JPH0644200B2 - Audio recording / playback device - Google Patents

Description

TECHNICAL FIELD The present invention relates to an audio recording / reproducing apparatus for recording quantized audio data in a magnetic disk device and reproducing the quantized audio data from the magnetic disk device.

(Prior Art) Conventionally, this type of audio recording / reproducing apparatus has been used for C radio such as AM
It is often used for audio recording / playback means of the M-unified editing device. In this case, an audio recording / reproducing S / N ratio of about 53 dB and a frequency characteristic of about 20 to 10 KHz are often used. In order to satisfy such a condition, a pulse signal sampled at a frequency of 20 to 22 KHz is used for 12
It is considered that it is necessary to quantize into a binary code of about bits and to record and reproduce. However, when recording a signal having a strong correlation between successive sampling values such as an audio signal, it is known that the amount of information may be about 6 bits per sample.

An excellent method for compressing and encoding this binary code is ADPCM (Adaptive Differential Quantization).
PCM) code. In this ADPCM code, a voice signal is divided into certain sections, the maximum value of its amplitude is recorded as a maximum value signal, and the section is indicated by a difference proportional to the maximum value signal.

Although this ADPCM code is a very effective method for compressing information, when the recorded ADPCM code is converted into a binary code and further converted into a voice signal, the recorded information is erroneously recorded and the ADPCM code is converted into a binary code. It may be converted into an audio signal by the means for converting into. In this case, if the correct voice is not reproduced after the portion where the information is erroneously transferred and the limit of the reproduction characteristic is exceeded by the means for converting the binary code into the voice signal, the voice waveform is clipped and reproduced. There was a drawback that the sound was distorted.

(Object of the invention) An object of the present invention is to eliminate the above drawbacks and to insert a cross code indicating that a voice signal has crossed a reference potential into an ADPCM code group to generate an erroneous voice waveform. Another object of the present invention is to provide a voice recording / reproducing apparatus which can obtain a correct voice waveform again from the position of the cross code.

(Structure of the Invention) The structure of the audio recording / reproducing apparatus of the present invention comprises: AD conversion means for converting audio into a binary code; ADPCM conversion means for converting the converted binary code into an ADPCM code; Crossing code generating means for detecting a crossing point with a potential and extracting a crossing code insertion timing;
Cross code insertion means for time-divisionally inserting a cross code in the M code, a recording device, a writing means for writing the quantized audio data in which the cross code is inserted into the recording device, and the quantization from the recording device. Reading means for reading voice data; cross code detecting means for detecting a cross code from the read data; and binary code converting means for converting the ADPCM code into a binary code while correcting the level of the ADPCM code by the detected cross code. A DA converting means for converting the binary code into an audio signal is included, and the audio is reproduced by using the cross code inserted at the time of recording with the cross code as a reference point.

(Example) Next, the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram of an embodiment of the present invention. In this embodiment, AD conversion means 1 for converting a voice signal into a binary code.
ADPCM code conversion means 2 for converting this binary code into an N-bit ADPCM code, and a cross code generation for generating a zero crossing code 2 ^N-1 when the most significant bit of the binary code is examined and a zero level is crossed. Means 3, cross code inserting means 4 for inserting the cross code and the absolute value of the immediately preceding voice data, a writing circuit 5 for writing the quantized voice data in a magnetic disk device 6 as a recording device, and a magnetic disk device. 6
A read circuit 7 for reading the quantized voice data from the magnetic disk device 6, a cross detection code means 8 for detecting a cross code from the voice data, and a binary conversion means 9 for converting the read ADPCM code into a binary code. And a DA conversion means 10 for converting this binary code into an analog signal and converting it into an audio signal, an input terminal 11 and an output terminal 12.

First, the audio signal input from the input terminal 11 is converted into a binary code by the AD conversion means 1. This binary code is converted into an ADPCM code by the ADPCM conversion means 2 and transferred to the cross code insertion means 4. On the other hand, the cross code detection means 3 detects the ground potential from the binary code, detects the insertion point (timing) of the cross code, and notifies the cross code insertion means 4 of the detection point. At this time, the cross code detecting means 3 checks the most significant bit of the binary code at each two points, and if the code is inverted, it is detected as a cross point.

