JP2002246913A - Data processing device, data processing method and digital audio mixer - Google Patents

Abstract

PROBLEM TO BE SOLVED: To dispense with round off processing in operational processing and to prevent adverse effects caused by it. SOLUTION: Input data Da to Dd are converted from a fixed decimal mode into a floating decimal mode through fixed/floating converters 103a 103d. Data Dae to Dde of a floating decimal mode are subjected to floating point arithmetic processing (equalizing, dynamics, adding, and fader processings) through signals processors 104a to 104d, an adder 105, and a signal processor 106 for obtaining data De0. The data De0 is subjected to limiter processing or compressor processing through an L/C part 107, and its maximum is limited to the maximum of data of fixed decimal mode or below, data outputted from the L/C 107 are converted from a floating decimal mode into fixed decimal mode through a floating/fixed converter 108 to obtain an output data Do. An operation is carried out in a floating decimal mode, so that a dynamic range can be expanded markedly, and round-down processing can be dispensed with in operation.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data processing apparatus and a data processing method suitable for processing audio data and the like, and a digital audio mixer using the same. Specifically, the input fixed-point format data is converted into floating-point format data, the floating-point format data is subjected to floating-point arithmetic processing, and the floating-point format data obtained by the arithmetic processing is converted to floating-point format data. The present invention relates to a data processing device or the like that converts data into fixed-point format and outputs the data, thereby eliminating the need for truncation processing in the course of arithmetic processing and preventing adverse effects due to the truncation processing.

[0002]

2. Description of the Related Art Hitherto, a digital signal processing device for performing a limiter process or a compressor (compression) process on digital audio data in a fixed-point format has been proposed (Japanese Patent Laid-Open No. Hei 6-334559). In this digital signal processing device, a gain value to be multiplied by input digital audio data is obtained by inputting logarithmic data obtained by expressing the input digital audio data in logarithmic expression or floating point as an address in a gain table. ing.

[0003]

When the fixed-point format digital audio data input to the digital signal processing device is obtained by fixed-point arithmetic processing, the value of the fixed-point arithmetic is maximized. If the value is exceeded, it is generally truncated and clipped to the maximum value.

Therefore, when performing the limiter processing or the compressor processing in the above-described digital signal processing apparatus, the value is fixed to a specified maximum value (0 [dB]) by fixed-point arithmetic.
As for the portion exceeding the maximum value, limiter processing or compressor processing is performed on the maximum value, as shown in FIG.

[0005] Therefore, in view of the slope of the input / output of the limiter process or the compressor process, even if the output value does not exceed the maximum value, the input value exceeds the specified maximum value. (Depicted by the broken line in FIG. 6), and the sound obtained from the digital audio data after the limiter processing or the compressor processing is distorted.

The sound distortion will be described with reference to a sound signal waveform. Assuming that the input audio signal to be subjected to the limiter processing or the compressor processing is clipped at the maximum value (0 [dB]) by the truncation processing to deform the waveform as shown in FIG. As shown in FIG. 7B, the output audio signal obtained by subjecting the audio signal to the limiter process or the compressor process has the waveform deformed by the clip as described above. Therefore, the sound based on the output sound signal is greatly distorted due to the influence of the clip. SUMMARY OF THE INVENTION An object of the present invention is to provide a data processing device and the like which can eliminate the need for truncation processing in the course of arithmetic processing and can prevent adverse effects due to the truncation processing.

[0007]

A data processing apparatus according to the present invention comprises: first data format conversion means for converting input fixed-point format data to floating-point format data; and a first data format. Data processing means for performing floating-point arithmetic processing on floating-point data obtained by conversion by the conversion means, and converting floating-point data obtained by processing by the data processing means to fixed-point data And a second data format conversion means for outputting the data.

[0008] A data processing method according to the present invention comprises:
A first step of converting input fixed-point format data to floating-point format data, and a second step of performing floating-point arithmetic processing on the floating-point format data obtained by the conversion in the first step Step and this second
And a third step of converting the floating-point format data obtained by the processing in the above step into fixed-point format data and outputting the data.

According to the present invention, fixed-point data, such as audio data and video data, is input, and the fixed-point data is converted to floating-point data. Then, floating point arithmetic processing is performed on the data in the floating point format. For example,
When the data is audio data, equalization processing, dynamics processing, fader processing, and the like are performed as floating-point arithmetic processing. In the floating-point arithmetic processing, the dynamic range can be greatly expanded as compared with the fixed-point arithmetic processing. Therefore, in the floating-point arithmetic processing, the truncation processing becomes unnecessary.

[0010] The floating-point format data obtained by the floating-point processing is converted into fixed-point format data and output. In this case, for example, floating-point format data is converted into fixed-point format data after its maximum value is suppressed to at least the maximum value of fixed-point format data by limiter processing or compressor processing.

