1 /* GStreamer Wavpack parser
2 * Copyright (C) 2012 Mark Nauwelaerts <mark.nauwelaerts@collabora.co.uk>
3 * Copyright (C) 2012 Nokia Corporation. All rights reserved.
4 * Contact: Stefan Kost <stefan.kost@nokia.com>
6 * This library is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Library General Public
8 * License as published by the Free Software Foundation; either
9 * version 2 of the License, or (at your option) any later version.
11 * This library is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * Library General Public License for more details.
16 * You should have received a copy of the GNU Library General Public
17 * License along with this library; if not, write to the
18 * Free Software Foundation, Inc., 51 Franklin St, Fifth Floor,
19 * Boston, MA 02110-1301, USA.
22 * SECTION:element-wavpackparse
23 * @short_description: Wavpack parser
24 * @see_also: #GstAmrParse, #GstAACParse
26 * This is an Wavpack parser.
29 * <title>Example launch line</title>
31 * gst-launch-1.0 filesrc location=abc.wavpack ! wavpackparse ! wavpackdec ! audioresample ! audioconvert ! autoaudiosink
42 #include "gstwavpackparse.h"
44 #include <gst/base/base.h>
45 #include <gst/pbutils/pbutils.h>
46 #include <gst/audio/audio.h>
48 GST_DEBUG_CATEGORY_STATIC (wavpack_parse_debug);
49 #define GST_CAT_DEFAULT wavpack_parse_debug
51 static GstStaticPadTemplate src_template = GST_STATIC_PAD_TEMPLATE ("src",
54 GST_STATIC_CAPS ("audio/x-wavpack, "
55 "depth = (int) [ 1, 32 ], "
56 "channels = (int) [ 1, 8 ], "
57 "rate = (int) [ 6000, 192000 ], " "framed = (boolean) TRUE; "
58 "audio/x-wavpack-correction, " "framed = (boolean) TRUE")
61 static GstStaticPadTemplate sink_template = GST_STATIC_PAD_TEMPLATE ("sink",
64 GST_STATIC_CAPS ("audio/x-wavpack"));
66 static void gst_wavpack_parse_finalize (GObject * object);
68 static gboolean gst_wavpack_parse_start (GstBaseParse * parse);
69 static gboolean gst_wavpack_parse_stop (GstBaseParse * parse);
70 static GstFlowReturn gst_wavpack_parse_handle_frame (GstBaseParse * parse,
71 GstBaseParseFrame * frame, gint * skipsize);
72 static GstCaps *gst_wavpack_parse_get_sink_caps (GstBaseParse * parse,
74 static GstFlowReturn gst_wavpack_parse_pre_push_frame (GstBaseParse * parse,
75 GstBaseParseFrame * frame);
77 #define gst_wavpack_parse_parent_class parent_class
78 G_DEFINE_TYPE (GstWavpackParse, gst_wavpack_parse, GST_TYPE_BASE_PARSE);
81 gst_wavpack_parse_class_init (GstWavpackParseClass * klass)
83 GstBaseParseClass *parse_class = GST_BASE_PARSE_CLASS (klass);
84 GstElementClass *element_class = GST_ELEMENT_CLASS (klass);
85 GObjectClass *object_class = G_OBJECT_CLASS (klass);
87 GST_DEBUG_CATEGORY_INIT (wavpack_parse_debug, "wavpackparse", 0,
88 "Wavpack audio stream parser");
90 object_class->finalize = gst_wavpack_parse_finalize;
92 parse_class->start = GST_DEBUG_FUNCPTR (gst_wavpack_parse_start);
93 parse_class->stop = GST_DEBUG_FUNCPTR (gst_wavpack_parse_stop);
94 parse_class->handle_frame =
95 GST_DEBUG_FUNCPTR (gst_wavpack_parse_handle_frame);
96 parse_class->get_sink_caps =
97 GST_DEBUG_FUNCPTR (gst_wavpack_parse_get_sink_caps);
98 parse_class->pre_push_frame =
99 GST_DEBUG_FUNCPTR (gst_wavpack_parse_pre_push_frame);
101 gst_element_class_add_pad_template (element_class,
102 gst_static_pad_template_get (&sink_template));
103 gst_element_class_add_pad_template (element_class,
104 gst_static_pad_template_get (&src_template));
106 gst_element_class_set_static_metadata (element_class,
107 "Wavpack audio stream parser", "Codec/Parser/Audio",
108 "Wavpack parser", "Mark Nauwelaerts <mark.nauwelaerts@collabora.co.uk>");
112 gst_wavpack_parse_reset (GstWavpackParse * wvparse)
114 wvparse->channels = -1;
115 wvparse->channel_mask = 0;
116 wvparse->sample_rate = -1;
118 wvparse->total_samples = 0;
119 wvparse->sent_codec_tag = FALSE;
123 gst_wavpack_parse_init (GstWavpackParse * wvparse)
125 gst_wavpack_parse_reset (wvparse);
126 GST_PAD_SET_ACCEPT_INTERSECT (GST_BASE_PARSE_SINK_PAD (wvparse));
127 GST_PAD_SET_ACCEPT_TEMPLATE (GST_BASE_PARSE_SINK_PAD (wvparse));
131 gst_wavpack_parse_finalize (GObject * object)
133 G_OBJECT_CLASS (parent_class)->finalize (object);
137 gst_wavpack_parse_start (GstBaseParse * parse)
139 GstWavpackParse *wvparse = GST_WAVPACK_PARSE (parse);
141 GST_DEBUG_OBJECT (parse, "starting");
143 gst_wavpack_parse_reset (wvparse);
145 /* need header at least */
146 gst_base_parse_set_min_frame_size (GST_BASE_PARSE (wvparse),
147 sizeof (WavpackHeader));
149 /* inform baseclass we can come up with ts, based on counters in packets */
150 gst_base_parse_set_has_timing_info (GST_BASE_PARSE_CAST (wvparse), TRUE);
151 gst_base_parse_set_syncable (GST_BASE_PARSE_CAST (wvparse), TRUE);
157 gst_wavpack_parse_stop (GstBaseParse * parse)
159 GST_DEBUG_OBJECT (parse, "stopping");
165 gst_wavpack_get_default_channel_mask (gint nchannels)
167 gint channel_mask = 0;
169 /* Set the default channel mask for the given number of channels.
170 * It's the same as for WAVE_FORMAT_EXTENDED:
171 * http://www.microsoft.com/whdc/device/audio/multichaud.mspx
175 channel_mask |= 0x00400;
176 channel_mask |= 0x00200;
178 channel_mask |= 0x00100;
180 channel_mask |= 0x00080;
181 channel_mask |= 0x00040;
183 channel_mask |= 0x00020;
184 channel_mask |= 0x00010;
186 channel_mask |= 0x00008;
188 channel_mask |= 0x00004;
190 channel_mask |= 0x00002;
191 channel_mask |= 0x00001;
194 /* For mono use front center */
195 channel_mask |= 0x00004;
204 const guint32 ms_mask;
205 const GstAudioChannelPosition gst_pos;
206 } layout_mapping[] = {
208 0x00001, GST_AUDIO_CHANNEL_POSITION_FRONT_LEFT}, {
209 0x00002, GST_AUDIO_CHANNEL_POSITION_FRONT_RIGHT}, {
210 0x00004, GST_AUDIO_CHANNEL_POSITION_FRONT_CENTER}, {
211 0x00008, GST_AUDIO_CHANNEL_POSITION_LFE1}, {
212 0x00010, GST_AUDIO_CHANNEL_POSITION_REAR_LEFT}, {
213 0x00020, GST_AUDIO_CHANNEL_POSITION_REAR_RIGHT}, {
214 0x00040, GST_AUDIO_CHANNEL_POSITION_FRONT_LEFT_OF_CENTER}, {
215 0x00080, GST_AUDIO_CHANNEL_POSITION_FRONT_RIGHT_OF_CENTER}, {
216 0x00100, GST_AUDIO_CHANNEL_POSITION_REAR_CENTER}, {
217 0x00200, GST_AUDIO_CHANNEL_POSITION_SIDE_LEFT}, {
218 0x00400, GST_AUDIO_CHANNEL_POSITION_SIDE_RIGHT}, {
219 0x00800, GST_AUDIO_CHANNEL_POSITION_TOP_CENTER}, {
220 0x01000, GST_AUDIO_CHANNEL_POSITION_TOP_FRONT_LEFT}, {
221 0x02000, GST_AUDIO_CHANNEL_POSITION_TOP_FRONT_CENTER}, {
222 0x04000, GST_AUDIO_CHANNEL_POSITION_TOP_FRONT_RIGHT}, {
223 0x08000, GST_AUDIO_CHANNEL_POSITION_TOP_REAR_LEFT}, {
224 0x10000, GST_AUDIO_CHANNEL_POSITION_TOP_REAR_CENTER}, {
225 0x20000, GST_AUDIO_CHANNEL_POSITION_TOP_REAR_RIGHT}
228 #define MAX_CHANNEL_POSITIONS G_N_ELEMENTS (layout_mapping)
231 gst_wavpack_get_channel_positions (gint num_channels, gint layout,
232 GstAudioChannelPosition * pos)
236 if (num_channels == 1 && layout == 0x00004) {
237 pos[0] = GST_AUDIO_CHANNEL_POSITION_MONO;
242 for (i = 0; i < MAX_CHANNEL_POSITIONS; ++i) {
243 if ((layout & layout_mapping[i].ms_mask) != 0) {
244 if (p >= num_channels) {
245 GST_WARNING ("More bits set in the channel layout map than there "
246 "are channels! Broken file");
249 if (layout_mapping[i].gst_pos == GST_AUDIO_CHANNEL_POSITION_INVALID) {
250 GST_WARNING ("Unsupported channel position (mask 0x%08x) in channel "
251 "layout map - ignoring those channels", layout_mapping[i].ms_mask);
252 /* what to do? just ignore it and let downstream deal with a channel
253 * layout that has INVALID positions in it for now ... */
255 pos[p] = layout_mapping[i].gst_pos;
260 if (p != num_channels) {
261 GST_WARNING ("Only %d bits set in the channel layout map, but there are "
262 "supposed to be %d channels! Broken file", p, num_channels);
269 static const guint32 sample_rates[] = {
270 6000, 8000, 9600, 11025, 12000, 16000, 22050,
271 24000, 32000, 44100, 48000, 64000, 88200, 96000, 192000
274 #define CHECK(call) { \
279 /* caller ensures properly sync'ed with enough data */
281 gst_wavpack_parse_frame_metadata (GstWavpackParse * parse, GstBuffer * buf,
282 gint skip, WavpackHeader * wph, WavpackInfo * wpi)
288 g_return_val_if_fail (wph != NULL || wpi != NULL, FALSE);
289 g_return_val_if_fail (gst_buffer_get_size (buf) >=
290 skip + sizeof (WavpackHeader), FALSE);
292 gst_buffer_map (buf, &map, GST_MAP_READ);
294 gst_byte_reader_init (&br, map.data + skip, wph->ckSize + 8);
295 /* skip past header */
296 gst_byte_reader_skip_unchecked (&br, sizeof (WavpackHeader));
298 /* get some basics from header */
299 i = (wph->flags >> 23) & 0xF;
301 wpi->rate = (i < G_N_ELEMENTS (sample_rates)) ? sample_rates[i] : 44100;
302 wpi->width = ((wph->flags & 0x3) + 1) * 8;
304 wpi->channels = (wph->flags & 0x4) ? 1 : 2;
305 if (!wpi->channel_mask)
306 wpi->channel_mask = 5 - wpi->channels;
308 /* need to dig metadata blocks for some more */
309 while (gst_byte_reader_get_remaining (&br)) {
316 CHECK (gst_byte_reader_get_uint8 (&br, &id));
317 CHECK (gst_byte_reader_get_uint8 (&br, &c));
319 CHECK (gst_byte_reader_get_uint16_le (&br, &size2));
324 if (id & ID_ODD_SIZE)
327 CHECK (gst_byte_reader_get_data (&br, size + (size & 1), &data));
328 gst_byte_reader_init (&mbr, data, size);
