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));
130 gst_wavpack_parse_finalize (GObject * object)
132 G_OBJECT_CLASS (parent_class)->finalize (object);
136 gst_wavpack_parse_start (GstBaseParse * parse)
138 GstWavpackParse *wvparse = GST_WAVPACK_PARSE (parse);
140 GST_DEBUG_OBJECT (parse, "starting");
142 gst_wavpack_parse_reset (wvparse);
144 /* need header at least */
145 gst_base_parse_set_min_frame_size (GST_BASE_PARSE (wvparse),
146 sizeof (WavpackHeader));
148 /* inform baseclass we can come up with ts, based on counters in packets */
149 gst_base_parse_set_has_timing_info (GST_BASE_PARSE_CAST (wvparse), TRUE);
150 gst_base_parse_set_syncable (GST_BASE_PARSE_CAST (wvparse), TRUE);
156 gst_wavpack_parse_stop (GstBaseParse * parse)
158 GST_DEBUG_OBJECT (parse, "stopping");
164 gst_wavpack_get_default_channel_mask (gint nchannels)
166 gint channel_mask = 0;
168 /* Set the default channel mask for the given number of channels.
169 * It's the same as for WAVE_FORMAT_EXTENDED:
170 * http://www.microsoft.com/whdc/device/audio/multichaud.mspx
174 channel_mask |= 0x00400;
175 channel_mask |= 0x00200;
177 channel_mask |= 0x00100;
179 channel_mask |= 0x00080;
180 channel_mask |= 0x00040;
182 channel_mask |= 0x00020;
183 channel_mask |= 0x00010;
185 channel_mask |= 0x00008;
187 channel_mask |= 0x00004;
189 channel_mask |= 0x00002;
190 channel_mask |= 0x00001;
193 /* For mono use front center */
194 channel_mask |= 0x00004;
203 const guint32 ms_mask;
204 const GstAudioChannelPosition gst_pos;
205 } layout_mapping[] = {
207 0x00001, GST_AUDIO_CHANNEL_POSITION_FRONT_LEFT}, {
208 0x00002, GST_AUDIO_CHANNEL_POSITION_FRONT_RIGHT}, {
209 0x00004, GST_AUDIO_CHANNEL_POSITION_FRONT_CENTER}, {
210 0x00008, GST_AUDIO_CHANNEL_POSITION_LFE1}, {
211 0x00010, GST_AUDIO_CHANNEL_POSITION_REAR_LEFT}, {
212 0x00020, GST_AUDIO_CHANNEL_POSITION_REAR_RIGHT}, {
213 0x00040, GST_AUDIO_CHANNEL_POSITION_FRONT_LEFT_OF_CENTER}, {
214 0x00080, GST_AUDIO_CHANNEL_POSITION_FRONT_RIGHT_OF_CENTER}, {
215 0x00100, GST_AUDIO_CHANNEL_POSITION_REAR_CENTER}, {
216 0x00200, GST_AUDIO_CHANNEL_POSITION_SIDE_LEFT}, {
217 0x00400, GST_AUDIO_CHANNEL_POSITION_SIDE_RIGHT}, {
218 0x00800, GST_AUDIO_CHANNEL_POSITION_TOP_CENTER}, {
219 0x01000, GST_AUDIO_CHANNEL_POSITION_TOP_FRONT_LEFT}, {
220 0x02000, GST_AUDIO_CHANNEL_POSITION_TOP_FRONT_CENTER}, {
221 0x04000, GST_AUDIO_CHANNEL_POSITION_TOP_FRONT_RIGHT}, {
222 0x08000, GST_AUDIO_CHANNEL_POSITION_TOP_REAR_LEFT}, {
223 0x10000, GST_AUDIO_CHANNEL_POSITION_TOP_REAR_CENTER}, {
224 0x20000, GST_AUDIO_CHANNEL_POSITION_TOP_REAR_RIGHT}
227 #define MAX_CHANNEL_POSITIONS G_N_ELEMENTS (layout_mapping)
230 gst_wavpack_get_channel_positions (gint num_channels, gint layout,
231 GstAudioChannelPosition * pos)
235 if (num_channels == 1 && layout == 0x00004) {
236 pos[0] = GST_AUDIO_CHANNEL_POSITION_MONO;
241 for (i = 0; i < MAX_CHANNEL_POSITIONS; ++i) {
242 if ((layout & layout_mapping[i].ms_mask) != 0) {
243 if (p >= num_channels) {
244 GST_WARNING ("More bits set in the channel layout map than there "
245 "are channels! Broken file");
248 if (layout_mapping[i].gst_pos == GST_AUDIO_CHANNEL_POSITION_INVALID) {
249 GST_WARNING ("Unsupported channel position (mask 0x%08x) in channel "
250 "layout map - ignoring those channels", layout_mapping[i].ms_mask);
251 /* what to do? just ignore it and let downstream deal with a channel
252 * layout that has INVALID positions in it for now ... */
254 pos[p] = layout_mapping[i].gst_pos;
259 if (p != num_channels) {
260 GST_WARNING ("Only %d bits set in the channel layout map, but there are "
261 "supposed to be %d channels! Broken file", p, num_channels);
268 static const guint32 sample_rates[] = {
269 6000, 8000, 9600, 11025, 12000, 16000, 22050,
270 24000, 32000, 44100, 48000, 64000, 88200, 96000, 192000
273 #define CHECK(call) { \
278 /* caller ensures properly sync'ed with enough data */
280 gst_wavpack_parse_frame_metadata (GstWavpackParse * parse, GstBuffer * buf,
281 gint skip, WavpackHeader * wph, WavpackInfo * wpi)
287 g_return_val_if_fail (wph != NULL || wpi != NULL, FALSE);
288 g_return_val_if_fail (gst_buffer_get_size (buf) >=
289 skip + sizeof (WavpackHeader), FALSE);
291 gst_buffer_map (buf, &map, GST_MAP_READ);
293 gst_byte_reader_init (&br, map.data + skip, wph->ckSize + 8);
294 /* skip past header */
295 gst_byte_reader_skip_unchecked (&br, sizeof (WavpackHeader));
297 /* get some basics from header */
298 i = (wph->flags >> 23) & 0xF;
300 wpi->rate = (i < G_N_ELEMENTS (sample_rates)) ? sample_rates[i] : 44100;
301 wpi->width = ((wph->flags & 0x3) + 1) * 8;
303 wpi->channels = (wph->flags & 0x4) ? 1 : 2;
304 if (!wpi->channel_mask)
305 wpi->channel_mask = 5 - wpi->channels;
307 /* need to dig metadata blocks for some more */
308 while (gst_byte_reader_get_remaining (&br)) {
315 CHECK (gst_byte_reader_get_uint8 (&br, &id));
316 CHECK (gst_byte_reader_get_uint8 (&br, &c));
318 CHECK (gst_byte_reader_get_uint16_le (&br, &size2));
323 if (id & ID_ODD_SIZE)
326 CHECK (gst_byte_reader_get_data (&br, size + (size & 1), &data));
327 gst_byte_reader_init (&mbr, data, size);
330 case ID_WVC_BITSTREAM:
331 GST_LOG_OBJECT (parse, "correction bitstream");
332 wpi->correction = TRUE;
334 case ID_WV_BITSTREAM:
335 case ID_WVX_BITSTREAM:
339 CHECK (gst_byte_reader_get_uint24_le (&mbr, &wpi->rate));
340 GST_LOG_OBJECT (parse, "updated with custom rate %d", wpi->rate);
342 GST_DEBUG_OBJECT (parse, "unexpected size for SAMPLE_RATE metadata");
345 case ID_CHANNEL_INFO:
351 CHECK (gst_byte_reader_get_uint16_le (&mbr, &channels));
352 channels = channels & 0xFFF;
353 CHECK (gst_byte_reader_get_uint24_le (&mbr, &mask));
355 CHECK (gst_byte_reader_get_uint8 (&mbr, &c));
357 while (gst_byte_reader_get_uint8 (&mbr, &c))
358 mask |= (((guint32) c) << 8);
360 GST_DEBUG_OBJECT (parse, "unexpected size for CHANNEL_INFO metadata");
363 wpi->channels = channels;
364 wpi->channel_mask = mask;
368 GST_LOG_OBJECT (parse, "unparsed ID 0x%x", id);
373 gst_buffer_unmap (buf, &map);
380 gst_buffer_unmap (buf, &map);
381 GST_DEBUG_OBJECT (parse, "short read while parsing metadata");
382 /* let's look the other way anyway */
387 /* caller ensures properly sync'ed with enough data */
389 gst_wavpack_parse_frame_header (GstWavpackParse * parse, GstBuffer * buf,
390 gint skip, WavpackHeader * _wph)
393 WavpackHeader wph = { {0,}, 0, };
397 g_return_val_if_fail (gst_buffer_get_size (buf) >=
398 skip + sizeof (WavpackHeader), FALSE);
400 gst_buffer_map (buf, &map, GST_MAP_READ);
401 gst_byte_reader_init (&br, map.data, map.size);
404 gst_byte_reader_skip_unchecked (&br, skip + 4);
407 hdl &= gst_byte_reader_get_uint32_le (&br, &wph.ckSize);
408 hdl &= gst_byte_reader_get_uint16_le (&br, &wph.version);
409 hdl &= gst_byte_reader_get_uint8 (&br, &wph.track_no);
410 hdl &= gst_byte_reader_get_uint8 (&br, &wph.index_no);
411 hdl &= gst_byte_reader_get_uint32_le (&br, &wph.total_samples);
412 hdl &= gst_byte_reader_get_uint32_le (&br, &wph.block_index);
413 hdl &= gst_byte_reader_get_uint32_le (&br, &wph.block_samples);
414 hdl &= gst_byte_reader_get_uint32_le (&br, &wph.flags);
415 hdl &= gst_byte_reader_get_uint32_le (&br, &wph.crc);
418 GST_WARNING_OBJECT (parse, "Error reading header");
421 GST_LOG_OBJECT (parse, "size %d", wph.ckSize);
422 GST_LOG_OBJECT (parse, "version 0x%x", wph.version);
423 GST_LOG_OBJECT (parse, "total samples %d", wph.total_samples);
424 GST_LOG_OBJECT (parse, "block index %d", wph.block_index);
425 GST_LOG_OBJECT (parse, "block samples %d", wph.block_samples);
426 GST_LOG_OBJECT (parse, "flags 0x%x", wph.flags);
427 GST_LOG_OBJECT (parse, "crc 0x%x", wph.flags);
429 if (!parse->total_samples && wph.block_index == 0 && wph.total_samples != -1) {
430 GST_DEBUG_OBJECT (parse, "determined duration of %u samples",
432 parse->total_samples = wph.total_samples;
438 gst_buffer_unmap (buf, &map);
444 gst_wavpack_parse_handle_frame (GstBaseParse * parse,
445 GstBaseParseFrame * frame, gint * skipsize)
447 GstWavpackParse *wvparse = GST_WAVPACK_PARSE (parse);
448 GstBuffer *buf = frame->buffer;
449 GstByteReader reader;
451 guint rate, chans, width, mask;
452 gboolean lost_sync, draining, final;
455 WavpackInfo wpi = { 0, };
458 if (G_UNLIKELY (gst_buffer_get_size (buf) < sizeof (WavpackHeader)))
461 gst_buffer_map (buf, &map, GST_MAP_READ);
462 gst_byte_reader_init (&reader, map.data, map.size);
464 /* scan for 'wvpk' marker */
465 off = gst_byte_reader_masked_scan_uint32 (&reader, 0xffffffff, 0x7776706b,
468 GST_LOG_OBJECT (parse, "possible sync at buffer offset %d", off);
470 /* didn't find anything that looks like a sync word, skip */
472 *skipsize = map.size - 3;
476 /* possible frame header, but not at offset 0? skip bytes before sync */
482 /* make sure the values in the frame header look sane */
483 gst_wavpack_parse_frame_header (wvparse, buf, 0, &wph);
484 frmsize = wph.ckSize + 8;
486 /* need the entire frame for parsing */
487 if (gst_byte_reader_get_remaining (&reader) < frmsize)
490 /* got a frame, now we can dig for some more metadata */
491 GST_LOG_OBJECT (parse, "got frame");
492 gst_wavpack_parse_frame_metadata (wvparse, buf, 0, &wph, &wpi);
494 lost_sync = GST_BASE_PARSE_LOST_SYNC (parse);
495 draining = GST_BASE_PARSE_DRAINING (parse);
497 while (!(final = (wph.flags & FLAG_FINAL_BLOCK)) || (lost_sync && !draining)) {
500 GST_LOG_OBJECT (wvparse, "checking next frame syncword; "
501 "lost_sync: %d, draining: %d, final: %d", lost_sync, draining, final);
503 if (!gst_byte_reader_skip (&reader, wph.ckSize + 8) ||
504 !gst_byte_reader_peek_uint32_be (&reader, &word)) {
505 GST_DEBUG_OBJECT (wvparse, "... but not sufficient data");
509 if (word != 0x7776706b) {
510 GST_DEBUG_OBJECT (wvparse, "0x%x not OK", word);
514 /* need to parse each frame/block for metadata if several ones */
518 GST_LOG_OBJECT (wvparse, "checking frame at offset %d (0x%x)",
520 av = gst_byte_reader_get_remaining (&reader);
521 if (av < sizeof (WavpackHeader)) {
522 frmsize += sizeof (WavpackHeader);
525 gst_wavpack_parse_frame_header (wvparse, buf, frmsize, &wph);
527 frmsize += wph.ckSize + 8;
528 if (av < wph.ckSize + 8)
530 gst_wavpack_parse_frame_metadata (wvparse, buf, off, &wph, &wpi);
531 /* could also check for matching block_index and block_samples ?? */
535 /* resynced if we make it here */
541 chans = wpi.channels;
542 mask = wpi.channel_mask;
544 GST_LOG_OBJECT (parse, "rate: %u, width: %u, chans: %u", rate, width, chans);
546 GST_BUFFER_TIMESTAMP (buf) =
547 gst_util_uint64_scale_int (wph.block_index, GST_SECOND, rate);
548 GST_BUFFER_DURATION (buf) =
549 gst_util_uint64_scale_int (wph.block_index + wph.block_samples,
550 GST_SECOND, rate) - GST_BUFFER_TIMESTAMP (buf);
552 if (G_UNLIKELY (wvparse->sample_rate != rate || wvparse->channels != chans
553 || wvparse->width != width || wvparse->channel_mask != mask)) {
556 if (wpi.correction) {
557 caps = gst_caps_new_simple ("audio/x-wavpack-correction",
558 "framed", G_TYPE_BOOLEAN, TRUE, NULL);
560 caps = gst_caps_new_simple ("audio/x-wavpack",
561 "channels", G_TYPE_INT, chans,
562 "rate", G_TYPE_INT, rate,
563 "depth", G_TYPE_INT, width, "framed", G_TYPE_BOOLEAN, TRUE, NULL);
566 mask = gst_wavpack_get_default_channel_mask (wvparse->channels);
568 GstAudioChannelPosition pos[64] =
569 { GST_AUDIO_CHANNEL_POSITION_INVALID, };
572 if (!gst_wavpack_get_channel_positions (chans, mask, pos)) {
