Audio Network - send data over sound in JavaScript

Audio Network at Facebook

Return to main page

THIS PROJECT IS STILL UNDER DEVELOPMENT
Source code available at GitHub

Roadmap

  • 1.1.0 Last update for old PSK code
  • 1.2.x Completely rewritten Physical Layer with FSK support. Old PSK code marked as deprecated
  • 1.3.x Data link layer with frames and checksums
  • 2.0.0 Move to WebPack, TypeScript, Jest, Travis, Coveralls
  • 2.0.0 Old PSK code removed from the project.
  • 2.x.x Transport layer that supports simplified TCP protocol
  • 2.x.x Application layer with simple text exchange protocol
  • 2.x.x ?? OFDM

v1.3.2 (2018-05-16)

  • BugFix for the same issue in other area: "The AudioContext was not allowed to start. It must be resume (or created) after a user gesture on the page. https://goo.gl/7K7WLu"

v1.3.1 (2018-05-10)

  • BugFix: "The AudioContext was not allowed to start. It must be resume (or created) after a user gesture on the page. https://goo.gl/7K7WLu"

v1.3.0 (2018-01-02)

  • adds Data Link Layer to the network stack
  • up to 7 bytes of data can travel inside frame with header and checksum (Fletcher-8)
  • frames tranmission progress listener
  • incomming frames listener
  • incomming frame candidates listener
  • 'TwoWaySync' that helps to exchange 'Sync' sequences and SampleRate values (this will be simplified even more in future versions)

v1.2.0 (2017-12-28)

  • completely rewritten Physical Layer core based on Frequency-Shift Keying
  • special "sync" sequence for symbol synchronization based on correlation technique
  • many new examples
  • old FSK code marked as deprecated
  • volume control issue on mobile devices solved by detaching from microphone
  • Injector aliases again enabled by default

v1.1.0 (2017-11-12) Last update for old PSK code

  • Dynamic dev files loading
  • Boot config added
  • Env variables added (isBrowser, isNode, isWebWorker)
  • NodeJs export added - you can now use AudioNetwork in NodeJs applications!
  • ActiveAudioContext abstraction
  • CarrierGenerate and CarrierRecovery refactor
  • Constellation diagram: do not redraw constellation if queue wasn't changed
  • Constellation diagram: queue elements approach changed (only one item per constellation point now)
  • WindowFunction service with blackmanNuttall method
  • Visualizer abstract classes (Abstract2DVisualizer, AbstractVisualizer)
  • Visualizer code refactor
  • rename audio-network-begin to audio-network-boot
  • WebWorker basic functionality added

v1.0.4 (2016-07-04)

  • new example where only CarrierRecovery/CarrierGenerate were used to send data using Amplitude Modulation (without PhysicalLayer module)
  • expand class names - this allows easier debug in dev tools (for example CTM is now ChannelTransmitManager)
  • introduce SimplePromise class (idea is to have code as less dependent to browser as possible in order to port it easily to other languages)
  • move all charts to dedicated namespace 'visualizer'
  • move CarrierRecovery and CarrierGenerate to 'common' namespace
  • move audio service to separate namespace 'audio' and change name of the service to ActiveAudioContext
  • move adapter classes to 'physical-layer-adapter'
  • move current audio service logic to factory and connect it to ActiveAudioContext
  • AudioContext class was renamed to SimpleAudioContext in order to avoid collision with real window.AudioContext
  • addToQueue is now taking object as a parameter instead array of objects
  • Jasmine unit test web runner added (currently still without tests but at least running environment was prepared)
  • rename example directories (000 for basic demos, 100 for physical layer demos, …)
  • add htaccess redirection to renamed examples
  • create new chart 'SampleChart', create new example where CarrierGenerate will be demonstrated

v1.0.3 (2016-05-30)

  • fix transition from FIRSTSYNCINIT -> IDLE, currently there are some SYMBOL/GUARD states which are not right at this point
  • default settings update (symbol duration vs guard interval, sync duration reduced from 3 sec to 2 sec, notification per seconds increased)
  • gulp 'serve' task for serving the app locally
  • add YouTube movie with demo
  • move common css away of example
  • readme moved to index.html, styles and pages structure refactored
  • simple demo added
  • updates at README.md
  • added CHANGELOG.md
  • licence moved to MIT
  • SEO updates to whole page

v1.0.2 (2016-05-16)

