A method for real-time speech compression and decompression for NB-IoT

Abstract

The invention discloses a method for real-time voice compression and decompression used in the narrowband Internet of Things. The compression method includes: continuously intercepting voice data with a total duration of t in real time, and then pressing t / 2 n The interval time to take points for voice data, take enough 2 n The points are taken as the Nth frame data; the Nth frame data is subjected to high-frequency pre-emphasis, the time domain data is converted into frequency domain data, and the power spectrum is calculated by modulus to obtain the spectrum. Select the amplitude of each harmonic in the spectrum, and use each harmonic Calculate the total attenuation coefficient by calculating the wave amplitude; divide the amplitude of each harmonic by the total attenuation coefficient to obtain the amplitude ratio of each harmonic; combine the Nth frame data and spectrum to analyze the fundamental wave to obtain the fundamental frequency value; The wave frequency value, the amplitude ratio of each harmonic and the total attenuation coefficient are arranged to generate a transmission sequence code sequence; among them, t is 20ms-60ms, n=7, 8, 9; N is a positive integer. The speech compression method of the present invention has strong real-time performance, high compression rate, and can approximate the communication effect of making a phone call.

Description

technical field [0001] The invention relates to the technical field of voice communication in the narrowband Internet of Things, and more particularly, to a method for real-time voice compression and decompression for the narrowband Internet of Things, and its application in a walkie-talkie. Background technique [0002] With the explosive development of information technology, today's world has entered the era of the Internet of Everything, and various Internet of Things technologies emerge in an endless stream. However, due to the inherent narrow bandwidth, low transmission rate, and long delay of NB-IoT. As a result, it is difficult to transmit voice signals using communication methods such as Zigbee, LoRa, and NB-IOT. How to effectively use the limited bandwidth resources of the narrowband Internet of Things to achieve voice transmission and reception with a data volume of more than 22KB / s has become an urgent need. solved problem. [0003] 1. Take LoRa as an example: ...

AI Technical Summary

Problems solved by technology

However, due to the inherent narrow bandwidth, low transmission rate and long delay of NB-IoT

As a result, there are great difficulties in transmitting voice signals using communication methods such as Zigbee, LoRa, and NB-IOT. How to effectively use the limited bandwidth resources of the narrowband Internet of Things to realize voice transmission and reception with a data volume of more than 22KB / s has become an urgent need solved problem

[0003] 1. Take LoRa as an example: when using LoRa technology to transmit data, if the transmitted code rate is less than 8B / s, the transmission distance can reach 25 kilometers (open space), and when the code rate reaches 960B / s, under the same conditions The transmission distance has dropped below 1Km, and the data volume of general voice communication reaches 22KB / s. Under uncompressed conditions, it is basically impossible to transmit through LoRa

[0004] 2. The theoretical bandwidth of Zigbee is 256KB / s, but the data packets of various protocols in the transmitted data actually occupy a large proportion, and the bandwidth that can be reserved for voice communication is very limited. If it is not suitable for voice data up to 22KB / s Compression, even if Zigbee can barely transmit, it will bring a lot of delay

[0006] (1) The real-time performance is not strong: MPEG2 needs to sample at least one complete voice syllable data before the compressed data is enough to ensure the quality of the call, and a complete syllable is at least 0.4-0.7s long, plus the call time Delay, usually the total call delay is at least 1s, and our real-time voice compression technology completes one frame compression every 50ms, so the real-time nature of the call is much stronger than MPEG2

[0007] (2) Insufficient compression ratio: the compression ratio of MPEG2 is 12:1--10:1, and the insufficient compression ratio cannot be used in real-time voice transmission technology in the narrowband Internet of Things

Images