Mirror/squeezelite-esp32
1
0
Fork 0
mirror of https://github.com/sle118/squeezelite-esp32.git synced 2025-12-08 12:37:01 +03:00
Code Issues Packages Projects Releases Wiki Activity

fix spdif for s3 and remove one override

Browse Source 
SPDIF on esp32 was partly incorrect due to word ordering and required i2s_hal override. This is not needed anymore as the "mistery" of SPDIF hack has been properly sorted out
This commit is contained in:
philippe44
2023-08-30 22:00:05 -07:00
parent e9ccd8eef7
commit 55053d5941
4 changed files with 35 additions and 296 deletions
Download Patch File Download Diff File Expand all files Collapse all files

34
components/squeezelite/output_i2s.c
View File

@@ -583,15 +583,11 @@ static void output_thread_i2s(void *arg) {
i2s_set_sample_rates(CONFIG_I2S_NUM, spdif.enabled ? i2s_config.sample_rate * 2 : i2s_config.sample_rate);
i2s_zero_dma_buffer(CONFIG_I2S_NUM);
#if BYTES_PER_FRAME == 4
equalizer_set_samplerate(output.current_sample_rate);
#endif
}
#if BYTES_PER_FRAME == 4
// run equalizer
equalizer_process(obuf, oframes * BYTES_PER_FRAME);
#endif
// we assume that here we have been able to entirely fill the DMA buffers
if (spdif.enabled) {
@@ -600,7 +596,7 @@ static void output_thread_i2s(void *arg) {
// need IRAM for speed but can't allocate a FRAME_BLOCK * 16, so process by smaller chunks
while (count < oframes) {
size_t chunk = min(SPDIF_BLOCK, oframes - count);
spdif_convert((ISAMPLE_T*) obuf + count * 2, chunk, (u32_t*) spdif.buf, &spdif.count);
spdif_convert((ISAMPLE_T*) obuf + count * 2, chunk, (u32_t*) spdif.buf, &spdif.count);
i2s_write(CONFIG_I2S_NUM, spdif.buf, chunk * 16, &obytes, portMAX_DELAY);
bytes += obytes / (16 / BYTES_PER_FRAME);
count += chunk;
@@ -709,12 +705,16 @@ static const u16_t spdif_bmclookup[256] = { //biphase mark encoded values (least
/*
SPDIF is supposed to be (before BMC encoding, from LSB to MSB)
PPPP AAAA SSSS SSSS SSSS SSSS SSSS VUCP
after BMC encoding, each bits becomes 2 hence this becomes a 64 bits word. The
the trick is to start not with a PPPP sequence but with an VUCP sequence to that
the 16 bits samples are aligned with a BMC word boundary. Note that the LSB of the
audio is transmitted first (not the MSB) and that ESP32 libray sends R then L,
contrary to what seems to be usually done, so (dst) order had to be changed
0.... 1... 191.. 0
BLFMRF MLFWRF MLFWRF BLFMRF (B,M,W=preamble-4, L/R=left/Right-24, F=Flags-4)
each xLF pattern is 32 bits
PPPP AAAA SSSS SSSS SSSS SSSS SSSS VUCP (P=preamble, A=auxiliary, S=sample-20bits, V=valid, U=user data, C=channel status, P=parity)
After BMC encoding, each bit becomes 2 hence this becomes a 64 bits word. The parity
is fixed by changing AAAA bits so that VUPC does not change. Then then trick is to
start not with a PPPP sequence but with an VUCP sequence to that the 16 bits samples
are aligned with a BMC word boundary. Input buffer is left first => LRLR...
The I2S interface must output first the B/M/W preamble which means that second
32 bits words must be first and so must be marked right channel.
*/
static void IRAM_ATTR spdif_convert(ISAMPLE_T *src, size_t frames, u32_t *dst, size_t *count) {
register u16_t hi, lo, aux;
@@ -729,7 +729,6 @@ static void IRAM_ATTR spdif_convert(ISAMPLE_T *src, size_t frames, u32_t *dst, s
// invert if last preceeding bit is 1
lo ^= ~((s16_t)hi) >> 16;
// first 16 bits
*dst++ = ((u32_t)lo << 16) | hi;
aux = 0xb333 ^ (((u32_t)((s16_t)lo)) >> 17);
#else
hi = spdif_bmclookup[(u8_t)(*src >> 24)];
@@ -738,39 +737,38 @@ static void IRAM_ATTR spdif_convert(ISAMPLE_T *src, size_t frames, u32_t *dst, s
// invert if last preceeding bit is 1
lo ^= ~((s16_t)hi) >> 16;
// first 16 bits
*dst++ = ((u32_t)lo << 16) | hi;
// we use 20 bits samples as we need to force parity
aux = spdif_bmclookup[(u8_t)(*src++ >> 12)];
aux = (u8_t) (aux ^ (~((s16_t)lo) >> 16));
aux |= (0xb3 ^ (((u16_t)((s8_t)aux)) >> 9)) << 8;
#endif
// VUCP-Bits: Valid, Subcode, Channelstatus, Parity = 0
// As parity is always 0, we can use fixed preambles
// set special preamble every 192 iteration
if (++cnt > 191) {
*dst++ = VUCP | (PREAMBLE_B << 16 ) | aux; //special preamble for one of 192 frames
cnt = 0;
} else {
*dst++ = VUCP | (PREAMBLE_M << 16) | aux;
}
}
// now write sample's 16 low bits
*dst++ = ((u32_t)lo << 16) | hi;
// then do right channel, no need to check PREAMBLE_B
#if BYTES_PER_FRAME == 4
hi = spdif_bmclookup[(u8_t)(*src >> 8)];
lo = spdif_bmclookup[(u8_t) *src++];
lo ^= ~((s16_t)hi) >> 16;
*dst++ = ((u32_t)lo << 16) | hi;
aux = 0xb333 ^ (((u32_t)((s16_t)lo)) >> 17);
#else
hi = spdif_bmclookup[(u8_t)(*src >> 24)];
lo = spdif_bmclookup[(u8_t)(*src >> 16)];
lo ^= ~((s16_t)hi) >> 16;
*dst++ = ((u32_t)lo << 16) | hi;
aux = spdif_bmclookup[(u8_t)(*src++ >> 12)];
aux = (u8_t) (aux ^ (~((s16_t)lo) >> 16));
aux |= (0xb3 ^ (((u16_t)((s8_t)aux)) >> 9)) << 8;
#endif
*dst++ = VUCP | (PREAMBLE_W << 16) | aux;
*dst++ = ((u32_t)lo << 16) | hi;
}
*count = cnt;