As described above, the input fixed-point format data is converted into floating-point format data, and the calculation is performed in the floating-point format with the dynamic range greatly expanded. Truncation processing is not required. Therefore, an adverse effect due to the truncation processing can be prevented. For example, in the case of handling audio data, it is possible to prevent a large distortion from occurring in the sound due to the clip due to the truncation processing.

Further, a digital audio mixer according to the present invention includes a first data format converter for converting a plurality of input fixed-point audio data into floating-point format data, and a first data format converter. A first signal processing unit for performing a first floating-point arithmetic process on each of a plurality of audio data in the floating-point format obtained by conversion by the conversion unit; and a first signal processing unit for processing the audio data. An adder for adding a plurality of audio data in a floating-point format in a floating-point format, and a second signal for performing a second floating-point arithmetic processing on the audio data in the floating-point format obtained by adding the audio data in the adder The processing unit and at least the maximum value of the audio data in the floating-point format obtained by processing in the second signal processing unit is converted into at least the fixed-point format audio data. A third signal processing unit for performing a third floating point arithmetic processing for suppressing the audio data to be equal to or less than the maximum value of the audio data, and a floating point format audio data obtained by processing the third signal processing unit in a fixed point format And a second data format conversion unit that converts the audio data into audio data and outputs the audio data.

According to the present invention, a plurality of audio data in a fixed-point format are input, and the plurality of audio data in the fixed-point format are respectively converted into audio data in a floating-point format. Then, floating-point arithmetic processing is performed on the plurality of audio data in the floating-point format. For example, equalization processing, dynamics processing, and the like are performed.

Then, a plurality of audio data in the floating-point format obtained by the arithmetic processing are added in the floating-point format, and the floating-point arithmetic processing is performed on the added data. For example, fader processing is performed. further,
Floating point arithmetic processing (limiter processing or compressor processing) for suppressing the maximum value of the floating point audio data obtained by this arithmetic processing to at least the maximum value of the fixed point audio data.
Is performed.

In the floating-point arithmetic processing, the dynamic range can be greatly expanded as compared with the fixed-point arithmetic processing. Therefore, in each of the above-described floating-point arithmetic processes, the truncation process becomes unnecessary. The floating-point audio data obtained by the floating-point processing is converted into fixed-point audio data and output.

As described above, a plurality of input fixed-point format audio data are converted into floating-point format audio data, and operations are performed in a floating-point format in which the dynamic range is dramatically expanded. The truncation processing in the calculation process is not required. Therefore, it is possible to prevent the audio from being greatly distorted due to the clip due to the truncation processing.

[0017]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 shows a configuration of a digital audio mixer 100 as an embodiment. Digital audio mixer 100
Is an input terminal 101 for inputting audio signals Sa to Sd as input analog audio signals, respectively.
a to 101d and an A / D converter 10 for converting these audio signals Sa to Sd into fixed-point audio data Da to Dd as digital data, respectively.
2a to 102d, and a fixed / floating converter 103 for converting these fixed-point audio data Da to Dd into floating-point audio data Dae to Dde, respectively.
a to 103d.

The digital audio mixer 10
0 is audio data Dae to Dde in floating-point format obtained by conversion by the fixed / floating converters 103a to 103d.
Signal processing units 104a to 104e which perform equalization processing as floating point arithmetic processing and dynamics processing for
104d. In this floating-point arithmetic processing, the dynamic range can be greatly expanded as compared with the fixed-point arithmetic processing, and the truncation processing becomes unnecessary. This is the same in the other floating point arithmetic processing described below.

Here, the dynamics processing refers to processing for dynamically changing the output gain in accordance with the value of input data, and the above-described processing of the limiter and the compressor is one of the dynamics processing.
The dynamics processing performed in 4d is, for example, processing of an expander or a gate.

The digital audio mixer 10
0 indicates an addition unit 105 that adds floating-point format audio data output from the signal processing units 104a to 104d, respectively, and a floating-point format audio data output from the addition unit 105. A signal processing unit 106 that performs fader processing as decimal point calculation processing.

The digital audio mixer 10
0 is a limiter / compressor for performing a limiter process or a compressor process on the floating-point format audio data De0 output from the signal processing unit 106 to suppress the maximum value thereof to at least the maximum value of the fixed-point format audio data. A compressor section (L / C) section 107.

FIG. 2 shows an example of the configuration of the L / C unit 107. The L / C unit 107 is supplied with an input terminal 107a to which the input data Din is input, and input data Din to be input to the input terminal 107a, and generates a coefficient corresponding to the value of the input data Din. Part 107b;
A multiplier 107c for multiplying the input data Din input to the input terminal 107 by the coefficient value c generated by the coefficient generator 107b to obtain output data Dout;
and an output terminal 107d for outputting ut.

