JP3013380B2 - Data compression encoder - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a PCM signal data compression encoding apparatus suitable for use in, for example, a DSP (Digital Signal Processor).

[Summary of the Invention]

The present invention blocks the input signal every fixed sample,
For each block, a filter is supplied to a first filter group having a plurality of filters, and a filter that provides the highest compression ratio is selected from the first filter group, and an output from the selected filter is compressed by a compression code. And an input signal, which is divided into blocks for each fixed sample, and supplied to a second filter group having a plurality of filters for each block to obtain the highest compression ratio among the second filter groups. And a system that selects a filter to be output, compresses and encodes an output from the selected filter, and outputs an output from the system having the first filter group and an output from the system having the second filter group. Switching means for switching and selecting any one of the first and second filters, and when encoding becomes impossible in the system having the first filter group, the switching means selects the second filter. By selecting switches the output from the system with the group to prevent the encoding impossible in advance,
Real-time encoding can be enabled, and the first group of filters includes a high-establishment filter that has good output signal quality but cannot encode.
The second filter group includes a low-probability filter that cannot be encoded, so that the encoding is normally performed by a system having the first filter group having good output signal quality, and the output is taken out. Data compression encoding that can provide an output of good quality and no encoding failure by extracting the encoded output through a system having the second set of filters that is less likely to be encoded when it occurs. Device.

[Conventional technology]

In general, as a method of reducing the transmission bit rate by compressing the number of bits of a supplied PCM signal, for example, a filter that can obtain the highest compression ratio in a block unit for each of a plurality of samples is selected from among a plurality of previously prepared filters. There is known a so-called filter-selection-type data compression-encoding device that selects from the following.

To such a filter selection type data compression encoding apparatus, for example, an input signal having 16 samples as one block is supplied for each block. The input signal for each block is supplied to a zero-order filter that outputs straight PCM, a first-order difference filter that outputs a first-order difference, and a second-order difference filter that outputs a second-order difference, which are the plurality of filters. Then, the maximum absolute value in the block is detected for each filter, and the block data that has passed through the filter with the minimum maximum absolute value in the block is selected.
It is requantized from bits to 4 bits and output. In addition,
At the time of the requantization, a so-called noise shaping process is performed in which a difference between an input and an output of the quantizer is fed back to the input side of the quantizer and added to data newly supplied to the quantizer. I have. At the time of the output, filter information indicating a filter through which the data of the recorded block has passed, range information indicating a range of the data, and the like are output together with the data, and the decoder side responds to the information. Data is reproduced.

In recent years, a secondary differential filter (hereinafter, referred to as a second secondary differential filter) having better sound quality than the secondary differential filter (hereinafter, referred to as a first secondary differential filter).
Has been developed, and the data compression encoding apparatus of the filter selection type is provided with the second secondary differential filter in addition to the three filters of the zero-order filter, the primary differential filter, and the first secondary differential filter. It became so.

As described above, the data compression encoding apparatus provided with the four filters also supplies the input signal supplied to each block to each of the four filters, and the maximum number of blocks within the block is determined for each block passing through the four filters. An absolute value is detected, data is selected based on the maximum absolute value in the block, re-quantized, and output.

[Problems to be solved by the invention]

However, among the four filters provided in the above-described data compression encoding apparatus, the second secondary difference filter preferably has excellent sound quality, but sometimes cannot perform encoding. This inability to encode is caused, for example, by overflow or the like. When the input signal is encoded in real time or when a long sound source (source) is encoded, the inability to encode causes serious damage.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-described problem, and has as its object to provide a data compression encoding device that prevents the above-described inability to encode and enables encoding with good sound quality.

[Means for solving the problem]

According to the present invention, an input signal is divided into blocks by a predetermined number of samples and supplied to a first filter group having a plurality of filters for each block, and a filter capable of obtaining the highest compression rate among the first filter groups is provided. A first system for selecting and outputting an output from the selected filter by compression-encoding by a quantizer and a noise shaping circuit, and blocking the input signal for each fixed sample; , A filter that provides the highest compression rate is selected from the second filter group, and the output from the selected filter is compressed by a quantizer and a noise shaping circuit. Switching means for switching and selecting one of the output from the first system and the output from the second system for outputting the second system by converting and outputting the second system It has said switching means,
The switching is performed in response to a quantization error fed back by the noise shaping circuit of the first system or the second system and an overflow signal indicating that an overflow has occurred at the time of addition of the input signal. , The first filter group includes a filter which has a high probability that the quality of the output signal is good but cannot be encoded.
The above group of filters is characterized by including a filter with a low probability of being incapable of encoding, thereby solving the above-mentioned problem.

