speaker (community library)


Name Value
Name speaker
Version 1.2.0
Installs 4182
Author Julien Vanier jvanier@gmail.com
URL https://github.com/monkbroc/particle-speaker
Download .tar.gz
All Versions 1.2.0, 1.1.0, 1.0.1

Generate audio output for a speaker

Example Build Testing

Device OS Version:

This table is generated from an automated build. Success only indicates that the code compiled successfully.

Library Read Me

This content is provided by the library maintainer and has not been validated or approved.


Build Status

Generate audio output for a speaker for Particle devices (Photon, Electron)


Connect a speaker amplifier like this one from Adafruit to the digital to analog converter DAC pin of a Photon or Electron and run this code to play a sawtooth wave.

#include "speaker.h"

uint16_t bufferSize = 128;
Speaker speaker(bufferSize);

void setup() {
uint16_t audioFrequency = 22050; // Hz

uint16_t audioSignal = 0;

void loop() {
if (speaker.ready()) {
uint16_t *buffer = speaker.getBuffer();
// Produces a 1 kHz sawtooth wave
for (uint16_t i = 0; i < bufferSize; i++) {
buffer[i] = audioSignal;
audioSignal += 2267;
if (audioSignal > 50000) {
audioSignal = 0;

1 kHz sawtooth signal (image removed)
A 1 kHz sawtooth signal played from a Photon

See complete example in the examples directory.

Currently the output is mono. Since the Photon and Electron have 2 DAC outputs, the library could be extended to support stereo output.



Speaker speaker(bufferSize);

Creates a speaker object with 2 buffers of the same size (double buffering).

While the library plays the sound in one buffer your application fills the second one.

The larger the buffer, the more delay there will be in between your application filling a buffer and it being played. The shorter the buffer, the less time your code has to fill the next buffer. Short buffers work best for real-time audio synthesis and longer buffer for playback from an SD card.

The application has bufferSize / audioFrequency seconds to fill the next buffer. For example, this is 2.9 ms at 44100 Hz with a 128 sample buffer.

The copy from memory to the DAC is done using direct memory access (DMA) so the CPU is free to do other tasks.



Sets up the DAC pin and TIM6 timer to trigger at the correct audio freqency. Common frequencies are 44100 Hz, 22050 Hz, 11025 Hz and 8000 Hz.

Starts playing the content of the buffer immediately so you may want to fill the audio buffer before calling speaker.begin. The buffer is zero by default so not filling the buffer first would still be OK.

Note: Do not call analogWrite(DAC, ...); when using this library since it completely takes over the DAC peripheral.`



Stops the audio playback.


bool readyForMoreAudio = speaker.ready();

Returns true once when the audio buffer is ready to be filled with more audio samples. Will return false when called again until the buffer has finished playing.


uint16_t *buffer = speaker.getBuffer();

Returns a pointer to an array of bufferSize audio samples.

The audio samples are 16 bit integers but the DAC on the Photon and Electron only has 12 bits to the least significant 4 bits are ignored.

You must only write to this array when speaker.ready() is true.

Resource Utilization

This library uses the DAC1 digital to analog converter, TIM6 basic timer and DMA1 stream 5 direct memory access.


Read the STM Application note AN3126 - Audio and waveform generation using the DAC in STM32 microcontrollers for more background on using the DAC and DMA for audio generation.


Copyright 2016 Julien Vanier

Released under the MIT license

Browse Library Files