JP3101118B2 - ADPCM codec - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION The present invention relates to a digital cordless
ADPCM codec used for telephones (PHP)
It relates to.

[0002]

2. Description of the Related Art Generally, an ADPCM codec has a transmitting side for converting an analog signal waveform into a digital compressed signal and a receiving side for converting a digital compressed signal into an analog signal waveform. The transmitting side includes an A / D unit for modulating an analog signal waveform into a PCM which is a digital signal;
Compression unit for the M signal to mu-Law companding compressed signal, composed of ADPCM encoder unit for encoding the mu-Law companding compressed signal to ADPC M signal. The receiving side, ADPCM decoder for decoding the ADPC M signal <br/> No. to mu-Law companding compressed signal, decompressing unit for decompressing the mu-Law companding compressed signal into a linear PCM signal, the PCM signal into an analog signal waveform It is composed of a D / A unit for decoding.

[0003]

As described above, since the conventional ADPCM codec is an AD (adaptive difference type) PCM, its signal pattern is obtained by converting the original linear PCM signal according to the amount of change in its amplitude. This is information obtained by converting the degree of the change into a coefficient. The ADPCM performs signal compression by replacing the coefficient with a coefficient as described above. At this time, when replacing a current signal waveform with a coefficient, the change is determined by comparing the signal waveform with a past signal waveform and converting the signal waveform into a coefficient. For this reason,
If the coefficient representing the past signal waveform is incorrect, the signal waveform to be currently decoded changes according to the change of the past signal waveform formation adaptive coefficient.

That is, the decoded signal waveform of the past is different from the original analog signal waveform before encoding, but the current signal waveform to be decoded is a decoded analog signal waveform different from the original analog signal waveform. Since the decoded analog signal waveform is compared with the component, the currently decoded analog signal waveform is different from the original analog signal waveform before encoding. At this time, among the adaptive coefficients of the ADPCM signal, if an error occurs in which a bit signal or the like that greatly changes the value of the coefficient is changed, the decoded analog signal waveform is greatly disturbed as described above, and sometimes unnaturally suddenly. An excessive signal (excessive impact sound) accompanied by a large signal or an abnormal signal (or a signal such as noise) may occur. When this signal is converted into sound by an acoustic element, if it is used as an earpiece, it will be perceived as an abnormally loud sound by human ears, and it will be very harsh, and in some cases it is expected that the ear may be hindered.

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and a digital limiter is applied to a decoded PCM digital signal before decoding and D / A conversion of the ADPCM signal before decoding. ADP that limits the level of the PCM signal that will appear as a signal so that no excessive signal is generated in the decoded analog signal waveform
It aims to provide a CM codec.

[0006]

According to the present invention, in order to achieve the above object, an ADPCM signal is converted into a linear PCM signal.
ADPCM decoder for coding and ADPCM decoder
Digital limit to the linear PCM signal output from the
And a digital limiter circuit
The digital limiter circuit determines that an error exists.
To a linear PCM signal corresponding to the ADPCM signal
It is characterized by applying a digital limiter .

[0007]

[Action] linear The ADPC M No. signal that is transmitted PCM
After decoding into a signal, the linear PCM signal is compared with a digital threshold level for comparing and judging the digital limiter, so that a linear PCM signal exceeding the threshold level becomes a signal level that does not become an excessive signal after decoding. After that, the decoded analog signal waveform after the D / A conversion is prevented from becoming an excessive signal.

[0008]

Embodiments will be described below with reference to the drawings. FIG. 1 is a block diagram for explaining an embodiment of an ADPCM codec according to the present invention. In FIG. 1, reference numeral 1 denotes a baseband processor (BB processor), and 2 denotes an ADPCM codec.
(Adaptive Differential Pulse Code Modulation) Codec, 3 is an ADPCM decoder, 4 is a digital limiter circuit, 5 is 8 bit μ-La
w A compression circuit, 6 is a D / A converter.

