JPH1051318A - Data converter - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent sound quality degradation by the hard clip of reproduced signals and to improve the distortion feeling of reproduced sound by outputting the data of a level not reaching saturation as they are without performing nonlinear compression, performing limitation to the area of a data part to be saturated and nonlinearly compressing the data. SOLUTION: The data inputted from a data IN 1 are stored in a memory 2. For the memory 2, in order to delay the data, a shift register and a FIFO memory are used. Output data for outputting signals delayed by the memory 2 as they are and compressed data for which they are nonlinearly compressed in a compression circuit 3 are obtained. It is detected that the data inputted to the memory 2 exceed a saturation level and the compressed data are selected in a selector circuit 4 and outputted. When the inputted data become lower than the saturation level, the output data of the memory 2 are selected in the selector circuit 4 and outputted after about the double of memory delay time. Thus, only the data provided with the saturation level are outputted as the nonlinearly compressed data.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a technique for reducing sound quality degradation due to hard clipping of a reproduced signal due to saturation of data when the level of data such as digital audio is increased in bit length conversion or the like and saturation of an analog amplifier or the like. Intercept,
The present invention relates to a data conversion device that improves the sense of distortion of reproduced sound.

[0002]

2. Description of the Related Art Conventionally, digital audio data recorded on a compact disk or the like has a signal-to-noise ratio (S / N).
Is recorded at the largest possible data level in order to increase the Generally, in master recording of music or the like, recording is performed with multiple bits, for example, high bits such as 20 bits or 24 bits. When the high bit data is used as predetermined bit data such as 16 bits of a CD, the lower bit of the high bit data recorded in the master recorder is often used by being discarded.

[0003] In master recording, it is common to dislike clips caused by saturation of the data level and to record at a low recording signal level. Therefore, the signal level recorded on a CD or the like is also an average output when reproduced. The sound pressure level will be low. When editing and recording such sources,
This is solved by paying attention to the recording so that the signal level is converted to the largest possible signal level.However, when the recording signal level saturates when the sound level of the music source is large, clipping occurs when DA conversion is performed. The sound quality is deteriorated due to abnormal noise. In order to solve this feeling of distortion, a soft clip method is used before the large-amplitude part of the analog data is converted into a digital signal so as to compress non-linearly and prevent abrupt clipping.

[0004]

In the above data compression, since the data is compressed from a level considerably smaller than the saturation level through a smooth nonlinear amplifier, the level of the data is not saturated but the peak portion of the data falls in the nonlinear region. Then, the long section data is compressed, which causes deterioration in the sound quality of music in this section.

According to the present invention, data at a level that does not reach saturation is not subjected to non-linear compression, but is output as it is, and limited to an area of a saturated data portion, and data is non-linearly compressed. .

[0006]

According to the present invention, there is provided a storage means for storing digital data in a memory, a compression means for compressing data output from the storage means, and wherein the level of the digital data exceeds a predetermined level. The data conversion device comprises a detecting means for detecting the above, and an output switching means for switching between the output of the storage means and the output of the compressing means based on the output of the detecting means.

According to the present invention, a storage means for storing digital data in a memory, a compression means for compressing data output from the storage means, and detecting that the level of the digital data exceeds a predetermined level. A data conversion apparatus comprising: a detection unit; and an output switching unit that switches the output of the storage unit and the output of the compression unit based on the output of the detection unit by cross-fading.

[0008]

BEST MODE FOR CARRYING OUT THE INVENTION In the vicinity of a section including a section where input digital data is clipped, data is switched between data passed through a compression means for making at least a gradual change in peak data and data not clipped.
As a result, the normal level data that is not clipped has a linear linear output characteristic without being compressed even if it fluctuates at a level close to saturation, and the data at the saturation level becomes compressed data that changes slowly.

The data converter of the present invention stores data in a memory and obtains output data that outputs a delayed signal as it is from the memory and compressed data that has been nonlinearly compressed by a compression circuit. Detecting that the data input to the memory exceeds the saturation level, the data selector selects and outputs the compressed data. When the input data falls below the saturation level, the data selector selects and outputs the output data of the memory means approximately twice after the memory delay time. As a result, only data having a level of saturation is output as non-linearly compressed data.

An embodiment of the present invention will be described with reference to the drawings.
FIG. 1 is a block diagram showing one embodiment of the data compression device of the present invention. FIG. 2 is a waveform diagram expressing the data of each part in FIG. 1 as an analog waveform.

The data input from the data IN 1 is stored in the memory 2. The memory 2 has a shift register or a FIFO (First-In) for the purpose of delaying data.
First-Out) memory is used. The storage capacity of the memory 2 is a capacity capable of storing data for a time required for non-linear compression or the like, and generally a capacity capable of storing data of several ms to several tens ms.

The data c delayed by the memory 2 is
The data is compressed by the compression circuit 3 to the saturation level V. For this compression circuit, a conversion circuit using a non-linear data table or the like can be used.

The input data 1 is input to a level detector 5, and the level detector 5 detects a section where the level of the input data exceeds a saturation level. The level detector 5 can be easily realized by a data comparator or the like. The output section of the signal b detected by the level detector 5 is expanded as shown by the output signal e via the hold circuit 6 for time expansion. The hold circuit 6 can be realized by a retriggerable monostable multivibrator.