Further, in the cross code insertion means 4, the quantization is performed according to the maximum value of the cross code and the binary code at the timing sent from the cross code generation means 3 in the ADPCM code group transferred from the ADPCM exchange means 2. A recording signal is created by inserting the absolute value of the immediately preceding audio data.

The cross code inserted at this time is "2 ^N-1 " when the audio data is represented by an N-bit ADPCM code.
This is the minimum value when the ADPCM-converted voice data is expressed by "2's complement", and it is from-(2 ^N-1 -1) to + (2
^{This is because N−1} −1) can be expressed by connecting them with N bits. The recording signal created by the cross code inserting means 4 is recorded on the magnetic disk device 6 via the means 5.

On the other hand, at the time of reproduction, the signal stored in the magnetic disk device 6 is transferred to the cross code detecting means 8 via the read circuit 7 to detect the cross code, and the ADPCM code with the cross code deleted and the cross code are detected. The timing is transferred to the binary conversion means 9. The binary conversion means 9 converts the ADPCM code sent from the crossing signal detection means 8 into the original binary code and corrects it at the crossing code detection timing. The binary code thus obtained is converted into an analog signal voice by the DA conversion means 10 and output to the output terminal 12.

The error correction when the binary conversion means 9 converts the ADPCM code into the binary code is performed as follows.

First, it is checked whether the transferred signal is a cross code, and if it is not a cross code, it is unconditionally converted into a binary code and transferred to the DA conversion means 10. At this time, if there is a cross code, the next AD
When the PCM code is once converted into a binary code and the most significant bit is changed, the binary code obtained by assuming that the zero crossing is performed is transferred to the DA conversion means 10. If the most significant bit has not changed, the ADPCM code causes a bit inversion during transfer, so correction is performed. In this case, a binary code is newly created from the absolute value of the audio data written next to the cross code and the ADPCM code immediately after the cross code, and the corrected binary code is transferred to the DA conversion means 10.

FIGS. 2 (A), (B), and (C) are a waveform diagram and a data diagram showing how the cross code is inserted and reproduced by the embodiment of FIG. In the figure, a to s are audio sample points,
Z ₁ and Z ₂ are the insertion points of the cross code, Fig. 2 (A) is the input waveform of the voice, Fig. 2 (B) is the corresponding quantized voice data, and Fig. 2 (C) is the output of the voice. It is a waveform. An insertion point Z ₁ shows a case where a cross point is inserted after i when the sample point i is zero, and an insertion point Z ₂ is a sample where the code changes when the sample point is not zero. A cross code is inserted between the points p and q. In this way, the ADPCM voice data with the cross code inserted (Fig. 2 (B) outputs the output waveform of the voice like the sample points of the voice from a'to s' (Fig. 2 (C)). The cross code inserted in the converted ADPCM voice data is deleted during voice reproduction and does not appear in the output waveform.

3 (A) and 3 (B) are time charts showing the state of signal correction according to the present embodiment. FIG. 3 (A) is a voice waveform in a conventional device, and FIG. 3 (B) is this embodiment. The audio waveform in the example is shown. It is assumed that the audio data transferred from the magnetic disk device 6 is erroneously transferred at the timing E.
If this voice data is incorrect, the conventional device (Fig. 3 (A))
In that case, the playback limit of the audio is exceeded and the waveform is clipped, and since ADPCM is used, such a state of waveform distortion continues until the end of this audio.

On the other hand, in the present embodiment, the waveform is clipped from the timing E until the zero-crossing code Z is detected, but after the zero-crossing code Z is detected, the waveform is reproduced to that point zero. After the zero-crossing code Z, correct sound can be reproduced as shown in FIG. 3 (B).

Next, the operation of this embodiment will be described in detail in association with the waveform diagram.

FIG. 4 is a flow chart showing a state at the time of recording by means 2, 3 and 4 of the present embodiment, and FIG. 5 is a waveform diagram at the time of code generation in FIG. Suppose the waveform to be recorded is given by the function (p).

First, in step 21, the maximum value M ₁ in the area T
To detect. In FIG. 5, M ₁ = (P4). Next, in step 22, the values of the points P _{1 to} P ₁₃ of the section T are set to M ₁
To normalize and convert to PCM. That is, when each point is P _N , (P _N ) / M ₁ is converted to PCM. The zero-crossing points are detected by detecting the points where the positive and negative signs of the PCM data thus obtained change. (Step 23). In the figure, point P ₇ -P
There is a crossover Z ₁ between ₈ . (Corresponding to 3 in the flow chart) On the other hand, the data obtained by encoding the points P _{1 to} P ₁₃ is C ₁ to C.
₁₃ , the difference value D _{n at} each point is D _n −C
It is given by _n−1 (n ≧ 2). The difference data D _n and the zero-crossing data Z _n thus obtained are, for example, the seventh
As shown in Figure (A), it is packaged (step 24),
It is recorded in the magnetic disk device 6.

FIG. 5 shows a flow chart during voice reproduction by means 8 and 9 of this embodiment, and FIGS. 7A and 7B show data and waveform charts at the time of reproduction. First, in step 31, a recording code is input, and in the ADPCM-converted audio data D ₁ and D ₂ , M _k
It is played by _{(D n -D n-1)} + D n-1. In the figure, when the cross code Z is detected after converting the data D ₆ , in step 32, if there is the cross code Z, the process branches to the step 34 side. If there is no crossing signal Z, it is converted to the binary code in step 33. Here, the next maximum value M ₂ is read once, and then the data D ₇ is read.
If the ADPCM data of this data D ₇ is converted into a binary code in step 35, and if the signal is judged to be a zero crossing in step 36, YES is obtained in the branch of step 36.
And the binary code is considered correct.

Next, generated any error, between the FIG. 7 (B) of the data D ₁₃ and D _14, and determines that it has not the zero crossing despite there is a zero crossing sign Z. In this case, since some error has occurred, the correction operation of step 37 is started.

In this case, the binary code of the data D ₁₃ has an error, and it is meaningless to calculate the next point based on this binary code. Therefore, the point corresponding to the data D ₁₃ is forcibly set to zero, the binary code is calculated, and the points thereafter are corrected. Even if the zero point is not completely corrected at this time,
Next, since the correction is performed again at the error detection zero point, the waveform is not greatly distorted beyond the reproduction limit.

In this method, the amount of recording signal for inserting the zero-crossing code increases, but the sampling rate of voice is 100 μm.
In contrast to sec / sample, since the data of the disk is taken in at 1 μsec / sample, the data does not affect the waveform.

(Effects of the Invention) As described above, the present invention inserts a cross code indicating an intersection at a position where the reference potential of the ADPCM code group intersects, thereby stably reproducing audio even when the ADPCM code is used. You can

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment of the present invention, FIG. 2 (A),
(B) and (C) are diagrams showing the input waveform, data, and output waveform of the present embodiment, and FIGS. 3 (A) and (B) are waveform diagrams showing examples of the conventional method and the waveform correction of the present embodiment, FIG. 4 is a flow chart when converting to a recording code of this embodiment, FIG. 5 is a waveform diagram explaining FIG. 4, FIG. 6 is a flow chart at the time of reproducing a binary code of this embodiment, and FIG. 7 (A). , (B) are data and waveform diagrams for explaining FIG. 6. In the figure, 1 ... (voice) AD conversion means, 2 ... ADPCM conversion means,
3 ... Cross code generation means, 4 ... Cross code insertion means, 5
...... Writing circuit, 6 ... Magnetic disk device, 7 ... Read circuit, 8 ... Cross code detecting means, 9 ... Binary code converting means, 10 ... DA converting means, 11 ... Input terminal, 12
...... Output terminals, 21 to 24 ...... Code conversion step, 31
~ 37 ... code reproduction step.

Claims (1)

[Claims] 1. An AD conversion means for converting voice into a binary code, an ADPCM conversion means for converting the converted binary code into an ADPCM code, and an intersection code insertion by detecting an intersection of the binary code and a ground potential. Cross-code generation means for extracting timing, cross-code insertion means for time-divisionally inserting cross-codes in the ADPCM code, recording device, and writing for writing electronic voice data in which the cross-codes are inserted in the recording device. Inserting means, reading means for reading the digitized voice data from the recording device, cross code detecting means for detecting a cross code from the read data, and level correction of the ADPCM code by the detected cross code. However, it includes a binary code converting means for converting the binary code into a binary code and a DA converting means for converting the binary code into an audio signal. Audio recording and reproducing apparatus characterized by reproducing audio based on the intersection code during reproduction by the code.