Could we make the handshake process simpler?

[christianxiao]

Hi, this project is really awesome! And it may be the best DSP and sound wave library written in JS! When I try to transmit data over sound just like what the doc says, I notice that the handshake process requires the receiver to listen to silence for a while. I guess we may make the handshake process simpler, just like the TCP or HTTP handshake: Server side:

while(true){
1. Always listen for a special sync freq or sound wave, 
2. If it receives, do something, for example, sent the HTML file
}

Client side:

1. Send the special sync freq
2. Wait for the server data.

I notice that there is an audio chat example, will that help to implement handshake? I'm really new to DSP and don't know whether this request is suitable. We will really appreciate that if you give us some suggestion. Thank you.

[robertrypula]

Hi, I'm glad that you liked this project :) Your suggestion is very good and to be honest I already implemented similar solution. Receiver is listening for special sequence on two frequencies. It allows to pick some information about the signal strength and additionally perform synchronization. Synchronization is important because according to the Nyquist the sampling frequency needs to be at least twice of the signal frequency. It real world application it means that we could have valid and stable symbol in odd or even sampling windows. Synchronization and sequence detection is based on the correlation technique.

Everything what I wrote above is still not included in the npm package and even mentioned in the project website. In general existing PSK solution was not working fine with mobile devices and it was in general very slow. It will be deprecated soon and replaced with new approach that will work on any device. It will use FSK modulation. Later I plan to extend it into OFDM technique.

So far you can check those examples: https://audio-network.rypula.pl/spectral-waterfall https://audio-network.rypula.pl/fsk-synchronization-demo https://audio-network.rypula.pl/physical-layer-simplest https://audio-network.rypula.pl/physical-layer-simple https://audio-network.rypula.pl/physical-layer-listeners-demo https://audio-network.rypula.pl/physical-layer-ascii-demo

Currently it's not possible to use new solution as production ready npm package.

The reason that main npm package seems to be dead is because for the last few months I'm working on articles about data transmission in JavaScript for polish 'Programista' magazine. In fall 2017 the last, third part should be published. Then I'll have more time to focus on next Audio Network lib release and the page in general. In fact it's my hobby project so my time is very limited. In late 2017 I should publish English version of the whole series. It will cover many topics like simple and intuitive Discrete Fourier Transform algorithm, Web Audio API and simple network stack based on some kind of mixture of OSI and TCP/IP model.

Answering your last question - the 'Audio Chat' example is also related to third part of the article about data transmission. Now it's not yet implemented but it should be ready in about 3 weeks :)

Thanks, Robert

[christianxiao]

Thank you for your patient reply and they are really helpful~ I will check your examples above and dive deep into them :)