decoding. All muxing/demuxing functionality is now available in libavformat/dv.[ch]. * dv1394.c and avidec.c were hooked up with general DV demuxer. * DVAUDIO is dead! Long live pcm_s16le! * DV audio is now always recognized -- which means we can now hear all those ducks quaking in pond.dv. Originally committed as revision 2319 to svn://svn.ffmpeg.org/ffmpeg/trunktags/v0.5
@@ -103,7 +103,6 @@ void avcodec_register_all(void) | |||
register_avcodec(&mpeg_xvmc_decoder); | |||
#endif | |||
register_avcodec(&dvvideo_decoder); | |||
register_avcodec(&dvaudio_decoder); | |||
register_avcodec(&mjpeg_decoder); | |||
register_avcodec(&mjpegb_decoder); | |||
register_avcodec(&mp2_decoder); | |||
@@ -1388,7 +1388,6 @@ extern AVCodec rv10_decoder; | |||
extern AVCodec svq1_decoder; | |||
extern AVCodec svq3_decoder; | |||
extern AVCodec dvvideo_decoder; | |||
extern AVCodec dvaudio_decoder; | |||
extern AVCodec wmav1_decoder; | |||
extern AVCodec wmav2_decoder; | |||
extern AVCodec mjpeg_decoder; | |||
@@ -25,34 +25,29 @@ | |||
#include "dsputil.h" | |||
#include "mpegvideo.h" | |||
#include "simple_idct.h" | |||
#define NTSC_FRAME_SIZE 120000 | |||
#define PAL_FRAME_SIZE 144000 | |||
#define TEX_VLC_BITS 9 | |||
#include "dvdata.h" | |||
typedef struct DVVideoDecodeContext { | |||
AVCodecContext *avctx; | |||
const DVprofile* sys; | |||
GetBitContext gb; | |||
VLC *vlc; | |||
int sampling_411; /* 0 = 420, 1 = 411 */ | |||
int width, height; | |||
uint8_t *current_picture[3]; /* picture structure */ | |||
AVFrame picture; | |||
int linesize[3]; | |||
DCTELEM block[5*6][64] __align8; | |||
/* FIXME: the following is extracted from DSP */ | |||
uint8_t dv_zigzag[2][64]; | |||
uint8_t idct_permutation[64]; | |||
void (*get_pixels)(DCTELEM *block, const uint8_t *pixels, int line_size); | |||
void (*fdct)(DCTELEM *block); | |||
/* XXX: move it to static storage ? */ | |||
uint8_t dv_shift[2][22][64]; | |||
void (*idct_put[2])(uint8_t *dest, int line_size, DCTELEM *block); | |||
} DVVideoDecodeContext; | |||
#include "dvdata.h" | |||
static VLC dv_vlc; | |||
#define TEX_VLC_BITS 9 | |||
/* XXX: also include quantization */ | |||
static RL_VLC_ELEM *dv_rl_vlc[1]; | |||
static VLC_TYPE dv_vlc_codes[15][23]; | |||
static void dv_build_unquantize_tables(DVVideoDecodeContext *s) | |||
{ | |||
@@ -85,6 +80,7 @@ static int dvvideo_decode_init(AVCodecContext *avctx) | |||
if (!done) { | |||
int i; | |||
VLC dv_vlc; | |||
done = 1; | |||
@@ -114,6 +110,12 @@ static int dvvideo_decode_init(AVCodecContext *avctx) | |||
dv_rl_vlc[0][i].level = level; | |||
dv_rl_vlc[0][i].run = run; | |||
} | |||
memset(dv_vlc_codes, 0xff, sizeof(dv_vlc_codes)); | |||
for (i = 0; i < NB_DV_VLC - 1; i++) { | |||
if (dv_vlc_run[i] < 15 && dv_vlc_level[i] < 23 && dv_vlc_len[i] < 15) | |||
dv_vlc_codes[dv_vlc_run[i]][dv_vlc_level[i]] = i; | |||
} | |||
} | |||
/* ugly way to get the idct & scantable */ | |||
@@ -124,6 +126,9 @@ static int dvvideo_decode_init(AVCodecContext *avctx) | |||
if (DCT_common_init(&s2) < 0) | |||
return -1; | |||
s->get_pixels = s2.dsp.get_pixels; | |||
s->fdct = s2.dsp.fdct; | |||
s->idct_put[0] = s2.dsp.idct_put; | |||
memcpy(s->idct_permutation, s2.dsp.idct_permutation, 64); | |||
memcpy(s->dv_zigzag[0], s2.intra_scantable.permutated, 64); | |||
@@ -134,11 +139,11 @@ static int dvvideo_decode_init(AVCodecContext *avctx) | |||
/* XXX: do it only for constant case */ | |||
dv_build_unquantize_tables(s); | |||
return 0; | |||
} | |||
//#define VLC_DEBUG | |||
// #define VLC_DEBUG | |||
typedef struct BlockInfo { | |||
const uint8_t *shift_table; | |||
@@ -450,29 +455,29 @@ static inline void dv_decode_video_segment(DVVideoDecodeContext *s, | |||
v = *mb_pos_ptr++; | |||
mb_x = v & 0xff; | |||
mb_y = v >> 8; | |||
y_ptr = s->current_picture[0] + (mb_y * s->linesize[0] * 8) + (mb_x * 8); | |||
if (s->sampling_411) | |||
c_offset = (mb_y * s->linesize[1] * 8) + ((mb_x >> 2) * 8); | |||
y_ptr = s->picture.data[0] + (mb_y * s->picture.linesize[0] * 8) + (mb_x * 8); | |||
if (s->sys->pix_fmt == PIX_FMT_YUV411P) | |||
c_offset = (mb_y * s->picture.linesize[1] * 8) + ((mb_x >> 2) * 8); | |||
else | |||
c_offset = ((mb_y >> 1) * s->linesize[1] * 8) + ((mb_x >> 1) * 8); | |||
c_offset = ((mb_y >> 1) * s->picture.linesize[1] * 8) + ((mb_x >> 1) * 8); | |||
for(j = 0;j < 6; j++) { | |||
idct_put = s->idct_put[mb->dct_mode]; | |||
if (j < 4) { | |||
if (s->sampling_411 && mb_x < (704 / 8)) { | |||
if (s->sys->pix_fmt == PIX_FMT_YUV411P && mb_x < (704 / 8)) { | |||
/* NOTE: at end of line, the macroblock is handled as 420 */ | |||
idct_put(y_ptr + (j * 8), s->linesize[0], block); | |||
idct_put(y_ptr + (j * 8), s->picture.linesize[0], block); | |||
} else { | |||
idct_put(y_ptr + ((j & 1) * 8) + ((j >> 1) * 8 * s->linesize[0]), | |||
s->linesize[0], block); | |||
idct_put(y_ptr + ((j & 1) * 8) + ((j >> 1) * 8 * s->picture.linesize[0]), | |||
s->picture.linesize[0], block); | |||
} | |||
} else { | |||
if (s->sampling_411 && mb_x >= (704 / 8)) { | |||
if (s->sys->pix_fmt == PIX_FMT_YUV411P && mb_x >= (704 / 8)) { | |||
uint8_t pixels[64], *c_ptr, *c_ptr1, *ptr; | |||
int y, linesize; | |||
/* NOTE: at end of line, the macroblock is handled as 420 */ | |||
idct_put(pixels, 8, block); | |||
linesize = s->linesize[6 - j]; | |||
c_ptr = s->current_picture[6 - j] + c_offset; | |||
linesize = s->picture.linesize[6 - j]; | |||
c_ptr = s->picture.data[6 - j] + c_offset; | |||
ptr = pixels; | |||
for(y = 0;y < 8; y++) { | |||
/* convert to 411P */ | |||
@@ -486,8 +491,8 @@ static inline void dv_decode_video_segment(DVVideoDecodeContext *s, | |||
} | |||
} else { | |||
/* don't ask me why they inverted Cb and Cr ! */ | |||
idct_put(s->current_picture[6 - j] + c_offset, | |||
s->linesize[6 - j], block); | |||
idct_put(s->picture.data[6 - j] + c_offset, | |||
s->picture.linesize[6 - j], block); | |||
} | |||
} | |||
block += 64; | |||
@@ -496,7 +501,6 @@ static inline void dv_decode_video_segment(DVVideoDecodeContext *s, | |||
} | |||
} | |||
/* NOTE: exactly one frame must be given (120000 bytes for NTSC, | |||
144000 bytes for PAL) */ | |||
static int dvvideo_decode_frame(AVCodecContext *avctx, | |||
@@ -504,115 +508,35 @@ static int dvvideo_decode_frame(AVCodecContext *avctx, | |||
uint8_t *buf, int buf_size) | |||
{ | |||
DVVideoDecodeContext *s = avctx->priv_data; | |||
int sct, dsf, apt, ds, nb_dif_segs, vs, width, height, i, packet_size; | |||
uint8_t *buf_ptr; | |||
int ds, vs; | |||
const uint16_t *mb_pos_ptr; | |||
/* parse id */ | |||
init_get_bits(&s->gb, buf, buf_size*8); | |||
sct = get_bits(&s->gb, 3); | |||
if (sct != 0) | |||
return -1; | |||
skip_bits(&s->gb, 5); | |||
get_bits(&s->gb, 4); /* dsn (sequence number */ | |||
get_bits(&s->gb, 1); /* fsc (channel number) */ | |||
skip_bits(&s->gb, 3); | |||
get_bits(&s->gb, 8); /* dbn (diff block number 0-134) */ | |||
dsf = get_bits(&s->gb, 1); /* 0 = NTSC 1 = PAL */ | |||
if (get_bits(&s->gb, 1) != 0) | |||
return -1; | |||
skip_bits(&s->gb, 11); | |||
apt = get_bits(&s->gb, 3); /* apt */ | |||
get_bits(&s->gb, 1); /* tf1 */ | |||
skip_bits(&s->gb, 4); | |||
get_bits(&s->gb, 3); /* ap1 */ | |||
get_bits(&s->gb, 1); /* tf2 */ | |||
skip_bits(&s->gb, 4); | |||
get_bits(&s->gb, 3); /* ap2 */ | |||
get_bits(&s->gb, 1); /* tf3 */ | |||
skip_bits(&s->gb, 4); | |||
get_bits(&s->gb, 3); /* ap3 */ | |||
/* init size */ | |||
width = 720; | |||
if (dsf) { | |||
avctx->frame_rate = 25; | |||
avctx->frame_rate_base = 1; | |||
packet_size = PAL_FRAME_SIZE; | |||
height = 576; | |||
nb_dif_segs = 12; | |||
} else { | |||
avctx->frame_rate = 30000; | |||
avctx->frame_rate_base = 1001; | |||
packet_size = NTSC_FRAME_SIZE; | |||
height = 480; | |||
nb_dif_segs = 10; | |||
} | |||
/* NOTE: we only accept several full frames */ | |||
if (buf_size < packet_size) | |||
return -1; | |||
/* NTSC[dsf == 0] is always 720x480, 4:1:1 | |||
* PAL[dsf == 1] is always 720x576, 4:2:0 for IEC 68134[apt == 0] | |||
* but for the SMPTE 314M[apt == 1] it is 720x576, 4:1:1 | |||
*/ | |||
s->sampling_411 = !dsf || apt; | |||
if (s->sampling_411) { | |||
mb_pos_ptr = dsf ? dv_place_411P : dv_place_411; | |||
avctx->pix_fmt = PIX_FMT_YUV411P; | |||
} else { | |||
mb_pos_ptr = dv_place_420; | |||
avctx->pix_fmt = PIX_FMT_YUV420P; | |||
} | |||
avctx->width = width; | |||
avctx->height = height; | |||
/* Once again, this is pretty complicated by the fact that the same | |||
* field is used differently by IEC 68134[apt == 0] and | |||
* SMPTE 314M[apt == 1]. | |||
*/ | |||
if (buf[VAUX_TC61_OFFSET] == 0x61 && | |||
((apt == 0 && (buf[VAUX_TC61_OFFSET + 2] & 0x07) == 0x07) || | |||
(apt == 1 && (buf[VAUX_TC61_OFFSET + 2] & 0x07) == 0x02))) | |||
avctx->aspect_ratio = 16.0 / 9.0; | |||
else | |||
avctx->aspect_ratio = 4.0 / 3.0; | |||
s->sys = dv_frame_profile(buf); | |||
if (!s->sys || buf_size < s->sys->frame_size) | |||
return -1; /* NOTE: we only accept several full frames */ | |||
if(s->picture.data[0]) | |||
avctx->release_buffer(avctx, &s->picture); | |||
s->picture.reference= 0; | |||
s->picture.reference = 0; | |||
avctx->pix_fmt = s->sys->pix_fmt; | |||
if(avctx->get_buffer(avctx, &s->picture) < 0) { | |||
fprintf(stderr, "get_buffer() failed\n"); | |||
return -1; | |||
} | |||
for(i=0;i<3;i++) { | |||
s->current_picture[i] = s->picture.data[i]; | |||
s->linesize[i] = s->picture.linesize[i]; | |||
if (!s->current_picture[i]) | |||
return -1; | |||
} | |||
s->width = width; | |||
s->height = height; | |||
/* for each DIF segment */ | |||
buf_ptr = buf; | |||
for (ds = 0; ds < nb_dif_segs; ds++) { | |||
buf_ptr += 6 * 80; /* skip DIF segment header */ | |||
mb_pos_ptr = s->sys->video_place; | |||
for (ds = 0; ds < s->sys->difseg_size; ds++) { | |||
buf += 6 * 80; /* skip DIF segment header */ | |||
for(vs = 0; vs < 27; vs++) { | |||
if ((vs % 3) == 0) { | |||
/* skip audio block */ | |||
buf_ptr += 80; | |||
} | |||
dv_decode_video_segment(s, buf_ptr, mb_pos_ptr); | |||
buf_ptr += 5 * 80; | |||
if ((vs % 3) == 0) | |||
buf += 80; /* skip audio block */ | |||
dv_decode_video_segment(s, buf, mb_pos_ptr); | |||
buf += 5 * 80; | |||
mb_pos_ptr += 5; | |||
} | |||
} | |||
@@ -623,7 +547,7 @@ static int dvvideo_decode_frame(AVCodecContext *avctx, | |||
*data_size = sizeof(AVFrame); | |||
*(AVFrame*)data= s->picture; | |||
return packet_size; | |||
return s->sys->frame_size; | |||
} | |||
static int dvvideo_decode_end(AVCodecContext *avctx) | |||
@@ -645,158 +569,3 @@ AVCodec dvvideo_decoder = { | |||
CODEC_CAP_DR1, | |||
NULL | |||
}; | |||
typedef struct DVAudioDecodeContext { | |||
AVCodecContext *avctx; | |||
GetBitContext gb; | |||
} DVAudioDecodeContext; | |||
static int dvaudio_decode_init(AVCodecContext *avctx) | |||
{ | |||
// DVAudioDecodeContext *s = avctx->priv_data; | |||
return 0; | |||
} | |||
static uint16_t dv_audio_12to16(uint16_t sample) | |||
{ | |||
uint16_t shift, result; | |||
sample = (sample < 0x800) ? sample : sample | 0xf000; | |||
shift = (sample & 0xf00) >> 8; | |||
if (shift < 0x2 || shift > 0xd) { | |||
result = sample; | |||
} else if (shift < 0x8) { | |||
shift--; | |||
result = (sample - (256 * shift)) << shift; | |||
} else { | |||
shift = 0xe - shift; | |||
result = ((sample + ((256 * shift) + 1)) << shift) - 1; | |||
} | |||
return result; | |||
} | |||
/* NOTE: exactly one frame must be given (120000 bytes for NTSC, | |||
144000 bytes for PAL) | |||
There's a couple of assumptions being made here: | |||
1. By default we silence erroneous (0x8000/16bit 0x800/12bit) | |||
audio samples. We can pass them upwards when ffmpeg will be ready | |||
to deal with them. | |||
2. We don't do software emphasis. | |||
3. Audio is always returned as 16bit linear samples: 12bit | |||
nonlinear samples are converted into 16bit linear ones. | |||
*/ | |||
static int dvaudio_decode_frame(AVCodecContext *avctx, | |||
void *data, int *data_size, | |||
uint8_t *buf, int buf_size) | |||
{ | |||
DVVideoDecodeContext *s = avctx->priv_data; | |||
const uint16_t (*unshuffle)[9]; | |||
int smpls, freq, quant, sys, stride, difseg, ad, dp, nb_dif_segs, i; | |||
uint16_t lc, rc; | |||
uint8_t *buf_ptr; | |||
/* parse id */ | |||
init_get_bits(&s->gb, &buf[AAUX_AS_OFFSET], 5*8); | |||
i = get_bits(&s->gb, 8); | |||
if (i != 0x50) { /* No audio ? */ | |||
*data_size = 0; | |||
return buf_size; | |||
} | |||
get_bits(&s->gb, 1); /* 0 - locked audio, 1 - unlocked audio */ | |||
skip_bits(&s->gb, 1); | |||
smpls = get_bits(&s->gb, 6); /* samples in this frame - min. samples */ | |||
skip_bits(&s->gb, 8); | |||
skip_bits(&s->gb, 2); | |||
sys = get_bits(&s->gb, 1); /* 0 - 60 fields, 1 = 50 fields */ | |||
skip_bits(&s->gb, 5); | |||
get_bits(&s->gb, 1); /* 0 - emphasis on, 1 - emphasis off */ | |||
get_bits(&s->gb, 1); /* 0 - reserved, 1 - emphasis time constant 50/15us */ | |||
freq = get_bits(&s->gb, 3); /* 0 - 48KHz, 1 - 44,1kHz, 2 - 32 kHz */ | |||
quant = get_bits(&s->gb, 3); /* 0 - 16bit linear, 1 - 12bit nonlinear */ | |||
if (quant > 1) | |||
return -1; /* Unsupported quantization */ | |||
avctx->sample_rate = dv_audio_frequency[freq]; | |||
avctx->channels = 2; | |||
avctx->bit_rate = avctx->channels * avctx->sample_rate * 16; | |||
// What about: | |||
// avctx->frame_size = | |||
*data_size = (dv_audio_min_samples[sys][freq] + smpls) * | |||
avctx->channels * 2; | |||
if (sys) { | |||
nb_dif_segs = 12; | |||
stride = 108; | |||
unshuffle = dv_place_audio50; | |||
} else { | |||
nb_dif_segs = 10; | |||
stride = 90; | |||
unshuffle = dv_place_audio60; | |||
} | |||
/* for each DIF segment */ | |||
buf_ptr = buf; | |||
for (difseg = 0; difseg < nb_dif_segs; difseg++) { | |||
buf_ptr += 6 * 80; /* skip DIF segment header */ | |||
for (ad = 0; ad < 9; ad++) { | |||
for (dp = 8; dp < 80; dp+=2) { | |||
if (quant == 0) { /* 16bit quantization */ | |||
i = unshuffle[difseg][ad] + (dp - 8)/2 * stride; | |||
((short *)data)[i] = (buf_ptr[dp] << 8) | buf_ptr[dp+1]; | |||
if (((unsigned short *)data)[i] == 0x8000) | |||
((short *)data)[i] = 0; | |||
} else { /* 12bit quantization */ | |||
if (difseg >= nb_dif_segs/2) | |||
goto out; /* We're not doing 4ch at this time */ | |||
lc = ((uint16_t)buf_ptr[dp] << 4) | | |||
((uint16_t)buf_ptr[dp+2] >> 4); | |||
rc = ((uint16_t)buf_ptr[dp+1] << 4) | | |||
((uint16_t)buf_ptr[dp+2] & 0x0f); | |||
lc = (lc == 0x800 ? 0 : dv_audio_12to16(lc)); | |||
rc = (rc == 0x800 ? 0 : dv_audio_12to16(rc)); | |||
i = unshuffle[difseg][ad] + (dp - 8)/3 * stride; | |||
((short *)data)[i] = lc; | |||
i = unshuffle[difseg+nb_dif_segs/2][ad] + (dp - 8)/3 * stride; | |||
((short *)data)[i] = rc; | |||
++dp; | |||
} | |||
} | |||
buf_ptr += 16 * 80; /* 15 Video DIFs + 1 Audio DIF */ | |||
} | |||
} | |||
out: | |||
return buf_size; | |||
} | |||
static int dvaudio_decode_end(AVCodecContext *avctx) | |||
{ | |||
// DVAudioDecodeContext *s = avctx->priv_data; | |||
return 0; | |||
} | |||
AVCodec dvaudio_decoder = { | |||
"dvaudio", | |||
CODEC_TYPE_AUDIO, | |||
CODEC_ID_DVAUDIO, | |||
sizeof(DVAudioDecodeContext), | |||
dvaudio_decode_init, | |||
NULL, | |||
dvaudio_decode_end, | |||
dvaudio_decode_frame, | |||
0, | |||
NULL | |||
}; |
@@ -21,11 +21,34 @@ | |||
* @file dvdata.h | |||
* Constants for DV codec. | |||
*/ | |||
/* | |||
* DVprofile is used to express the differences between various | |||
* DV flavors. For now it's primarily used for differentiating | |||
* 525/60 and 625/50, but the plans are to use it for various | |||
* DV specs as well (e.g. SMPTE314M vs. IEC 61834). | |||
*/ | |||
typedef struct DVprofile { | |||
int dsf; /* value of the dsf in the DV header */ | |||
int frame_size; /* total size of one frame in bytes */ | |||
int difseg_size; /* number of DIF segments */ | |||
int frame_rate; | |||
int frame_rate_base; | |||
int ltc_divisor; /* FPS from the LTS standpoint */ | |||
int height; /* picture height in pixels */ | |||
int width; /* picture width in pixels */ | |||
const uint16_t *video_place; /* positions of all DV macro blocks */ | |||
enum PixelFormat pix_fmt; /* picture pixel format */ | |||
int audio_stride; /* size of audio_shuffle table */ | |||
int audio_min_samples[3];/* min ammount of audio samples */ | |||
/* for 48Khz, 44.1Khz and 32Khz */ | |||
int audio_samples_dist[5];/* how many samples are supposed to be */ | |||
/* in each frame in a 5 frames window */ | |||
const uint16_t (*audio_shuffle)[9]; /* PCM shuffling table */ | |||
} DVprofile; | |||
#define NB_DV_VLC 409 | |||
#define AAUX_AS_OFFSET (80*6 + 80*16*3 + 3) | |||
#define AAUX_ASC_OFFSET (80*6 + 80*16*4 + 3) | |||
#define VAUX_TC61_OFFSET (80*5 + 48 + 5) | |||
static const uint16_t dv_vlc_bits[409] = { | |||
0x0000, 0x0002, 0x0007, 0x0008, 0x0009, 0x0014, 0x0015, 0x0016, | |||
@@ -283,7 +306,7 @@ static const uint8_t dv_248_areas[64] = { | |||
1,2,2,3,3,3,3,3, | |||
}; | |||
static uint8_t dv_quant_shifts[22][4] = { | |||
static const uint8_t dv_quant_shifts[22][4] = { | |||
{ 3,3,4,4 }, | |||
{ 3,3,4,4 }, | |||
{ 2,3,3,4 }, | |||
@@ -1240,7 +1263,7 @@ static const uint16_t dv_place_411[1350] = { | |||
0x0834, 0x2320, 0x2f44, 0x3810, 0x1658, | |||
}; | |||
static const uint16_t dv_place_audio60[10][9] = { | |||
static const uint16_t dv_audio_shuffle525[10][9] = { | |||
{ 0, 30, 60, 20, 50, 80, 10, 40, 70 }, /* 1st channel */ | |||
{ 6, 36, 66, 26, 56, 86, 16, 46, 76 }, | |||
{ 12, 42, 72, 2, 32, 62, 22, 52, 82 }, | |||
@@ -1254,7 +1277,7 @@ static const uint16_t dv_place_audio60[10][9] = { | |||
{ 25, 55, 85, 15, 45, 75, 5, 35, 65 }, | |||
}; | |||
static const uint16_t dv_place_audio50[12][9] = { | |||
static const uint16_t dv_audio_shuffle625[12][9] = { | |||
{ 0, 36, 72, 26, 62, 98, 16, 52, 88}, /* 1st channel */ | |||
{ 6, 42, 78, 32, 68, 104, 22, 58, 94}, | |||
{ 12, 48, 84, 2, 38, 74, 28, 64, 100}, | |||
@@ -1271,10 +1294,77 @@ static const uint16_t dv_place_audio50[12][9] = { | |||
}; | |||
static const int dv_audio_frequency[3] = { | |||
48000, 44100, 32000, | |||
48000, 44100, 32000, | |||
}; | |||
static const int dv_audio_min_samples[2][3] = { | |||
{ 1580, 1452, 1053 }, /* 60 fields */ | |||
{ 1896, 1742, 1264 }, /* 50 fileds */ | |||
static const DVprofile dv_profiles[] = { | |||
{ .dsf = 0, | |||
.frame_size = 120000, /* IEC 61834, SMPTE-314M - 525/60 (NTSC) */ | |||
.difseg_size = 10, | |||
.frame_rate = 30000, | |||
.ltc_divisor = 30, | |||
.frame_rate_base = 1001, | |||
.height = 480, | |||
.width = 720, | |||
.video_place = dv_place_411, | |||
.pix_fmt = PIX_FMT_YUV411P, | |||
.audio_stride = 90, | |||
.audio_min_samples = { 1580, 1452, 1053 }, /* for 48, 44.1 and 32Khz */ | |||
.audio_samples_dist = { 1602, 1601, 1602, 1601, 1602 }, | |||
.audio_shuffle = dv_audio_shuffle525, | |||
}, | |||
{ .dsf = 1, | |||
.frame_size = 144000, /* IEC 61834 - 625/50 (PAL) */ | |||
.difseg_size = 12, | |||
.frame_rate = 25, | |||
.frame_rate_base = 1, | |||
.ltc_divisor = 25, | |||
.height = 576, | |||
.width = 720, | |||
.video_place = dv_place_420, | |||
.pix_fmt = PIX_FMT_YUV420P, | |||
.audio_stride = 108, | |||
.audio_min_samples = { 1896, 1742, 1264 }, /* for 48, 44.1 and 32Khz */ | |||
.audio_samples_dist = { 1920, 1920, 1920, 1920, 1920 }, | |||
.audio_shuffle = dv_audio_shuffle625, | |||
}, | |||
{ .dsf = 1, | |||
.frame_size = 144000, /* SMPTE-314M - 625/50 (PAL) */ | |||
.difseg_size = 12, | |||
.frame_rate = 25, | |||
.frame_rate_base = 1, | |||
.ltc_divisor = 25, | |||
.height = 576, | |||
.width = 720, | |||
.video_place = dv_place_411P, | |||
.pix_fmt = PIX_FMT_YUV411P, | |||
.audio_stride = 108, | |||
.audio_min_samples = { 1896, 1742, 1264 }, /* for 48, 44.1 and 32Khz */ | |||
.audio_samples_dist = { 1920, 1920, 1920, 1920, 1920 }, | |||
.audio_shuffle = dv_audio_shuffle625, | |||
} | |||
}; | |||
static inline const DVprofile* dv_frame_profile(uint8_t* frame) | |||
{ | |||
if ((frame[3] & 0x80) == 0) { /* DSF flag */ | |||
return &dv_profiles[0]; | |||
} | |||
else if ((frame[5] & 0x07) == 0) { /* APT flag */ | |||
return &dv_profiles[1]; | |||
} | |||
else | |||
return &dv_profiles[2]; | |||
} | |||
static inline const DVprofile* dv_codec_profile(AVCodecContext* codec) | |||
{ | |||
if (codec->width != 720) { | |||
return NULL; | |||
} | |||
else if (codec->height == 480) { | |||
return &dv_profiles[0]; | |||
} | |||
else | |||
return &dv_profiles[1]; | |||
} |
@@ -13,7 +13,7 @@ PPOBJS= | |||
# mux and demuxes | |||
OBJS+=mpeg.o mpegts.o mpegtsenc.o ffm.o crc.o img.o raw.o rm.o \ | |||
avienc.o avidec.o wav.o swf.o au.o gif.o mov.o mpjpeg.o dvcore.o dv.o \ | |||
avienc.o avidec.o wav.o swf.o au.o gif.o mov.o mpjpeg.o dv.o \ | |||
yuv4mpeg.o 4xm.o flvenc.o flvdec.o movenc.o psxstr.o idroq.o ipmovie.o \ | |||
nut.o wc3movie.o mp3.o | |||
@@ -18,21 +18,10 @@ | |||
*/ | |||
#include "avformat.h" | |||
#include "avi.h" | |||
#include "dv.h" | |||
//#define DEBUG | |||
static const struct AVI1Handler { | |||
enum CodecID vcid; | |||
enum CodecID acid; | |||
uint32_t tag; | |||
} AVI1Handlers[] = { | |||
{ CODEC_ID_DVVIDEO, CODEC_ID_DVAUDIO, MKTAG('d', 'v', 's', 'd') }, | |||
{ CODEC_ID_DVVIDEO, CODEC_ID_DVAUDIO, MKTAG('d', 'v', 'h', 'd') }, | |||
{ CODEC_ID_DVVIDEO, CODEC_ID_DVAUDIO, MKTAG('d', 'v', 's', 'l') }, | |||
/* This is supposed to be the last one */ | |||
{ CODEC_ID_NONE, CODEC_ID_NONE, 0 }, | |||
}; | |||
typedef struct AVIIndex { | |||
unsigned char tag[4]; | |||
unsigned int flags, pos, len; | |||
@@ -40,14 +29,11 @@ typedef struct AVIIndex { | |||
} AVIIndex; | |||
typedef struct { | |||
int64_t riff_end; | |||
int64_t movi_end; | |||
int type; | |||
uint8_t *buf; | |||
int buf_size; | |||
int stream_index; | |||
int64_t riff_end; | |||
int64_t movi_end; | |||
offset_t movi_list; | |||
AVIIndex *first, *last; | |||
void* dv_demux; | |||
} AVIContext; | |||
#ifdef DEBUG | |||
@@ -97,11 +83,6 @@ static int avi_read_header(AVFormatContext *s, AVFormatParameters *ap) | |||
stream_index = -1; | |||
codec_type = -1; | |||
frame_period = 0; | |||
avi->type = 2; | |||
avi->buf = av_malloc(1); | |||
if (!avi->buf) | |||
return -1; | |||
avi->buf_size = 1; | |||
for(;;) { | |||
if (url_feof(pb)) | |||
goto fail; | |||
@@ -134,7 +115,7 @@ static int avi_read_header(AVFormatContext *s, AVFormatParameters *ap) | |||
url_fskip(pb, 4 * 4); | |||
n = get_le32(pb); | |||
for(i=0;i<n;i++) { | |||
st = av_new_stream(s, 0); | |||
st = av_new_stream(s, i); | |||
if (!st) | |||
goto fail; | |||
} | |||
@@ -144,24 +125,36 @@ static int avi_read_header(AVFormatContext *s, AVFormatParameters *ap) | |||
/* stream header */ | |||
stream_index++; | |||
tag1 = get_le32(pb); | |||
handler = get_le32(pb); /* codec tag */ | |||
switch(tag1) { | |||
case MKTAG('i', 'a', 'v', 's'): | |||
case MKTAG('i', 'v', 'a', 's'): | |||
/* | |||
* After some consideration -- I don't think we | |||
* have to support anything but DV in a type1 AVIs. | |||
*/ | |||
if (s->nb_streams != 1) | |||
goto fail; | |||
avi->type = 1; | |||
avi->stream_index = 0; | |||
if (handler != MKTAG('d', 'v', 's', 'd') && | |||
handler != MKTAG('d', 'v', 'h', 'd') && | |||
handler != MKTAG('d', 'v', 's', 'l')) | |||
goto fail; | |||
avi->dv_demux = dv_init_demux(s, stream_index, stream_index + 1); | |||
if (!avi->dv_demux) | |||
goto fail; | |||
stream_index++; | |||
case MKTAG('v', 'i', 'd', 's'): | |||
codec_type = CODEC_TYPE_VIDEO; | |||
if (stream_index >= s->nb_streams) { | |||
url_fskip(pb, size - 4); | |||
url_fskip(pb, size - 8); | |||
break; | |||
} | |||
st = s->streams[stream_index]; | |||
handler = get_le32(pb); /* codec tag */ | |||
get_le32(pb); /* flags */ | |||
get_le16(pb); /* priority */ | |||
get_le16(pb); /* language */ | |||
@@ -186,29 +179,6 @@ static int avi_read_header(AVFormatContext *s, AVFormatParameters *ap) | |||
st->codec.frame_rate_base * AV_TIME_BASE / | |||
st->codec.frame_rate; | |||
if (avi->type == 1) { | |||
AVStream *st; | |||
st = av_new_stream(s, 0); | |||
if (!st) | |||
goto fail; | |||
stream_index++; | |||
for (i=0; AVI1Handlers[i].tag != 0; ++i) | |||
if (AVI1Handlers[i].tag == handler) | |||
break; | |||
if (AVI1Handlers[i].tag != 0) { | |||
s->streams[0]->codec.codec_type = CODEC_TYPE_VIDEO; | |||
s->streams[0]->codec.codec_id = AVI1Handlers[i].vcid; | |||
s->streams[1]->codec.codec_type = CODEC_TYPE_AUDIO; | |||
s->streams[1]->codec.codec_id = AVI1Handlers[i].acid; | |||
} else { | |||
goto fail; | |||
} | |||
} | |||
url_fskip(pb, size - 9 * 4); | |||
break; | |||
case MKTAG('a', 'u', 'd', 's'): | |||
@@ -218,12 +188,11 @@ static int avi_read_header(AVFormatContext *s, AVFormatParameters *ap) | |||
codec_type = CODEC_TYPE_AUDIO; | |||
if (stream_index >= s->nb_streams) { | |||
url_fskip(pb, size - 4); | |||
url_fskip(pb, size - 8); | |||
break; | |||
} | |||
st = s->streams[stream_index]; | |||
get_le32(pb); /* tag */ | |||
get_le32(pb); /* flags */ | |||
get_le16(pb); /* priority */ | |||
get_le16(pb); /* language */ | |||
@@ -244,7 +213,7 @@ static int avi_read_header(AVFormatContext *s, AVFormatParameters *ap) | |||
break; | |||