573 GST_WARNING_OBJECT (wvparse, "Failed to determine channel layout");
575 gst_audio_channel_positions_to_mask (pos, chans, FALSE, &gmask);
577 gst_caps_set_simple (caps,
578 "channel-mask", GST_TYPE_BITMASK, gmask, NULL);
583 gst_pad_set_caps (GST_BASE_PARSE_SRC_PAD (parse), caps);
584 gst_caps_unref (caps);
586 wvparse->sample_rate = rate;
587 wvparse->channels = chans;
588 wvparse->width = width;
589 wvparse->channel_mask = mask;
591 if (wvparse->total_samples) {
592 GST_DEBUG_OBJECT (wvparse, "setting duration");
593 gst_base_parse_set_duration (GST_BASE_PARSE (wvparse),
594 GST_FORMAT_TIME, gst_util_uint64_scale_int (wvparse->total_samples,
595 GST_SECOND, wvparse->sample_rate), 0);
599 /* return to normal size */
600 gst_base_parse_set_min_frame_size (parse, sizeof (WavpackHeader));
601 gst_buffer_unmap (buf, &map);
603 return gst_base_parse_finish_frame (parse, frame, frmsize);
606 gst_buffer_unmap (buf, &map);
607 GST_LOG_OBJECT (wvparse, "skipping %d", *skipsize);
611 gst_buffer_unmap (buf, &map);
612 GST_LOG_OBJECT (wvparse, "need at least %u", frmsize);
613 gst_base_parse_set_min_frame_size (parse, frmsize);
619 remove_fields (GstCaps * caps)
623 n = gst_caps_get_size (caps);
624 for (i = 0; i < n; i++) {
625 GstStructure *s = gst_caps_get_structure (caps, i);
627 gst_structure_remove_field (s, "framed");
632 gst_wavpack_parse_get_sink_caps (GstBaseParse * parse, GstCaps * filter)
634 GstCaps *peercaps, *templ;
637 templ = gst_pad_get_pad_template_caps (GST_BASE_PARSE_SINK_PAD (parse));
639 GstCaps *fcopy = gst_caps_copy (filter);
640 /* Remove the fields we convert */
641 remove_fields (fcopy);
642 peercaps = gst_pad_peer_query_caps (GST_BASE_PARSE_SRC_PAD (parse), fcopy);
643 gst_caps_unref (fcopy);
645 peercaps = gst_pad_peer_query_caps (GST_BASE_PARSE_SRC_PAD (parse), NULL);
648 /* Remove the framed field */
649 peercaps = gst_caps_make_writable (peercaps);
650 remove_fields (peercaps);
652 res = gst_caps_intersect_full (peercaps, templ, GST_CAPS_INTERSECT_FIRST);
653 gst_caps_unref (peercaps);
654 gst_caps_unref (templ);
660 GstCaps *intersection;
663 gst_caps_intersect_full (filter, res, GST_CAPS_INTERSECT_FIRST);
664 gst_caps_unref (res);
672 gst_wavpack_parse_pre_push_frame (GstBaseParse * parse,
673 GstBaseParseFrame * frame)
675 GstWavpackParse *wavpackparse = GST_WAVPACK_PARSE (parse);
677 if (!wavpackparse->sent_codec_tag) {
681 taglist = gst_tag_list_new_empty ();
684 caps = gst_pad_get_current_caps (GST_BASE_PARSE_SRC_PAD (parse));
685 gst_pb_utils_add_codec_description_to_tag_list (taglist,
686 GST_TAG_AUDIO_CODEC, caps);
687 gst_caps_unref (caps);
689 gst_pad_push_event (GST_BASE_PARSE_SRC_PAD (wavpackparse),
690 gst_event_new_tag (taglist));
692 /* also signals the end of first-frame processing */
693 wavpackparse->sent_codec_tag = TRUE;