Your keyboard can make sounds! If you've got a spare pin you can hook up a simple speaker and make it beep. You can use those beeps to indicate layer transitions, modifiers, special keys, or just to play some funky 8bit tunes.
To activate this feature, add AUDIO_ENABLE = yes
to your rules.mk
.
On Atmega32U4 based boards, up to two simultaneous tones can be rendered. With one speaker connected to a PWM capable pin on PORTC driven by timer 3 and the other on one of the PWM pins on PORTB driven by timer 1.
The following pins can be configured as audio outputs in config.h
- for one speaker set either one out of:
#define AUDIO_PIN C4
#define AUDIO_PIN C5
#define AUDIO_PIN C6
#define AUDIO_PIN B5
#define AUDIO_PIN B6
#define AUDIO_PIN B7
and optionally, for a second speaker, one of:
#define AUDIO_PIN_ALT B5
#define AUDIO_PIN_ALT B6
#define AUDIO_PIN_ALT B7
per speaker is - for example with a piezo buzzer - the black lead to Ground, and the red lead connected to the selected AUDIO_PIN for the primary; and similarly with AUDIO_PIN_ALT for the secondary.
for more technical details, see the notes on Audio driver.
Most STM32 MCUs have DAC peripherals, with a notable exception of the STM32F1xx series. Generally, the DAC peripheral drives pins A4 or A5. To enable DAC-based audio output on STM32 devices, add AUDIO_DRIVER = dac_basic
to rules.mk
and set in config.h
either:
#define AUDIO_PIN A4
or #define AUDIO_PIN A5
the other DAC channel can optionally be used with a secondary speaker, just set:
#define AUDIO_PIN_ALT A4
or #define AUDIO_PIN_ALT A5
Do note though that the dac_basic driver is only capable of reproducing one tone per speaker/channel at a time, for more tones simultaneously, try the dac_additive driver.
for two piezos, for example configured as AUDIO_PIN A4
and AUDIO_PIN_ALT A5
would be: red lead to A4 and black to Ground, and similarly with the second one: A5 = red, and Ground = black
another alternative is to drive one piezo with both DAC pins - for an extra "push".
wiring red to A4 and black to A5 (or the other way round) and add #define AUDIO_PIN_ALT_AS_NEGATIVE
to config.h
The Proton-C comes (optionally) with one 'builtin' piezo, which is wired to A4+A5.
For this board config.h
would include these defines:
#define AUDIO_PIN A5
#define AUDIO_PIN_ALT A4
#define AUDIO_PIN_ALT_AS_NEGATIVE
Another option, besides dac_basic (which produces sound through a square-wave), is to use the DAC to do additive wave synthesis.
With a number of predefined wave-forms or by providing your own implementation to generate samples on the fly.
To use this feature set AUDIO_DRIVER = dac_additive
in your rules.mk
, and select in config.h
EITHER #define AUDIO_PIN A4
or #define AUDIO_PIN A5
.
The used waveform defaults to sine, but others can be selected by adding one of the following defines to config.h
:
#define AUDIO_DAC_SAMPLE_WAVEFORM_SINE
#define AUDIO_DAC_SAMPLE_WAVEFORM_TRIANGLE
#define AUDIO_DAC_SAMPLE_WAVEFORM_TRAPEZOID
#define AUDIO_DAC_SAMPLE_WAVEFORM_SQUARE
Should you rather choose to generate and use your own sample-table with the DAC unit, implement uint16_t dac_value_generate(void)
with your keyboard - for an example implementation see keyboards/planck/keymaps/synth_sample or keyboards/planck/keymaps/synth_wavetable
if the DAC pins are unavailable (or the MCU has no usable DAC at all, like STM32F1xx); PWM can be an alternative. Note that there is currently only one speaker/pin supported.
set in rules.mk
:
AUDIO_DRIVER = pwm_software
and in config.h
:
#define AUDIO_PIN C13
(can be any pin) to have the selected pin output a pwm signal, generated from a timer callback which toggles the pin in software.
the usual piezo wiring: red goes to the selected AUDIO_PIN, black goes to ground.
OR if you can chose to drive one piezo with two pins, for example #define AUDIO_PIN B1
, #define AUDIO_PIN_ALT B2
in config.h
, with #define AUDIO_PIN_ALT_AS_NEGATIVE
- then the red lead could go to B1, the black to B2.
STM32F1xx have to fall back to using PWM, but can do so in hardware; but again on currently only one speaker/pin.
AUDIO_DRIVER = pwm_hardware
in rules.mk
, and in config.h
:
#define AUDIO_PIN A8
#define AUDIO_PWM_DRIVER PWMD1
#define AUDIO_PWM_CHANNEL 1
(as well as #define AUDIO_PWM_PAL_MODE 42
if you are on STM32F2 or larger)
which will use Timer 1 to directly drive pin PA8 through the PWM hardware (TIM1_CH1 = PA8).
