Back to

A-FSK transmission for the upcoming SCK V2.0


Hello everyone!

First of all! Merry Christmas!

Image credit: Taringa

As the last post for this exciting year, we are going to talk about a new feature we are planning to introduce in the upcoming version of the SCK: a FSK communication protocol via Audio (A-FSK). You might have read about this technique and it’s usage in the Amazon Dash configuration process, and on the post today we are going to describe very briefly the work in progress for this feature.

So! FSK stands for Frequency Shift Keying, which is a form of transmission through frequency variations on the carrying waveforms. It’s major counterpart is the so called ASK, or Amplitude Shift Keying, in which the transmission is carried out via amplitude variations. A very simple form of ASK is OOK, which stands for On-Off-Keying, in which the amplitude of the carrier wave oscillates between a value and nothingness:

Image credit: Electric Stack Exchange

As in many other situations, there is a trade off between the options on the table: ASK or FSK? Maybe another one? The main disadvantage of ASK it is said to have a higher probability of error with respect to FSK, since noise interference affects amplitude of the transmitted wave. FSK, on the other hand, it is said to have a lower bandwidth efficiency. However, since we have talked about FFT quite a lot now, we thought FSK would be our best bet and also, because maybe bandwidth is not such a big deal after all as we can see below.

Then, the idea is to implement an algorithm that is able to identify if the sound transmitted from an emitter (i.e. a smartphone) contains a series of reference frequencies in certain known spots. Following this principle, our aim is to transmit a byte per sound wave, hence, in a sound wave containing up to 8 possible carrier frequencies that might or not be activated. The activation (or not) of these frequencies in the analysed spectrum will yield a 1 or a 0, that we can use on a bit mask and extract 8-bit ASCII characters codes:

Image credit: Martin Melhus

The emitter could be based on the Web Audio API, as the example from Martin Melhus above from his project on a Web Audio Modem. Finally, the receiver would be our beloved I2S Mems microphone that we have been talking about for so long now, doing a FFT algorithm and detecting the peaks in it, identifying the carrier frequencies activation.

That’s all for today! If you feel curious about it, we are currently developing the implementation of all this and, if you want to have a look at our approach, you can do so in the github repository!

Let us know your thoughts!

1 Like