static void
_parse_wav(ScanData s, Buffer *buf)
{
uint32_t offset = 12;
s->type_name = "wav";
mediascan_add_StreamData(s, 1);
while ( offset < s->size - 8 ) {
char chunk_id[5];
uint32_t chunk_size;
// Verify we have at least 8 bytes
if ( !buffer_check_load(buf, s->fp, 8, BLOCK_SIZE) ) {
return;
}
strncpy( chunk_id, (char *)buffer_ptr(buf), 4 );
chunk_id[4] = '\0';
buffer_consume(buf, 4);
chunk_size = buffer_get_int_le(buf);
// Adjust for padding
if ( chunk_size % 2 ) {
chunk_size++;
}
offset += 8;
LOG_DEBUG("%s size %d\n", chunk_id, chunk_size);
// Seek past data, everything else we parse
// XXX: Are there other large chunks we should ignore?
if ( !strcmp( chunk_id, "data" ) ) {
s->audio_offset = offset;
s->audio_size = chunk_size;
// Calculate duration, unless we already know it (i.e. from 'fact')
if ( !s->duration_ms ) {
if (s->bitrate) {
s->duration_ms = (chunk_size / (s->bitrate / 8.)) * 1000;
}
}
// sanity check size, this is inside the data chunk code
// to support setting audio_offset even when the data size is wrong
if (chunk_size > s->size - offset) {
LOG_DEBUG("data size > file_size, skipping\n");
return;
}
// Seek past data if there are more chunks after it
if ( s->size > offset + chunk_size ) {
fseek(s->fp, offset + chunk_size, SEEK_SET);
}
buffer_clear(buf);
}
else if ( !strcmp( chunk_id, "id3 " ) || !strcmp( chunk_id, "ID3 " ) || !strcmp( chunk_id, "ID32" ) ) {
// Read header to verify version
unsigned char *bptr = buffer_ptr(buf);
if (
(bptr[0] == 'I' && bptr[1] == 'D' && bptr[2] == '3') &&
bptr[3] < 0xff && bptr[4] < 0xff &&
bptr[6] < 0x80 && bptr[7] < 0x80 && bptr[8] < 0x80 && bptr[9] < 0x80
) {
// Start parsing ID3 from offset
//parse_id3(infile, file, info, tags, offset, file_size);
}
// Seek past ID3 and clear buffer
fseek(s->fp, offset + chunk_size, SEEK_SET);
buffer_clear(buf);
}
else {
// sanity check size
if (chunk_size > s->size - offset) {
LOG_DEBUG("chunk_size > file_size, skipping\n");
return;
}
// Make sure we have enough data
if ( !buffer_check_load(buf, s->fp, chunk_size, BLOCK_SIZE) ) {
return;
}
if ( !strcmp( chunk_id, "fmt " ) ) {
_parse_wav_fmt(s, buf, chunk_size);
}
else if ( !strcmp( chunk_id, "LIST" ) ) {
//_parse_wav_list(buf, chunk_size, tags);
}
else if ( !strcmp( chunk_id, "PEAK" ) ) {
_parse_wav_peak(s, buf, chunk_size, 0);
}
else if ( !strcmp( chunk_id, "fact" ) ) {
// A 4-byte fact chunk in a non-PCM wav is the number of samples
// Use it to calculate duration
if ( chunk_size == 4 ) {
uint32_t num_samples = buffer_get_int_le(buf);
if (s->streams[0].samplerate) {
s->duration_ms = (num_samples * 1000) / s->streams[0].samplerate;
}
}
else {
// Unknown, skip it
buffer_consume(buf, chunk_size);
}
}
else {
if (
!strcmp(chunk_id, "SAUR") // Wavosour data chunk
|| !strcmp(chunk_id, "otom") // Wavosaur?