The extension of the output section by the hold circuit 6 is as follows.
Delay time t due to storage and readout time of memory 2
2t, which is twice as large as The selector circuit 4 selects the compressed data d obtained from the compression circuit 3 by the output e of the hold circuit 6, and outputs the output data f as the data OUT7.

In the selector circuit 4, it is preferable to perform not only simple switching but also cross-fading. Since the levels of the data c and the compressed data d change suddenly, the selector circuit 4 switches the data by smoothly cross-fading, and smoothly connects the data to obtain the output data f as the data OUT7. The description of the selector circuit for performing the cross-fade is well known and thus will be omitted. However, it can be realized by a circuit that adds the fade-out signal and the fade-in signal in a superimposed manner.

The operation of the data converter according to this embodiment will be described with reference to the waveform diagram shown in FIG. The data input from the data IN1 is digital data expressed as an analog-converted waveform as shown in a, and compresses the level to a saturation level V or less. For this reason, the level detector 5 detects a section exceeding the saturation level V. On the other hand, the data a is stored and read out in the memory 2 so that the time t
After a delay, the data is non-linearly compressed by the compression circuit 3 so that the maximum level of the data becomes equal to or lower than the saturation level V as shown by d, and is compressed so that the waveform changes smoothly.

The compressed data d is compressed to a level lower than the saturation level V in a portion exceeding the saturation level V, but is non-linearly compressed. Level compression is performed in a non-linear manner so that The output “b” of the level detector 5 is extended and held for about 2t−t′2 (t′2 is a cross-fade period) by the hold circuit 6 and is used as a switching signal “e” of the selector circuit 4. A switching signal e is generated, and this signal is
, The output data f output from the selector circuit 4 is cross-fade from the data c to the compressed data d compressed by the compression circuit 3 in the t′1 section from the leading edge, and when the switching signal e becomes L The output data f output from the selector circuit 4 is switched from the compressed data d to the original data c by cross-fading in a period t'2 from the subsequent edge.

Here, when the saturation level V is detected,
The time t is before the time when the data c exceeds the saturation level V through the memory, and the data output switched by the selector circuit 4 is cross-fade within t′1 from the leading edge of the switching signal e as indicated by f. Switching is performed smoothly, and data is compressed.

When a time T elapses from the point when the level detector 5 detects a section in which the level of the data a input from the data IN1 exceeds the saturation level V, and the level of the input data a becomes lower than the saturation level V, the memory 2 Is held by the hold circuit 6 until the point of time 2t-t'2, which is about twice the delay time t, and the switching signal e is output. Switching signal e
Becomes L, the selector circuit 4 outputs data that is cross-fade from the leading edge to the output data c of the memory 2 within t′2 and is not smoothly compressed. Since the switching end timing at this time is T + 2t, the relationship is extended by substantially the same time t as the leading edge of the switching. The sections t'1 and t'2 of the crossfade are set to be shorter than the delay time t in order to make the connected data have a smooth waveform change.

Although the detection level of the data a by the level detector 5 has been described as the saturation level V, the same effect can be obtained by setting the detection level to a level close to the saturation level.

As a result, the data f output as the data OUT7 is not distorted as it approaches the saturation level V and becomes non-linear in the past. Is compressed only in the T + 2t section near. When the output data f is DA converted and amplified by an audio amplifier or the like to obtain a sound output, the audio data including the clipped portion and amplified to increase the average bell suppresses distortion in a data clip section and has a high S / N ratio. This is suitable for bit-down editing for editing and recording a recording source such as a DA converter or a CD. The data conversion device of the present invention is used in a data compression circuit or a clipping circuit such as an amplifier. The data conversion device is connected to a subsequent stage because there is no need to limit a section in which distortion due to data compression occurs and there is no fear of generating a sharp clip noise. In such a device, the amplification degree can be increased without worrying about the occurrence of noise.

[0022]

According to the present invention, the signal below the level that is saturated and clipped does not change, and only the signal in the vicinity of the section exceeding the clipped saturation level is compressed. Can be smoothly compressed to prevent generation of abnormal noise during reproduction.

[Brief description of the drawings]

FIG. 1 is a block diagram of a data conversion apparatus according to an embodiment of the present invention.

FIG. 2 is an imaginary diagram of a waveform of each unit based on an analog waveform.

[Description of Signs] 1 data IN 2 memory 3 compression circuit 4 selector circuit 5 level detector 6 hold circuit 7 data OUT

Claims (2)

[Claims] A storage unit for storing digital data in a memory; a compression unit for compressing data output from the storage unit; a detection unit for detecting that the level of the digital data has exceeded a predetermined level; A data conversion device comprising output switching means for switching between the output of the storage means and the output of the compression means based on the output of the detection means. 2. A storage means for storing digital data in a memory, a compression means for compressing data output from the storage means, a detection means for detecting that the level of the digital data has exceeded a predetermined level, A data conversion device comprising output switching means for switching the output of the storage means and the output of the compression means based on the output of the detection means by cross-fading.