In the L / C unit 107a, as shown in FIG. 3, when the value of the input data Din is equal to or less than the predetermined value La, the input data Din and the output data Dout have the same value. When the value of the input data Din exceeds the predetermined value La, the output data Din is proportional to the value of the input data Din.
Is suppressed. In FIG. 3, the maximum value of the input data Din is represented by 0 [dB].

Although there is no clear definition of the limiter and the compressor, generally the limiter and the compressor region (FIG. 3)
, Where the slope of the straight line in the region where the value of the input data Din exceeds the predetermined value La) is k, the limiter is defined when 1 / k ≧ 10 is satisfied, and the compressor is defined when 1 <1 / k <10 is satisfied. Say

Returning to FIG. 1, the digital audio mixer 100 converts the floating point format audio data De1 output from the L / C unit 107 into fixed point format audio data Do.
08, and D for converting the fixed-point format audio data Do into an audio signal So as an analog signal.
A / A converter 109 and an output terminal 110 for outputting the audio signal So.

Next, the operation of the digital audio mixer 100 shown in FIG. 1 will be described. Input terminals 101a-1
Audio signals Sa to Sd as input analog audio signals are respectively input to 01d. These audio signals Sa to Sd are supplied to A / D converters 102a to 102d, respectively, and are converted into fixed-point audio data Da to Dd as digital data.

These audio data Da to Dd are supplied to fixed / floating converters 103a to 103d, respectively, and are converted into audio data Dae to Dde in floating point format. These audio data Dae to Dde
Are supplied to the signal processing units 104a to 104d, respectively, and are subjected to floating-point arithmetic processing, that is, equalization processing and dynamics processing.

As described above, the signal processing units 104a to 104d
Are supplied to the adder 105 and are added in the floating point format. The addition data from the addition unit 105 is
The signal is supplied to the signal processing unit 106 and subjected to floating-point arithmetic processing, that is, fader processing.

The audio data De0 in the floating-point format obtained by processing in the signal processing unit 106 is supplied to the L / C unit 107.
To perform a limiter process or a compressor process, and suppress the maximum value thereof to at least the maximum value of the fixed-point format audio data. As a result, the amplitude can be suppressed without clipping due to the truncation processing.

The floating-point audio data De1 output from the L / C unit 107 is supplied to a floating / fixed conversion unit 108 and converted into fixed-point audio data Do. In this case, as described above, since the maximum value of the floating-point format audio data De1 is suppressed to be equal to or less than the maximum value of the fixed-point format audio data by the L / C unit 107, the floating-point format audio data De1 is fixed. It can be well represented as audio data in decimal format.

The fixed-point format audio data Do converted and output by the floating / fixed conversion unit 108 is
The signal is supplied to the A converter 109 and converted into an analog audio signal So. Then, this audio signal S
o is the output terminal 11 as an output analog audio signal.
Output to 0.

As described above, in the digital audio mixer 100 shown in FIG.
04a to 104d, the addition unit 105, the signal processing unit 106, and the L / C unit 107 perform floating-point arithmetic processing. In this floating-point arithmetic processing, the dynamic range can be significantly increased as compared with the fixed-point arithmetic processing. Therefore, the signal processing units 104a to 104d, the addition unit 105,
The signal processing unit 106 and the L / C unit 107 do not need to clip to the maximum value by the truncation process even when the value exceeds the specified maximum value (0 [dB]).

Therefore, when the limiter processing or the compressor processing is performed in the L / C unit 107, as shown in FIG. 4, even if the part exceeds the specified maximum value (0 [dB]), the original limiter processing or the compressor processing is performed. Processing is performed along the processing curve. Therefore, the sound based on the audio signal So output from the output terminal 110 has no distortion due to the clip.

The sound distortion will be described using a sound signal waveform. In the digital audio mixer 100 shown in FIG. 1, the input audio signal to be subjected to the limiter process or the compressor process in the L / C unit 107 may exceed the prescribed maximum value (0 [dB]) as shown in FIG. 5A. Clipping by truncation processing is not performed, and there is no waveform deformation due to clipping. Therefore, as a matter of course, an output audio signal obtained by performing a limiter process or a compressor process on the input audio signal is also
As shown in FIG. 5B, there is no waveform deformation due to the clip. Therefore, the sound based on the output sound signal has no distortion due to clip processing.

Although the four-input digital audio mixer 100 has been described in the above embodiment, the number of inputs is not limited to this, and the content of signal processing, the order of signal processing, and the like are also described above. However, the present invention is not limited to the above. Further, in the above-described embodiment, the case where each signal processing is realized by hardware has been described, but it is needless to say that these processing can be realized by software.