[Action]

The data compression encoding apparatus according to the present invention includes a system having a first filter group and a system having a second filter group.
The input signals are processed in parallel by the two systems, and when the encoding failure occurs in the system having the first filter group, this is not output and the switching means outputs the signal from the system having the second filter group. Since the output signal can be selected, real-time encoding can be performed without re-encoding.

Normally, encoding is performed by a system having the first filter group having good output signal quality, the output is taken out, and when the encoding is impossible, the second filter group having a low probability of being incapable of encoding is provided. By taking out the encoded output through the system, it is possible to supply an output with good quality and no encoding failure.

〔Example〕

Hereinafter, an embodiment of a data compression encoding apparatus according to the present invention will be described with reference to the drawings.

FIG. 1 is a functional block diagram showing functions of an embodiment of a data compression encoding apparatus according to the present invention in a block manner.

From this input terminal 1, for example, an audio PCM signal in which one sample is 16 bits (one word) and 16 samples are one block is supplied.

As the audio PCM signal, for example, in addition to digital data from a device such as a CD (compact disk) player or a DAT (digital audio tape recorder) or a memory of sound source data, an audio signal picked up by a microphone or the like is A-D (Analog-digital) converted digital data and the like.

The audio PCM signal via the input terminal 1 is supplied to a first filter group 2 and a second filter group 3, respectively.

The first filter group 2 is, for example, a zero-order difference filter
4, a first-order difference filter 5 and a first second-order difference filter 6. The second filter group includes a zero-order difference filter 7, a first-order difference filter 8, and a second-order difference filter. 9.

The zero-order difference filter 4 and the zero-order difference filter 7
And the first-order difference filter 5 and the first-order difference filter 8 have common characteristics, respectively, but the first-secondary difference filter 6 and the second-order difference filter 9 have different characteristics. I have.

That is, the first secondary difference filter 6 has an advantage that the sound quality at the time of reproduction is superior to the audio PCM signal encoded via the second secondary difference filter 9. Has the drawback that sometimes encoding is impossible. On the other hand, the second secondary difference filter 9 is inferior to the first secondary difference filter in terms of sound quality, but has an advantage that encoding is almost impossible, and for example, a so-called CD. -Level B of I standard
Are used.

As described above, for example, in the case of performing real-time encoding, if the encoding failure such as overflow occurs, the damage is large.
A first filter provided with filters having different characteristics.
And a second filter group. Usually, a first filter group having a filter having good sound quality is used. When data overflows through the first filter group, a second filter group is used. Data from groups were used.

That is, the audio PCM signal supplied to the first filter group 2 is supplied to a zero-order difference filter 4, a first-order difference filter 5, and a first second-order difference filter 6, respectively.
The audio PCM signal supplied to the second filter group 3 is supplied to a zero-order difference filter 7, a first-order difference filter 8, and a second second-order difference filter 9, respectively.

The audio PCM signal of one block supplied to the zero-order difference filter 4 has a zero-order difference, and is supplied to the range detector 10 as a so-called straight PCM signal.

This range detector 10 supplies all data in the block of the straight PCM signal to the selector 18,
The maximum absolute value in the block is detected and supplied to a comparison (minimum value detection) block 16.

The audio PCM signal of one block supplied to the primary difference filter 5 has a primary difference in the block, and is supplied to the range detector 11 as primary difference data.

The range detector 11 supplies all the first-order difference filters of one block to the selector 18, detects the maximum absolute value in the block, and supplies it to the comparison block 16.

The audio PCM signal of one block supplied to the first secondary difference filter 6 has a secondary difference in the block and is supplied to the range detector 12.

The range detector 12 supplies all the secondary differential data of one block to the selector 18, detects the maximum absolute value in the block, and supplies it to the comparison block 16.

The comparison block 16 compares the maximum absolute value in the block of the supplied straight PCM data, the maximum absolute value in the block of the primary difference data, and the maximum absolute value in the block of the secondary difference data and has a minimum value. And supplies the range of the detected minimum value as range information to the quantizer 20 and the multiplexer 25.
Of the two filters, which filter has passed is detected, and this is supplied as filter information to the selector 18, the noise shaping circuit 22, and the multiplexer 25.

The selector 18 selects an audio PCM signal through the filter specified by the filter information from the comparison block 16 by, for example, switching a switch, and supplies the selected one block of audio PCM signal to the quantizer 20. I do.

The quantizer 20 re-quantizes the supplied 16-bit audio PCM signal to, for example, 4 bits, and supplies the 4-bit audio PCM signal to the multiplexer 25 via the noise shaping circuit 22.

This requantization is performed in accordance with the range information supplied from the comparison block 16.
Four bits are extracted from the bit where "1" is first set in the bits.

The data, information, and the like supplied to the multiplexer 25 are extracted via the selected terminal 24a of the switch 24.

Next, the zero-order difference filter 7 of the second filter group 3
The one-block audio PCM signal supplied to
A zero-order difference is calculated and supplied to the range detector 13 as a so-called straight PCM signal.

This range detector 13 supplies all data in the block of the straight PCM signal to the selector 19,
The maximum absolute value in the block is detected and supplied to a comparison (minimum value detection) block 17.

The audio PCM signal of one block supplied to the primary difference filter 8 has a primary difference in the block, and is supplied to the range detector 14 as primary difference data.

The range detector 14 supplies all the primary difference data of one block to the selector 19, detects the maximum absolute value in the block, and supplies it to the comparison block 17.

The audio PCM signal of one block supplied to the second secondary difference filter 9 has a secondary difference in the block and is supplied to the range detector 15.

The range detector 15 supplies all the secondary difference data of one block to the selector 19, detects the maximum absolute value in the block, and supplies it to the comparison block 17.

The comparison block 17 compares the maximum absolute value in the block of the supplied straight PCM data, the maximum absolute value in the block of the primary difference data, and the maximum absolute value in the block of the secondary difference data and has the smallest value. Is supplied to the quantizer 21 and the multiplexer 26 as range information as the range information, and the minimum value
Of the two filters, which filter has passed is detected, and this is supplied to the selector 19, the noise shaping circuit 23, and the multiplexer 26 as filter information.

The selector 19 selects an audio PCM signal through the filter specified by the filter information from the comparison block 17 by, for example, switching a switch, and supplies the selected one block of audio PCM signal to the quantizer 21. I do.

The quantizer 21 was supplied in the same manner as the quantizer 20 described above.
The 16-bit audio PCM signal is requantized to, for example, 4 bits, and the 4-bit audio data is supplied to the multiplexer 26 via the noise shaping circuit 23.

The data, information, and the like supplied to the multiplexer 26 are extracted via the selected terminal 24b of the switch 24.

Here, the noise shaping circuits 22 and 23 are an error between the input and the output of the quantizer 20 or the quantizer 21.
The so-called quantization error is fed back to the input of the quantizer 20 or the quantizer 21 according to the filter information from the comparison block 16 or 17. That is, the data of the lower 4 bits after the 4th bit extracted at the time of the requantization is fed back to the input of the quantizer 20 or the quantizer 21 and newly transmitted to the quantizer 20 or the quantizer 21. It is added to the supplied data.

However, the quantization error returned by the noise shaping and the new data are added to generate a carry, and “1” shifts to a bit higher than the extracted bit specified by the range information. There is a case where a so-called overflow which is one of the non-encodings may occur.

When this overflow occurs, a direct current error component occurs during reproduction, which causes problems such as failure in good reproduction.
This is particularly noticeable when real-time encoding is required, such as when the input signal immediately reproduces an audio signal picked up by the microphone.

As described above, the overflow is unlikely to occur in the system of the second filter group 3, but usually, the overflow sometimes occurs in the system of the first filter group 2 from which data is extracted.

For this reason, the noise shaping circuit 22 of the system having the first filter group supplies an overflow signal indicating that the overflow has occurred to the switch 24 when the overflow has occurred.

When this overflow signal is supplied, switch 24
Switches the switch from the selected terminal 24a to the selected terminal 24b.

As described above, the audio data supplied to the selected terminal 24b is data that passes through the second filter group that is unlikely to be encoded. By thus switching the switch when an overflow occurs, The data in which the overflow has occurred can be selected by avoiding the data in which the overflow has occurred.

Audio data via such a switch 24 is
The audio is extracted via the selected audio output terminal 27.

As output data taken out from the output terminal 27, for example, one block is as shown in FIG. 2. One byte of header information (compression parameter information or auxiliary information, etc.) RF and eight bytes of sample data Data D _{A0 to} D _B3 . The header information RF is 4
It consists of bit range information, 2-bit mode selection information or filter selection information, two 1-bit flag information items, and an undefined area for adding some information in the future. . As described above, the data of one sample is compressed from, for example, 16 bits to 4 bits, and the data D _{A0 to} D _B3 include the 4-bit data D _{A0H to} D _{B3L for} 16 samples.

Then, such data and each information are recorded on a recording medium or the like, or transmitted directly to a decoding side to be reproduced.

