process contains an inverse MDCT process.
The signal processing device of claim 5, wherein

said first process contains an inverse quantizing process inversely quantizing said compressed and encoded signal, and

said second process contains a sub-band synthesis filter bank process.
The signal processing device of claim 1, wherein

division is made for said first process and said second process so that the calculation period necessary for said first process is N times the calculation period necessary for said second process.
A signal processing method converting a first digital signal into a second digital signal by employing first to Nth signal processing sub sections and a signal processing main section, comprising the following steps of

completing a first process within a period (N * T) in each (i+1)th signal processing sub sections for (N * t + i)th frame signals, wherein i and t are integers, N is a natural number >= 2, T is a real number, and 0 ≤ i < N, of the first digital signal one after another; and

converting a signal processed in said (i + 1)th signal processing sub section into the second digital signal by completing a second process within a time T in said signal processing main section.
A signal processing device for converting a first digital signal into a second digital signal by performing a first process and then a second process for each frame signal of the first digital signal framed for each predetermined time interval, comprising: a signal processing main section (50,70) which is provided with frame signals of said first digital signal and which completes said first process within a period T, wherein T is a real number; and first to Nth signal processing sub sections (51,52,53,71,72) wherein said (i+1)th signal processing sub section is provided with the (N * t + i)th frame signal which had been processed in said signal processing main section, wherein i and t are integers, N is a natural number >=2, and 0 <= i < N, and converts said frame signal into said second digital signal by completing said second process within a period (N * T). The signal processing device of claim 10, wherein said signal processing device further comprising: a distribution section inputting said digital frame signal outputted from said signal processing main section into one of said first to Nth signal processing sub sections for each frame interval one after another; and a selection section selectively outputting one of the after-process signal outputted from one of said first to Nth signal processing sub sections for each frame interval one after another. The signal processing device of claim 10, wherein said signal processing device further comprising: a first memory storing said frame signal of said first digital signal one after another; a second memory storing said frame signal of said second digital signal one after another; and a distribution and selection section which sends a signal for which said first process is performed by said signal processing main section to said (i + 1)th signal processing sub section, receives a signal for which said second process has been performed for the (N * (t - 1) + i)th frame signal from said (i + 1)th signal processing sub section, and outputs said signal to said main signal processing section; and wherein said signal processing main section is connected to said first and second memories, picks out a frame signal from said first memory for each time interval T one after another, performs said first process for said (N * t + i)th frame signal to output said frame signal to said distribution and selection section, and stores this after-process signal received from said distribution and selection section in said second memory; and

wherein said first to Nth signal processing sub sections which performs said second process for said frame signal received from said distribution and selection section, and sends the after-process signal to said distribution and selection section.
The signal processing device of claim 10, wherein

said first process contains a process employing information generated in the past frame time; and

said second process excludes a process employing information generated in the past frame time.
The signal processing device of claim 13, wherein

said first digital signal is a PCM signal of an audio signal;

said second digital signal is a compressed and encoded signal of an audio signal;

said first process contains a process converting said PCM signal into information of a frequency spectrum; and

said second process contains a process encoding and compressing said information of said frequency spectrum.
The signal processing device of claim 14, wherein

said first process contains a MDCT process; and

said second process contains a Huffman coding process.
The signal processing device of claim 14, wherein

said first process contains a sub-band analysis filter bank process; and

said second process contains a quantizing process.
The signal processing device of claim 10, wherein

division is made for said first process and said second process so that the calculation period necessary for said second process is N times the calculation period necessary for said first process.
A signal processing method converting a first digital signal into a second digital signal by employing first to Nth signal processing sub sections and a signal processing main section, comprising the following steps of: completing said first process within a period T in said main processing section for frame signals of the first digital signal framed for each predetermined time interval, and converting (N * t + i)th frame signal processed by said signal processing main section and given to said (1 + i)th signal processing sub section one after another, wherein i and t are integers, N is a natural number >=2, T is a real number and 0 <= i < N, into said second digital signal by completing said second process within a period (N *T) in said first to Nth signal processing sub sections. A portable type apparatus comprising: an audio signal input section inputting an encoded audio signal; a signal processing device of claim 5 decoding said encoded audio signal; and an audio signal output section outputting said decoded audio signal. A portable type apparatus comprising: an audio signal input section inputting an audio signal; a signal processing device of claim 14 encoding said audio signal; and an memory holding said encoded audio signal.
Anspruch[fr]
Dispositif de traitement de signaux pour convertir un premier signal digital dans un second signal digital en réalisant un premier procès et puis un second procès pour chaque signal de cadre dudit premier signal digital encadré pour chaque intervalle de temps prédéterminé, comprenant: - une première jusqu'à Nème sous sections de traitement de signaux (11, 12, 13, 31, 32) où chacune des (i+1)èmes sous sections de traitement de signaux est pourvue avec le (N ' t + i)ème cadre de signal, où i et t sont des entiers, N est un nombre naturel >=2, et 0 ≤ i < N, dudit premier signal digital et desquels chacun complète ledit premier procès avec une période (N *T) où T est un nombre réel; et - une section principale de traitement de signaux (10, 30) qui convertit un signal traité dans ladite (i+1)ème sous section de traitement de signaux dans ledit second signal digital en complétant ledit second procès dans une période T. Dispositif de traitement de signaux selon la revendication 1, où ledit dispositif de traitement de signaux comprenant de plus: - une section de distribution introduisant ledit premier signal digital à l'une desdites première jusqu'à Nème sous sections de traitement de signaux pour chaque intervalle de cadre l'un après l'autre; et - une section de sélection faisant sortir sélectivement l'un des signaux après-procès sorti desdites première jusqu'à Nème sous sections de traitement de signaux pour chaque intervalle de cadre l'un après l'autre pour faire entrer le signal à ladite section principale de traitement de signaux. Dispositif de traitement de signaux selon la revendication 1, où ledit dispositif de traitement de signaux comprenant de plus: - une première mémoire stockant ledit signal de cadre dudit premier signal digital l'un après l'autre; - une seconde mémoire stockant ledit signal de cadre dudit second signal digital l'un après l'autre; et - une section de distribution et de sélection qui envoie le (N * t + i)ème signal de cadre, où i et t sont des entiers, et 0 ≤ i < N, obtenu de ladite section principale de traitement de signaux à ladite (i + 1)ème sous section de traitement de signaux, reçoit le signal que ledit premier procès l'a réalisé pour le (N *(t -1) + i)ème signal de cadre de la (i + 1)ème sous section de traitement de signaux, et fait sortir ledit signal à ladite section principale de traitement de signaux; et - où chacune de ladite première jusqu'à Nème sous section de traitement de signaux est connectée à ladite section de distribution et sélection, réalise le premier procès pour ledit signal de cadre reçu de ladite section de distribution et sélection, et envoie le signal après-procès à ladite section de distribution et séjection; et - où ladite section principale de traitement de signaux est connectée auxdites première et seconde mémoires, sélectionne ledit signal de cadre de ladite première mémoire pour chaque intervalle de temps T l'un après l'autre pour faire sortir ledit signal de cadre à ladite section de distribution et sélection, et réalise ledit second procès pour ledit signal reçu de ladite section de distribution et sélection pour stocker ce signal après-procès dans ladite seconde mémoire. Dispositif de traitement de signaux selon la revendication 1, où - ledit second procès contient un procès utilisant l'information de générée dans le temps de cadre passé; et - ledit premier procès exclut un procès utilisant l'information générée dans le temps de cadre passé. Dispositif de traitement de signaux selon la revendication 4, où - ledit premier signal digital est un signal comprimé et codé d'un signal audio; - ledit second signal digital et un signal PCM d'un signal audio; - ledit premier procès contient un procès sélectionnant l'information à partir du signal comprimé et codé pour convertir l'information dans l'information d'un spectre de fréquences; et - ledit second procès contient un procès convertissant ladite information dudit spectre de fréquences dans ledit signal PCM de base de temps. Dispositif de traitement de signaux selon la revendication 5, où - ledit premier procès contient un procès de décodage d'un code de longueur variable; et - ledit second procès contient un procès inverse MDCT. Dispositif de traitement de signaux selon la revendication 5, où - ledit premier procès contient un procès de quantification inverse quantifiant inversement ledit signal comprimé et codé; et - ledit second procès contient un procès de filtrage de synthèses sous bande. Dispositif de traitement de signaux selon la revendication 1, où - la division est faite pour ledit premier procès et ledit second procès de sorte que la période de calcul nécessaire pour ledit premier procès soit de N fois la période de calcul nécessaire pour ledit second procès. Une méthode de traitement de signaux convertissant un premier signal digital dans un second signal digital en utilisant les première jusqu'à Nème sous sections de traitement de signaux et une section principale de traitement de signaux, comprenant les suivantes étapes de - compléter un premier procès dans une période (N *T) en chacune des (i + 1)èmes sous sections de traitement de signaux pour les (N* t + i)èmes signaux de cadre, où i et t sont des entiers, N est un nombre naturel >=2, T est un nombre réel, et 0 ≤ i <N, du premier signal digital l'un après l'autre; et - convertir un signal traité dans ladite (i + 1)ème sous section de traitement de signaux dans le second signal digital en complétant un second procès dans un temps T dans ladite section principale de traitement de signaux. Dispositif de traitement de signaux pour convertir un premier signal digital dans un second signal digital en réalisant un premier procès et puis un second procès pour chaque signal de cadre du premier signal digital encadré pour chaque intervalle de temps prédéterminé, comprenant: - une section principale de traitement de signaux (50, 70), qui est pourvue avec des signaux de cadre dudit premier signal digital et qui complète ledit premier procès dans une période T, où T est un nombre réel; et - les première jusqu'à Nème sous sections de traitement de signaux (51, 52, 53, 71,7 2), où ladite (i + 1)ème sous section de traitement de signaux est pourvue avec le (N* t + i)ème signal de cadre qui a été traité dans ladite section principale de traitement de signaux, où i et t sont des entiers, N est un nombre naturel >=2, et 0 <= i < N et convertit ledit signal de cadre dans ledit second signal digital en complétant ledit second procès dans une période (N * T). Le dispositif de traitement de signaux selon la revendication 10, où ledit dispositif de traitement de signaux comprenant de plus: - une section de distribution introduisant ledit signal de cadre digital sorti de ladite section principale de traitement de signaux dans l'une desdites première jusqu'à Nème sous sections de traitement de signaux pour chaque intervalle de cadre l'un après l'autre; et - une section de sélection faisant sortir sélectivement l'un du signal après-procès sorti de l'une desdites première jusqu'à Nième sous sections de traitement de signaux pour chaque intervalle de cadre l'un après l'autre. Le dispositif de traitement de signaux selon la revendication 10, où ledit dispositif de traitement de signaux comprenant de plus: - une première mémoire stockant ledit signal de cadre dudit premier signal digital l'un après l'autre; - une seconde mémoire stockant ledit signal de cadre dudit second signal digital l'un après l'autre; et - une section de distribution et sélection qui envoie un signal pour lequel ledit premier procès est réalisé par ladite section principale de traitement de signaux à ladite (i + 1)ème sous section de traitement de signaux, reçoit un signal pour lequel ledit second procès a été réalisé pour le (N * (t-1) + i)ème signal de cadre de ladite (i + 1)ème sous section de traitement de signaux et fait sortir ledit signal à ladite section principale de traitement de signaux; et - où ladite section principale de traitement de signaux est connectée audites première et seconde mémoires, sélectionne un signal de cadre de ladite première mémoire pour chaque intervalle de temps T l'un après l'autre, réalise ledit premier procès pour ledit (N* t + i)ème signal de cadre pour faire sortir ledit signal de cadre à ladite section de distribution et sélection, et stocke ce signal après-procès reçu de ladite section de distribution et sélection dans ladite seconde mémoire; et - où lesdites première à Nème sous sections de traitement de signaux qui réalisent ledit second procès pour ledit signal de cadre reçu de ladite section de distribution et sélection, et envoie le signal après-procès à ladite section de distribution et sélection. Le dispositif de traitement de signaux selon la revendication 10, où - ledit premier procès contient un procès utilisant l'information générée dans le temps de cadre passé; et - ledit second procès exclut un procès utilisant l'information générée dans le temps de cadre passé. Le dispositif de traitement de signaux selon la revendication 13, où - ledit premier signal digital est un signal PCM d'un signal audio; - ledit second signal digital est un signal comprimé et codé d'un signal audio; - ledit premier procès contient un procès convertissant ledit signal PCM dans une information d'un spectre de fréquences; et - ledit second procès contient un procès codant et compressant ladite information dudit spectre de fréquences. Le dispositif de traitement de signaux selon la revendication 14, où - ledit premier procès contient un procès MDCT, et - ledit second procès contient un procès de codage Huffman. Le dispositif de traitement de signaux selon la revendication 14, où - ledit premier procès contient un procès de filtrage analyses sous bande; et - ledit second procès contient un procès de quantification. Le dispositif de traitement de signaux selon la revendication 10, où - la division est faite pour ledit premier procès et ledit second procès de sorte que la période de calcul nécessaire pour ledit second procès est de N fois la période de calcul nécessaire pour ledit premier procès. Procédé de traitement de signaux convertissant un premier signal digital dans un second signal digital en utilisant les première jusqu'à Nème sous sections de traitement de signaux et une section principale de traitement de signaux, comprenant les suivantes étapes de: - compléter ledit premier procès dans une période T dans ladite section principale de traitement de signaux pour les signaux de cadre du premier signal digital encadré pour chaque intervalle de temps prédéterminé; et - convertir le (N * t + i)ème signal de cadre traité par ladite section principale de traitement de signaux et donné à ladite (1 + i)ème sous section de traitement de signaux l'un après l'autre, où i et t sont des entiers, N est un nombre naturel >=2, T est un nombre réel et 0 <= i <N, dans ledit second signal digital en complétant ledit second procès dans une période (N * T) dans lesdites première jusqu'à Nème sous sections de traitement de signaux. Dispositif de type portable, comprenant: - une section d'entrée de signaux audio permettant entrer un signal audio codé; - un dispositif de traitement de signaux de la revendication 5 décodant ledit signal audio codé; et - une section de sortie de signaux audio faisant sortir ledit signal audio décodé. Dispositif de type portable, comprenant: - une section d'entrée de signaux audio permettant entrer un signal audio; - un dispositif de traitement de signaux selon la revendication 14 codant ledit signal audio; et - une mémoire comprenant ledit signal audio codé.






IPC
A Täglicher Lebensbedarf
B Arbeitsverfahren; Transportieren
C Chemie; Hüttenwesen
D Textilien; Papier
E Bauwesen; Erdbohren; Bergbau
F Maschinenbau; Beleuchtung; Heizung; Waffen; Sprengen
G Physik
H Elektrotechnik

Anmelder
Datum

Patentrecherche

Patent Zeichnungen (PDF)

Copyright © 2008 Patent-De Alle Rechte vorbehalten. eMail: info@patent-de.com