case MKTAG('s', 't', 'r', 'f'): | |||
/* stream header */ | |||
if (stream_index >= s->nb_streams || avi->type == 1) { | |||
if (stream_index >= s->nb_streams || avi->dv_demux) { | |||
url_fskip(pb, size); | |||
} else { | |||
st = s->streams[stream_index]; | |||
@@ -305,7 +274,6 @@ static int avi_read_header(AVFormatContext *s, AVFormatParameters *ap) | |||
/* check stream number */ | |||
if (stream_index != s->nb_streams - 1) { | |||
fail: | |||
av_free(avi->buf); | |||
for(i=0;i<s->nb_streams;i++) { | |||
av_freep(&s->streams[i]->codec.extradata); | |||
av_freep(&s->streams[i]); | |||
@@ -316,31 +284,21 @@ static int avi_read_header(AVFormatContext *s, AVFormatParameters *ap) | |||
return 0; | |||
} | |||
static void __destruct_pkt(struct AVPacket *pkt) | |||
{ | |||
pkt->data = NULL; pkt->size = 0; | |||
return; | |||
} | |||
static int avi_read_packet(AVFormatContext *s, AVPacket *pkt) | |||
{ | |||
AVIContext *avi = s->priv_data; | |||
ByteIOContext *pb = &s->pb; | |||
int n, d[8], size, i; | |||
void* dstr; | |||
memset(d, -1, sizeof(int)*8); | |||
if (avi->type == 1 && avi->stream_index) { | |||
/* duplicate DV packet */ | |||
av_init_packet(pkt); | |||
pkt->data = avi->buf; | |||
pkt->size = avi->buf_size; | |||
pkt->destruct = __destruct_pkt; | |||
pkt->stream_index = avi->stream_index; | |||
avi->stream_index = !avi->stream_index; | |||
return 0; | |||
if (avi->dv_demux) { | |||
size = dv_get_packet(avi->dv_demux, pkt); | |||
if (size >= 0) | |||
return size; | |||
} | |||
for(i=url_ftell(pb); !url_feof(pb); i++) { | |||
int j; | |||
@@ -387,26 +345,24 @@ static int avi_read_packet(AVFormatContext *s, AVPacket *pkt) | |||
&& n < s->nb_streams | |||
&& i + size <= avi->movi_end) { | |||
if (avi->type == 1) { | |||
uint8_t *tbuf = av_realloc(avi->buf, size + FF_INPUT_BUFFER_PADDING_SIZE); | |||
if (!tbuf) | |||
return -1; | |||
avi->buf = tbuf; | |||
avi->buf_size = size; | |||
av_init_packet(pkt); | |||
pkt->data = avi->buf; | |||
pkt->size = avi->buf_size; | |||
pkt->destruct = __destruct_pkt; | |||
avi->stream_index = n; | |||
} else { | |||
av_new_packet(pkt, size); | |||
} | |||
av_new_packet(pkt, size); | |||
get_buffer(pb, pkt->data, size); | |||
if (size & 1) | |||
if (size & 1) { | |||
get_byte(pb); | |||
pkt->stream_index = n; | |||
pkt->flags |= PKT_FLAG_KEY; // FIXME: We really should read index for that | |||
return 0; | |||
size++; | |||
} | |||
if (avi->dv_demux) { | |||
dstr = pkt->destruct; | |||
size = dv_produce_packet(avi->dv_demux, pkt, | |||
pkt->data, pkt->size); | |||
pkt->destruct = dstr; | |||
} else { | |||
pkt->stream_index = n; | |||
pkt->flags |= PKT_FLAG_KEY; // FIXME: We really should read | |||
// index for that | |||
} | |||
return size; | |||
} | |||
} | |||
return -1; | |||
@@ -416,7 +372,6 @@ static int avi_read_close(AVFormatContext *s) | |||
{ | |||
int i; | |||
AVIContext *avi = s->priv_data; | |||
av_free(avi->buf); | |||
for(i=0;i<s->nb_streams;i++) { | |||
AVStream *st = s->streams[i]; | |||
@@ -424,6 +379,9 @@ static int avi_read_close(AVFormatContext *s) | |||
av_free(st->codec.extradata); | |||
} | |||
if (avi->dv_demux) | |||
av_free(avi->dv_demux); | |||
return 0; | |||
} | |||
@@ -1,4 +1,10 @@ | |||
/* | |||
* General DV muxer/demuxer | |||
* Copyright (c) 2003 Roman Shaposhnick | |||
* | |||
* Many thanks to Dan Dennedy <dan@dennedy.org> for providing wealth | |||
* of DV technical info. | |||
* | |||
* Raw DV format | |||
* Copyright (c) 2002 Fabrice Bellard. | |||
* | |||
@@ -16,38 +22,696 @@ | |||
* License along with this library; if not, write to the Free Software | |||
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA | |||
*/ | |||
#include <time.h> | |||
#include "avformat.h" | |||
#include "dvcore.h" | |||
#include "dvdata.h" | |||
typedef struct DVDemuxContext { | |||
int is_audio; | |||
uint8_t buf[144000]; | |||
int size; | |||
AVPacket audio_pkt; | |||
AVStream *vst; | |||
AVStream *ast; | |||
} DVDemuxContext; | |||
/* raw input */ | |||
static int dv_read_header(AVFormatContext *s, | |||
AVFormatParameters *ap) | |||
typedef struct DVMuxContext { | |||
const DVprofile* sys; /* Current DV profile. E.g.: 525/60, 625/50 */ | |||
uint8_t frame_buf[144000]; /* frame under contruction */ | |||
FifoBuffer audio_data; /* Fifo for storing excessive amounts of PCM */ | |||
int frames; /* Number of a current frame */ | |||
time_t start_time; /* Start time of recording */ | |||
uint8_t aspect; /* Aspect ID 0 - 4:3, 7 - 16:9 */ | |||
int has_audio; /* frame under contruction has audio */ | |||
int has_video; /* frame under contruction has video */ | |||
} DVMuxContext; | |||
enum dv_section_type { | |||
dv_sect_header = 0x1f, | |||
dv_sect_subcode = 0x3f, | |||
dv_sect_vaux = 0x56, | |||
dv_sect_audio = 0x76, | |||
dv_sect_video = 0x96, | |||
}; | |||
enum dv_pack_type { | |||
dv_header525 = 0x3f, /* see dv_write_pack for important details on */ | |||
dv_header625 = 0xbf, /* these two packs */ | |||
dv_timecode = 0x13, | |||
dv_audio_source = 0x50, | |||
dv_audio_control = 0x51, | |||
dv_audio_recdate = 0x52, | |||
dv_audio_rectime = 0x53, | |||
dv_video_source = 0x60, | |||
dv_video_control = 0x61, | |||
dv_viedo_recdate = 0x62, | |||
dv_video_rectime = 0x63, | |||
dv_unknown_pack = 0xff, | |||
}; | |||
/* | |||
* The reason why the following three big ugly looking tables are | |||
* here is my lack of DV spec IEC 61834. The tables were basically | |||
* constructed to make code that places packs in SSYB, VAUX and | |||
* AAUX blocks very simple and table-driven. They conform to the | |||
* SMPTE 314M and the output of my personal DV camcorder, neither | |||
* of which is sufficient for a reliable DV stream producing. Thus | |||
* while code is still in development I'll be gathering input from | |||
* people with different DV equipment and modifying the tables to | |||
* accommodate all the quirks. Later on, if possible, some of them | |||
* will be folded into smaller tables and/or switch-if logic. For | |||
* now, my only excuse is -- they don't eat up that much of a space. | |||
*/ | |||
static const int dv_ssyb_packs_dist[12][6] = { | |||
{ 0x13, 0x13, 0x13, 0x13, 0x13, 0x13 }, | |||
{ 0x13, 0x13, 0x13, 0x13, 0x13, 0x13 }, | |||
{ 0x13, 0x13, 0x13, 0x13, 0x13, 0x13 }, | |||
{ 0x13, 0x13, 0x13, 0x13, 0x13, 0x13 }, | |||
{ 0x13, 0x13, 0x13, 0x13, 0x13, 0x13 }, | |||
{ 0x13, 0x13, 0x13, 0x13, 0x13, 0x13 }, | |||
{ 0x13, 0x62, 0x63, 0x13, 0x62, 0x63 }, | |||
{ 0x13, 0x62, 0x63, 0x13, 0x62, 0x63 }, | |||
{ 0x13, 0x62, 0x63, 0x13, 0x62, 0x63 }, | |||
{ 0x13, 0x62, 0x63, 0x13, 0x62, 0x63 }, | |||
{ 0x13, 0x62, 0x63, 0x13, 0x62, 0x63 }, | |||
{ 0x13, 0x62, 0x63, 0x13, 0x62, 0x63 }, | |||
}; | |||
static const int dv_vaux_packs_dist[12][15] = { | |||
{ 0x60, 0x61, 0x62, 0x63, 0xff, 0xff, 0xff, 0xff, 0xff, | |||
0x60, 0x61, 0x62, 0x63, 0xff, 0xff }, | |||
{ 0x60, 0x61, 0x62, 0x63, 0xff, 0xff, 0xff, 0xff, 0xff, | |||
0x60, 0x61, 0x62, 0x63, 0xff, 0xff }, | |||
{ 0x60, 0x61, 0x62, 0x63, 0xff, 0xff, 0xff, 0xff, 0xff, | |||
0x60, 0x61, 0x62, 0x63, 0xff, 0xff }, | |||
{ 0x60, 0x61, 0x62, 0x63, 0xff, 0xff, 0xff, 0xff, 0xff, | |||
0x60, 0x61, 0x62, 0x63, 0xff, 0xff }, | |||
{ 0x60, 0x61, 0x62, 0x63, 0xff, 0xff, 0xff, 0xff, 0xff, | |||
0x60, 0x61, 0x62, 0x63, 0xff, 0xff }, | |||
{ 0x60, 0x61, 0x62, 0x63, 0xff, 0xff, 0xff, 0xff, 0xff, | |||
0x60, 0x61, 0x62, 0x63, 0xff, 0xff }, | |||
{ 0x60, 0x61, 0x62, 0x63, 0xff, 0xff, 0xff, 0xff, 0xff, | |||
0x60, 0x61, 0x62, 0x63, 0xff, 0xff }, | |||
{ 0x60, 0x61, 0x62, 0x63, 0xff, 0xff, 0xff, 0xff, 0xff, | |||
0x60, 0x61, 0x62, 0x63, 0xff, 0xff }, | |||
{ 0x60, 0x61, 0x62, 0x63, 0xff, 0xff, 0xff, 0xff, 0xff, | |||
0x60, 0x61, 0x62, 0x63, 0xff, 0xff }, | |||
{ 0x60, 0x61, 0x62, 0x63, 0xff, 0xff, 0xff, 0xff, 0xff, | |||
0x60, 0x61, 0x62, 0x63, 0xff, 0xff }, | |||
{ 0x60, 0x61, 0x62, 0x63, 0xff, 0xff, 0xff, 0xff, 0xff, | |||
0x60, 0x61, 0x62, 0x63, 0xff, 0xff }, | |||
{ 0x60, 0x61, 0x62, 0x63, 0xff, 0xff, 0xff, 0xff, 0xff, | |||
0x60, 0x61, 0x62, 0x63, 0xff, 0xff }, | |||
}; | |||
static const int dv_aaux_packs_dist[12][9] = { | |||
{ 0xff, 0xff, 0xff, 0x50, 0x51, 0x52, 0x53, 0xff, 0xff }, | |||
{ 0x50, 0x51, 0x52, 0x53, 0xff, 0xff, 0xff, 0xff, 0xff }, | |||
{ 0xff, 0xff, 0xff, 0x50, 0x51, 0x52, 0x53, 0xff, 0xff }, | |||
{ 0x50, 0x51, 0x52, 0x53, 0xff, 0xff, 0xff, 0xff, 0xff }, | |||