In the above embodiment, the output from the multiplexer 25 or the multiplexer 26 is switched and selected according to the overflow signal from the noise shaping circuit 22.
The multiplexer 25 and the multiplexer 26 are united and provided with a selector function, and the audio data via the system having the first filter group or the system having the second filter group are provided in response to the overflow signal. Audio data may be selected.

As is apparent from the above description, the data compression encoding apparatus according to the present embodiment includes a system having the first filter group 2 having good sound quality, and a second filter having a slightly inferior sound quality but hardly causing encoding failure. Two systems including a system having a group 3 are provided, input signals of the same block are simultaneously supplied to the two systems, and data is output through a system having the first filter group having good sound quality. When the encoding failure occurs in the system having the first filter group, by outputting the data through the system having the second filter group 3 which hardly causes the encoding failure, the encoding inability is detected beforehand. Therefore, for example, data having an overflow or the like is not output as it is. In addition, since the signal for each block is processed in parallel by the system having the two filter groups, re-encoding becomes unnecessary, and real-time encoding can be performed.

In the above-described embodiment, when the encoding becomes impossible, the switch is automatically switched to select the system of the first filter group or the system of the second filter group.
In the case of a small system that uses P but has only a small memory or a real-time system that needs to reproduce encoded data immediately, one of the above two systems can be used as needed. Of course, only the system (for example, the system having the second filter group, etc.) may be used, or the switch may be manually switched if necessary. .

[Effects of the Invention] The data compression encoding apparatus according to the present invention blocks an input signal for each fixed sample, and supplies the input signal to a first filter group having a plurality of filters for each block.
A first system that selects a filter that can obtain the highest compression rate from among the filter groups of the above, compresses and encodes an output from the selected filter by a quantizer and a noise shaping circuit, and outputs the first system. Each sample is divided into blocks, and each block is supplied to a second filter group having a plurality of filters, and a filter capable of obtaining the highest compression ratio is selected from the second filter group. And a second system for compressing and encoding the output from the second system by a quantizer and a noise shaping circuit and outputting one of the output from the first system and the output from the second system. And a switching unit for adding the input signal to the quantization error fed back by the noise shaping circuit of the first system or the second system. By switching in response to the overflow signal indicating that the overflow occurs,
Even when an overflow occurs, it is not necessary to re-encode, and it is possible to switch to an output in which an overflow does not occur in real time, and it is possible to perform real-time encoding while preventing non-encoding from occurring.

Further, in the data compression encoding apparatus according to the present invention, the first filter group includes a filter having a high probability that output signal quality is good but encoding cannot be performed.
Includes a low probability of becoming unencodable, so that the encoding is normally performed by a system having the first filter group having good output signal quality, and the output is taken out. By taking the encoded output through a system with a low probability of the second set of filters becoming unencodable,
It is possible to supply an output with good quality and no occurrence of encoding failure.

[Brief description of the drawings]

FIG. 1 is a functional block diagram showing functions of a data compression encoding apparatus according to the present invention in a block diagram, and FIG. 2 is a schematic diagram showing a format of output data of the data compression encoding apparatus according to the present invention. . DESCRIPTION OF SYMBOLS 1 ... Input terminal 2 ... 1st filter group 3 ... 2nd filter group 4, 7 ... 0th-order difference filter 5, 8 ... 1st-order difference filter 6 ... 1st-secondary difference filter 9 ...... Second second-order difference filters 10, 11, 12 ... Range detectors 13, 14, 15 ... Range detectors 16, 17 ... Comparison blocks 18, 19 ... Selectors 20, 21 ... Quantizers 22, 23 Noise shaping circuit 24 Switch 25, 26 Multiplexer 27 Output terminal

Claims (2)

(57) [Claims] An input signal is divided into blocks by a predetermined number of samples and supplied to a first filter group having a plurality of filters for each block, and a filter capable of obtaining the highest compression ratio among the first filter groups. And a first system that compresses and encodes an output from the selected filter by a quantizer and a noise shaping circuit and outputs the result. The filter is supplied to a second filter group having a filter, and a filter that provides the highest compression ratio is selected from the second filter group, and an output from the selected filter is compressed by a quantizer and a noise shaping circuit. A second system for encoding and outputting; and a switch for selecting and outputting one of the output from the first system and the output from the second system. A switching stage, wherein the switching means responds to a quantization error fed back by the noise shaping circuit of the first system or the second system and an overflow signal indicating that an overflow at the time of addition of the input signal has occurred. A data compression encoding device characterized in that the data compression encoding device is switched. 2. The first filter group includes a filter having a good output signal quality but a high probability of being unable to encode, and the second filter group includes a filter having a low probability of being unable to encode. The data compression encoding apparatus according to claim 1, wherein: