forksand
/
qmk_firmware
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

refined a bit

Browse Source
arm-dac-work
Jack Humbert 6 years ago
parent 208bee10f2
commit 474d100b56
5 changed files with 127 additions and 104 deletions
Show Stats Download Patch File Download Diff File

3
keyboards/planck/ez/config.h
Unescape View File

@ -51,6 +51,9 @@
#undef AUDIO_VOICES
#undef C6_AUDIO
#define A5_AUDIO
#define DAC_OFF_VALUE 4095
/* Debounce reduces chatter (unintended double-presses) - set 0 if debouncing is not needed */
#define DEBOUNCE 6

2
keyboards/planck/rev6/config.h
Unescape View File

@ -51,6 +51,8 @@
#undef AUDIO_VOICES
#undef C6_AUDIO
#define A5_AUDIO
/* Debounce reduces chatter (unintended double-presses) - set 0 if debouncing is not needed */
#define DEBOUNCE 6

4
keyboards/planck/rev6/rev6.c
Unescape View File

@ -16,7 +16,9 @@
#include "rev6.h"
void matrix_init_kb(void) {
matrix_init_user();
palSetPadMode(GPIOA, 4, PAL_MODE_OUTPUT_PUSHPULL );
palSetPad(GPIOA, 4);
matrix_init_user();
}
void matrix_scan_kb(void) {