{ 0xff, 0xff, 0xff, 0x50, 0x51, 0x52, 0x53, 0xff, 0xff }, | |||
{ 0x50, 0x51, 0x52, 0x53, 0xff, 0xff, 0xff, 0xff, 0xff }, | |||
{ 0xff, 0xff, 0xff, 0x50, 0x51, 0x52, 0x53, 0xff, 0xff }, | |||
{ 0x50, 0x51, 0x52, 0x53, 0xff, 0xff, 0xff, 0xff, 0xff }, | |||
{ 0xff, 0xff, 0xff, 0x50, 0x51, 0x52, 0x53, 0xff, 0xff }, | |||
{ 0x50, 0x51, 0x52, 0x53, 0xff, 0xff, 0xff, 0xff, 0xff }, | |||
{ 0xff, 0xff, 0xff, 0x50, 0x51, 0x52, 0x53, 0xff, 0xff }, | |||
{ 0x50, 0x51, 0x52, 0x53, 0xff, 0xff, 0xff, 0xff, 0xff }, | |||
}; | |||
static inline uint16_t dv_audio_12to16(uint16_t sample) | |||
{ | |||
uint16_t shift, result; | |||
sample = (sample < 0x800) ? sample : sample | 0xf000; | |||
shift = (sample & 0xf00) >> 8; | |||
if (shift < 0x2 || shift > 0xd) { | |||
result = sample; | |||
} else if (shift < 0x8) { | |||
shift--; | |||
result = (sample - (256 * shift)) << shift; | |||
} else { | |||
shift = 0xe - shift; | |||
result = ((sample + ((256 * shift) + 1)) << shift) - 1; | |||
} | |||
return result; | |||
} | |||
static int dv_audio_frame_size(const DVprofile* sys, int frame) | |||
{ | |||
AVStream *vst, *ast; | |||
DVDemuxContext *c = s->priv_data; | |||
return sys->audio_samples_dist[frame % (sizeof(sys->audio_samples_dist)/ | |||
sizeof(sys->audio_samples_dist[0]))]; | |||
} | |||
vst = av_new_stream(s, 0); | |||
if (!vst) | |||
return AVERROR_NOMEM; | |||
vst->codec.codec_type = CODEC_TYPE_VIDEO; | |||
vst->codec.codec_id = CODEC_ID_DVVIDEO; | |||
vst->codec.bit_rate = 25000000; | |||
static int dv_write_pack(enum dv_pack_type pack_id, DVMuxContext *c, uint8_t* buf) | |||
{ | |||
struct tm tc; | |||
time_t ct; | |||
int ltc_frame; | |||
ast = av_new_stream(s, 1); | |||
if (!ast) | |||
return AVERROR_NOMEM; | |||
buf[0] = (uint8_t)pack_id; | |||
switch (pack_id) { | |||
case dv_header525: /* I can't imagine why these two weren't defined as real */ | |||
case dv_header625: /* packs in SMPTE314M -- they definitely look like ones */ | |||
buf[1] = 0xf8 | /* reserved -- always 1 */ | |||
(0 & 0x07); /* APT: Track application ID */ | |||
buf[2] = (0 << 7) | /* TF1: audio data is 0 - valid; 1 - invalid */ | |||
(0x0f << 3) | /* reserved -- always 1 */ | |||
(0 & 0x07); /* AP1: Audio application ID */ | |||
buf[3] = (0 << 7) | /* TF2: video data is 0 - valid; 1 - invalid */ | |||
(0x0f << 3) | /* reserved -- always 1 */ | |||
(0 & 0x07); /* AP2: Video application ID */ | |||
buf[4] = (0 << 7) | /* TF3: subcode(SSYB) is 0 - valid; 1 - invalid */ | |||
(0x0f << 3) | /* reserved -- always 1 */ | |||
(0 & 0x07); /* AP3: Subcode application ID */ | |||
break; | |||
case dv_timecode: | |||
ct = (time_t)(c->frames / ((float)c->sys->frame_rate / | |||
(float)c->sys->frame_rate_base)); | |||
localtime_r(&ct, &tc); | |||
/* | |||
* LTC drop-frame frame counter drops two frames (0 and 1) every | |||
* minute, unless it is exactly divisible by 10 | |||
*/ | |||
ltc_frame = (c->frames + 2*ct/60 - 2*ct/600) % c->sys->ltc_divisor; | |||
buf[1] = (0 << 7) | /* Color fame: 0 - unsync; 1 - sync mode */ | |||
(1 << 6) | /* Drop frame timecode: 0 - nondrop; 1 - drop */ | |||
((ltc_frame / 10) << 4) | /* Tens of frames */ | |||
(ltc_frame % 10); /* Units of frames */ | |||
buf[2] = (1 << 7) | /* Biphase mark polarity correction: 0 - even; 1 - odd */ | |||
((tc.tm_sec / 10) << 4) | /* Tens of seconds */ | |||
(tc.tm_sec % 10); /* Units of seconds */ | |||
buf[3] = (1 << 7) | /* Binary group flag BGF0 */ | |||
((tc.tm_min / 10) << 4) | /* Tens of minutes */ | |||
(tc.tm_min % 10); /* Units of minutes */ | |||
buf[4] = (1 << 7) | /* Binary group flag BGF2 */ | |||
(1 << 6) | /* Binary group flag BGF1 */ | |||
((tc.tm_hour / 10) << 4) | /* Tens of hours */ | |||
(tc.tm_hour % 10); /* Units of hours */ | |||
break; | |||
case dv_audio_source: /* AAUX source pack */ | |||
buf[1] = (0 << 7) | /* locked mode */ | |||
(1 << 6) | /* reserved -- always 1 */ | |||
(dv_audio_frame_size(c->sys, c->frames) - | |||
c->sys->audio_min_samples[0]); | |||
/* # of samples */ | |||
buf[2] = (0 << 7) | /* multi-stereo */ | |||
(0 << 5) | /* #of audio channels per block: 0 -- 1 channel */ | |||
(0 << 4) | /* pair bit: 0 -- one pair of channels */ | |||
0; /* audio mode */ | |||
buf[3] = (1 << 7) | /* res */ | |||
(1 << 6) | /* multi-language flag */ | |||
(c->sys->dsf << 5) | /* system: 60fields/50fields */ | |||
0; /* definition: 0 -- SD (525/625) */ | |||
buf[4] = (1 << 7) | /* emphasis: 1 -- off */ | |||
(0 << 6) | /* emphasis time constant: 0 -- reserved */ | |||
(0 << 3) | /* frequency: 0 -- 48Khz, 1 -- 44,1Khz, 2 -- 32Khz */ | |||
0; /* quantization: 0 -- 16bit linear, 1 -- 12bit nonlinear */ | |||
break; | |||
case dv_audio_control: | |||
buf[1] = (0 << 6) | /* copy protection: 0 -- unrestricted */ | |||
(1 << 4) | /* input source: 1 -- digital input */ | |||
(3 << 2) | /* compression: 3 -- no information */ | |||
0; /* misc. info/SMPTE emphasis off */ | |||
buf[2] = (1 << 7) | /* recording start point: 1 -- no */ | |||
(1 << 6) | /* recording end point: 1 -- no */ | |||
(1 << 3) | /* recording mode: 1 -- original */ | |||
7; | |||
buf[3] = (1 << 7) | /* direction: 1 -- forward */ | |||
0x20; /* speed */ | |||
buf[4] = (1 << 7) | /* reserved -- always 1 */ | |||
0x7f; /* genre category */ | |||
break; | |||
case dv_audio_recdate: | |||
case dv_viedo_recdate: /* VAUX recording date */ | |||
ct = c->start_time + (time_t)(c->frames / | |||
((float)c->sys->frame_rate / (float)c->sys->frame_rate_base)); | |||
localtime_r(&ct, &tc); | |||
buf[1] = 0xff; /* ds, tm, tens of time zone, units of time zone */ | |||
/* 0xff is very likely to be "unknown" */ | |||
buf[2] = (3 << 6) | /* reserved -- always 1 */ | |||
((tc.tm_mday / 10) << 4) | /* Tens of day */ | |||
(tc.tm_mday % 10); /* Units of day */ | |||
buf[3] = /* we set high 4 bits to 0, shouldn't we set them to week? */ | |||
(tc.tm_mon % 10); /* Units of month */ | |||
buf[4] = (((tc.tm_year % 100) / 10) << 4) | /* Tens of year */ | |||
(tc.tm_year % 10); /* Units of year */ | |||
break; | |||
case dv_audio_rectime: /* AAUX recording time */ | |||
case dv_video_rectime: /* VAUX recording time */ | |||
ct = c->start_time + (time_t)(c->frames / | |||
((float)c->sys->frame_rate / (float)c->sys->frame_rate_base)); | |||
localtime_r(&ct, &tc); | |||
buf[1] = (3 << 6) | /* reserved -- always 1 */ | |||
0x3f; /* tens of frame, units of frame: 0x3f - "unknown" ? */ | |||
buf[2] = (1 << 7) | /* reserved -- always 1 */ | |||
((tc.tm_sec / 10) << 4) | /* Tens of seconds */ | |||
(tc.tm_sec % 10); /* Units of seconds */ | |||
buf[3] = (1 << 7) | /* reserved -- always 1 */ | |||
((tc.tm_min / 10) << 4) | /* Tens of minutes */ | |||
(tc.tm_min % 10); /* Units of minutes */ | |||
buf[4] = (3 << 6) | /* reserved -- always 1 */ | |||
((tc.tm_hour / 10) << 4) | /* Tens of hours */ | |||
(tc.tm_hour % 10); /* Units of hours */ | |||
break; | |||
case dv_video_source: | |||
buf[1] = 0xff; /* reserved -- always 1 */ | |||
buf[2] = (1 << 7) | /* B/W: 0 - b/w, 1 - color */ | |||
(1 << 6) | /* following CLF is valid - 0, invalid - 1 */ | |||
(3 << 4) | /* CLF: color frames id (see ITU-R BT.470-4) */ | |||
0xf; /* reserved -- always 1 */ | |||
buf[3] = (3 << 6) | /* reserved -- always 1 */ | |||
(c->sys->dsf << 5) | /* system: 60fields/50fields */ | |||
0; /* signal type video compression */ | |||
buf[4] = 0xff; /* VISC: 0xff -- no information */ | |||
break; | |||
case dv_video_control: | |||
buf[1] = (0 << 6) | /* Copy generation management (CGMS) 0 -- free */ | |||
0x3f; /* reserved -- always 1 */ | |||
buf[2] = 0xc8 | /* reserved -- always b11001xxx */ | |||
c->aspect; | |||
buf[3] = (1 << 7) | /* Frame/field flag 1 -- frame, 0 -- field */ | |||
(1 << 6) | /* First/second field flag 0 -- field 2, 1 -- field 1 */ | |||
(1 << 5) | /* Frame change flag 0 -- same picture as before, 1 -- different */ | |||
(1 << 4) | /* 1 - interlaced, 0 - noninterlaced */ | |||
0xc; /* reserved -- always b1100 */ | |||
buf[4] = 0xff; /* reserved -- always 1 */ | |||
break; | |||
default: | |||
buf[1] = buf[2] = buf[3] = buf[4] = 0xff; | |||
} | |||
return 5; | |||
} | |||
ast->codec.codec_type = CODEC_TYPE_AUDIO; | |||
ast->codec.codec_id = CODEC_ID_DVAUDIO; | |||
c->is_audio = 0; | |||
static inline int dv_write_dif_id(enum dv_section_type t, uint8_t seq_num, | |||
