Sound coding technology 1.G.721 ADPCM encoder ADPCM is a waveform coding technique that uses high correlation between samples and quantization order adaptation to compress data. 2. Subband coding (SBC) The main process of subband coding is: The benefits of using separate encoding for each subband: The G.711 standard with a sampling rate of 8 kHz, 8 bits / sample, and a data rate of 64 kb / s is a codec standard developed by CCITT for a voice signal frequency of (300 to 3400) Hz. This is a narrowband audio signal coding. Modern voice coding technology has been able to reduce the data rate without significantly reducing the sound quality. The 8KHz sampling rate recommended by CCITT, 4 bits per sample, the G.721 standard at 32 kb / s, and G.723, the extended standard of G.721, all illustrate the progress of voice compression coding technology. G.722 is the audio signal encoding and decoding standard recommended by CCITT. The standard describes the coding principle, algorithm and calculation details of the audio signal with a bandwidth of 7kHz and a data rate of 64kb / s. 4.G.722 SB-ADPCM codec In order to meet the growing urgent need for video conference calls, CITT formulated the G.722 recommended standard for this in 1988, which is called 7KHz audio signal coding with a data rate of 64kb / s-7kHz Audio-coding with 64kb / s. This The standard improves the quality of voice signals from phone quality to AM radio broadcast quality, while the data transmission rate remains at 64kb / s. 5. Linear predictive coding (LPC) Linear predictive coding is a very important coding method. In principle, LPC analyzes the voice waveform to generate the parameters of the channel excitation and transfer function. The encoding of the sound waveform is actually converted into the encoding of these parameters, which greatly reduces the amount of sound data. At the receiving end, the parameters analyzed by LPC are used to reconstruct the speech through a speech synthesizer. Board-To-Board Connectors,Machine-Pin Board-To-Board Connectors,Gold Plated Board-To-Board Connectors,8 Pin Board-To-Board Connectors Shenzhen Jinyicheng Electronci Technology Co.,Ltd. , https://www.jycconnectors.com
CCITT has formulated the G.721 recommended standard for this purpose. This standard is called 32kb / s Adaptive Differential Pulse Code ModulaTIon to 24 and 40kb / s for Digital Circuit MulTIplicaTIon Equipment Application. The data rate of the encoder using this standard can be reduced to 40kb / s and 24kb / s.
The G.721 ADPCM standard recommended by CCITT is a transcoding system. It uses ADPCM conversion technology to realize the conversion between 64kb / s A-law or μ-law PCM rate and 32kb / s rate.
â‘ Use a set of band-pass filters (BPF) to divide the frequency band of the input audio signal into several consecutive frequency bands, each frequency band is called a sub-band.
â‘¡A separate coding scheme is used to encode the audio signal in each subband.
â‘¢When transmitting on the channel, the codes of each sub-band are combined.
â‘£When decoding at the receiving end, the codes of each sub-band are decoded separately, and then they are combined to restore the original audio signal.
· First. Adaptive control of each sub-band signal, the size of the quantization step (quantization step) can be adjusted according to the energy level of each sub-band. Subbands with higher energy levels are dequantized with large quantization steps to reduce the total quantization noise.
· Second, according to the importance of each sub-band signal in perception, different number of bits can be assigned to each sub-band to represent each sample value. For example, in the low-frequency sub-band, in order to protect the structure of tones and formants, a smaller quantization order and more quantization stages are required, that is, more bits are allocated to represent the sample value. Friction and noise-like sounds in speech usually appear in high-frequency subbands, and fewer bits are assigned to it.
3. Sub-band-adaptive differential pulse code modulation (SB-ADPCM)
The main goal of G.722 is to maintain a data rate of 64kb / s, and the quality of the audio signal is significantly higher than that of G.711. The G.722 standard increases the sampling frequency of audio signals from 8kHz to 16KHz, which is twice the sampling rate of G.711PCM, so the frequency of the signal to be encoded is extended from the original 3.4kHz to 7kHz. This has greatly improved the quality of audio signals, from the voice quality of digital phones to the quality of AM radio broadcasting. As far as the quality of dialogue signals is concerned, there is not much improvement in increasing the sampling rate, but for signals such as music, the quality is greatly improved.
The G.722 coding and decoding system adopts self-adaptive differential pulse code modulation technology, and divides the frequency band into two equal-bandwidth sub-generations, namely high-frequency sub-band and low-frequency sub-band. The signal in each subband of equal bandwidth is encoded with ADPCM.
Broadband speech refers to speech with a bandwidth of (50 ～ 7000) Hz. This kind of speech has a significant increase in intelligibility and naturalness compared to speech with a bandwidth of (300 ～ 3400) Hz. It is also easier to recognize the other party's speech people.
The synthesizer is actually a discrete time-varying time-varying linear filter, which represents a human voice generation system model. Time-varying linear filters are used as both predictors and synthesizers. When analyzing voice waveforms, it is mainly used as a predictor. As the voice waveform changes, the model parameters and excitation conditions are periodically adapted to the new requirements.