218
quantum/audio/audio_arm.c
Unescape View File

@ -26,19 +26,54 @@
// -----------------------------------------------------------------------------
/**
* Size of the dac_buffer arrays. All must be the same size.
*/
#define DAC_BUFFER_SIZE 256U
/**
* Highest value allowed by our 12bit DAC.
*/
#ifndef DAC_SAMPLE_MAX
#define DAC_SAMPLE_MAX 4095U
#endif
/**
* Effective bitrate of the DAC. 44.1khz is the standard for most audio - any
* lower will sacrifice perceptible audio quality. Any higher will limit the
* number of simultaneous voices. In most situations, a tenth (1/10) of the
* sample rate is where notes become unbearable.
*/
#ifndef DAC_SAMPLE_RATE
#define DAC_SAMPLE_RATE 44100U
#endif
/**
* The number of voices (in polyphony) that are supported. If too high a value
* is used here, the keyboard will freeze and glitch-out when that many voices
* are being played.
*/
#ifndef DAC_VOICES_MAX
#define DAC_VOICES_MAX 2
#endif
/**
* The default value of the DAC when not playing anything. Certain hardware
* setups may require a high (DAC_SAMPLE_MAX) or low (0) value here.
*/
#ifndef DAC_OFF_VALUE
#define DAC_OFF_VALUE DAC_SAMPLE_MAX / 2
#endif
int voices = 0;
int voice_place = 0;
float frequency = 0;
float frequency_alt = 0;
int volume = 0;
long position = 0;
float frequencies[8] = {0, 0, 0, 0, 0, 0, 0, 0};
int volumes[8] = {0, 0, 0, 0, 0, 0, 0, 0};
bool sliding = false;
float place = 0;
uint8_t * sample;
uint16_t sample_length = 0;
@ -77,43 +112,6 @@ bool glissando = true;
#endif
float startup_song[][2] = STARTUP_SONG;
/** Size of the dac_buffer arrays. All must be the same size. */
#define DAC_BUFFER_SIZE 256U
/** Highest value allowed by our 12bit DAC */
#ifndef DAC_SAMPLE_MAX
#define DAC_SAMPLE_MAX 4095U
#endif
/** Effective bitrate of the DAC. 44.1khz is the standard for most audio - any
* lower will sacrifice perceptible audio quality. Any higher will limit the
* number of simultaneous voices.
*/
#ifndef DAC_SAMPLE_RATE
#define DAC_SAMPLE_RATE 44100U
#endif
/** The number of voices (in polyphony) that are supported. Certain voices will
* glitch out at different values - most (the look-ups) survive 5.
*/
#ifndef DAC_VOICES_MAX
#define DAC_VOICES_MAX 5
#endif
/** The default value of the DAC when not playing anything. Certain hardware
* setups may require a high (DAC_SAMPLE_MAX) or low (0) value here.
*/
#ifndef DAC_OFF_VALUE
#define DAC_OFF_VALUE DAC_SAMPLE_MAX / 2
#endif
GPTConfig gpt7cfg1 = {
.frequency = DAC_SAMPLE_RATE,
.callback = NULL,
.cr2 = TIM_CR2_MMS_1, /* MMS = 010 = TRGO on Update Event. */
.dier = 0U
};
static const dacsample_t dac_buffer[DAC_BUFFER_SIZE] = {
// 256 values, max 4095
0x800,0x832,0x864,0x896,0x8c8,0x8fa,0x92c,0x95e,
@ -137,12 +135,12 @@ static const dacsample_t dac_buffer[DAC_BUFFER_SIZE] = {
0x4f0,0x4c2,0x494,0x467,0x43a,0x40e,0x3e3,0x3b8,
0x38e,0x365,0x33c,0x314,0x2ed,0x2c6,0x2a0,0x27c,
0x258,0x235,0x212,0x1f1,0x1d1,0x1b1,0x193,0x175,
0x159,0x13e,0x123,0x10a,0xf2,0xdb,0xc5,0xb0,
0x9c,0x89,0x78,0x67,0x58,0x4a,0x3d,0x32,
0x27,0x1e,0x16,0xf,0xa,0x6,0x2,0x1,
0x0,0x1,0x2,0x6,0xa,0xf,0x16,0x1e,
0x27,0x32,0x3d,0x4a,0x58,0x67,0x78,0x89,
0x9c,0xb0,0xc5,0xdb,0xf2,0x10a,0x123,0x13e,
0x159,0x13e,0x123,0x10a,0xf2, 0xdb, 0xc5, 0xb0,
0x9c, 0x89, 0x78, 0x67, 0x58, 0x4a, 0x3d, 0x32,
0x27, 0x1e, 0x16, 0xf, 0xa, 0x6, 0x2, 0x1,
0x0, 0x1, 0x2, 0x6, 0xa, 0xf, 0x16, 0x1e,
0x27, 0x32, 0x3d, 0x4a, 0x58, 0x67, 0x78, 0x89,
0x9c, 0xb0, 0xc5, 0xdb, 0xf2, 0x10a,0x123,0x13e,
0x159,0x175,0x193,0x1b1,0x1d1,0x1f1,0x212,0x235,
0x258,0x27c,0x2a0,0x2c6,0x2ed,0x314,0x33c,0x365,
0x38e,0x3b8,0x3e3,0x40e,0x43a,0x467,0x494,0x4c2,
@ -152,7 +150,7 @@ static const dacsample_t dac_buffer[DAC_BUFFER_SIZE] = {
