xref: /third_party/libsnd/src/ogg_speex.c (revision b815c7f3)
1/*
2** Copyright (C) 2008-2016 Erik de Castro Lopo <erikd@mega-nerd.com>
3**
4** This program is free software ; you can redistribute it and/or modify
5** it under the terms of the GNU Lesser General Public License as published by
6** the Free Software Foundation ; either version 2.1 of the License, or
7** (at your option) any later version.
8**
9** This program is distributed in the hope that it will be useful,
10** but WITHOUT ANY WARRANTY ; without even the implied warranty of
11** MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.	 See the
12** GNU Lesser General Public License for more details.
13**
14** You should have received a copy of the GNU Lesser General Public License
15** along with this program ; if not, write to the Free Software
16** Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
17*/
18
19
20#include "sfconfig.h"
21
22#include <stdio.h>
23#include <fcntl.h>
24#include <string.h>
25#include <ctype.h>
26#include <time.h>
27#include <math.h>
28
29#if HAVE_UNISTD_H
30#include <unistd.h>
31#else
32#include "sf_unistd.h"
33#endif
34
35#include "sndfile.h"
36#include "sfendian.h"
37#include "common.h"
38
39#if (ENABLE_EXPERIMENTAL_CODE && HAVE_EXTERNAL_XIPH_LIBS)
40
41#include <ogg/ogg.h>
42
43#include <speex/speex.h>
44#include <speex/speex_stereo.h>
45#include <speex/speex_header.h>
46#include <speex/speex_callbacks.h>
47
48#include "ogg.h"
49
50#define	OGG_SPX_READ_SIZE	200
51
52typedef struct
53{	SpeexBits bits ;
54
55	int32_t serialno ;
56
57	int frame_size, granule_frame_size, nframes ;
58	int force_mode ;
59
60	SpeexStereoState stereo ;
61	SpeexHeader header ;
62
63	void * state ;
64} SPX_PRIVATE ;
65
66static int	spx_read_header (SF_PRIVATE * psf) ;
67static int	spx_close (SF_PRIVATE *psf) ;
68static void *spx_header_read (SF_PRIVATE * psf, ogg_packet *op, spx_int32_t enh_enabled, int force_mode) ;
69static void spx_print_comments (const char *comments, int length) ;
70
71int
72ogg_speex_open (SF_PRIVATE *psf)
73{	OGG_PRIVATE* odata = psf->container_data ;
74	SPX_PRIVATE* spx = calloc (1, sizeof (SPX_PRIVATE)) ;
75	int	error = 0 ;
76
77	if (odata == NULL)
78	{	psf_log_printf (psf, "%s : odata is NULL???\n", __func__) ;
79		free (spx) ;
80		return SFE_INTERNAL ;
81		} ;
82
83	psf->codec_data = spx ;
84	if (spx == NULL)
85		return SFE_MALLOC_FAILED ;
86
87	if (psf->file.mode == SFM_RDWR)
88		return SFE_BAD_MODE_RW ;
89
90	if (psf->file.mode == SFM_READ)
91	{	/* Call this here so it only gets called once, so no memory is leaked. */
92		ogg_sync_init (&odata->osync) ;
93
94		if ((error = spx_read_header (psf)))
95			return error ;
96
97#if 0
98		psf->read_short		= spx_read_s ;
99		psf->read_int		= spx_read_i ;
100		psf->read_float		= spx_read_f ;
101		psf->read_double	= spx_read_d ;
102		psf->sf.frames		= spx_length (psf) ;
103#endif
104		} ;
105
106	psf->codec_close = spx_close ;
107
108	if (psf->file.mode == SFM_WRITE)
109	{
110#if 0
111		/* Set the default spx quality here. */
112		vdata->quality = 0.4 ;
113
114		psf->write_header	= spx_write_header ;
115		psf->write_short	= spx_write_s ;
116		psf->write_int		= spx_write_i ;
117		psf->write_float	= spx_write_f ;
118		psf->write_double	= spx_write_d ;
119#endif
120
121		psf->sf.frames = SF_COUNT_MAX ; /* Unknown really */
122		psf->strings.flags = SF_STR_ALLOW_START ;
123		} ;
124
125	psf->bytewidth = 1 ;
126	psf->blockwidth = psf->bytewidth * psf->sf.channels ;
127
128#if 0
129	psf->seek = spx_seek ;
130	psf->command = spx_command ;
131#endif
132
133	/* FIXME, FIXME, FIXME : Hack these here for now and correct later. */
134	psf->sf.format = SF_FORMAT_OGG | SF_FORMAT_SPEEX ;
135	psf->sf.sections = 1 ;
136
137	psf->datalength = 1 ;
138	psf->dataoffset = 0 ;
139	/* End FIXME. */
140
141	return error ;
142} /* ogg_speex_open */
143
144#define le_short (x)	(x)
145
146static int
147spx_read_header (SF_PRIVATE * psf)
148{	static SpeexStereoState STEREO_INIT = SPEEX_STEREO_STATE_INIT ;
149
150	OGG_PRIVATE* odata = psf->container_data ;
151	SPX_PRIVATE* spx = psf->codec_data ;
152
153	ogg_int64_t page_granule = 0 ;
154	int stream_init = 0 ;
155	int	page_nb_packets = 0 ;
156	int packet_count = 0 ;
157	int enh_enabled = 1 ;
158	int force_mode = -1 ;
159	char * data ;
160	int nb_read ;
161	int lookahead ;
162
163printf ("%s %d\n", __func__, __LINE__) ;
164
165	psf_log_printf (psf, "Speex header\n") ;
166	odata->eos = 0 ;
167
168	/* Reset ogg stuff which has already been used in src/ogg.c. */
169	ogg_stream_reset (&odata->ostream) ;
170	ogg_sync_reset (&odata->osync) ;
171
172	/* Seek to start of stream. */
173	psf_fseek (psf, 0, SEEK_SET) ;
174
175	/* Initialize. */
176	ogg_sync_init (&odata->osync) ;
177	speex_bits_init (&spx->bits) ;
178
179	/* Set defaults. */
180	psf->sf.channels = -1 ;
181	psf->sf.samplerate = 0 ;
182	spx->stereo = STEREO_INIT ;
183
184	/* Get a pointer to the ogg buffer and read data into it. */
185	data = ogg_sync_buffer (&odata->osync, OGG_SPX_READ_SIZE) ;
186	nb_read = psf_fread (data, 1, OGG_SPX_READ_SIZE, psf) ;
187	ogg_sync_wrote (&odata->osync, nb_read) ;
188
189	/* Now we chew on Ogg packets. */
190	while (ogg_sync_pageout (&odata->osync, &odata->opage) == 1)
191	{	if (stream_init == 0)
192		{	ogg_stream_init (&odata->ostream, ogg_page_serialno (&odata->opage)) ;
193			stream_init = 1 ;
194			} ;
195
196		if (ogg_page_serialno (&odata->opage) != odata->ostream.serialno)
197		{	/* so all streams are read. */
198			ogg_stream_reset_serialno (&odata->ostream, ogg_page_serialno (&odata->opage)) ;
199			} ;
200
201		/*Add page to the bitstream*/
202		ogg_stream_pagein (&odata->ostream, &odata->opage) ;
203		page_granule = ogg_page_granulepos (&odata->opage) ;
204		page_nb_packets = ogg_page_packets (&odata->opage) ;
205
206		/*Extract all available packets*/
207		while (odata->eos == 0 && ogg_stream_packetout (&odata->ostream, &odata->opacket) == 1)
208		{	if (odata->opacket.bytes >= 8 && memcmp (odata->opacket.packet, "Speex   ", 8) == 0)
209			{	spx->serialno = odata->ostream.serialno ;
210				} ;
211
212			if (spx->serialno == -1 || odata->ostream.serialno != spx->serialno)
213				break ;
214
215			if (packet_count == 0)
216			{	spx->state = spx_header_read (psf, &odata->opacket, enh_enabled, force_mode) ;
217				if (! spx->state)
218					break ;
219
220				speex_decoder_ctl (spx->state, SPEEX_GET_LOOKAHEAD, &lookahead) ;
221				if (spx->nframes == 0)
222					spx->nframes = 1 ;
223				}
224			else if (packet_count == 1)
225			{	spx_print_comments ((const char*) odata->opacket.packet, odata->opacket.bytes) ;
226				}
227			else if (packet_count < 2 + spx->header.extra_headers)
228			{	/* Ignore extra headers */
229				}
230			packet_count ++ ;
231			} ;
232		} ;
233
234	psf_log_printf (psf, "End\n") ;
235
236	psf_log_printf (psf, "packet_count %d\n", packet_count) ;
237	psf_log_printf (psf, "page_nb_packets %d\n", page_nb_packets) ;
238	psf_log_printf (psf, "page_granule %lld\n", page_granule) ;
239
240	return 0 ;
241} /* spx_read_header */
242
243static int
244spx_close (SF_PRIVATE *psf)
245{	SPX_PRIVATE* spx = psf->codec_data ;
246
247	if (spx->state)
248		speex_decoder_destroy (spx->state) ;
249
250	if (spx)
251		speex_bits_destroy (&spx->bits) ;
252
253	return 0 ;
254} /* spx_close */
255
256
257
258static void *
259spx_header_read (SF_PRIVATE * psf, ogg_packet *op, spx_int32_t enh_enabled, int force_mode)
260{	SPX_PRIVATE* spx = psf->codec_data ;
261	void *st ;
262	const SpeexMode *mode ;
263	SpeexHeader *tmp_header ;
264	int modeID ;
265	SpeexCallback callback ;
266
267	tmp_header = speex_packet_to_header ((char*) op->packet, op->bytes) ;
268	if (tmp_header == NULL)
269	{	psf_log_printf (psf, "Cannot read Speex header\n") ;
270		return NULL ;
271		} ;
272
273	memcpy (&spx->header, tmp_header, sizeof (spx->header)) ;
274	free (tmp_header) ;
275	tmp_header = NULL ;
276
277	if (spx->header.mode >= SPEEX_NB_MODES || spx->header.mode < 0)
278	{	psf_log_printf (psf, "Mode number %d does not (yet/any longer) exist in this version\n", spx->header.mode) ;
279		return NULL ;
280		} ;
281
282	modeID = spx->header.mode ;
283	if (force_mode != -1)
284		modeID = force_mode ;
285
286	mode = speex_lib_get_mode (modeID) ;
287
288	if (spx->header.speex_version_id > 1)
289	{	psf_log_printf (psf, "This file was encoded with Speex bit-stream version %d, which I don't know how to decode\n", spx->header.speex_version_id) ;
290		return NULL ;
291		} ;
292
293	if (mode->bitstream_version < spx->header.mode_bitstream_version)
294	{	psf_log_printf (psf, "The file was encoded with a newer version of Speex. You need to upgrade in order to play it.\n") ;
295		return NULL ;
296		} ;
297
298	if (mode->bitstream_version > spx->header.mode_bitstream_version)
299	{	psf_log_printf (psf, "The file was encoded with an older version of Speex. You would need to downgrade the version in order to play it.\n") ;
300		return NULL ;
301		} ;
302
303	st = speex_decoder_init (mode) ;
304	if (!st)
305	{	psf_log_printf (psf, "Decoder initialization failed.\n") ;
306		return NULL ;
307		} ;
308
309	speex_decoder_ctl (st, SPEEX_SET_ENH, &enh_enabled) ;
310	speex_decoder_ctl (st, SPEEX_GET_FRAME_SIZE, &spx->frame_size) ;
311	spx->granule_frame_size = spx->frame_size ;
312
313	if (!psf->sf.samplerate)
314		psf->sf.samplerate = spx->header.rate ;
315	/* Adjust rate if --force-* options are used */
316	if (force_mode != -1)
317	{	if (spx->header.mode < force_mode)
318		{	psf->sf.samplerate <<= (force_mode - spx->header.mode) ;