|| !strcmp(chunk_id, "PAD ") // Padding
) {
// Known chunks to skip
}
else {
// Warn about unknown chunks so we can investigate them
LOG_DEBUG("Unhandled WAV chunk %s size %d (skipped)\n", chunk_id, chunk_size);
}
buffer_consume(buf, chunk_size);
}
}
offset += chunk_size;
}
}
void
_parse_wav(PerlIO *infile, Buffer *buf, char *file, uint32_t file_size, HV *info, HV *tags)
{
uint32_t offset = 12;
while ( offset < file_size - 8 ) {
char chunk_id[5];
uint32_t chunk_size;
// Verify we have at least 8 bytes
if ( !_check_buf(infile, buf, 8, WAV_BLOCK_SIZE) ) {
return;
}
strncpy( chunk_id, (char *)buffer_ptr(buf), 4 );
chunk_id[4] = '\0';
buffer_consume(buf, 4);
chunk_size = buffer_get_int_le(buf);
// Adjust for padding
if ( chunk_size % 2 ) {
chunk_size++;
}
offset += 8;
DEBUG_TRACE("%s size %d\n", chunk_id, chunk_size);
// Seek past data, everything else we parse
// XXX: Are there other large chunks we should ignore?
if ( !strcmp( chunk_id, "data" ) ) {
SV **bitrate;
my_hv_store( info, "audio_offset", newSVuv(offset) );
my_hv_store( info, "audio_size", newSVuv(chunk_size) );
// Calculate duration, unless we already know it (i.e. from 'fact')
if ( !my_hv_fetch( info, "song_length_ms" ) ) {
bitrate = my_hv_fetch( info, "bitrate" );
if (bitrate != NULL) {
my_hv_store( info, "song_length_ms", newSVuv( (chunk_size / (SvIV(*bitrate) / 8.)) * 1000 ) );
}
}
// sanity check size, this is inside the data chunk code
// to support setting audio_offset even when the data size is wrong
if (chunk_size > file_size - offset) {
DEBUG_TRACE("data size > file_size, skipping\n");
return;
}
// Seek past data if there are more chunks after it
if ( file_size > offset + chunk_size ) {
PerlIO_seek(infile, offset + chunk_size, SEEK_SET);
}
buffer_clear(buf);
}
else if ( !strcmp( chunk_id, "id3 " ) || !strcmp( chunk_id, "ID3 " ) || !strcmp( chunk_id, "ID32" ) ) {
// Read header to verify version
unsigned char *bptr = buffer_ptr(buf);
if (
(bptr[0] == 'I' && bptr[1] == 'D' && bptr[2] == '3') &&
bptr[3] < 0xff && bptr[4] < 0xff &&
bptr[6] < 0x80 && bptr[7] < 0x80 && bptr[8] < 0x80 && bptr[9] < 0x80
) {
// Start parsing ID3 from offset
parse_id3(infile, file, info, tags, offset, file_size);
}
// Seek past ID3 and clear buffer
PerlIO_seek(infile, offset + chunk_size, SEEK_SET);
buffer_clear(buf);
}
else {
// sanity check size
if (chunk_size > file_size - offset) {
DEBUG_TRACE("chunk_size > file_size, skipping\n");
return;
}
// Make sure we have enough data
if ( !_check_buf(infile, buf, chunk_size, WAV_BLOCK_SIZE) ) {
return;
}
if ( !strcmp( chunk_id, "fmt " ) ) {
_parse_wav_fmt(buf, chunk_size, info);
}
else if ( !strcmp( chunk_id, "LIST" ) ) {
_parse_wav_list(buf, chunk_size, tags);
}
else if ( !strcmp( chunk_id, "PEAK" ) ) {
_parse_wav_peak(buf, chunk_size, info, 0);
}
else if ( !strcmp( chunk_id, "fact" ) ) {
// A 4-byte fact chunk in a non-PCM wav is the number of samples
// Use it to calculate duration
if ( chunk_size == 4 ) {
uint32_t num_samples = buffer_get_int_le(buf);
SV **samplerate = my_hv_fetch( info, "samplerate" );
if (samplerate != NULL) {
my_hv_store( info, "song_length_ms", newSVuv( (num_samples * 1000) / SvIV(*samplerate) ) );
}
}
else {
// Unknown, skip it
buffer_consume(buf, chunk_size);
}
}
else {
if (
!strcmp(chunk_id, "SAUR") // Wavosour data chunk
|| !strcmp(chunk_id, "otom") // Wavosaur?
|| !strcmp(chunk_id, "PAD ") // Padding
) {
// Known chunks to skip
}
else {
// Warn about unknown chunks so we can investigate them
PerlIO_printf(PerlIO_stderr(), "Unhandled WAV chunk %s size %d (skipped)\n", chunk_id, chunk_size);
}
buffer_consume(buf, chunk_size);
}
}
offset += chunk_size;
}
}