  • huge readme update, npm keywords update

v1.0.1 (2016-05-10)

  • Carrier example bug fix. Versioning information added for main example

v1.0.0 (2016-05-10)

  • fix typo: physial-layer into physical-layer
  • keep view type after reinitialization
  • remove audio-network prefix from main classes names
  • change name: dftTimeSpan -> dftWindowTimeSpan
  • move general configuration to some common service
  • refactor NO_SIGNAL state
  • change input TX to LOOPBACK
  • move templates code to dedicated files
  • introduce Dependency Injection
  • prepare release version + some code minification
  • measure CPU load by measuring times before and after execution

Finished stuff before release was even planned

  • BUG FIXED for some dftWindowTime after converting it to numbers of samples CarrierRecovery queue size is fractional
  • real/imm delete
  • compute at getCarrierDetail
  • index passed into handler
  • decibel power/amplitude check
  • load wav file
  • add getOutput* methods
  • add outputTxEnable/outputTxDisable methods
  • check inverse phase shift issue
  • add script node sample to config
  • rewrite main API
    • move code to factory
    • cleanup inside main service
    • internal notifyHandler for constellation update, external for user purposes
    • add rx method outside the factory
    • destroy constellation
    • ability to change frequency
    • fix recorded file loading logic
    • fix history point colors
    • add ability to choose destination source
  • move script processor node to receive manager
  • move sin/cos to internal Math service (to give ability to quickly add lookup tables)
  • fix layout
  • add phase offset input to align symbol '0'
  • add html generation as js + ofdm support
  • change send logic (add amplitude, symbolCount, symbol to each OFDM block)
  • change send sequence logic (use format: '5.5.2.0 1.2.4.1')
  • add DFT time span to config
  • internal loop for notifications
    • add script node block time (from audiocontext)
    • add sample offset time from script node block time
  • add symbol config to rx
  • prefill amplitude value basing on ofdm size at channel tx
  • add symbol detection to rx
  • move example code related to html generation to separate file with comment
  • add buttons for symbols (TX)
  • add squares with symbol number (RX)
  • update sequence textarea after pskSize change
  • rename delay-loop-handler
  • after psk change only related tx/rx should be updated
  • add rx/tx to channel headers
  • change 'frame' to 'packet'
  • add 'Send sync signal' to TX
  • ability to hide some of the widgets: 'Easy' and 'Advanced' view
  • fix styles
    • add source button active class
    • improve responsive design
  • add margin to sections
  • use first symbol of each packet to fine tune phase offset (add checkbox for that feature)
  • add inter-packet gap duration
  • add quick configs like: 'baud-5, ofdm-1, psk-2' or 'baud-20, ofdm-16, psk-2' or …
    • add checkbox for tx/rx config
    • add callback to destroy
    • add bit speed information at UI
  • refactor all transmit and receive logic (move it to physical layer internals)
    • [TX] remove symbol generation from template-util
    • [TX] symbol shouldn't have any guard interval or/and interpacket gap
    • [RX] add setTimes* methods (maybe it's worth to add some error margin - times inside are for internal adapter use)
    • [RX] add setSyncPreamble(true/false) method
    • [RX] add packet receive handler packetReceived(data)
    • [RX] add multiple channels support (introduce new class in the middle)
    • set threshold to very low value (-100 dB) to force idle state for a while
    • compute average noise level power at idle state
    • after noise level is set raise threshold xx dB above noise level
    • so far do not collect symbol and packet data (wait for sync)
    • run sync on the TX side
    • sync state will be detected - grab average max signal strength
    • subtract 10 decibels from max signal and enable symbol/packet collecting
    • [RX] add method to reset receiver state machine (to follow steps above again)
    • [RX] double check state times at receive adapter
    • move sync time to main adapter class
    • check symbol, guard times
    • average sample sizes should be config dependent (mostly for samplesPerSecond setting)
    • [RX] grab/clean packet data and notify packet handler