A description will be given of the case of the ADPCM transmission system in a digital cordless telephone (PHP). The baseband processor 1 receives the input CRC (Cyclic
Redundancy Check) The coded bit pattern signal is converted into a 4 bit 32 kbps ADPCM signal, and it is determined whether or not there is an error in the bit pattern sequence at that time. The information is used as a CRC signal. , When there is no output, the output is "L". The output from the baseband processor 1 and the ADPCM signal and the CRC signal are completely synchronized, and these two signals are input to the ADPCM codec 2 having the digital limiter circuit 4. The ADPCM codec 2 decodes 4-bit ADPCM signals into 14-bit linear PCM signals.
digital limiter circuit 4 for applying digital limiter to bit linear PCM signal, compression circuit 5 for converting 14-bit linear PCM signal to 8-bit μ-law compressed signal, D / A converter circuit for D / A converting μ-Law compressed signal to analog signal 6.

When the 4-bit ADPCM signal is input to the ADPCM decoder 3, the 4-bit ADPCM signal becomes 14
The bit is converted into a linear PCM signal and transmitted to the digital limiter circuit 4. The 14-bit linear PCM signal is transmitted, and the digital limiter circuit 4 simultaneously receives the CRC signal output from the baseband processor 1. The threshold level as the limit level in the digital limiter circuit 4 is determined in advance by external control or the like before these signals are input, and the determined threshold level and the 14-bit linear PCM are determined.
The signals will be compared. Determination condition of the digital limiter circuit 4, when an error is "L" of the CRC signal, i.e., when an error in CRC encoded bit pattern signal is not, in the case of this there is no error in the 14bit linear PCM signal, feed In such a case, the digital limiter circuit 4 does not limit even if the 14-bit linear signal exceeds the threshold level in such a case, and the input 14-bit linear PC
The M signal is transmitted to the compression circuit 5 faithfully.

On the other hand, when the error of the CRC signal is “H”,
That is, when there is an error in the 14-bit linear PCM signal, the digital limiter circuit 4 outputs the 14-bit linear PCM signal.
Compare the signal with the threshold level. At this time, 14bit
When the linear PCM signal is below the threshold level, the input signal is faithfully transmitted to the compression circuit 5 as described above. 1
When the 4-bit linear PCM signal exceeds the threshold level, the 14-bit linear PCM signal is manipulated, for example, by shifting the 14-bit PCM code by 3 bits so that the value becomes smaller, and reducing the value of the inputted 14-bit linear PCM signal. And transmits it to the compression circuit 5.

Thereafter, the 14-bit linear 14-bit signal transmitted from the digital limiter circuit 4 is converted into an 8-bit signal by the compression circuit 5 in accordance with μ-Law based on CCITT G.711 or the like.
The signal is converted into a PCM signal and transmitted to the D / A converter circuit 6. In the D / A converter circuit 6, the 8 bit
The PCM signal is once converted to a 14-bit linear PCM signal, and the 14-bit linear PCM signal is decoded and converted to an analog signal and output. At this time, since the analog signal generally has a step-like signal waveform, a low-pass filter (LPF) is used to remove these step-like components as harmonics and smooth the signal waveform. It is often inserted immediately after analog signal output.

[0013]

As is apparent from the above description, the present invention has the following effects. In other words, when an error occurs in the transmitted ADPCM signal and the resulting analog signal becomes excessive when decoded, the digital limiter unit appropriately limits the PCM signal decoded from the ADPCM signal before decoding. By applying the signal, the signal is decoded into an analog signal, and when the sound is heard by the acoustic element, the sound is heard without being excessively loud, so that the sound is not disturbed and the ear is not disturbed. . In addition, the detection method can appropriately correct only a portion where the ADPCM signal has an error by using the CRC signal. Therefore, even a signal having a large signal level but not abnormal is not corrected. Both the minimum signal and the maximum signal are decoded smoothly. It becomes very convenient.

[Brief description of the drawings]

FIG. 1 is a configuration diagram illustrating an embodiment of an ADPCM codec according to the present invention.

[Explanation of symbols]

1 ... baseband processor (BB processor), 2
… Is ADPCM (Adaptive Differential Pulse Code M)
odulation: adaptive differential pulse code modulation) codec, 3
... ADPCM decoder, 4 ... Digital limiter circuit,
5 ... 8bit μ-Law compression circuit, 6 ... D / A converter.

Claims (1)

(57) [Claims] An ADPCM signal is converted to a linear PCM signal.
ADPCM decoder for coding and ADPCM decoder
Digital limit to the linear PCM signal output from the
And a digital limiter circuit
The digital limiter circuit determines that an error exists.
To a linear PCM signal corresponding to the ADPCM signal
An ADPCM codec characterized by applying a digital limiter .