Should you want to use the pwm-hardware on another pin and timer - be ready to dig into the STM32 data-sheet to pick the right TIMx_CHy and pin-alternate function.
Since most drivers can only render one tone per speaker at a time (with the one exception: arm dac-additive) there also exists a "workaround-feature" that does time-slicing/multiplexing - which does what the name implies: cycle through a set of active tones (e.g. when playing chords in Music Mode) at a given rate, and put one tone at a time out through the one/few speakers that are available.
To enable this feature, and configure a starting-rate, add the following defines to config.h
:
#define AUDIO_ENABLE_TONE_MULTIPLEXING
#define AUDIO_TONE_MULTIPLEXING_RATE_DEFAULT 10
The audio core offers interface functions to get/set/change the tone multiplexing rate from within keymap.c
.
There's a couple of different sounds that will automatically be enabled without any other configuration:
STARTUP_SONG // plays when the keyboard starts up (audio.c)
GOODBYE_SONG // plays when you press the RESET key (quantum.c)
AG_NORM_SONG // plays when you press AG_NORM (quantum.c)
AG_SWAP_SONG // plays when you press AG_SWAP (quantum.c)
CG_NORM_SONG // plays when you press CG_NORM (quantum.c)
CG_SWAP_SONG // plays when you press CG_SWAP (quantum.c)
MUSIC_ON_SONG // plays when music mode is activated (process_music.c)
MUSIC_OFF_SONG // plays when music mode is deactivated (process_music.c)
CHROMATIC_SONG // plays when the chromatic music mode is selected (process_music.c)
GUITAR_SONG // plays when the guitar music mode is selected (process_music.c)
VIOLIN_SONG // plays when the violin music mode is selected (process_music.c)
MAJOR_SONG // plays when the major music mode is selected (process_music.c)
You can override the default songs by doing something like this in your config.h
:
#ifdef AUDIO_ENABLE
#define STARTUP_SONG SONG(STARTUP_SOUND)
#endif
A full list of sounds can be found in quantum/audio/song_list.h - feel free to add your own to this list! All available notes can be seen in quantum/audio/musical_notes.h.
To play a custom sound at a particular time, you can define a song like this (near the top of the file):
float my_song[][2] = SONG(QWERTY_SOUND);
And then play your song like this:
PLAY_SONG(my_song);
Alternatively, you can play it in a loop like this:
PLAY_LOOP(my_song);
It's advised that you wrap all audio features in #ifdef AUDIO_ENABLE
/ #endif
to avoid causing problems when audio isn't built into the keyboard.
The available keycodes for audio are:
AU_ON
- Turn Audio Feature onAU_OFF
- Turn Audio Feature offAU_TOG
- Toggle Audio Feature state
!> These keycodes turn all of the audio functionality on and off. Turning it off means that audio feedback, audio clicky, music mode, etc. are disabled, completely.
the 'speed' at which SONGs are played is dictated by the set Tempo, which is measured in beats-per-minute. Note lengths are defined relative to that.
The initial/default tempo is set to 120 bpm, but can be configured by setting TEMPO_DEFAULT
in config.c
.
There is also a set of functions to modify the tempo from within the user/keymap code:
void audio_set_tempo(uint8_t tempo);
void audio_increase_tempo(uint8_t tempo_change);
void audio_decrease_tempo(uint8_t tempo_change);
For ARM devices, you can adjust the DAC sample values. If your board is too loud for you or your coworkers, you can set the max using AUDIO_DAC_SAMPLE_MAX
in your config.h
:
#define AUDIO_DAC_SAMPLE_MAX 4095U
the DAC usually runs in 12Bit mode, hence a volume of 100% = 4095U
Note: this only adjusts the volume aka 'works' if you stick to WAVEFORM_SQUARE, since its samples are generated on the fly - any other waveform uses a hardcoded/precomputed sample-buffer.
Aka "audio effects", different ones can be enabled by setting in config.h
these defines:
#define AUDIO_VOICES
to enable the feature, and #define AUDIO_VOICE_DEFAULT something
to select a specific effect
for details see quantum/audio/voices.h and .c
The music mode maps your columns to a chromatic scale, and your rows to octaves. This works best with ortholinear keyboards, but can be made to work with others. All keycodes less than 0xFF
get blocked, so you won't type while playing notes - if you have special keys/mods, those will still work. A work-around for this is to jump to a different layer with KC_NOs before (or after) enabling music mode.
Recording is experimental due to some memory issues - if you experience some weird behavior, unplugging/replugging your keyboard will fix things.
