2 * copyright (c) 2002 Mark Hills <mark@pogo.org.uk>
4 * This file is part of FFmpeg.
6 * FFmpeg is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Lesser General Public
8 * License as published by the Free Software Foundation; either
9 * version 2.1 of the License, or (at your option) any later version.
11 * FFmpeg 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 * Lesser General Public License for more details.
16 * You should have received a copy of the GNU Lesser General Public
17 * License along with FFmpeg; if not, write to the Free Software
18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
23 * Ogg Vorbis codec support via libvorbisenc.
24 * @author Mark Hills <mark@pogo.org.uk>
27 #include <vorbis/vorbisenc.h>
29 #include "libavutil/opt.h"
31 #include "bytestream.h"
37 #define OGGVORBIS_FRAME_SIZE 64
39 #define BUFFER_SIZE (1024*64)
41 typedef struct OggVorbisContext {
46 uint8_t buffer[BUFFER_SIZE];
57 static const AVOption options[]={
58 {"iblock", "Sets the impulse block bias", offsetof(OggVorbisContext, iblock), FF_OPT_TYPE_DOUBLE, {.dbl = 0}, -15, 0, AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_ENCODING_PARAM},
61 static const AVClass class = { "libvorbis", av_default_item_name, options, LIBAVUTIL_VERSION_INT };
63 static av_cold int oggvorbis_init_encoder(vorbis_info *vi, AVCodecContext *avccontext) {
64 OggVorbisContext *context = avccontext->priv_data ;
67 if(avccontext->flags & CODEC_FLAG_QSCALE) {
68 /* variable bitrate */
69 if(vorbis_encode_setup_vbr(vi, avccontext->channels,
70 avccontext->sample_rate,
71 avccontext->global_quality / (float)FF_QP2LAMBDA / 10.0))
74 int minrate = avccontext->rc_min_rate > 0 ? avccontext->rc_min_rate : -1;
75 int maxrate = avccontext->rc_min_rate > 0 ? avccontext->rc_max_rate : -1;
77 /* constant bitrate */
78 if(vorbis_encode_setup_managed(vi, avccontext->channels,
79 avccontext->sample_rate, minrate, avccontext->bit_rate, maxrate))
82 /* variable bitrate by estimate, disable slow rate management */
83 if(minrate == -1 && maxrate == -1)
84 if(vorbis_encode_ctl(vi, OV_ECTL_RATEMANAGE2_SET, NULL))
88 /* cutoff frequency */
89 if(avccontext->cutoff > 0) {
90 cfreq = avccontext->cutoff / 1000.0;
91 if(vorbis_encode_ctl(vi, OV_ECTL_LOWPASS_SET, &cfreq))
96 vorbis_encode_ctl(vi, OV_ECTL_IBLOCK_SET, &context->iblock);
99 if (avccontext->channels == 3 &&
100 avccontext->channel_layout != (AV_CH_LAYOUT_STEREO|AV_CH_FRONT_CENTER) ||
101 avccontext->channels == 4 &&
102 avccontext->channel_layout != AV_CH_LAYOUT_2_2 &&
103 avccontext->channel_layout != AV_CH_LAYOUT_QUAD ||
104 avccontext->channels == 5 &&
105 avccontext->channel_layout != AV_CH_LAYOUT_5POINT0 &&
106 avccontext->channel_layout != AV_CH_LAYOUT_5POINT0_BACK ||
107 avccontext->channels == 6 &&
108 avccontext->channel_layout != AV_CH_LAYOUT_5POINT1 &&
109 avccontext->channel_layout != AV_CH_LAYOUT_5POINT1_BACK ||
110 avccontext->channels == 7 &&
111 avccontext->channel_layout != (AV_CH_LAYOUT_5POINT1|AV_CH_BACK_CENTER) ||
112 avccontext->channels == 8 &&
113 avccontext->channel_layout != AV_CH_LAYOUT_7POINT1) {
114 if (avccontext->channel_layout) {
116 av_get_channel_layout_string(name, sizeof(name), avccontext->channels,
117 avccontext->channel_layout);
118 av_log(avccontext, AV_LOG_ERROR, "%s not supported by Vorbis: "
119 "output stream will have incorrect "
120 "channel layout.\n", name);
122 av_log(avccontext, AV_LOG_WARNING, "No channel layout specified. The encoder "
123 "will use Vorbis channel layout for "
124 "%d channels.\n", avccontext->channels);
128 return vorbis_encode_setup_init(vi);
131 /* How many bytes are needed for a buffer of length 'l' */
132 static int xiph_len(int l) { return (1 + l / 255 + l); }
134 static av_cold int oggvorbis_encode_init(AVCodecContext *avccontext) {
135 OggVorbisContext *context = avccontext->priv_data ;
136 ogg_packet header, header_comm, header_code;
140 vorbis_info_init(&context->vi) ;
141 if(oggvorbis_init_encoder(&context->vi, avccontext) < 0) {
142 av_log(avccontext, AV_LOG_ERROR, "oggvorbis_encode_init: init_encoder failed\n") ;
145 vorbis_analysis_init(&context->vd, &context->vi) ;
146 vorbis_block_init(&context->vd, &context->vb) ;
148 vorbis_comment_init(&context->vc);
149 vorbis_comment_add_tag(&context->vc, "encoder", LIBAVCODEC_IDENT) ;
151 vorbis_analysis_headerout(&context->vd, &context->vc, &header,
152 &header_comm, &header_code);
154 avccontext->extradata_size=
155 1 + xiph_len(header.bytes) + xiph_len(header_comm.bytes) +
157 p = avccontext->extradata =
158 av_malloc(avccontext->extradata_size + FF_INPUT_BUFFER_PADDING_SIZE);
161 offset += av_xiphlacing(&p[offset], header.bytes);
162 offset += av_xiphlacing(&p[offset], header_comm.bytes);
163 memcpy(&p[offset], header.packet, header.bytes);
164 offset += header.bytes;
165 memcpy(&p[offset], header_comm.packet, header_comm.bytes);
166 offset += header_comm.bytes;
167 memcpy(&p[offset], header_code.packet, header_code.bytes);
168 offset += header_code.bytes;
169 assert(offset == avccontext->extradata_size);
171 /* vorbis_block_clear(&context->vb);
172 vorbis_dsp_clear(&context->vd);
173 vorbis_info_clear(&context->vi);*/
174 vorbis_comment_clear(&context->vc);
176 avccontext->frame_size = OGGVORBIS_FRAME_SIZE ;
178 avccontext->coded_frame= avcodec_alloc_frame();
179 avccontext->coded_frame->key_frame= 1;
185 static int oggvorbis_encode_frame(AVCodecContext *avccontext,
186 unsigned char *packets,
187 int buf_size, void *data)
189 OggVorbisContext *context = avccontext->priv_data ;
191 signed short *audio = data ;
195 const int samples = avccontext->frame_size;
197 int c, channels = context->vi.channels;
199 buffer = vorbis_analysis_buffer(&context->vd, samples) ;
200 for (c = 0; c < channels; c++) {
201 int co = (channels > 8) ? c :
202 ff_vorbis_encoding_channel_layout_offsets[channels-1][c];
203 for(l = 0 ; l < samples ; l++)
204 buffer[c][l]=audio[l*channels+co]/32768.f;
206 vorbis_analysis_wrote(&context->vd, samples) ;
209 vorbis_analysis_wrote(&context->vd, 0) ;
213 while(vorbis_analysis_blockout(&context->vd, &context->vb) == 1) {
214 vorbis_analysis(&context->vb, NULL);
215 vorbis_bitrate_addblock(&context->vb) ;
217 while(vorbis_bitrate_flushpacket(&context->vd, &op)) {
218 /* i'd love to say the following line is a hack, but sadly it's
219 * not, apparently the end of stream decision is in libogg. */
220 if(op.bytes==1 && op.e_o_s)
222 if (context->buffer_index + sizeof(ogg_packet) + op.bytes > BUFFER_SIZE) {
223 av_log(avccontext, AV_LOG_ERROR, "libvorbis: buffer overflow.");
226 memcpy(context->buffer + context->buffer_index, &op, sizeof(ogg_packet));
227 context->buffer_index += sizeof(ogg_packet);
228 memcpy(context->buffer + context->buffer_index, op.packet, op.bytes);
229 context->buffer_index += op.bytes;
230 // av_log(avccontext, AV_LOG_DEBUG, "e%d / %d\n", context->buffer_index, op.bytes);
235 if(context->buffer_index){
236 ogg_packet *op2= (ogg_packet*)context->buffer;
237 op2->packet = context->buffer + sizeof(ogg_packet);
240 avccontext->coded_frame->pts= av_rescale_q(op2->granulepos, (AVRational){1, avccontext->sample_rate}, avccontext->time_base);
241 //FIXME we should reorder the user supplied pts and not assume that they are spaced by 1/sample_rate
244 av_log(avccontext, AV_LOG_ERROR, "libvorbis: buffer overflow.");
248 memcpy(packets, op2->packet, l);
249 context->buffer_index -= l + sizeof(ogg_packet);
250 memmove(context->buffer, context->buffer + l + sizeof(ogg_packet), context->buffer_index);
251 // av_log(avccontext, AV_LOG_DEBUG, "E%d\n", l);
258 static av_cold int oggvorbis_encode_close(AVCodecContext *avccontext) {
259 OggVorbisContext *context = avccontext->priv_data ;
260 /* ogg_packet op ; */
262 vorbis_analysis_wrote(&context->vd, 0) ; /* notify vorbisenc this is EOF */
264 vorbis_block_clear(&context->vb);
265 vorbis_dsp_clear(&context->vd);
266 vorbis_info_clear(&context->vi);
268 av_freep(&avccontext->coded_frame);
269 av_freep(&avccontext->extradata);
275 AVCodec ff_libvorbis_encoder = {
279 sizeof(OggVorbisContext),
280 oggvorbis_encode_init,
281 oggvorbis_encode_frame,
282 oggvorbis_encode_close,
283 .capabilities= CODEC_CAP_DELAY,
284 .sample_fmts = (const enum AVSampleFormat[]){AV_SAMPLE_FMT_S16,AV_SAMPLE_FMT_NONE},
285 .long_name= NULL_IF_CONFIG_SMALL("libvorbis Vorbis"),