In the above embodiment, the digital audio mixer 1 for handling audio data is used.
00, the present invention can of course be applied to a data processing device that handles other data, for example, video data. Further, in the above-described embodiment, the A / D converters 102a to 102d and the D / A converter 109 are arranged on the input side and the output side of the digital audio mixer 100 to correspond to the analog input and the analog output. Is a digital input,
Those corresponding to digital output can be similarly considered.

[0037]

According to the present invention, the input fixed-point format data is converted into floating-point format data, and the floating-point format data is subjected to floating-point arithmetic processing. The obtained data in the floating-point format is converted into data in the fixed-point format and output. The truncation process does not need to be performed in the course of the arithmetic processing, and adverse effects due to the truncation process can be prevented.
For example, according to the apparatus that handles audio data, it is possible to prevent a large distortion from being generated in sound due to a clip resulting from the truncation processing.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a digital audio mixer as an embodiment.

FIG. 2 is a block diagram illustrating a configuration of a limiter / compressor unit.

FIG. 3 is a diagram for explaining input / output characteristics of a limiter and a compressor.

FIG. 4 is a diagram for explaining a limiter and compressor process.

FIG. 5 is a waveform diagram for explaining a limiter and compressor process (floating point calculation).

FIG. 6 is a diagram for explaining a conventional limiter and compressor process.

FIG. 7 is a waveform diagram for explaining conventional limiter and compressor processing (fixed point calculation).

[Explanation of symbols]

100: Digital audio mixer, 101a
... to 101d ... input terminals, 102a to 102d ...
A / D converters, 103a to 103d: fixed / floating conversion units, 104a to 104d, 106: signal processing units, 105: addition units, 107: limiter / compressor units (L / C units) , 108: floating / fixed converter, 109: D / A converter, 110: output terminal

Claims (10)

[Claims] 1. A first data format converter for converting input fixed-point format data into floating-point format data, and a floating-point format data obtained by conversion by the first data format converter. Signal processing means for performing floating-point arithmetic processing on the data, and second data format conversion means for converting floating-point data obtained by the signal processing means into fixed-point data and outputting the data. A data processing device comprising: 2. The signal processing means includes a limiter or a compressor at a final stage that suppresses a maximum value of floating-point format data obtained in the immediately preceding process to at least a maximum value of the fixed-point format data. The data processing device according to claim 1, wherein 3. The data processing according to claim 1, wherein said data is audio data, and said signal processing means performs at least one of equalization processing, dynamics processing and fader processing. apparatus. 4. A first step of converting input fixed-point format data into floating-point format data, and a floating-point operation on the floating-point format data obtained by the conversion in the first step. Data comprising: a second step of performing processing; and a third step of converting floating-point format data obtained by processing in the second step into fixed-point format data and outputting the data. Processing method. 5. The floating-point arithmetic processing in the second step includes a step of suppressing a maximum value of the finally obtained floating-point format data to be equal to or less than a maximum value of the fixed-point format data. The data processing method according to claim 4, wherein: 6. The data processing apparatus according to claim 1, wherein the data is audio data, and the floating point arithmetic processing in the second step includes at least one of equalizing processing, dynamics processing, and fader processing. 5. The data processing method according to 4. 7. A first data format conversion unit for converting a plurality of input fixed-point format audio data into floating-point format audio data, and a data format obtained by the first data format conversion unit. A first signal processing unit for performing a first floating-point operation on each of a plurality of audio data in a floating-point format; and a plurality of audio data in a floating-point format obtained by processing in the first signal processing unit , A second signal processing unit for performing a second floating-point arithmetic process on audio data in the floating-point format obtained by adding the second data in the floating-point format, The maximum value of the audio data in the floating-point format obtained by processing in the signal processing unit is suppressed to at least the maximum value of the audio data in the fixed-point format. A third signal processing unit for performing a third floating-point operation process for converting floating-point audio data obtained by processing in the third signal processing unit into fixed-point audio data and outputting And a second data format conversion unit. 8. The digital audio system according to claim 7, wherein the first signal processing unit performs an equalization process and a dynamics process, and the second signal processing unit performs a fader process. mixer. 9. Converting a plurality of input analog audio signals into audio data in a fixed-point format,
8. The digital audio mixer according to claim 7, further comprising an analog / digital converter for obtaining a plurality of audio data in the fixed-point format to be input to the first data format converter. 10. A digital / analog converter for converting the fixed-point format audio data output from the second data format converter into an analog signal to obtain an output analog audio signal. The digital audio mixer according to claim 7, wherein: