Apparatus and method for compressing a digital input signal in more than one compression mode

I claim:

1. Apparatus for deriving a compressed digital signal from a digital input signal by compressing the digital input signal in a selected one of at least two compression modes, the digital input signal being compressed with a different compression ratio in each of the at least two compression modes, the compressed digital signal having a bit rate, the bit rate of the compressed digital signal being different in each of the compression modes, the apparatus comprising:

receiving means for receiving the digital input signal at a fixed sampling frequency, the sampling frequency being invariable between the compression modes;

a low-pass filter having a cut-off frequency set according to the selected one of the compression modes, the low-pass filter receiving the digital input signal and providing a bandwidth-limited signal having a bandwidth defined by the cut-off frequency of the low-pass filter; and

compressor means for deriving the compressed digital signal from the bandwidth-limited signal.

2. The apparatus of claim 1, wherein:

the apparatus additionally comprises block dividing means for dividing the bandwidth-limited signal in time into blocks, each block having a block length, the block lengths of the blocks having a maximum block length depending on the selected one of the compression modes; and

the compressor means derives the compressed digital signal from the blocks of the bandwidth-limited signal.

3. The apparatus of claim 1, wherein the compressor means includes:

frequency dividing means for deriving spectral coefficients from the bandwidth-limited signal; and

quantizing means for quantizing the spectral coefficients grouped by frequency into bands, the bands having a wider bandwidth towards higher frequencies.

4. The apparatus of claim 3, wherein:

the quantizing means includes a bit allocating means for allocating among the bands quantizing bits for quantizing the spectral coefficients in each band, the bit allocating means allocating no quantizing bits to bands having a frequency above the cut-off frequency of the low-pass filter; and

the quantizing means includes means for generating sub information for each band below the cut-off frequency of the low-pass filter.

5. The apparatus of claim 3 wherein the frequency dividing means includes an orthogonal transform circuit.

6. The apparatus of claim 3, wherein:

the apparatus additionally comprises:

filter means for dividing the band-limited signal in frequency into a frequency range signal in each of plural frequency ranges, and

block dividing means for dividing each frequency range signal in time into blocks; and

the frequency dividing means includes means for orthogonally transforming the blocks of each frequency range signal to provide the spectral coefficients.

7. The apparatus of claim 6, wherein:

the filter means includes down sampling means for downsampling the frequency range signal in the frequency range wherein the cut-off frequency of the low-pass filter lies, the downsampling means generating a downsampled frequency range signal; and

the block dividing means and the frequency dividing means operate on the downsampled frequency range signal.

8. The apparatus of claim 6, wherein the filter means divides the bandwidth-limited signal in frequency into frequency ranges including a lowest frequency range adjacent a next-lowest frequency range, the lowest frequency range and the next-lowest frequency range having an equal bandwidth.

9. The apparatus of claim 8, wherein the filter means divides the bandwidth-limited signal in frequency into frequency ranges additionally including a highest frequency range, the highest frequency range having a bandwidth greater than the lowest frequency range.

10. Apparatus for deriving a compressed digital signal from a digital input signal by compressing the digital input signal in a selected one of at least two compression modes, the digital input signal being compressed with a different compression ratio in each of the at least two compression modes, the compressed digital signal having a bit rate, the bit rate of the compressed digital signal being different in each of the compression modes, the apparatus comprising:

means for receiving the digital input signal at a fixed sampling frequency, the sampling frequency being invariable between the compression modes;

block dividing means for dividing the digital input signal in time into blocks, each block having a block length, the block lengths of the blocks having a maximum block length depending on the selected one of the compression modes, the maximum block length being greater in the compression modes wherein the digital output signal has a lower bit rate; and

compressor means for deriving the compressed digital signal from the blocks of the digital input signal.

11. The apparatus of claim 10, wherein the compressor means includes:

frequency dividing means for deriving spectral coefficients from the digital input signal; and

quantizing means for quantizing the spectral coefficients grouped by frequency into bands, the bands having a wider bandwidth towards higher frequencies.

12. The apparatus of claim 11, wherein the frequency dividing means includes an orthogonal transform circuit.

13. The apparatus of claim 11, wherein:

the apparatus additionally comprises filter means for dividing the digital input signal in frequency into a frequency range signal in each of plural frequency ranges;

the block dividing means is for dividing each frequency range signal in time into blocks, each block having a block length, the block lengths of the blocks having a maximum block length depending on the selected one of the compression modes; and

the frequency dividing means includes means for orthogonally transforming the blocks of each frequency range signal to provide the spectral coefficients.

14. The apparatus of claim 13, wherein:

the filter means includes down sampling means for downsampling one of the frequency range signals, the downsampling means generating a downsampled frequency range signal; and

the dividing means and the frequency dividing means operate on the downsampled frequency range signal.

15. The apparatus of claim 13, wherein the filter means divides the digital input signal in frequency into frequency ranges including a lowest frequency range adjacent a next-lowest frequency range, the lowest frequency range and the next-lowest frequency range having an equal bandwidth.

16. The apparatus of claim 15, wherein the filter means divides the bandwidth-limited signal in frequency into frequency ranges additionally including a highest frequency range, the highest frequency range having a bandwidth greater than the lowest frequency range.

17. The apparatus of claim 13, wherein the block length whereinto the block dividing means divides each frequency range signal has a minimum block length equal to a fraction of the maximum block length, the minimum block length whereinto the block dividing means divides the frequency range signal in at least the highest frequency range being the same irrespective of the compression mode.

18. Method for deriving a compressed digital signal from a digital input signal by compressing the digital input signal in a selected one of at least two compression modes, the digital signal being compressed with a different compression ratio in each of the at least two compression modes, the compressed digital signal having a bit rate, the bit rate of the compressed digital signal being different in each of the compression modes, the method comprising the steps of:

receiving the digital input signal at a fixed sampling frequency, the sampling frequency being invariable between the compression modes;

subjecting the digital input signal to low-pass filtering with a cut-off frequency set according to the selected one of the compression modes to provide a bandwidth-limited signal having a bandwidth defined by the cut-off frequency of the low-pass filtering; and

deriving the compressed digital signal from the bandwidth-limited signal.

19. The method of claim 18, wherein:

the method additionally comprises the step of dividing the bandwidth-limited signal in time into blocks, each block having a block length, the block lengths of the blocks having a maximum block length depending on the selected one of the compression modes; and

in the deriving step, the compressed digital signal is derived from the blocks of the bandwidth-limited signal.

20. The method of claim 18, wherein the step of deriving the compressed digital signal includes the steps of:

deriving spectral coefficients from the bandwidth-limited signal; and

quantizing the spectral coefficients grouped by frequency into bands, the bands having a wider bandwidth towards higher frequencies.

21. The method of claim 20, wherein:

the method additionally comprises the step of providing quantizing bits;

the step of quantizing the spectral coefficients includes the step of allocating the quantizing bits among the bands for quantizing the spectral coefficients in each band, but allocating no quantizing bits to bands having a frequency above the cut-off frequency set according to the selected one of the compression modes; and

the quantizing means includes means for generating sub information for each band having a frequency below the cut-off frequency set according to the selected one of the compression modes.

22. The method of claim 20, wherein the step of deriving spectral coefficients from the bandwidth-limited signal includes the step of orthogonally transforming the bandwidth-limited signal.

23. The method of claim 20, wherein:

the method additionally comprises the steps of:

dividing the band-limited signal in frequency into a frequency range signal in each of plural frequency ranges, and

dividing each frequency range signal in time into blocks; and

the step of deriving spectral coefficients from the bandwidth-limited signal includes the step of orthogonally transforming the blocks of each frequency range signal to provide the spectral coefficients.

24. The method of claim 23, wherein:

the step of dividing the band-limited signal in frequency includes the step of downsampling the frequency range signal in the frequency range wherein the cut-off frequency of the selected one of the compression modes lies to generate a downsampled frequency range signal; and

the steps of dividing each frequency range signal in time into blocks, and of deriving spectral coefficients from the bandwidth-limited signal operate on the downsampled frequency range signal.

25. The method of claim 23, wherein, in the step of dividing the band-limited signal in frequency, the bandwidth-limited signal is divided in frequency into frequency ranges including a lowest frequency range adjacent a next-lowest frequency range, the lowest frequency range and the next-lowest frequency range having an equal bandwidth.

26. The method of claim 25, wherein, in the step of dividing the band-limited signal in frequency, the bandwidth-limited signal is divided in frequency into frequency ranges additionally including a highest frequency range, the highest frequency range having a bandwidth greater than the lowest frequency range.

27. The method of claim 18, additionally comprising the steps of:

providing a recording medium; and

recording the compressed signal on the recording medium.

28. The method of claim 27, wherein, in the step of providing a recording medium, an optical disc is provided as the recording medium.

29. The method of claim 27, wherein, in the step of providing a recording medium, a semiconductor memory is provided as the recording medium.

30. Method for deriving a compressed digital signal from a digital input signal by compressing the digital input signal using a selected one of at least two compression modes, the digital input signal being compressed with a different compression ratio in each of the at least two compression modes, the compressed digital signal having a bit rate, the bit rate of the compressed digital signal being different in each of the compression modes, the method comprising steps of:

receiving the digital input signal at a fixed sampling frequency, the sampling frequency being invariable between the compression modes;

dividing the digital input signal in time into blocks, each block having a block length, the block lengths of the blocks having a maximum block length depending on the selected one of the compression modes, the maximum block length being greater in the compression modes wherein the digital output signal has a lower bit rate; and

deriving the compressed digital signal from the blocks of the digital input signal.

31. The method of claim 30, wherein the step of deriving the compressed signal includes the steps of:

deriving spectral coefficients from the digital input signal; and

quantizing the spectral coefficients grouped by frequency into bands, the bands having a wider bandwidth towards higher frequencies.

32. The method of claim 31, wherein the step of deriving spectral coefficients from the digital input signal includes the step of orthogonally transforming the digital input signal.

33. The method of claim 31, wherein:

the method additionally comprises the step of dividing the digital input signal in frequency into a frequency range signal in each of plural frequency ranges;

in the step of dividing the digital input signal in time, each frequency range signal is divided in time into blocks, each block having a block length, the block lengths of the blocks having a maximum block length depending on the selected one of the compression modes; and

the step of deriving spectral coefficients from the digital input signal includes the step of orthogonally transforming the blocks of each frequency range signal to provide the spectral coefficients.

34. The method of claim 33, wherein:

the method additionally includes the step of downsampling one of the frequency range signals to generate a downsampled frequency range signal; and

the steps of dividing the digital input signal in time and deriving spectral coefficients from the digital input signal operate on the downsampled frequency range signal.

35. The method of claim 33, wherein, in the step of dividing the digital input signal in frequency, the digital input signal is divided in frequency into frequency ranges including a lowest frequency range adjacent a next-lowest frequency range, the lowest frequency range and the next-lowest frequency range having an equal bandwidth.

36. The method of claim 35, wherein, in the step of dividing the digital input signal in frequency, the digital input signal is divided in frequency into frequency ranges additionally including a highest frequency range, the highest frequency range having a bandwidth greater than the lowest frequency range.

37. The method of claim 30, wherein, in the step of dividing each frequency range signal in time, each frequency range signal is divided into blocks additionally having a minimum block length equal to a fraction of the maximum block length, the minimum block length of the blocks in at least the highest frequency range being the same irrespective of the compression mode.

38. The method of claim 30, additionally comprising the steps of:

providing a recording medium; and

recording the compressed signal on the recording medium.

39. The method of claim 38, wherein, in the step of providing a recording medium, an optical disc is provided as the recording medium.

40. The method of claim 38, wherein, in the step of providing a recording medium, a semiconductor memory is provided as the recording medium.

Description Submit all comments and votes

FIELD OF THE INVENTION

This invention relates to an apparatus and method for deriving a compressed digital signal from a digital input signal in which the compressed digital signal can be compressed in more than one compression mode having different bit rates.

BACKGROUND OF THE INVENTION

The inventor's assignee has proposed in, e.g., U.S. Pat. Nos. 5,243,588 and 5,244,705, and U.S. patent application Ser. No. 07/736,046, now abandoned on Mar. 3, 1994, the disclosures of which are incorporated herein by reference, a technique for compressing a digital audio input signal and recording the resulting compressed recording signal in bursts with a predetermined number of bits of the compressed recording signal as a recording unit.

With this technique, the compressed recording signal is an adaptive differential PCM (ADPCM) audio signal, and a magneto-optical disc is used as the recording medium for recording the compressed recording signal according to the so-called CD-I (CD-Interactive) or CD-ROM XA recording signal format. The compressed recording signal is recorded in bursts on the magneto-optical disc, with, e.g., 32 sectors of the compressed recording signal plus several linking sectors as a recording unit. The linking sectors are used to accommodate the additional signal generated by interleaving the compressed recording signal in the 32 sectors.

A recording and reproducing apparatus for a magneto-optical disc may employ one of several recording and reproduction modes for the compressed recording signal. In the CD-I and CD-XA formats, recording modes A, B, and C have been defined in which an uncompressed PCM audio signal, similar to that recorded on a normal Compact Disk (CD), but with a lower sampling frequency, is compressed to provide the compressed recording signal for recording on the magneto-optical disc. Recording mode A has a sampling frequency of 37.8 kHz, and the PCM audio signal is compressed by a compression ratio of two; recording mode B has the same sampling frequency as mode A and compression ratio of four; and recording mode C has a sampling frequency of 18.9 kHz, and a compression ratio of eight. In recording mode B, for example, the PCM audio input signal is compressed by a compression ratio of four, so that the playback time of a compact disc on which a mode B recording signal is recorded is four times that of a disc recorded according to the standard CD format (CD-DA format). Using a recording mode in which the PCM audio signal is compressed enables the size of the recording and reproducing apparatus to be reduced, because a recording or playback time comparable with that of a standard 12 cm disc can be provided by a smaller-sized disc.

The velocity of the recording track relative to the pickup head (the "recording velocity") of the smaller-sized disc on which a recording mode B compressed signal is recorded is chosen to be the same as that of a standard CD. This means that the bit rate of the compressed recording signal reproduced from the disc is four times the bit rate required by the mode B decoder. This allows the same recording unit of the compressed recording signal to be read from the disc four times, but only one of the four readings of the recording unit of the compressed recording signal is fed into the decoder.

The compressed recording signal is recorded on the disc on a spiral track. When reproducing the track, the pickup head is caused to execute a radial track jump on each complete revolution of the disc. The track jump returns the head to its original position on the track. Causing the head to execute four track jumps causes the head to read the same part of the track four times. This method of reproducing the compressed recording signal recorded on the track is advantageous, especially when used in a small-sized portable apparatus, since it enables satisfactory reproduction to be obtained even if only one of the four readings of the recording unit of the compressed recording signal is free of errors. This method of reproducing the compressed recording signal from the disc therefore provides a strong immunity against reproduction errors caused by physical disturbances and the like.

In future, semiconductor memories are expected to be used as a medium for recording digital audio signals. To enable semiconductor memories to provide a usable playing time, it is necessary to increase the compression ratio further by using variable bit rate compression encoding, such as entropy encoding. Specifically, it is anticipated that audio signals will be recorded and/or reproduced using IC cards employing semiconductor memories. A compressed recording signal that has been compressed using a variable bit rate compression technique will be recorded on and reproduced from the IC card.

Although it is expected that, in future, with progress in semiconductor technology, the playing time provided by an IC card will increase, and the cost of the IC card will decrease, compared with the playing time and cost of a present-day IC card, the IC card, which has barely started to be supplied to the market, is at present expensive and has a short playing time. Therefore, it is thought that an IC card might be used early on by transferring to it part of the contents of another, less expensive, larger capacity, recording medium, such as a magneto-optical disc. Signal exchange and re-recording operations would be conducted between the IC card and the magneto-optical disc. Specifically, a desired one or more selections recorded on the magneto-optical disc would be copied to the IC card. The copied selections would then be replaced by other selection(s) when desired. By repeatedly exchanging the selections recorded on the IC card, a variety of selections may be played on a portable IC card player using a small number of available IC cards.

Different applications require different bandwidths and signal-to-noise ratios for recording and reproducing audio signals. For example, when an audio signal is to be recorded and reproduced with high-fidelity quality, a bandwidth extending to 15 kHz or 20 kHz, and a large signal-to-noise ratio are required. To provide these characteristics using a system in which a compressed digital recording signal is recorded on a recording medium and reproduced therefrom, the compressed recording signal must have a relatively high bit rate. For example, a bit rate in the range of 256 kbps to 64 kbps per audio channel is required. On the other hand, when a digital audio signal representing speech is to be recorded and reproduced, a bandwidth extending to 5 kHz or 7 kHz is more than adequate, and a lower signal-to-noise ratio may be acceptable. Such characteristics may be provided using a bit rate in the range of 64 kbps to several kbps. Lower bit rates increase the recording time of the recording medium. Thus, to record different types of audio signals while making optimum use of the recording capacity of the recording medium, the recording/reproducing apparatus should be capable of recording and reproducing at different bit rates as economically as possible.

Conventional recording and reproducing apparatus using, for example the above-mentioned recording modes A, B, and C operate at several different sampling frequencies to provide recording modes with different bandwidths and signal-to-noise ratios. To operate at different sampling frequencies requires a complex sampling frequency signal generating circuit, and increased complexity in the LSI signal processing circuits. Moreover, when the sampling frequencies of the compression modes are different, switching the encoder between the different recording modes is difficult.

When a compressed recording signal recorded on a high-capacity magneto-optical disc with a high bit rate is to be convened so that it can be recorded on a low-capacity IC card using a low bit rate recording mode, the compressed recording signal must be expanded back to an uncompressed PCM signal, which must then be compressed again using a low bit rate recording mode. This requires a large amount of signal processing, which economically-viable signal processing LSIs may be unable to carry out in real time.

Additionally, in the low bit rate recording modes, the reduction in the number of bits available to represent the audio signal can lead to a deterioration of sound quality. For example, if the bandwidth is narrowed, and the bandwidth of bands into which the spectral coefficients are grouped is the same at all frequencies, dividing the audio frequency range of 0 Hz to 20 kHz into 32 bands makes the bandwidth of each band approximately 700 Hz. This is many times the bandwidth of the low-frequency critical bands, which is typically about 100 Hz, and is larger that the bandwidth of critical bands throughout most of the middle and low frequencies. This mismatch between the bandwidth of these equal bandwidth bands and the bandwidth of the critical bands at low and middle frequencies significantly reduces the efficiency of the compression process.

OBJECTS AND SUMMARY OF THE INVENTION

It is an object of the invention to provide an encoder capable of compressing a digital input signal in one of plural compression modes in which the complication of having a sampling frequency generating circuit that generates plural sampling frequencies, and the consequent increase in the scale of the LSI, are avoided.

It is a further object of the invention to provide an encoder capable of compressing a digital input signal in a compression mode providing a compressed signal having a low bit rate for recording on a recording medium with a limited storage capacity, such as an IC card.

It is a yet further object of the invention to provide an encoder for compressing a digital input signal in a compression mode providing a compressed signal having a low bit rate in which the deterioration of sound quality resulting from using the low bit rate is minimized.

Accordingly, the invention provides an apparatus for deriving a compressed digital signal from a digital input signal by compressing the digital input signal in a selected one of at least two compression modes. The compressed digital signal has a different bit rate in each of the compression modes. The apparatus includes a receiving circuit that receives the digital input signal at the same sampling frequency in all compression modes. The apparatus also includes a low-pass filter that has a cut-off frequency set according to the selected one of the compression modes. The low-pass filter receives the digital input signal and provides a bandwidth-limited signal. Finally, the apparatus includes a compressor circuit that derives the compressed digital signal from the bandwidth-limited signal.

The apparatus may additionally comprise a circuit that divides the bandwidth-limited signal in time into blocks. Each block has a block length, and the block lengths of the blocks have a maximum block length that depends on the selected one of the compression modes. In this case, the compressor circuit derives the compressed digital signal from the blocks of the bandwidth-limited signal.

The compressor circuit may include a frequency dividing circuit that derives spectral coefficients from the bandwidth-limited signal, and a quantizing circuit that quantizes the spectral coefficients grouped by frequency into bands. The bands have a wider bandwidth towards higher frequencies. The frequency dividing circuit may include an orthogonal transform circuit.

The apparatus may additionally include a filter circuit that divides the band-limited signal in frequency into a frequency range signal in each of plural frequency ranges, and a block dividing circuit that divides each frequency range signal in time into blocks. In this case, the frequency circuit includes a circuit that orthogonally transforms the blocks of each frequency range signal to provide the spectral coefficients.

The invention also provides an apparatus for deriving a compressed digital signal from a digital input signal by compressing the digital input signal in a selected one of at least two compression modes. The compressed digital signal has a different bit rate in each of the compression modes. The apparatus comprises a circuit that receives the digital input signal at the same sampling frequency in each of the compression modes. The apparatus also includes a block dividing circuit that divides the digital input signal in time into blocks. Each block has a block length; and the block lengths of the blocks have a maximum block length that depends on the selected one of the compression modes. The maximum block length is greater in the compression modes in which the digital output signal has a lower bit rate. Finally, the apparatus includes a compressor circuit that derives the compressed digital signal from the blocks of the digital input signal.

The apparatus may additionally include a frequency dividing circuit that derives spectral coefficients from the digital input signal, and a quantizing circuit that quantizes the spectral coefficients grouped by frequency into bands. The bands have a wider bandwidth towards higher frequencies. The frequency dividing circuit may include an orthogonal transform circuit.

The apparatus may additionally comprise a filter circuit that divides the digital input signal in frequency into a frequency range signal in each of plural frequency ranges. In this case, the block dividing circuit divides each frequency range signal in time into blocks. Each block has a block length; and the block lengths of the blocks have a maximum block length that depends on the selected one of the compression modes. Also, in this case, the frequency dividing circuit includes a circuit that orthogonally transforms the blocks of each frequency range signal to provide the spectral coefficients.

The invention further provides a method for deriving a compressed digital signal from a digital input signal by compressing the digital input signal in a selected one of at least two compression modes. The compressed digital signal has a different bit rate in each of the compression modes. In the method, the digital input signal is received at the same sampling frequency in all the compression modes. The digital input signal is subject to low-pass filtering with a cut-off frequency set according to the selected one of the compression modes to provide a bandwidth-limited signal. Finally, the compressed digital signal is derived from the bandwidth-limited signal.

The method may additionally include dividing the bandwidth-limited signal in time into blocks. Each block has a block length; and the block lengths of the blocks have a maximum block length depending on the selected one of the compression modes. In this case, the compressed digital signal is derived from the blocks of the bandwidth-limited signal.

Deriving the compressed digital signal may include deriving spectral coefficients from the bandwidth-limited signal, and quantizing the spectral coefficients grouped by frequency into bands having a wider bandwidth towards higher frequencies.

The method may also additionally include dividing the band-limited signal in frequency into a frequency range signal in each of plural frequency ranges, and dividing each frequency range signal in time into blocks. In this case, deriving spectral coefficients from the bandwidth-limited signal may include orthogonally transforming the blocks of each frequency range signal to provide the spectral coefficients.

The method may also include providing a recording medium and recording the compressed signal on the recording medium.

Finally, the invention provides a method for deriving a compressed digital signal from a digital input signal by compressing the digital input signal using a selected one of at least two compression modes. The compressed digital signal has a different bit rate in each of the compression modes. In the method, the digital input signal is received at the same sampling frequency in all compression modes. The digital input signal is divided in time into blocks. Each block has a block length; and the block lengths of the blocks have a maximum block length depending on the selected one of the compression modes. The maximum block length is greater in the compression modes in which the digital output signal has a lower bit rate. Finally, the compressed digital signal is derived from the blocks of the digital input signal.

Deriving the compressed signal may include deriving spectral coefficients from the digital input signal and quantizing the spectral coefficients grouped by frequency into bands having a wider bandwidth towards higher frequencies.

The method may additionally comprise dividing the digital input signal in frequency into a frequency range signal in each of plural frequency ranges. In this case, each frequency range signal is divided in time into blocks. Each block has a block length; and the block lengths of the blocks have a maximum block length depending on the selected one of the compression modes. Also, in this case, deriving spectral coefficients from the digital input signal may include orthogonally transforming the blocks of each frequency range signal to provide the spectral coefficients.

The method may also include providing a recording medium and recording the compressed signal on the recording medium.

The encoder according to the invention uses the same sampling frequency regardless of the bit rate of each compression mode. This saves having to use a complex sampling frequency signal generating circuit capable of generating plural sampling frequencies, and allows the scale of the LSI to be reduced.

Moreover, when the sampling frequencies of the respective compression modes are the same, conversion of signals between the different compression modes can be carried out more easily than if the compression modes use different sampling frequencies. When the compressed signal compressed in a high bit rate compression mode and recorded on a large-capacity recording medium, such as a magneto-optical disc, is to be transferred to a small-capacity recording medium, such as an IC card, in a lower bit-rate compression mode, it is not necessary to cancel the compression of the high-bit rate compression mode completely and to fully expand the compressed signal. Instead, it is possible, simply with additional processing, to convert the compression mode to a lower bit rate compression mode. This reduces amount of signal processing required, and allows the process to be carried out in real-time or faster than real-time.

Also, in the encoder according to the invention, in the lower bit rate compression modes, signal processing operations are not carried out above the upper frequency limit of the compression mode. This saves a number of signal processing operations, which can be used to provide additional signal processing to improve the sound quality in the low bit rate compression modes.

The frequency range division filters located prior to the orthogonal transform circuits in the encoder according to the invention may be used to save having to perform signal processing in the high frequency range. In compression modes in which the entire high frequency range is unnecessary, no signal processing need be carried out in the high frequency range. Even in compression mode B, in which the high frequency range is partly used, the number of signal processing operations is reduced by downsampling the frequency range signal in the high frequency range.

In the lower bit rate compression modes, the number of bits usable for quantizing the spectral coefficients is reduced, giving rise to the need for preventing deterioration in sound quality. In the encoder according to the present invention, the maximum block length of the blocks of the frequency range signals subject to orthogonal transform processing and quantizing is increased, which improves the compression efficiency. With an increased maximum block length, the process of orthogonally transforming the frequency range signals from the time domain to the frequency domain can be carried out more accurately. Also, and the amount of sub-information, such as scale factors and word length data in the compressed signal can be reduced.

If the spectral coefficients were divided by frequency into bands having an equal bandwidth, dividing the frequency range of 0 Hz to 22 kHz into 32 bands would generate bands with a bandwidth of about 700 Hz. This is many times wider than the critical bandwidth of about 100 Hz at low frequencies, and is wider than the critical bandwidth at middle frequencies. This mismatch would considerably reduce the compression efficiency.

Therefore, in the encoder according to the present invention, the bandwidth of the bands into which the spectral coefficients resulting from the orthogonal transform are divided for allocating quantizing bits is selected to be broader towards higher frequencies, at least in most bands, so as to correspond more closely with the critical bandwidths. This prevents a lowering of the compression efficiency, as would be the case if the spectral coefficients were uniformly divided in frequency.

In the low bit rate compression modes, quantizing bits and sub information are not allocated to bands at and above the upper bandwidth limit of the digital input signal to avoid wasting bits.

Also, to minimize the deterioration of sound quality resulting from using a lower bit rate compression mode, the maximum block length of the blocks of the frequency range signals subject to orthogonal transform processing and quantizing is increased as the bit rate is reduced.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a block diagram of a practical example of a recording and reproducing apparatus for a compressed recording signal and including an encoder according to the present invention.

FIG. 2 illustrates the contents recorded in a magneto-optical disc and an IC card.

FIG. 3 shows an example of the appearance of the front panel of the recording and reproducing apparatus.

FIG. 4 is a block diagram showing a practical example of the encoder according to the invention for compressing a digital audio input signal.

FIG. 5 shows the frame and block structure in which the frequency range signals derived from the digital audio input signal are processed in four different compression modes.

FIG. 6 is block diagram of the allowable noise calculation circuit 20 shown in FIG. 4.

FIG. 7 shows a simplified bark spectrum, and the masking range of each band.

FIG. 8 shows a masking spectrum.

FIG. 9 shows a synthesis of the minimum audible level curve and the masking spectrum.

FIG. 10 shows how a frame of 11.6 ms is divided in frequency and in time into 52 bands in consideration of the bandwidths of the critical bands and the efficiency of the block floating processing applied to the bands.

FIG. 11 illustrates how the frame length for mode B is increased compared with that of mode A on account of the reduced the bandwidth and bit rate of mode B.

FIG. 12 is a simplified block diagram of the encoder according to the invention showing details of the downsampling.

FIG. 13 is a block circuit diagram showing a practical example of a decoder for expanding a compressed recording signal generated by an encoder according to the invention.

DETAILED DESCRIPTION OF THE INVENTION

A preferred embodiment of the present invention will now be described with reference to the accompanying drawings.

1. OVERVIEW OF THE RECORDING/REPRODUCING APPARATUS

FIG. 1 shows the schematic arrangement of an embodiment of an