static const dacsample_t dac_buffer_triangle[DAC_BUFFER_SIZE] = {
// 256 values, max 4095
0x20,0x40,0x60,0x80,0xa0,0xc0,0xe0,0x100,
0x20, 0x40, 0x60, 0x80, 0xa0, 0xc0, 0xe0, 0x100,
0x120,0x140,0x160,0x180,0x1a0,0x1c0,0x1e0,0x200,
0x220,0x240,0x260,0x280,0x2a0,0x2c0,0x2e0,0x300,
0x320,0x340,0x360,0x380,0x3a0,0x3c0,0x3e0,0x400,
@ -183,64 +181,68 @@ static const dacsample_t dac_buffer_triangle[DAC_BUFFER_SIZE] = {
0x3e0,0x3c0,0x3a0,0x380,0x360,0x340,0x320,0x300,
0x2e0,0x2c0,0x2a0,0x280,0x260,0x240,0x220,0x200,
0x1e0,0x1c0,0x1a0,0x180,0x160,0x140,0x120,0x100,
0xe0,0xc0,0xa0,0x80,0x60,0x40,0x20,0x0
0xe0, 0xc0, 0xa0, 0x80, 0x60, 0x40, 0x20, 0x0
};
// static const dacsample_t dac_buffer_square[DAC_BUFFER_SIZE] = {
// // First half is max, second half is 0
// [0 ... DAC_BUFFER_SIZE/2-1] = DAC_SAMPLE_MAX,
// [DAC_BUFFER_SIZE/2 ... DAC_BUFFER_SIZE -1] = 0,
// };
static const dacsample_t dac_buffer_square[DAC_BUFFER_SIZE] = {
// First half is max, second half is 0
[0 ... DAC_BUFFER_SIZE/2-1] = DAC_SAMPLE_MAX,
[DAC_BUFFER_SIZE/2 ... DAC_BUFFER_SIZE -1] = 0,
};
dacsample_t dac_buffer_lr[DAC_BUFFER_SIZE] = { DAC_OFF_VALUE };
static dacsample_t dac_buffer_empty[DAC_BUFFER_SIZE] = { DAC_OFF_VALUE };
float dac_if[8] = {0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0};
/*
* DAC streaming callback.
/**
* DAC streaming callback. Does all of the main computing for sound synthesis.
*/
static void end_cb1(DACDriver * dacp, dacsample_t * samples, size_t rows) {
static void dac_end(DACDriver * dacp, dacsample_t * sample_p, size_t sample_count) {
(void)dacp;
(void)dac_buffer;
// (void)dac_buffer_triangle;
(void)dac_buffer_square;
uint8_t working_voices = voices;
if (working_voices > DAC_VOICES_MAX)
working_voices = DAC_VOICES_MAX;
for (uint8_t s = 0; s < rows; s++) {
for (uint8_t s = 0; s < sample_count; s++) {
if (working_voices > 0) {
uint16_t sample_sum = 0;
for (uint8_t i = 0; i < working_voices; i++) {
dac_if[i] = dac_if[i] + ((frequencies[i]*(float)DAC_BUFFER_SIZE)/(float)DAC_SAMPLE_RATE*1.5);
dac_if[i] = dac_if[i] + ((frequencies[i]*DAC_BUFFER_SIZE)/DAC_SAMPLE_RATE);
// Needed because % doesn't work with floats
// 0.5 less than the size because we use round() later
while(dac_if[i] >= (DAC_BUFFER_SIZE - 0.5))
while (dac_if[i] >= (DAC_BUFFER_SIZE))
dac_if[i] = dac_if[i] - DAC_BUFFER_SIZE;
uint16_t dac_i = (uint16_t)dac_if[i];
// Wavetable generation/lookup
// sine
// sample_sum += dac_buffer[(uint16_t)round(dac_if[i])] / working_voices;
// triangle wave (5 voices)
sample_sum += dac_buffer_triangle[(uint16_t)round(dac_if[i])] / working_voices;
// rising triangle (4 voices)
// SINE
sample_sum += dac_buffer[dac_i] / working_voices / 3;
// TRIANGLE
sample_sum += dac_buffer_triangle[dac_i] / working_voices / 3;
// RISING TRIANGLE
// sample_sum += (uint16_t)round((dac_if[i] * DAC_SAMPLE_MAX) / DAC_BUFFER_SIZE / working_voices );
// square (max 5 voices)
// SQUARE
// sample_sum += ((dac_if[i] > (DAC_BUFFER_SIZE / 2)) ? DAC_SAMPLE_MAX / working_voices: 0);
sample_sum += dac_buffer_square[dac_i] / working_voices / 3;
}
samples[s] = sample_sum;
sample_p[s] = sample_sum;
} else {
samples[s] = DAC_OFF_VALUE;
sample_p[s] = DAC_OFF_VALUE;
}
}
if (playing_notes) {
note_position += rows;
note_position += sample_count;
// end of the note
if ((note_position >= (note_length*420))) {
// End of the note - 35 is arbitary here, but gets us close to AVR's timing
if ((note_position >= (note_length*DAC_SAMPLE_RATE/35))) {
stop_note((*notes_pointer)[current_note][0]);
current_note++;
if (current_note >= notes_count) {
@ -255,34 +257,52 @@ static void end_cb1(DACDriver * dacp, dacsample_t * samples, size_t rows) {
envelope_index = 0;
note_length = ((*notes_pointer)[current_note][1] / 4) * (((float)note_tempo) / 100);
