unsigned drefs_count;
MOV_dref_t *drefs;
int dref_id;
+ int wrong_dts; ///< dts are wrong due to negative ctts
} MOVStreamContext;
typedef struct MOVContext {
case CODEC_ID_PCM_S32BE:
st->codec->codec_id = CODEC_ID_PCM_S32LE;
break;
+ case CODEC_ID_PCM_F32BE:
+ st->codec->codec_id = CODEC_ID_PCM_F32LE;
+ break;
+ case CODEC_ID_PCM_F64BE:
+ st->codec->codec_id = CODEC_ID_PCM_F64LE;
+ break;
default:
break;
}
return 0;
}
+/**
+ * Compute codec id for 'lpcm' tag.
+ * See CoreAudioTypes and AudioStreamBasicDescription at Apple.
+ */
+static enum CodecID mov_get_lpcm_codec_id(int bps, int flags)
+{
+ if (flags & 1) { // floating point
+ if (flags & 2) { // big endian
+ if (bps == 32) return CODEC_ID_PCM_F32BE;
+ else if (bps == 64) return CODEC_ID_PCM_F64BE;
+ } else {
+ if (bps == 32) return CODEC_ID_PCM_F32LE;
+ else if (bps == 64) return CODEC_ID_PCM_F64LE;
+ }
+ } else {
+ if (flags & 2) {
+ if (bps == 8)
+ // signed integer
+ if (flags & 4) return CODEC_ID_PCM_S8;
+ else return CODEC_ID_PCM_U8;
+ else if (bps == 16) return CODEC_ID_PCM_S16BE;
+ else if (bps == 24) return CODEC_ID_PCM_S24BE;
+ else if (bps == 32) return CODEC_ID_PCM_S32BE;
+ } else {
+ if (bps == 8)
+ if (flags & 4) return CODEC_ID_PCM_S8;
+ else return CODEC_ID_PCM_U8;
+ else if (bps == 16) return CODEC_ID_PCM_S16LE;
+ else if (bps == 24) return CODEC_ID_PCM_S24LE;
+ else if (bps == 32) return CODEC_ID_PCM_S32LE;
+ }
+ }
+ return CODEC_ID_NONE;
+}
+
static int mov_read_stsd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom)
{
AVStream *st = c->fc->streams[c->fc->nb_streams-1];
st->codec->codec_name[codec_name[0]] = 0;
}
- st->codec->bits_per_sample = get_be16(pb); /* depth */
+ st->codec->bits_per_coded_sample = get_be16(pb); /* depth */
st->codec->color_table_id = get_be16(pb); /* colortable id */
dprintf(c->fc, "depth %d, ctab id %d\n",
- st->codec->bits_per_sample, st->codec->color_table_id);
+ st->codec->bits_per_coded_sample, st->codec->color_table_id);
/* figure out the palette situation */
- color_depth = st->codec->bits_per_sample & 0x1F;
- color_greyscale = st->codec->bits_per_sample & 0x20;
+ color_depth = st->codec->bits_per_coded_sample & 0x1F;
+ color_greyscale = st->codec->bits_per_coded_sample & 0x20;
/* if the depth is 2, 4, or 8 bpp, file is palettized */
if ((color_depth == 2) || (color_depth == 4) ||
if (color_greyscale) {
int color_index, color_dec;
/* compute the greyscale palette */
- st->codec->bits_per_sample = color_depth;
+ st->codec->bits_per_coded_sample = color_depth;
color_count = 1 << color_depth;
color_index = 255;
color_dec = 256 / (color_count - 1);
} else
st->codec->palctrl = NULL;
} else if(st->codec->codec_type==CODEC_TYPE_AUDIO) {
- int bits_per_sample;
+ int bits_per_sample, flags;
uint16_t version = get_be16(pb);
st->codec->codec_id = id;
st->codec->channels = get_be16(pb); /* channel count */
dprintf(c->fc, "audio channels %d\n", st->codec->channels);
- st->codec->bits_per_sample = get_be16(pb); /* sample size */
+ st->codec->bits_per_coded_sample = get_be16(pb); /* sample size */
sc->audio_cid = get_be16(pb);
get_be16(pb); /* packet size = 0 */
st->codec->sample_rate = ((get_be32(pb) >> 16));
+ //Read QT version 1 fields. In version 0 these do not exist.
+ dprintf(c->fc, "version =%d, isom =%d\n",version,c->isom);
+ if(!c->isom) {
+ if(version==1) {
+ sc->samples_per_frame = get_be32(pb);
+ get_be32(pb); /* bytes per packet */
+ sc->bytes_per_frame = get_be32(pb);
+ get_be32(pb); /* bytes per sample */
+ } else if(version==2) {
+ get_be32(pb); /* sizeof struct only */
+ st->codec->sample_rate = av_int2dbl(get_be64(pb)); /* float 64 */
+ st->codec->channels = get_be32(pb);
+ get_be32(pb); /* always 0x7F000000 */
+ st->codec->bits_per_coded_sample = get_be32(pb); /* bits per channel if sound is uncompressed */
+ flags = get_be32(pb); /* lcpm format specific flag */
+ sc->bytes_per_frame = get_be32(pb); /* bytes per audio packet if constant */
+ sc->samples_per_frame = get_be32(pb); /* lpcm frames per audio packet if constant */
+ if (format == MKTAG('l','p','c','m'))
+ st->codec->codec_id = mov_get_lpcm_codec_id(st->codec->bits_per_coded_sample, flags);
+ }
+ }
+
switch (st->codec->codec_id) {
case CODEC_ID_PCM_S8:
case CODEC_ID_PCM_U8:
- if (st->codec->bits_per_sample == 16)
+ if (st->codec->bits_per_coded_sample == 16)
st->codec->codec_id = CODEC_ID_PCM_S16BE;
break;
case CODEC_ID_PCM_S16LE:
case CODEC_ID_PCM_S16BE:
- if (st->codec->bits_per_sample == 8)
+ if (st->codec->bits_per_coded_sample == 8)
st->codec->codec_id = CODEC_ID_PCM_S8;
- else if (st->codec->bits_per_sample == 24)
- st->codec->codec_id = CODEC_ID_PCM_S24BE;
+ else if (st->codec->bits_per_coded_sample == 24)
+ st->codec->codec_id =
+ st->codec->codec_id == CODEC_ID_PCM_S16BE ?
+ CODEC_ID_PCM_S24BE : CODEC_ID_PCM_S24LE;
break;
/* set values for old format before stsd version 1 appeared */
case CODEC_ID_MACE3:
break;
}
- //Read QT version 1 fields. In version 0 these do not exist.
- dprintf(c->fc, "version =%d, isom =%d\n",version,c->isom);
- if(!c->isom) {
- if(version==1) {
- sc->samples_per_frame = get_be32(pb);
- get_be32(pb); /* bytes per packet */
- sc->bytes_per_frame = get_be32(pb);
- get_be32(pb); /* bytes per sample */
- } else if(version==2) {
- get_be32(pb); /* sizeof struct only */
- st->codec->sample_rate = av_int2dbl(get_be64(pb)); /* float 64 */
- st->codec->channels = get_be32(pb);
- get_be32(pb); /* always 0x7F000000 */
- get_be32(pb); /* bits per channel if sound is uncompressed */
- get_be32(pb); /* lcpm format specific flag */
- get_be32(pb); /* bytes per audio packet if constant */
- get_be32(pb); /* lpcm frames per audio packet if constant */
- }
- }
-
bits_per_sample = av_get_bits_per_sample(st->codec->codec_id);
if (bits_per_sample) {
- st->codec->bits_per_sample = bits_per_sample;
+ st->codec->bits_per_coded_sample = bits_per_sample;
sc->sample_size = (bits_per_sample >> 3) * st->codec->channels;
}
} else if(st->codec->codec_type==CODEC_TYPE_SUBTITLE){
break;
#endif
/* no ifdef since parameters are always those */
- case CODEC_ID_AMR_WB:
- st->codec->sample_rate= 16000;
- st->codec->channels= 1; /* really needed */
- break;
case CODEC_ID_QCELP:
case CODEC_ID_AMR_NB:
+ case CODEC_ID_AMR_WB:
st->codec->frame_size= sc->samples_per_frame;