331 case ID_WVC_BITSTREAM:
332 GST_LOG_OBJECT (parse, "correction bitstream");
333 wpi->correction = TRUE;
335 case ID_WV_BITSTREAM:
336 case ID_WVX_BITSTREAM:
340 CHECK (gst_byte_reader_get_uint24_le (&mbr, &wpi->rate));
341 GST_LOG_OBJECT (parse, "updated with custom rate %d", wpi->rate);
343 GST_DEBUG_OBJECT (parse, "unexpected size for SAMPLE_RATE metadata");
346 case ID_CHANNEL_INFO:
352 CHECK (gst_byte_reader_get_uint16_le (&mbr, &channels));
353 channels = channels & 0xFFF;
354 CHECK (gst_byte_reader_get_uint24_le (&mbr, &mask));
356 CHECK (gst_byte_reader_get_uint8 (&mbr, &c));
358 while (gst_byte_reader_get_uint8 (&mbr, &c))
359 mask |= (((guint32) c) << 8);
361 GST_DEBUG_OBJECT (parse, "unexpected size for CHANNEL_INFO metadata");
364 wpi->channels = channels;
365 wpi->channel_mask = mask;
369 GST_LOG_OBJECT (parse, "unparsed ID 0x%x", id);
374 gst_buffer_unmap (buf, &map);
381 gst_buffer_unmap (buf, &map);
382 GST_DEBUG_OBJECT (parse, "short read while parsing metadata");
383 /* let's look the other way anyway */
388 /* caller ensures properly sync'ed with enough data */
390 gst_wavpack_parse_frame_header (GstWavpackParse * parse, GstBuffer * buf,
391 gint skip, WavpackHeader * _wph)
394 WavpackHeader wph = { {0,}, 0, };
398 g_return_val_if_fail (gst_buffer_get_size (buf) >=
399 skip + sizeof (WavpackHeader), FALSE);
401 gst_buffer_map (buf, &map, GST_MAP_READ);
402 gst_byte_reader_init (&br, map.data, map.size);
405 gst_byte_reader_skip_unchecked (&br, skip + 4);
408 hdl &= gst_byte_reader_get_uint32_le (&br, &wph.ckSize);
409 hdl &= gst_byte_reader_get_uint16_le (&br, &wph.version);
410 hdl &= gst_byte_reader_get_uint8 (&br, &wph.track_no);
411 hdl &= gst_byte_reader_get_uint8 (&br, &wph.index_no);
412 hdl &= gst_byte_reader_get_uint32_le (&br, &wph.total_samples);
413 hdl &= gst_byte_reader_get_uint32_le (&br, &wph.block_index);
414 hdl &= gst_byte_reader_get_uint32_le (&br, &wph.block_samples);
415 hdl &= gst_byte_reader_get_uint32_le (&br, &wph.flags);
416 hdl &= gst_byte_reader_get_uint32_le (&br, &wph.crc);
419 GST_WARNING_OBJECT (parse, "Error reading header");
422 GST_LOG_OBJECT (parse, "size %d", wph.ckSize);
423 GST_LOG_OBJECT (parse, "version 0x%x", wph.version);
424 GST_LOG_OBJECT (parse, "total samples %d", wph.total_samples);
425 GST_LOG_OBJECT (parse, "block index %d", wph.block_index);
426 GST_LOG_OBJECT (parse, "block samples %d", wph.block_samples);
427 GST_LOG_OBJECT (parse, "flags 0x%x", wph.flags);
428 GST_LOG_OBJECT (parse, "crc 0x%x", wph.flags);
430 if (!parse->total_samples && wph.block_index == 0 && wph.total_samples != -1) {
431 GST_DEBUG_OBJECT (parse, "determined duration of %u samples",
433 parse->total_samples = wph.total_samples;
439 gst_buffer_unmap (buf, &map);
445 gst_wavpack_parse_handle_frame (GstBaseParse * parse,
446 GstBaseParseFrame * frame, gint * skipsize)
448 GstWavpackParse *wvparse = GST_WAVPACK_PARSE (parse);
449 GstBuffer *buf = frame->buffer;
450 GstByteReader reader;
452 guint rate, chans, width, mask;