319			spx->granule_frame_size >>= (force_mode - spx->header.mode) ;
320			} ;
321		if (spx->header.mode > force_mode)
322		{	psf->sf.samplerate >>= (spx->header.mode - force_mode) ;
323			spx->granule_frame_size <<= (spx->header.mode - force_mode) ;
324			} ;
325		} ;
326
327	speex_decoder_ctl (st, SPEEX_SET_SAMPLING_RATE, &psf->sf.samplerate) ;
328
329	spx->nframes = spx->header.frames_per_packet ;
330
331	if (psf->sf.channels == -1)
332		psf->sf.channels = spx->header.nb_channels ;
333
334	if (! (psf->sf.channels == 1))
335	{	psf->sf.channels = 2 ;
336		callback.callback_id = SPEEX_INBAND_STEREO ;
337		callback.func = speex_std_stereo_request_handler ;
338		callback.data = &spx->stereo ;
339		speex_decoder_ctl (st, SPEEX_SET_HANDLER, &callback) ;
340		} ;
341
342	spx->header.speex_version [sizeof (spx->header.speex_version) - 1] = 0 ;
343
344	psf_log_printf (psf, "  Encoder ver   : %s\n  Frames/packet : %d\n",
345					spx->header.speex_version, spx->header.frames_per_packet) ;
346
347	if (spx->header.bitrate > 0)
348		psf_log_printf (psf, "  Bit rate	  : %d\n", spx->header.bitrate) ;
349
350	psf_log_printf (psf, "  Sample rate   : %d\n  Mode		  : %s\n  VBR		   : %s\n  Channels	  : %d\n",
351					psf->sf.samplerate, mode->modeName, (spx->header.vbr ? "yes" : "no"), psf->sf.channels) ;
352
353	psf_log_printf (psf, "  Extra headers : %d\n", spx->header.extra_headers) ;
354
355	return st ;
356} /* spx_header_read */
357
358
359static void
360spx_print_comments (const char *c, int length)
361{
362	const char *end ;
363	int len, i, nb_fields ;
364
365printf ("%s %d\n", __func__, __LINE__) ;
366	if (length < 8)
367	{	fprintf (stderr, "Invalid/corrupted comments\n") ;
368		return ;
369		}
370	end = c + length ;
371	len = readint (c, 0) ;
372	c += 4 ;
373	if (len < 0 || c + len > end)
374	{	fprintf (stderr, "Invalid/corrupted comments\n") ;
375		return ;
376		}
377	(void) fwrite (c, 1, len, stderr) ;
378	c += len ;
379	fprintf (stderr, "\n") ;
380	if (c + 4 > end)
381	{	fprintf (stderr, "Invalid/corrupted comments\n") ;
382		return ;
383		}
384	nb_fields = readint (c, 0) ;
385	c += 4 ;
386	for (i = 0 ; i < nb_fields ; i++)
387	{	if (c + 4 > end)
388		{	fprintf (stderr, "Invalid/corrupted comments\n") ;
389			return ;
390			} ;
391		len = readint (c, 0) ;
392		c += 4 ;
393		if (len < 0 || c + len > end)
394		{	fprintf (stderr, "Invalid/corrupted comments\n") ;
395			return ;
396			}
397		(void) fwrite (c, 1, len, stderr) ;
398		c += len ;
399		fprintf (stderr, "\n") ;
400		} ;
401	return ;
402} /* spx_print_comments */
403
404
405/*
406encoded_speex_frames = (frames_per_packet * Packets)
407					 = 1 * 272
408					 = 272
409
410audio_samples = encoded_speex_frames * frame_size
411			  = 272 * 640
412			  = 174080
413
414duration = audio_samples / rate
415		 = 174080 / 44100
416		 = 3.947
417*/
418
419#else /* ENABLE_EXPERIMENTAL_CODE && HAVE_EXTERNAL_XIPH_LIBS */
420
421int
422ogg_speex_open (SF_PRIVATE *psf)
423{
424	psf_log_printf (psf, "This version of libsndfile was compiled without Ogg/Speex support.\n") ;
425	return SFE_UNIMPLEMENTED ;
426} /* ogg_speex_open */
427
428#endif
429