00001
00002
00003
00004
00005
00006
00007
00008
00009
00010
00011
00012
00013
00014
00015
00016
00017
00018
00019
00020
00027 #include <vorbis/vorbisenc.h>
00028
00029 #include "libavutil/opt.h"
00030 #include "avcodec.h"
00031 #include "bytestream.h"
00032 #include "vorbis.h"
00033 #include "libavutil/mathematics.h"
00034
00035 #undef NDEBUG
00036 #include <assert.h>
00037
00038 #define OGGVORBIS_FRAME_SIZE 64
00039
00040 #define BUFFER_SIZE (1024 * 64)
00041
00042 typedef struct OggVorbisContext {
00043 AVClass *av_class;
00044 vorbis_info vi;
00045 vorbis_dsp_state vd;
00046 vorbis_block vb;
00047 uint8_t buffer[BUFFER_SIZE];
00048 int buffer_index;
00049 int eof;
00050
00051
00052 vorbis_comment vc;
00053 ogg_packet op;
00054
00055 double iblock;
00056 } OggVorbisContext;
00057
00058 static const AVOption options[] = {
00059 { "iblock", "Sets the impulse block bias", offsetof(OggVorbisContext, iblock), AV_OPT_TYPE_DOUBLE, { .dbl = 0 }, -15, 0, AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM },
00060 { NULL }
00061 };
00062 static const AVClass class = { "libvorbis", av_default_item_name, options, LIBAVUTIL_VERSION_INT };
00063
00064 static const char * error(int oggerr, int *averr)
00065 {
00066 switch (oggerr) {
00067 case OV_EFAULT: *averr = AVERROR(EFAULT); return "internal error";
00068 case OV_EIMPL: *averr = AVERROR(EINVAL); return "not supported";
00069 case OV_EINVAL: *averr = AVERROR(EINVAL); return "invalid request";
00070 default: *averr = AVERROR(EINVAL); return "unknown error";
00071 }
00072 }
00073
00074 static av_cold int oggvorbis_init_encoder(vorbis_info *vi, AVCodecContext *avccontext)
00075 {
00076 OggVorbisContext *context = avccontext->priv_data;
00077 double cfreq;
00078 int r;
00079
00080 if (avccontext->flags & CODEC_FLAG_QSCALE) {
00081
00082 float quality = avccontext->global_quality / (float)FF_QP2LAMBDA;
00083 r = vorbis_encode_setup_vbr(vi, avccontext->channels,
00084 avccontext->sample_rate,
00085 quality / 10.0);
00086 if (r) {
00087 av_log(avccontext, AV_LOG_ERROR,
00088 "Unable to set quality to %g: %s\n", quality, error(r, &r));
00089 return r;
00090 }
00091 } else {
00092 int minrate = avccontext->rc_min_rate > 0 ? avccontext->rc_min_rate : -1;
00093 int maxrate = avccontext->rc_min_rate > 0 ? avccontext->rc_max_rate : -1;
00094
00095
00096 r = vorbis_encode_setup_managed(vi, avccontext->channels,
00097 avccontext->sample_rate, minrate,
00098 avccontext->bit_rate, maxrate);
00099 if (r) {
00100 av_log(avccontext, AV_LOG_ERROR,
00101 "Unable to set CBR to %d: %s\n", avccontext->bit_rate,
00102 error(r, &r));
00103 return r;
00104 }
00105
00106
00107 if (minrate == -1 && maxrate == -1)
00108 if (vorbis_encode_ctl(vi, OV_ECTL_RATEMANAGE2_SET, NULL))
00109 return AVERROR(EINVAL);
00110 }
00111
00112
00113 if (avccontext->cutoff > 0) {
00114 cfreq = avccontext->cutoff / 1000.0;
00115 if (vorbis_encode_ctl(vi, OV_ECTL_LOWPASS_SET, &cfreq))
00116 return AVERROR(EINVAL);
00117 }
00118
00119 if (context->iblock) {
00120 vorbis_encode_ctl(vi, OV_ECTL_IBLOCK_SET, &context->iblock);
00121 }
00122
00123 if (avccontext->channels == 3 &&
00124 avccontext->channel_layout != (AV_CH_LAYOUT_STEREO|AV_CH_FRONT_CENTER) ||
00125 avccontext->channels == 4 &&
00126 avccontext->channel_layout != AV_CH_LAYOUT_2_2 &&
00127 avccontext->channel_layout != AV_CH_LAYOUT_QUAD ||
00128 avccontext->channels == 5 &&
00129 avccontext->channel_layout != AV_CH_LAYOUT_5POINT0 &&
00130 avccontext->channel_layout != AV_CH_LAYOUT_5POINT0_BACK ||
00131 avccontext->channels == 6 &&
00132 avccontext->channel_layout != AV_CH_LAYOUT_5POINT1 &&
00133 avccontext->channel_layout != AV_CH_LAYOUT_5POINT1_BACK ||
00134 avccontext->channels == 7 &&
00135 avccontext->channel_layout != (AV_CH_LAYOUT_5POINT1|AV_CH_BACK_CENTER) ||
00136 avccontext->channels == 8 &&
00137 avccontext->channel_layout != AV_CH_LAYOUT_7POINT1) {
00138 if (avccontext->channel_layout) {
00139 char name[32];
00140 av_get_channel_layout_string(name, sizeof(name), avccontext->channels,
00141 avccontext->channel_layout);