453 gboolean lost_sync, draining, final;
456 WavpackInfo wpi = { 0, };
459 if (G_UNLIKELY (gst_buffer_get_size (buf) < sizeof (WavpackHeader)))
462 gst_buffer_map (buf, &map, GST_MAP_READ);
463 gst_byte_reader_init (&reader, map.data, map.size);
465 /* scan for 'wvpk' marker */
466 off = gst_byte_reader_masked_scan_uint32 (&reader, 0xffffffff, 0x7776706b,
469 GST_LOG_OBJECT (parse, "possible sync at buffer offset %d", off);
471 /* didn't find anything that looks like a sync word, skip */
473 *skipsize = map.size - 3;
477 /* possible frame header, but not at offset 0? skip bytes before sync */
483 /* make sure the values in the frame header look sane */
484 gst_wavpack_parse_frame_header (wvparse, buf, 0, &wph);
485 frmsize = wph.ckSize + 8;
487 /* need the entire frame for parsing */
488 if (gst_byte_reader_get_remaining (&reader) < frmsize)
491 /* got a frame, now we can dig for some more metadata */
492 GST_LOG_OBJECT (parse, "got frame");
493 gst_wavpack_parse_frame_metadata (wvparse, buf, 0, &wph, &wpi);
495 lost_sync = GST_BASE_PARSE_LOST_SYNC (parse);
496 draining = GST_BASE_PARSE_DRAINING (parse);
498 while (!(final = (wph.flags & FLAG_FINAL_BLOCK)) || (lost_sync && !draining)) {
501 GST_LOG_OBJECT (wvparse, "checking next frame syncword; "
502 "lost_sync: %d, draining: %d, final: %d", lost_sync, draining, final);
504 if (!gst_byte_reader_skip (&reader, wph.ckSize + 8) ||
505 !gst_byte_reader_peek_uint32_be (&reader, &word)) {
506 GST_DEBUG_OBJECT (wvparse, "... but not sufficient data");
510 if (word != 0x7776706b) {
511 GST_DEBUG_OBJECT (wvparse, "0x%x not OK", word);
515 /* need to parse each frame/block for metadata if several ones */
519 GST_LOG_OBJECT (wvparse, "checking frame at offset %d (0x%x)",
521 av = gst_byte_reader_get_remaining (&reader);
522 if (av < sizeof (WavpackHeader)) {
523 frmsize += sizeof (WavpackHeader);
526 gst_wavpack_parse_frame_header (wvparse, buf, frmsize, &wph);
528 frmsize += wph.ckSize + 8;
529 if (av < wph.ckSize + 8)
531 gst_wavpack_parse_frame_metadata (wvparse, buf, off, &wph, &wpi);
532 /* could also check for matching block_index and block_samples ?? */
536 /* resynced if we make it here */
542 chans = wpi.channels;
543 mask = wpi.channel_mask;
545 GST_LOG_OBJECT (parse, "rate: %u, width: %u, chans: %u", rate, width, chans);
547 GST_BUFFER_TIMESTAMP (buf) =
548 gst_util_uint64_scale_int (wph.block_index, GST_SECOND, rate);
549 GST_BUFFER_DURATION (buf) =
550 gst_util_uint64_scale_int (wph.block_index + wph.block_samples,
551 GST_SECOND, rate) - GST_BUFFER_TIMESTAMP (buf);
553 if (G_UNLIKELY (wvparse->sample_rate != rate || wvparse->channels != chans
554 || wvparse->width != width || wvparse->channel_mask != mask)) {
557 if (wpi.correction) {
558 caps = gst_caps_new_simple ("audio/x-wavpack-correction",
559 "framed", G_TYPE_BOOLEAN, TRUE, NULL);
561 caps = gst_caps_new_simple ("audio/x-wavpack",
562 "channels", G_TYPE_INT, chans,
563 "rate", G_TYPE_INT, rate,
564 "depth", G_TYPE_INT, width, "framed", G_TYPE_BOOLEAN, TRUE, NULL);
