Making it possible to send Bitcoin via voice

How do birds communicate? With chirping. Now machines can communicate with each other using sound. A new breed of companies is now creating something called the Internet of Acoustics.

  A brief history of sound

Let's review. Sound has been used as a signal for a long time. In the earliest times, horns, drums and bells sounded the alarm, mobilized assemblies, directed armies and city workers; the ringing of bells throughout the city guided our lives. These sounds were primarily for communication, and musicality was secondary.

Analog sonic codes can be found in unlikely places: Composers from Mozart to Schumann hide musical jokes with unique audio logic in their music; underwater data demodulators guide naval ships; telephone networks make dial tones noisy.

For many people (including me), the first experience of the Internet was the sound of a modem emitting something like the squeaking of a white-headed bird, the PCM encoding game on a cassette tape - these sounds were not the data itself but a byproduct of data transmission, and they were not designed to travel through the air or be heard by the human ear, but were designed for cables.

With small audio clips being decoded in real time on mobile devices, we have entered the digital age with modern systems for encoding URLs into recorded sequences for wireless transmission. In this weightless digital world, it is easy to forget that information is also an object, how can things not make a sound? In this era, machines can also make sounds, and they can communicate anything from pictures to payment methods.

  The era of acoustic networking is coming

  This is acoustic networking. First, machines can send links to remote networks via sound, and second, lookup tables inside the machine don't need another network: the sound itself is the network. Today's technology is good enough to send small amounts of data in the audible range or beyond, and good enough to encode each link created into a few seconds of audio or less.

  This is a bold idea

Sound exists in many forms, the first and most easily overlooked is the online form.

Think about it: there are more tiny, cheap speakers on the planet than there are people. Why not take advantage of this ubiquitous, everyday technology? We see a huge opportunity to connect a vast number of devices simply and directly.

Voice can go places that other networks can't, and voice can become a valuable part of the network ecosystem alongside existing protocols. We can easily re-engineer ATMs, TVs, toys, radios, and tablets - anything that can transmit sound can also transmit data.

  How does it work?

There are many technologies that can build acoustic networking, which generally include three aspects:

Description : Distinguishing one signal from another by using selected existing signal features, also known as "audio fingerprinting". Examples include music recognition services like Shazam and Soundhound. Geeks, forgive me for giving these two examples: while these technologies can also be called true acoustic networking technologies in a broad sense, even music can be reshaped for peer-to-peer communication.

Addition : Retrievable but human-imperceptible features (also called "audio watermarks") can be added to any audio signal. Examples include using codes for listener tracking in radio broadcasts. One approach uses tiny, imperceptible, rapid echoes that are usually ignored by the human ear. This approach is used by companies like InFrasonics, mufin and Civolution. Its bitrate is roughly equivalent to the fingerprint method above.

Encoding : In addition to extracting a sound from a large online library of known sounds, or quickly adding watermarks, the complete audio signal (pitch, timbre, scale or amplitude) itself is the code. Chirp and LISNR multi-pitch systems are examples of codes that are now strong enough to avoid noise interference, distortion and compression. The recently released Google Tone is based on another multi-frequency encoding method. The "clear hearing" of a purely coded signal has the advantage of faster speed than the first two techniques.

However……

All approaches have relative strengths and weaknesses. Often there is a conflict between practicality and engineering. A signal can be rich in data, but unpleasant to the ear. A signal can be embedded in a string quartet without being noticed, but the data encoded is minuscule. A signal can be extremely fragile in the face of real-world noise, reverberation, etc. A signal can be both unpleasant to the ear and unreliable.

Sending data via sound is necessarily slow. Sending data via pleasant sounds is even slower -- so the key is to send pointers rather than files. Sonic data raises security issues: one-to-many data sharing is quite useful, but not secure (depending on how it's designed); what if you only want to share with one person? This is an interesting challenge -- and people are working on new solutions as we speak.

Finally, and most importantly, do we really need the Internet of Sound? I believe so. Sound exists in many forms, and the first and most easily overlooked is the network form - it is a bridge across the last few meters and a medium to reach places that other networks cannot reach.

  Be careful

  Sound itself has a huge impact. Using it with care and respect, while also being mindful of the sonic environment, is critical. From headphone leakage to street noise to sonic user interfaces, we already have too much casual and rampant noise pollution that is dismissed in retrospect as just the sound of key presses.

As audio geeks and sound designers, we put a lot of thought into the sounds we send, and we obviously shape our audio sounds after nature, especially bird language. Acoustic networking requires a humanistic approach to communication: one that respects the user by putting the human ear first and foremost.

  Sound is everywhere

It’s time to get excited. Honestly the opportunities for voice networking are incredibly large, from existing applications to easy resource provisioning. There is potential to reimagine voice, music and data sharing right now, and our company is quite active in this area.

New hardware will accelerate its understanding: cheaper and cheaper DSP chips for mobile phones, and devices that need to be connected for always-on reception and hands-free user interfaces, are increasing the reach and efficiency of acoustic networking products. We are already seeing audio used in payments such as sending bitcoins over the radio, and in classrooms through picture or web sharing to send maps to mobile phones. We are just getting started. In short, from non-smartphones to doorbells, wherever there is sound, there is data.

The era of acoustic networking has arrived, and one thing is certain - you haven't heard anything yet.