uint8_t dif_num, uint8_t* buf) | |||
{ | |||
buf[0] = (uint8_t)t; /* Section type */ | |||
buf[1] = (seq_num<<4) | /* DIF seq number 0-9 for 525/60; 0-11 for 625/50 */ | |||
(0 << 3) | /* FSC: for 50Mb/s 0 - first channel; 1 - second */ | |||
7; /* reserved -- always 1 */ | |||
buf[2] = dif_num; /* DIF block number Video: 0-134, Audio: 0-8 */ | |||
return 3; | |||
} | |||
return 0; | |||
static inline int dv_write_ssyb_id(uint8_t syb_num, uint8_t fr, uint8_t* buf) | |||
{ | |||
if (syb_num == 0 || syb_num == 6) { | |||
buf[0] = (fr<<7) | /* FR ID 1 - first half of each channel; 0 - second */ | |||
(0<<4) | /* AP3 (Subcode application ID) */ | |||
0x0f; /* reserved -- always 1 */ | |||
} | |||
else if (syb_num == 11) { | |||
buf[0] = (fr<<7) | /* FR ID 1 - first half of each channel; 0 - second */ | |||
0x7f; /* reserved -- always 1 */ | |||
} | |||
else { | |||
buf[0] = (fr<<7) | /* FR ID 1 - first half of each channel; 0 - second */ | |||
(0<<4) | /* APT (Track application ID) */ | |||
0x0f; /* reserved -- always 1 */ | |||
} | |||
buf[1] = 0xf0 | /* reserved -- always 1 */ | |||
(syb_num & 0x0f); /* SSYB number 0 - 11 */ | |||
buf[2] = 0xff; /* reserved -- always 1 */ | |||
return 3; | |||
} | |||
static void dv_format_frame(DVMuxContext *c, uint8_t* buf) | |||
{ | |||
int i, j, k; | |||
for (i = 0; i < c->sys->difseg_size; i++) { | |||
memset(buf, 0xff, 80 * 6); /* First 6 DIF blocks are for control data */ | |||
/* DV header: 1DIF */ | |||
buf += dv_write_dif_id(dv_sect_header, i, 0, buf); | |||
buf += dv_write_pack((c->sys->dsf ? dv_header625 : dv_header525), c, buf); | |||
buf += 72; /* unused bytes */ | |||
/* DV subcode: 2DIFs */ | |||
for (j = 0; j < 2; j++) { | |||
buf += dv_write_dif_id( dv_sect_subcode, i, j, buf); | |||
for (k = 0; k < 6; k++) { | |||
buf += dv_write_ssyb_id(k, (i < c->sys->difseg_size/2), buf); | |||
buf += dv_write_pack(dv_ssyb_packs_dist[i][k], c, buf); | |||
} | |||
buf += 29; /* unused bytes */ | |||
} | |||
/* DV VAUX: 3DIFs */ | |||
for (j = 0; j < 3; j++) { | |||
buf += dv_write_dif_id(dv_sect_vaux, i, j, buf); | |||
for (k = 0; k < 15 ; k++) | |||
buf += dv_write_pack(dv_vaux_packs_dist[i][k], c, buf); | |||
buf += 2; /* unused bytes */ | |||
} | |||
/* DV Audio/Video: 135 Video DIFs + 9 Audio DIFs */ | |||
for (j = 0; j < 135; j++) { | |||
if (j%15 == 0) { | |||
buf += dv_write_dif_id(dv_sect_audio, i, j/15, buf); | |||
buf += dv_write_pack(dv_aaux_packs_dist[i][j/15], c, buf); | |||
buf += 72; /* shuffled PCM audio */ | |||
} | |||
buf += dv_write_dif_id(dv_sect_video, i, j, buf); | |||
buf += 77; /* 1 video macro block: 1 bytes control | |||
4 * 14 bytes Y 8x8 data | |||
10 bytes Cr 8x8 data | |||
10 bytes Cb 8x8 data */ | |||
} | |||
} | |||
} | |||
static void dv_inject_audio(DVMuxContext *c, const uint8_t* pcm, uint8_t* frame_ptr) | |||
{ | |||
int i, j, d, of; | |||
for (i = 0; i < c->sys->difseg_size; i++) { | |||
frame_ptr += 6 * 80; /* skip DIF segment header */ | |||
for (j = 0; j < 9; j++) { | |||
for (d = 8; d < 80; d+=2) { | |||
of = c->sys->audio_shuffle[i][j] + (d - 8)/2 * c->sys->audio_stride; | |||
frame_ptr[d] = pcm[of*2+1]; // FIXME: may be we have to admit | |||
frame_ptr[d+1] = pcm[of*2]; // that DV is a big endian PCM | |||
} | |||
frame_ptr += 16 * 80; /* 15 Video DIFs + 1 Audio DIF */ | |||
} | |||
} | |||
} | |||
static void dv_inject_video(DVMuxContext *c, const uint8_t* video_data, uint8_t* frame_ptr) | |||
{ | |||
int i, j; | |||
int ptr = 0; | |||
for (i = 0; i < c->sys->difseg_size; i++) { | |||
ptr += 6 * 80; /* skip DIF segment header */ | |||
for (j = 0; j < 135; j++) { | |||
if (j%15 == 0) | |||
ptr += 80; /* skip Audio DIF */ | |||
ptr += 3; | |||
memcpy(frame_ptr + ptr, video_data + ptr, 77); | |||
ptr += 77; | |||
} | |||
} | |||
} | |||
/* | |||
* This is the dumbest implementation of all -- it simply looks at | |||
* a fixed offset and if pack isn't there -- fails. We might want | |||
* to have a fallback mechanism for complete search of missing packs. | |||
*/ | |||
static const uint8_t* dv_extract_pack(uint8_t* frame, enum dv_pack_type t) | |||
{ | |||
int offs; | |||
switch (t) { | |||
case dv_audio_source: | |||
offs = (80*6 + 80*16*3 + 3); | |||
break; | |||
case dv_audio_control: | |||
offs = (80*6 + 80*16*4 + 3); | |||
break; | |||
case dv_video_control: | |||
offs = (80*5 + 48 + 5); | |||
break; | |||
default: | |||
return NULL; | |||
} | |||
return (frame[offs] == t ? &frame[offs] : NULL); | |||
} | |||
/* | |||
* There's a couple of assumptions being made here: | |||
* 1. By default we silence erroneous (0x8000/16bit 0x800/12bit) audio samples. | |||
* We can pass them upwards when ffmpeg will be ready to deal with them. | |||
* 2. We don't do software emphasis. | |||
* 3. Audio is always returned as 16bit linear samples: 12bit nonlinear samples | |||
* are converted into 16bit linear ones. | |||
*/ | |||
static int dv_extract_audio(uint8_t* frame, uint8_t* pcm) | |||
{ | |||
int size, i, j, d, of, smpls, freq, quant; | |||
uint16_t lc, rc; | |||
const DVprofile* sys; | |||
const uint8_t* as_pack; | |||
as_pack = dv_extract_pack(frame, dv_audio_source); | |||
if (!as_pack) /* No audio ? */ | |||
return 0; | |||
sys = dv_frame_profile(frame); | |||
smpls = as_pack[1] & 0x3f; /* samples in this frame - min. samples */ | |||
freq = (as_pack[4] >> 3) & 0x07; /* 0 - 48KHz, 1 - 44,1kHz, 2 - 32 kHz */ | |||
quant = as_pack[4] & 0x07; /* 0 - 16bit linear, 1 - 12bit nonlinear */ | |||
if (quant > 1) | |||
return -1; /* Unsupported quantization */ | |||
size = (sys->audio_min_samples[freq] + smpls) * 4; /* 2ch, 2bytes */ | |||
/* for each DIF segment */ | |||
for (i = 0; i < sys->difseg_size; i++) { | |||
frame += 6 * 80; /* skip DIF segment header */ | |||
for (j = 0; j < 9; j++) { | |||
for (d = 8; d < 80; d += 2) { | |||
if (quant == 0) { /* 16bit quantization */ | |||
of = sys->audio_shuffle[i][j] + (d - 8)/2 * sys->audio_stride; | |||
if (of*2 >= size) | |||
continue; | |||
pcm[of*2] = frame[d+1]; // FIXME: may be we have to admit | |||
pcm[of*2+1] = frame[d]; // that DV is a big endian PCM | |||
if (pcm[of*2+1] == 0x80 && pcm[of*2] == 0x00) | |||
pcm[of*2+1] = 0; | |||
} else { /* 12bit quantization */ | |||
if (i >= sys->difseg_size/2) | |||
goto out; /* We're not doing 4ch at this time */ | |||
lc = ((uint16_t)frame[d] << 4) | | |||
((uint16_t)frame[d+2] >> 4); | |||
rc = ((uint16_t)frame[d+1] << 4) | | |||
((uint16_t)frame[d+2] & 0x0f); | |||
lc = (lc == 0x800 ? 0 : dv_audio_12to16(lc)); | |||
rc = (rc == 0x800 ? 0 : dv_audio_12to16(rc)); | |||
of = sys->audio_shuffle[i][j] + (d - 8)/3 * sys->audio_stride; | |||
if (of*2 >= size) | |||
continue; | |||
pcm[of*2] = lc & 0xff; // FIXME: may be we have to admit | |||
pcm[of*2+1] = lc >> 8; // that DV is a big endian PCM | |||
of = sys->audio_shuffle[i+sys->difseg_size/2][j] + | |||
(d - 8)/3 * sys->audio_stride; | |||
pcm[of*2] = rc & 0xff; // FIXME: may be we have to admit | |||
pcm[of*2+1] = rc >> 8; // that DV is a big endian PCM | |||
++d; | |||
} | |||
} | |||
frame += 16 * 80; /* 15 Video DIFs + 1 Audio DIF */ | |||
} | |||
} | |||
out: | |||
return size; | |||
} | |||
static int dv_extract_audio_info(uint8_t* frame, AVCodecContext* avctx) | |||
{ | |||
const uint8_t* as_pack; | |||
const DVprofile* sys; | |||
int freq, smpls; | |||
sys = dv_frame_profile(frame); | |||
as_pack = dv_extract_pack(frame, dv_audio_source); | |||
if (!as_pack || !sys) /* No audio ? */ | |||
return -1; | |||
smpls = as_pack[1] & 0x3f; /* samples in this frame - min. samples */ | |||
freq = (as_pack[4] >> 3) & 0x07; /* 0 - 48KHz, 1 - 44,1kHz, 2 - 32 kHz */ | |||
avctx->sample_rate = dv_audio_frequency[freq]; | |||
avctx->channels = 2; | |||
avctx->bit_rate = avctx->channels * avctx->sample_rate * 16; | |||
return (sys->audio_min_samples[freq] + smpls) * 4; /* 2ch, 2bytes */; | |||
} | |||
static int dv_extract_video_info(uint8_t* frame, AVCodecContext* avctx) | |||
{ | |||
const DVprofile* sys; | |||
const uint8_t* vsc_pack; | |||
int apt; | |||
int size = 0; | |||
sys = dv_frame_profile(frame); | |||
if (sys) { | |||
apt = frame[4] & 0x07; | |||
avctx->frame_rate = sys->frame_rate; | |||
avctx->frame_rate_base = sys->frame_rate_base; | |||
avctx->width = sys->width; | |||
avctx->height = sys->height; | |||
avctx->pix_fmt = sys->pix_fmt; | |||