note_position = note_position - (note_length*420);
// note_position = 0;
// Skip forward in the next note's length if we've over shot the last, so
// the overall length of the song is the same
note_position = note_position - (note_length*DAC_SAMPLE_RATE/35);
}
}
}
/*
* DAC error callback.
*/
static void error_cb1(DACDriver *dacp, dacerror_t err) {
static void dac_error(DACDriver *dacp, dacerror_t err) {
(void)dacp;
(void)err;
chSysHalt("DAC failure");
chSysHalt("DAC failure. halp");
}
static const DACConfig dac1cfg1 = {
static const GPTConfig gpt6cfg1 = {
.frequency = DAC_SAMPLE_RATE * 3,
.callback = NULL,
.cr2 = TIM_CR2_MMS_1, /* MMS = 010 = TRGO on Update Event. */
.dier = 0U
};
static const DACConfig dac_conf = {
.init = DAC_SAMPLE_MAX,
.datamode = DAC_DHRM_12BIT_RIGHT
};
static const DACConversionGroup dacgrpcfg1 = {
/**
* @note The DAC_TRG(0) here selects the Timer 6 TRGO event, which is triggered
* on the rising edge after 3 APB1 clock cycles, causing our gpt6cfg1.frequency
* to be a third of what we expect.
*
* Here are all the values for DAC_TRG (TSEL in the ref manual)
* TIM15_TRGO 0b011
* TIM2_TRGO 0b100
* TIM3_TRGO 0b001
* TIM6_TRGO 0b000
* TIM7_TRGO 0b010
* EXTI9 0b110
* SWTRIG 0b111
*/
static const DACConversionGroup dac_conv_cfg = {
.num_channels = 1U,
.end_cb = end_cb1,
.error_cb = error_cb1,
.trigger = DAC_TRG(0)
.end_cb = dac_end,
.error_cb = dac_error,
.trigger = DAC_TRG(0b000)
};
void audio_init() {
@ -304,20 +324,20 @@ void audio_init() {
#endif
#endif // ARM EEPROM
palSetPadMode(GPIOA, 5, PAL_MODE_INPUT_ANALOG );
// palSetPadMode(GPIOA, 4, PAL_MODE_INPUT_ANALOG );
palSetPadMode(GPIOA, 4, PAL_MODE_OUTPUT_PUSHPULL );
palSetPad(GPIOA, 4);
// dacStart(&DACD1, &dac1cfg1);
// dacStartConversion(&DACD1, &dacgrpcfg1, dac_buffer_lr, 1);
dacStart(&DACD2, &dac1cfg1);
dacStartConversion(&DACD2, &dacgrpcfg1, dac_buffer_lr, DAC_BUFFER_SIZE);
#if defined(A4_AUDIO)
palSetPadMode(GPIOA, 4, PAL_MODE_INPUT_ANALOG );
dacStart(&DACD1, &dac_conf);
dacStartConversion(&DACD1, &dac_conv_cfg, dac_buffer_empty, DAC_BUFFER_SIZE);
#endif
#if defined(A5_AUDIO)
palSetPadMode(GPIOA, 5, PAL_MODE_INPUT_ANALOG );
dacStart(&DACD2, &dac_conf);
dacStartConversion(&DACD2, &dac_conv_cfg, dac_buffer_empty, DAC_BUFFER_SIZE);
#endif
gptStart(&GPTD6, &gpt7cfg1);
gptStart(&GPTD6, &gpt6cfg1);
gptStartContinuous(&GPTD6, 2U);
// gptStart(&GPTD7, &gpt7cfg1);
// gptStartContinuous(&GPTD7, 2U);
audio_initialized = true;
@ -337,13 +357,10 @@ void stop_all_notes() {
}
voices = 0;
gptStopTimer(&GPTD8);
playing_notes = false;
playing_note = false;
frequency = 0;
frequency_alt = 0;
volume = 0;
for (uint8_t i = 0; i < 8; i++)
{
@ -382,7 +399,6 @@ void stop_note(float freq) {
if (voices == 0) {
frequency = 0;
frequency_alt = 0;
volume = 0;
playing_note = false;
}
}
@ -428,6 +444,7 @@ void play_note(float freq, int vol) {
if (freq > 0) {
envelope_index = 0;
frequencies[voices] = freq;
dac_if[voices] = 0.0f;
volumes[voices] = vol;
voices++;
}
@ -450,7 +467,6 @@ void play_notes(float (*np)[][2], uint16_t n_count, bool n_repeat) {
notes_count = n_count;
notes_repeat = n_repeat;
place = 0;
current_note = 0;
note_length = ((*notes_pointer)[current_note][1] / 4) * (((float)note_tempo) / 100);

4
quantum/stm32/mcuconf.h
Unescape View File

@ -141,8 +141,8 @@
#define STM32_GPT_USE_TIM3 FALSE
#define STM32_GPT_USE_TIM4 FALSE
#define STM32_GPT_USE_TIM6 TRUE
#define STM32_GPT_USE_TIM7 TRUE
#define STM32_GPT_USE_TIM8 TRUE
#define STM32_GPT_USE_TIM7 FALSE
#define STM32_GPT_USE_TIM8 FALSE
#define STM32_GPT_TIM1_IRQ_PRIORITY 7
#define STM32_GPT_TIM2_IRQ_PRIORITY 7
#define STM32_GPT_TIM3_IRQ_PRIORITY 7

Write Preview
Loading…
Cancel
Save