- st->codec->sample_rate= 8000;
st->codec->channels= 1; /* really needed */
break;
case CODEC_ID_MP2:
int duration =get_be32(pb);
if (duration < 0) {
- av_log(c->fc, AV_LOG_ERROR, "negative ctts, ignoring\n");
- sc->ctts_count = 0;
- url_fskip(pb, 8 * (entries - i - 1));
- break;
+ sc->wrong_dts = 1;
+ st->codec->has_b_frames = 1;
}
sc->ctts_data[i].count = count;
sc->ctts_data[i].duration= duration;
- sc->time_rate= ff_gcd(sc->time_rate, duration);
+ sc->time_rate= ff_gcd(sc->time_rate, FFABS(duration));
}
return 0;
}
st->codec->width= 0; /* let decoder init width/height */
st->codec->height= 0;
break;
-#ifdef CONFIG_VORBIS_DECODER
- case CODEC_ID_VORBIS:
-#endif
- st->codec->sample_rate= 0; /* let decoder init parameters properly */
- break;
}
/* Do not need those anymore. */
uint32_t tag = get_le32(pb);
uint64_t next = url_ftell(pb) + tag_size - 8;
- if (next > end) // stop if tag_size is wrong
+ if (tag_size < 8 || next > end) // stop if tag_size is wrong
break;
switch (tag) {
static int mov_read_tkhd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom)
{
+ int i;
+ int width;
+ int height;
+ int64_t disp_transform[2];
+ int display_matrix[3][2];
AVStream *st = c->fc->streams[c->fc->nb_streams-1];
int version = get_byte(pb);
get_be16(pb); /* volume */
get_be16(pb); /* reserved */
- url_fskip(pb, 36); /* display matrix */
-
- /* those are fixed-point */
- get_be32(pb); /* track width */
- get_be32(pb); /* track height */
+ //read in the display matrix (outlined in ISO 14496-12, Section 6.2.2)
+ // they're kept in fixed point format through all calculations
+ // ignore u,v,z b/c we don't need the scale factor to calc aspect ratio
+ for (i = 0; i < 3; i++) {
+ display_matrix[i][0] = get_be32(pb); // 16.16 fixed point
+ display_matrix[i][1] = get_be32(pb); // 16.16 fixed point
+ get_be32(pb); // 2.30 fixed point (not used)
+ }
+ width = get_be32(pb); // 16.16 fixed point track width
+ height = get_be32(pb); // 16.16 fixed point track height
+
+ //transform the display width/height according to the matrix
+ // skip this if the display matrix is the default identity matrix
+ // to keep the same scale, use [width height 1<<16]
+ if (width && height &&
+ (display_matrix[0][0] != 65536 || display_matrix[0][1] ||
+ display_matrix[1][0] || display_matrix[1][1] != 65536 ||
+ display_matrix[2][0] || display_matrix[2][1])) {
+ for (i = 0; i < 2; i++)
+ disp_transform[i] =
+ (int64_t) width * display_matrix[0][i] +
+ (int64_t) height * display_matrix[1][i] +
+ ((int64_t) display_matrix[2][i] << 16);
+
+ //sample aspect ratio is new width/height divided by old width/height
+ st->sample_aspect_ratio = av_d2q(
+ ((double) disp_transform[0] * height) /
+ ((double) disp_transform[1] * width), INT_MAX);
+ }
return 0;
}
sc->sample_to_ctime_index++;
sc->sample_to_ctime_sample = 0;
}
+ if (sc->wrong_dts)
+ pkt->dts = AV_NOPTS_VALUE;
} else {
AVStream *st = s->streams[sc->ffindex];
int64_t next_dts = (sc->current_sample < sc->sample_count) ?
projects / ffmpeg / blobdiff