vsc_pack = dv_extract_pack(frame, dv_video_control); | |||
if (vsc_pack && (vsc_pack[2] & 0x07) == (apt?0x02:0x07)) | |||
avctx->aspect_ratio = 16.0 / 9.0; | |||
else | |||
avctx->aspect_ratio = 4.0 / 3.0; | |||
size = sys->frame_size; | |||
} | |||
return size; | |||
} | |||
/* | |||
* The following 6 functions constitute our interface to the world | |||
*/ | |||
int dv_assemble_frame(DVMuxContext *c, AVStream* st, | |||
const uint8_t* data, int data_size, uint8_t** frame) | |||
{ | |||
uint8_t pcm[8192]; | |||
int fsize, reqasize; | |||
*frame = &c->frame_buf[0]; | |||
if (c->has_audio && c->has_video) { /* must be a stale frame */ | |||
dv_format_frame(c, *frame); | |||
c->frames++; | |||
c->has_audio = c->has_video = 0; | |||
} | |||
if (st->codec.codec_type == CODEC_TYPE_VIDEO) { | |||
/* FIXME: we have to have more sensible approach than this one */ | |||
if (c->has_video) | |||
fprintf(stderr, "Can't process DV frame #%d. Insufficient audio data or severe sync problem.\n", c->frames); | |||
dv_inject_video(c, data, *frame); | |||
c->has_video = 1; | |||
data_size = 0; | |||
} | |||
reqasize = 4 * dv_audio_frame_size(c->sys, c->frames); | |||
fsize = fifo_size(&c->audio_data, c->audio_data.rptr); | |||
if (st->codec.codec_type == CODEC_TYPE_AUDIO || (c->has_video && fsize >= reqasize)) { | |||
if (fsize + data_size >= reqasize && !c->has_audio) { | |||
if (fsize >= reqasize) { | |||
fifo_read(&c->audio_data, &pcm[0], reqasize, &c->audio_data.rptr); | |||
} else { | |||
fifo_read(&c->audio_data, &pcm[0], fsize, &c->audio_data.rptr); | |||
memcpy(&pcm[fsize], &data[0], reqasize - fsize); | |||
data += reqasize - fsize; | |||
data_size -= reqasize - fsize; | |||
} | |||
dv_inject_audio(c, &pcm[0], *frame); | |||
c->has_audio = 1; | |||
} | |||
/* FIXME: we have to have more sensible approach than this one */ | |||
if (fifo_size(&c->audio_data, c->audio_data.rptr) + data_size >= 100*AVCODEC_MAX_AUDIO_FRAME_SIZE) | |||
fprintf(stderr, "Can't process DV frame #%d. Insufficient video data or severe sync problem.\n", c->frames); | |||
fifo_write(&c->audio_data, (uint8_t *)data, data_size, &c->audio_data.wptr); | |||
} | |||
return (c->has_audio && c->has_video) ? c->sys->frame_size : 0; | |||
} | |||
DVMuxContext* dv_init_mux(AVFormatContext* s) | |||
{ | |||
DVMuxContext *c; | |||
AVStream *vst; | |||
AVStream *ast; | |||
c = av_mallocz(sizeof(DVMuxContext)); | |||
if (!c) | |||
return NULL; | |||
if (s->nb_streams != 2) | |||
goto bail_out; | |||
/* We have to sort out where audio and where video stream is */ | |||
if (s->streams[0]->codec.codec_type == CODEC_TYPE_VIDEO && | |||
s->streams[1]->codec.codec_type == CODEC_TYPE_AUDIO) { | |||
vst = s->streams[0]; | |||
ast = s->streams[1]; | |||
} | |||
else if (s->streams[1]->codec.codec_type == CODEC_TYPE_VIDEO && | |||
s->streams[0]->codec.codec_type == CODEC_TYPE_AUDIO) { | |||
vst = s->streams[1]; | |||
ast = s->streams[0]; | |||
} else | |||
goto bail_out; | |||
/* Some checks -- DV format is very picky about its incoming streams */ | |||
if (vst->codec.codec_id != CODEC_ID_DVVIDEO || | |||
ast->codec.codec_id != CODEC_ID_PCM_S16LE) | |||
goto bail_out; | |||
if (ast->codec.sample_rate != 48000 || | |||
ast->codec.channels != 2) | |||
goto bail_out; | |||
c->sys = dv_codec_profile(&vst->codec); | |||
if (!c->sys) | |||
goto bail_out; | |||
/* Ok, everything seems to be in working order */ | |||
c->frames = 0; | |||
c->has_audio = c->has_video = 0; | |||
c->start_time = time(NULL); | |||
c->aspect = 0; /* 4:3 is the default */ | |||
if (vst->codec.aspect_ratio == 16.0 / 9.0) | |||
c->aspect = 0x07; | |||
if (fifo_init(&c->audio_data, 100*AVCODEC_MAX_AUDIO_FRAME_SIZE) < 0) | |||
goto bail_out; | |||
dv_format_frame(c, &c->frame_buf[0]); | |||
return c; | |||
bail_out: | |||
av_free(c); | |||
return NULL; | |||
} | |||
void dv_delete_mux(DVMuxContext *c) | |||
{ | |||
fifo_free(&c->audio_data); | |||
av_free(c); | |||
} | |||
DVDemuxContext* dv_init_demux(AVFormatContext *s, int vid, int aid) | |||
{ | |||
DVDemuxContext *c; | |||
c = av_mallocz(sizeof(DVDemuxContext)); | |||
if (!c) | |||
return NULL; | |||
c->vst = (vid < s->nb_streams) ? s->streams[vid] : av_new_stream(s, vid); | |||
c->ast = (aid < s->nb_streams) ? s->streams[aid] : av_new_stream(s, aid); | |||
if (!c->vst || !c->ast) | |||
goto fail; | |||
c->vst->codec.codec_type = CODEC_TYPE_VIDEO; | |||
c->vst->codec.codec_id = CODEC_ID_DVVIDEO; | |||
c->vst->codec.bit_rate = 25000000; | |||
c->ast->codec.codec_type = CODEC_TYPE_AUDIO; | |||
c->ast->codec.codec_id = CODEC_ID_PCM_S16LE; | |||
c->audio_pkt.size = 0; | |||
return c; | |||
fail: | |||
if (c->vst) | |||
av_free(c->vst); | |||
if (c->ast) | |||
av_free(c->ast); | |||
av_free(c); | |||
return NULL; | |||
} | |||
static void __destruct_pkt(struct AVPacket *pkt) | |||
@@ -56,43 +720,99 @@ static void __destruct_pkt(struct AVPacket *pkt) | |||
return; | |||
} | |||
static int dv_read_packet(AVFormatContext *s, AVPacket *pkt) | |||
int dv_get_packet(DVDemuxContext *c, AVPacket *pkt) | |||
{ | |||
int ret; | |||
DVDemuxContext *c = s->priv_data; | |||
if (!c->is_audio) { | |||
ret = get_buffer(&s->pb, c->buf, 4); | |||
if (ret <= 0) | |||
return -EIO; | |||
c->size = dv_frame_profile(&c->buf[0])->frame_size; | |||
ret = get_buffer(&s->pb, c->buf + 4, c->size - 4); | |||
if (ret <= 0) | |||
return -EIO; | |||
int size = -1; | |||
if (c->audio_pkt.size) { | |||
*pkt = c->audio_pkt; | |||
c->audio_pkt.size = 0; | |||
size = pkt->size; | |||
} | |||
return size; | |||
} | |||
int dv_produce_packet(DVDemuxContext *c, AVPacket *pkt, | |||
uint8_t* buf, int buf_size) | |||
{ | |||
int size; | |||
if (buf_size < 4 || buf_size < dv_frame_profile(buf)->frame_size) | |||
return -1; /* Broken frame, or not enough data */ | |||
/* Queueing audio packet */ | |||
/* FIXME: in case of no audio/bad audio we have to do something */ | |||
size = dv_extract_audio_info(buf, &c->ast->codec); | |||
if (av_new_packet(&c->audio_pkt, size) < 0) | |||
return AVERROR_NOMEM; | |||
dv_extract_audio(buf, c->audio_pkt.data); | |||
c->audio_pkt.size = size; | |||
c->audio_pkt.stream_index = c->ast->id; | |||
/* Now it's time to return video packet */ | |||
size = dv_extract_video_info(buf, &c->vst->codec); | |||
av_init_packet(pkt); | |||
pkt->destruct = __destruct_pkt; | |||
pkt->data = c->buf; | |||
pkt->size = c->size; | |||
pkt->stream_index = c->is_audio; | |||
pkt->data = buf; | |||
pkt->size = size; | |||
pkt->flags |= PKT_FLAG_KEY; | |||
pkt->stream_index = c->vst->id; | |||
c->is_audio = !c->is_audio; | |||
return c->size; | |||
return size; | |||
} | |||
/************************************************************ | |||
* Implementation of the easiest DV storage of all -- raw DV. | |||
************************************************************/ | |||
typedef struct RawDVContext { | |||
void* dv_demux; | |||
uint8_t buf[144000]; | |||
} RawDVContext; | |||
static int dv_read_header(AVFormatContext *s, | |||
AVFormatParameters *ap) | |||
{ | |||
RawDVContext *c = s->priv_data; | |||
c->dv_demux = dv_init_demux(s, 0, 1); | |||
return c->dv_demux ? 0 : -1; | |||
} | |||
static int dv_read_packet(AVFormatContext *s, AVPacket *pkt) | |||
{ | |||
int size; | |||
RawDVContext *c = s->priv_data; | |||
size = dv_get_packet(c->dv_demux, pkt); | |||
if (size < 0) { | |||
if (get_buffer(&s->pb, c->buf, 4) <= 0) | |||
return -EIO; | |||
size = dv_frame_profile(c->buf)->frame_size; | |||
if (get_buffer(&s->pb, c->buf + 4, size - 4) <= 0) | |||
return -EIO; | |||
size = dv_produce_packet(c->dv_demux, pkt, c->buf, size); | |||
} | |||
return size; | |||
} | |||
static int dv_read_close(AVFormatContext *s) | |||
{ | |||
RawDVContext *c = s->priv_data; | |||
av_free(c->dv_demux); | |||
return 0; | |||
} | |||
int dv_write_header(struct AVFormatContext *s) | |||
static int dv_write_header(AVFormatContext *s) | |||
{ | |||
DVMuxContext *c = s->priv_data; | |||
if (s->nb_streams != 2 || dv_core_init(c, s->streams) != 0) { | |||
s->priv_data = dv_init_mux(s); | |||
if (!s->priv_data) { | |||
fprintf(stderr, "Can't initialize DV format!\n" | |||
"Make sure that you supply exactly two streams:\n" | |||
" video: 25fps or 29.97fps, audio: 2ch/48Khz/PCM\n"); | |||
@@ -101,23 +821,20 @@ int dv_write_header(struct AVFormatContext *s) | |||
