AMSI Modulation & Demodulation Technology

In order to carry out end-to-end encrypted digital voice communication over cellular voice lines, it is necessary to realize highly reliable data communication with a minimum transmission rate of 2kbps to 3kbps over standard voice channels.At present, whether in cellular mobile network or fixed telephone network, due to the existence of vocoder algorithm, the traditional PSK, QAM, OFDM modulation technology and recently emerging voice-like modulation, wavelet modulation, cosine modulation and other schemes can not achieve more than 1.5kbps and practical reliability of data transmission. It does not have the data communication capability to realize digital voice encryption communication.

AMSI modulation and demodulation technology is an advanced digital signal modulation and demodulation technology, its working principle is similar to the traditional telephone Modem. At the sending end, the data through AMSI modulation is converted into analog modulation tone, like ordinary voice signal through the telephone, mobile phone, VoIP voice channel transmission; At the receiving end, the analog modulation tone is demodulated by AMSI and restored to a data stream. The difference between AMSI and the traditional telephone Modem is that the audio signal modulated by AMSI can still be accurately demodulated after being compressed and restored by the medium and low rate voice vocoder, that is, AMSI has enough penetration of the medium and low rate voice vocoder, thus providing a reliable basic data channel for digital voice encryption through standard voice channels such as mobile phones.

1.  AMSI technology currently penetrable voice coding channels: GSM EFR、UMTS AMR WB、UMTS AMR NB、AMR NB12.2、AMR WB24.4、SILK、OPUS、G.711… ；
2.  Basic performance of AMSI technology: It can provide 2Kbps to 4Kbps basic communication bandwidth, and the bit error rate is less than 0.2%

The conventional analog speech scrambling technology is usually to process the time domain signal of human voice, and then scramble the signal sequence in the transform domain, and then execute the reverse transformation, and finally produce scrambled speech. The disadvantage of this approach is that the scrambled speech usually no longer has human voice characteristics, such as pitch, resonance frequency, etc., so that when the scrambled speech passes through the conventional voice channel (such as micro reliance voice call), it will be determined as noise by the voice communication system and be reduced or even completely eliminated, resulting in the receiver can not restore the voice.

The analog scrambling algorithm integrates the speech construction technology, and uses the typical human voice model to process the scrambled signal, ensuring that the scrambled signal can deal with most noise reduction algorithms, and even in the call channel with AI noise reduction.