There are several main compression technologies for voice chips

Due to the huge amount of voice data, it is necessary to effectively compress the voice data, which allows us to record more voice content in the limited ROM space. Classified according to the type and principle of voice chips, there are two compression technologies for voice chips.

waveform compression technology

This technology mainly compresses to imitate the original sound waveform as much as possible. It is also called time domain compression algorithm. It uses mathematical methods to get as close as possible to the original sound waveform. The most representative of this main lossy compression algorithm is ADPCM compression. This lossy compression algorithm uses prediction methods to approximate the original waveform.

frequency compression technology

Frequency compression is used to first divide the audio data into several frames, and then convert the data of each frame from the time domain to the frequency domain. Finally, different frequencies are divided into different excitation sources and amplitude information is given. This compression algorithm uses the spectral approximation method as much as possible. Due to the use of multiple excitation sources for synthesis. Therefore, after compression, almost no original data is contained, and the row pair compression rate is high.

Summarize

Voice compression technology is generally divided into two types: lossless voice compression technology and lossy voice compression technology. Since lossless compression has a low compression rate and is expensive and difficult to store and transmit, only some professional HIFIs need to use it.