return 0; | |||
} | |||
int dv_write_packet(struct AVFormatContext *s, | |||
int stream_index, | |||
const uint8_t *buf, int size, int64_t pts) | |||
static int dv_write_packet(struct AVFormatContext *s, | |||
int stream_index, | |||
const uint8_t *buf, int size, int64_t pts) | |||
{ | |||
DVMuxContext *c = s->priv_data; | |||
uint8_t* frame; | |||
int fsize; | |||
if (stream_index == c->vst) | |||
dv_assemble_frame(c, buf, NULL, 0); | |||
else | |||
dv_assemble_frame(c, NULL, buf, size); | |||
if (c->has_audio && c->has_video) { | |||
put_buffer(&s->pb, &c->frame_buf[0], c->sys->frame_size); | |||
fsize = dv_assemble_frame((DVMuxContext *)s->priv_data, s->streams[stream_index], | |||
buf, size, &frame); | |||
if (fsize > 0) { | |||
put_buffer(&s->pb, frame, fsize); | |||
put_flush_packet(&s->pb); | |||
} | |||
return 0; | |||
} | |||
return fsize; | |||
} | |||
/* | |||
@@ -126,16 +843,16 @@ int dv_write_packet(struct AVFormatContext *s, | |||
* Currently we simply drop the last frame. I don't know whether this | |||
* is the best strategy of all | |||
*/ | |||
int dv_write_trailer(struct AVFormatContext *s) | |||
static int dv_write_trailer(struct AVFormatContext *s) | |||
{ | |||
dv_core_delete((DVMuxContext *)s->priv_data); | |||
dv_delete_mux((DVMuxContext *)s->priv_data); | |||
return 0; | |||
} | |||
static AVInputFormat dv_iformat = { | |||
"dv", | |||
"DV video format", | |||
sizeof(DVDemuxContext), | |||
sizeof(RawDVContext), | |||
NULL, | |||
dv_read_header, | |||
dv_read_packet, | |||
@@ -143,7 +860,7 @@ static AVInputFormat dv_iformat = { | |||
.extensions = "dv", | |||
}; | |||
AVOutputFormat dv_oformat = { | |||
static AVOutputFormat dv_oformat = { | |||
"dv", | |||
"DV video format", | |||
NULL, | |||
@@ -0,0 +1,32 @@ | |||
/* | |||
* General DV muxer/demuxer | |||
* Copyright (c) 2003 Roman Shaposhnick | |||
* | |||
* Many thanks to Dan Dennedy <dan@dennedy.org> for providing wealth | |||
* of DV technical info. | |||
* | |||
* Raw DV format | |||
* Copyright (c) 2002 Fabrice Bellard. | |||
* | |||
* This library is free software; you can redistribute it and/or | |||
* modify it under the terms of the GNU Lesser General Public | |||
* License as published by the Free Software Foundation; either | |||
* version 2 of the License, or (at your option) any later version. | |||
* | |||
* This library is distributed in the hope that it will be useful, | |||
* but WITHOUT ANY WARRANTY; without even the implied warranty of | |||
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | |||
* Lesser General Public License for more details. | |||
* | |||
* You should have received a copy of the GNU Lesser General Public | |||
* License along with this library; if not, write to the Free Software | |||
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA | |||
*/ | |||
void* dv_init_demux(AVFormatContext *s, int vid, int aid); | |||
int dv_get_packet(void*, AVPacket *); | |||
int dv_produce_packet(void*, AVPacket*, uint8_t*, int); | |||
void* dv_init_mux(AVFormatContext* s); | |||
int dv_assemble_frame(void *c, AVStream*, const uint8_t*, int, uint8_t**); | |||
void dv_delete_mux(void*); |
@@ -31,13 +31,11 @@ | |||
#undef DV1394_DEBUG | |||
#include "dv1394.h" | |||
#include "dv.h" | |||
struct dv1394_data { | |||
int fd; | |||
int channel; | |||
int width, height; | |||
int frame_rate; | |||
int frame_size; | |||
int format; | |||
void *ring; /* Ring buffer */ | |||
@@ -45,9 +43,9 @@ struct dv1394_data { | |||
int avail; /* Number of frames available for reading */ | |||
int done; /* Number of completed frames */ | |||
int stream; /* Current stream. 0 - video, 1 - audio */ | |||
int64_t pts; /* Current timestamp */ | |||
AVStream *vst, *ast; | |||
void* dv_demux; /* Generic DV muxing/demuxing context */ | |||
}; | |||
/* | |||
@@ -69,7 +67,6 @@ static int dv1394_reset(struct dv1394_data *dv) | |||
return -1; | |||
dv->avail = dv->done = 0; | |||
dv->stream = 0; | |||
return 0; | |||
} | |||
@@ -88,14 +85,9 @@ static int dv1394_read_header(AVFormatContext * context, AVFormatParameters * ap | |||
struct dv1394_data *dv = context->priv_data; | |||
const char *video_device; | |||
dv->vst = av_new_stream(context, 0); | |||
if (!dv->vst) | |||
return -ENOMEM; | |||
dv->ast = av_new_stream(context, 1); | |||
if (!dv->ast) { | |||
av_free(dv->vst); | |||
return -ENOMEM; | |||
} | |||
dv->dv_demux = dv_init_demux(context, 0, 1); | |||
if (!dv->dv_demux) | |||
goto failed; | |||
if (ap->standard && !strcasecmp(ap->standard, "pal")) | |||
dv->format = DV1394_PAL; | |||
@@ -107,17 +99,6 @@ static int dv1394_read_header(AVFormatContext * context, AVFormatParameters * ap | |||
else | |||
dv->channel = DV1394_DEFAULT_CHANNEL; | |||
dv->width = DV1394_WIDTH; | |||
if (dv->format == DV1394_NTSC) { | |||
dv->height = DV1394_NTSC_HEIGHT; | |||
dv->frame_size = DV1394_NTSC_FRAME_SIZE; | |||
dv->frame_rate = 30; | |||
} else { | |||
dv->height = DV1394_PAL_HEIGHT; | |||
dv->frame_size = DV1394_PAL_FRAME_SIZE; | |||
dv->frame_rate = 25; | |||
} | |||
/* Open and initialize DV1394 device */ | |||
video_device = ap->device; | |||
if (!video_device) | |||
@@ -140,21 +121,6 @@ static int dv1394_read_header(AVFormatContext * context, AVFormatParameters * ap | |||
goto failed; | |||
} | |||
dv->stream = 0; | |||
dv->vst->codec.codec_type = CODEC_TYPE_VIDEO; | |||
dv->vst->codec.codec_id = CODEC_ID_DVVIDEO; | |||
dv->vst->codec.width = dv->width; | |||
dv->vst->codec.height = dv->height; | |||
dv->vst->codec.frame_rate = dv->frame_rate; | |||
dv->vst->codec.frame_rate_base = 1; | |||
dv->vst->codec.bit_rate = 25000000; /* Consumer DV is 25Mbps */ | |||
dv->ast->codec.codec_type = CODEC_TYPE_AUDIO; | |||
dv->ast->codec.codec_id = CODEC_ID_DVAUDIO; | |||
dv->ast->codec.channels = 2; | |||
dv->ast->codec.sample_rate= 48000; | |||
av_set_pts_info(context, 48, 1, 1000000); | |||
if (dv1394_start(dv) < 0) | |||
@@ -164,55 +130,17 @@ static int dv1394_read_header(AVFormatContext * context, AVFormatParameters * ap | |||
failed: | |||
close(dv->fd); | |||
av_free(dv->vst); | |||
av_free(dv->ast); | |||
return -EIO; | |||
} | |||
static void __destruct_pkt(struct AVPacket *pkt) | |||
{ | |||
pkt->data = NULL; pkt->size = 0; | |||
return; | |||
} | |||
static inline int __get_frame(struct dv1394_data *dv, AVPacket *pkt) | |||
{ | |||
char *ptr = dv->ring + (dv->index * DV1394_PAL_FRAME_SIZE); | |||
if (dv->stream) { | |||
dv->index = (dv->index + 1) % DV1394_RING_FRAMES; | |||
dv->done++; dv->avail--; | |||
} else { | |||
dv->pts = av_gettime() & ((1LL << 48) - 1); | |||
} | |||
dv->format = ((ptr[3] & 0x80) == 0) ? DV1394_NTSC : DV1394_PAL; | |||
if (dv->format == DV1394_NTSC) { | |||
dv->frame_size = DV1394_NTSC_FRAME_SIZE; | |||
dv->vst->codec.height = dv->height = DV1394_NTSC_HEIGHT; | |||
dv->vst->codec.frame_rate = dv->frame_rate = 30; | |||
} else { | |||
dv->frame_size = DV1394_PAL_FRAME_SIZE; | |||
dv->vst->codec.height = dv->height = DV1394_PAL_HEIGHT; | |||
dv->vst->codec.frame_rate = dv->frame_rate = 25; | |||
} | |||
av_init_packet(pkt); | |||
pkt->destruct = __destruct_pkt; | |||
pkt->data = ptr; | |||
pkt->size = dv->frame_size; | |||
pkt->pts = dv->pts; | |||
pkt->stream_index = dv->stream; | |||
pkt->flags |= PKT_FLAG_KEY; | |||
dv->stream ^= 1; | |||
return dv->frame_size; | |||
} | |||
static int dv1394_read_packet(AVFormatContext *context, AVPacket *pkt) | |||
{ | |||
struct dv1394_data *dv = context->priv_data; | |||
int size; | |||
size = dv_get_packet(dv->dv_demux, pkt); | |||
if (size > 0) | |||
goto out; | |||
if (!dv->avail) { | |||
struct dv1394_status s; | |||
@@ -276,7 +204,16 @@ restart_poll: | |||
dv->done); | |||
#endif | |||
return __get_frame(dv, pkt); | |||
size = dv_produce_packet(dv->dv_demux, pkt, | |||
dv->ring + (dv->index * DV1394_PAL_FRAME_SIZE), | |||
DV1394_PAL_FRAME_SIZE); | |||
dv->index = (dv->index + 1) % DV1394_RING_FRAMES; | |||
dv->done++; dv->avail--; | |||
dv->pts = av_gettime() & ((1LL << 48) - 1); | |||
out: | |||
pkt->pts = dv->pts; | |||
return size; | |||
} | |||
static int dv1394_close(AVFormatContext * context) | |||
@@ -292,6 +229,7 @@ static int dv1394_close(AVFormatContext * context) | |||
perror("Failed to munmap DV1394 ring buffer"); | |||
close(dv->fd); | |||
av_free(dv->dv_demux); | |||
return 0; | |||
} | |||