567 mask = gst_wavpack_get_default_channel_mask (wvparse->channels);
569 GstAudioChannelPosition pos[64] =
570 { GST_AUDIO_CHANNEL_POSITION_INVALID, };
573 if (!gst_wavpack_get_channel_positions (chans, mask, pos)) {
574 GST_WARNING_OBJECT (wvparse, "Failed to determine channel layout");
576 gst_audio_channel_positions_to_mask (pos, chans, FALSE, &gmask);
578 gst_caps_set_simple (caps,
579 "channel-mask", GST_TYPE_BITMASK, gmask, NULL);
584 gst_pad_set_caps (GST_BASE_PARSE_SRC_PAD (parse), caps);
585 gst_caps_unref (caps);
587 wvparse->sample_rate = rate;
588 wvparse->channels = chans;
589 wvparse->width = width;
590 wvparse->channel_mask = mask;
592 if (wvparse->total_samples) {
593 GST_DEBUG_OBJECT (wvparse, "setting duration");
594 gst_base_parse_set_duration (GST_BASE_PARSE (wvparse),
595 GST_FORMAT_TIME, gst_util_uint64_scale_int (wvparse->total_samples,
596 GST_SECOND, wvparse->sample_rate), 0);
600 /* return to normal size */
601 gst_base_parse_set_min_frame_size (parse, sizeof (WavpackHeader));
602 gst_buffer_unmap (buf, &map);
604 return gst_base_parse_finish_frame (parse, frame, frmsize);
607 gst_buffer_unmap (buf, &map);
608 GST_LOG_OBJECT (wvparse, "skipping %d", *skipsize);
612 gst_buffer_unmap (buf, &map);
613 GST_LOG_OBJECT (wvparse, "need at least %u", frmsize);
614 gst_base_parse_set_min_frame_size (parse, frmsize);
620 remove_fields (GstCaps * caps)
624 n = gst_caps_get_size (caps);
625 for (i = 0; i < n; i++) {
626 GstStructure *s = gst_caps_get_structure (caps, i);
628 gst_structure_remove_field (s, "framed");
633 gst_wavpack_parse_get_sink_caps (GstBaseParse * parse, GstCaps * filter)
635 GstCaps *peercaps, *templ;
638 templ = gst_pad_get_pad_template_caps (GST_BASE_PARSE_SINK_PAD (parse));
640 GstCaps *fcopy = gst_caps_copy (filter);
641 /* Remove the fields we convert */
642 remove_fields (fcopy);
643 peercaps = gst_pad_peer_query_caps (GST_BASE_PARSE_SRC_PAD (parse), fcopy);
644 gst_caps_unref (fcopy);
646 peercaps = gst_pad_peer_query_caps (GST_BASE_PARSE_SRC_PAD (parse), NULL);
649 /* Remove the framed field */
650 peercaps = gst_caps_make_writable (peercaps);
651 remove_fields (peercaps);
653 res = gst_caps_intersect_full (peercaps, templ, GST_CAPS_INTERSECT_FIRST);
654 gst_caps_unref (peercaps);
655 gst_caps_unref (templ);
661 GstCaps *intersection;
664 gst_caps_intersect_full (filter, res, GST_CAPS_INTERSECT_FIRST);
665 gst_caps_unref (res);
673 gst_wavpack_parse_pre_push_frame (GstBaseParse * parse,
674 GstBaseParseFrame * frame)
676 GstWavpackParse *wavpackparse = GST_WAVPACK_PARSE (parse);
678 if (!wavpackparse->sent_codec_tag) {
682 taglist = gst_tag_list_new_empty ();
685 caps = gst_pad_get_current_caps (GST_BASE_PARSE_SRC_PAD (parse));
686 gst_pb_utils_add_codec_description_to_tag_list (taglist,
687 GST_TAG_AUDIO_CODEC, caps);
688 gst_caps_unref (caps);
690 gst_base_parse_merge_tags (parse, taglist, GST_TAG_MERGE_REPLACE);
691 gst_tag_list_unref (taglist);
693 /* also signals the end of first-frame processing */
694 wavpackparse->sent_codec_tag = TRUE;