Keycodes available:
MU_ON
- Turn music mode onMU_OFF
- Turn music mode offMU_TOG
- Toggle music modeMU_MOD
- Cycle through the music modes:CHROMATIC_MODE
- Chromatic scale, row changes the octaveGUITAR_MODE
- Chromatic scale, but the row changes the string (+5 st)VIOLIN_MODE
- Chromatic scale, but the row changes the string (+7 st)MAJOR_MODE
- Major scale
In music mode, the following keycodes work differently, and don't pass through:
LCTL
- start a recordingLALT
- stop recording/stop playingLGUI
- play recordingKC_UP
- speed-up playbackKC_DOWN
- slow-down playback
The pitch standard (PITCH_STANDARD_A
) is 440.0f by default - to change this, add something like this to your config.h
:
#define PITCH_STANDARD_A 432.0f
You can completely disable Music Mode as well. This is useful, if you're pressed for space on your controller. To disable it, add this to your config.h
:
#define NO_MUSIC_MODE
By default, MUSIC_MASK
is set to keycode < 0xFF
which means keycodes less than 0xFF
are turned into notes, and don't output anything. You can change this by defining this in your config.h
like this:
#define MUSIC_MASK keycode != KC_NO
Which will capture all keycodes - be careful, this will get you stuck in music mode until you restart your keyboard!
For a more advanced way to control which keycodes should still be processed, you can use music_mask_kb(keycode)
in <keyboard>.c
and music_mask_user(keycode)
in your keymap.c
:
bool music_mask_user(uint16_t keycode) {
switch (keycode) {
case RAISE:
case LOWER:
return false;
default:
return true;
}
}
Things that return false are not part of the mask, and are always processed.
By default, the Music Mode uses the columns and row to determine the scale for the keys. For a board that uses a rectangular matrix that matches the keyboard layout, this is just fine. However, for boards that use a more complicated matrix (such as the Planck Rev6, or many split keyboards) this would result in a very skewed experience.
However, the Music Map option allows you to remap the scaling for the music mode, so it fits the layout, and is more natural.
To enable this feature, add #define MUSIC_MAP
to your config.h
file, and then you will want to add a uint8_t music_map
to your keyboard's c
file, or your keymap.c
.
const uint8_t music_map[MATRIX_ROWS][MATRIX_COLS] = LAYOUT_ortho_4x12(
36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47,
24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35,
12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23,
0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11
);
You will want to use whichever LAYOUT
macro that your keyboard uses here. This maps it to the correct key location. Start in the bottom left of the keyboard layout, and move to the right, and then upwards. Fill in all the entries until you have a complete matrix.
You can look at the Planck Keyboard as an example of how to implement this.
This adds a click sound each time you hit a button, to simulate click sounds from the keyboard. And the sounds are slightly different for each keypress, so it doesn't sound like a single long note, if you type rapidly.
CK_TOGG
- Toggles the status (will play sound if enabled)CK_ON
- Turns on Audio Click (plays sound)CK_OFF
- Turns off Audio Click (doesn't play sound)CK_RST
- Resets the frequency to the default state (plays sound at default frequency)CK_UP
- Increases the frequency of the clicks (plays sound at new frequency)CK_DOWN
- Decreases the frequency of the clicks (plays sound at new frequency)
The feature is disabled by default, to save space. To enable it, add this to your config.h
:
#define AUDIO_CLICKY
You can configure the default, min and max frequencies, the stepping and built in randomness by defining these values:
Option | Default Value | Description |
---|---|---|
AUDIO_CLICKY_FREQ_DEFAULT |
440.0f | Sets the default/starting audio frequency for the clicky sounds. |
AUDIO_CLICKY_FREQ_MIN |
65.0f | Sets the lowest frequency (under 60f are a bit buggy). |
AUDIO_CLICKY_FREQ_MAX |
1500.0f | Sets the highest frequency. Too high may result in coworkers attacking you. |
AUDIO_CLICKY_FREQ_FACTOR |
1.18921f | Sets the stepping of UP/DOWN key codes. This is a multiplicative factor. The default steps the frequency up/down by a musical minor third. |
AUDIO_CLICKY_FREQ_RANDOMNESS |
0.05f | Sets a factor of randomness for the clicks, Setting this to 0f will make each click identical, and 1.0f will make this sound much like the 90's computer screen scrolling/typing effect. |
AUDIO_CLICKY_DELAY_DURATION |
1 | An integer note duration where 1 is 1/16th of the tempo, or a sixty-fourth note (see quantum/audio/musical_notes.h for implementation details). The main clicky effect will be delayed by this duration. Adjusting this to values around 6-12 will help compensate for loud switches. |
This is still a WIP, but check out quantum/process_keycode/process_midi.c
to see what's happening. Enable from the Makefile.
Key | Aliases | Description |
---|---|---|
AU_ON |
Audio mode on | |
AU_OFF |
Audio mode off | |
AU_TOG |
Toggles Audio mode | |
CLICKY_TOGGLE |
CK_TOGG |
Toggles Audio clicky mode |
CLICKY_UP |
CK_UP |
Increases frequency of the clicks |
CLICKY_DOWN |
CK_DOWN |
Decreases frequency of the clicks |
CLICKY_RESET |
CK_RST |
Resets frequency to default |
MU_ON |
Turns on Music Mode | |
MU_OFF |
Turns off Music Mode | |
MU_TOG |
Toggles Music Mode | |
MU_MOD |
Cycles through the music modes |