00142 av_log(avccontext, AV_LOG_ERROR, "%s not supported by Vorbis: "
00143 "output stream will have incorrect "
00144 "channel layout.\n", name);
00145 } else {
00146 av_log(avccontext, AV_LOG_WARNING, "No channel layout specified. The encoder "
00147 "will use Vorbis channel layout for "
00148 "%d channels.\n", avccontext->channels);
00149 }
00150 }
00151
00152 return vorbis_encode_setup_init(vi);
00153 }
00154
00155
00156 static int xiph_len(int l)
00157 {
00158 return 1 + l / 255 + l;
00159 }
00160
00161 static av_cold int oggvorbis_encode_init(AVCodecContext *avccontext)
00162 {
00163 OggVorbisContext *context = avccontext->priv_data;
00164 ogg_packet header, header_comm, header_code;
00165 uint8_t *p;
00166 unsigned int offset;
00167 int r;
00168
00169 vorbis_info_init(&context->vi);
00170 r = oggvorbis_init_encoder(&context->vi, avccontext);
00171 if (r < 0) {
00172 av_log(avccontext, AV_LOG_ERROR, "oggvorbis_encode_init failed\n");
00173 return r;
00174 }
00175 vorbis_analysis_init(&context->vd, &context->vi);
00176 vorbis_block_init(&context->vd, &context->vb);
00177
00178 vorbis_comment_init(&context->vc);
00179 vorbis_comment_add_tag(&context->vc, "encoder", LIBAVCODEC_IDENT);
00180
00181 vorbis_analysis_headerout(&context->vd, &context->vc, &header,
00182 &header_comm, &header_code);
00183
00184 avccontext->extradata_size =
00185 1 + xiph_len(header.bytes) + xiph_len(header_comm.bytes) +
00186 header_code.bytes;
00187 p = avccontext->extradata =
00188 av_malloc(avccontext->extradata_size + FF_INPUT_BUFFER_PADDING_SIZE);
00189 p[0] = 2;
00190 offset = 1;
00191 offset += av_xiphlacing(&p[offset], header.bytes);
00192 offset += av_xiphlacing(&p[offset], header_comm.bytes);
00193 memcpy(&p[offset], header.packet, header.bytes);
00194 offset += header.bytes;
00195 memcpy(&p[offset], header_comm.packet, header_comm.bytes);
00196 offset += header_comm.bytes;
00197 memcpy(&p[offset], header_code.packet, header_code.bytes);
00198 offset += header_code.bytes;
00199 assert(offset == avccontext->extradata_size);
00200
00201 #if 0
00202 vorbis_block_clear(&context->vb);
00203 vorbis_dsp_clear(&context->vd);
00204 vorbis_info_clear(&context->vi);
00205 #endif
00206 vorbis_comment_clear(&context->vc);
00207
00208 avccontext->frame_size = OGGVORBIS_FRAME_SIZE;
00209
00210 avccontext->coded_frame = avcodec_alloc_frame();
00211 avccontext->coded_frame->key_frame = 1;
00212
00213 return 0;
00214 }
00215
00216 static int oggvorbis_encode_frame(AVCodecContext *avccontext,
00217 unsigned char *packets,
00218 int buf_size, void *data)
00219 {
00220 OggVorbisContext *context = avccontext->priv_data;
00221 ogg_packet op;
00222 signed short *audio = data;
00223 int l;
00224
00225 if (data) {
00226 const int samples = avccontext->frame_size;
00227 float **buffer;
00228 int c, channels = context->vi.channels;
00229
00230 buffer = vorbis_analysis_buffer(&context->vd, samples);
00231 for (c = 0; c < channels; c++) {
00232 int co = (channels > 8) ? c :
00233 ff_vorbis_encoding_channel_layout_offsets[channels - 1][c];
00234 for (l = 0; l < samples; l++)
00235 buffer[c][l] = audio[l * channels + co] / 32768.f;
00236 }
00237 vorbis_analysis_wrote(&context->vd, samples);
00238 } else {
00239 if (!context->eof)
00240 vorbis_analysis_wrote(&context->vd, 0);
00241 context->eof = 1;
00242 }
00243
00244 while (vorbis_analysis_blockout(&context->vd, &context->vb) == 1) {
00245 vorbis_analysis(&context->vb, NULL);
00246 vorbis_bitrate_addblock(&context->vb);
00247
00248 while (vorbis_bitrate_flushpacket(&context->vd, &op)) {
00249
00250
00251 if (op.bytes == 1 && op.e_o_s)
00252 continue;
00253 if (context->buffer_index + sizeof(ogg_packet) + op.bytes > BUFFER_SIZE) {
00254 av_log(avccontext, AV_LOG_ERROR, "libvorbis: buffer overflow.\n");
00255 return AVERROR(EINVAL);
00256 }
00257 memcpy(context->buffer + context->buffer_index, &op, sizeof(ogg_packet));
00258 context->buffer_index += sizeof(ogg_packet);
00259 memcpy(context->buffer + context->buffer_index, op.packet, op.bytes);
00260 context->buffer_index += op.bytes;
00261
00262 }
00263 }
00264
00265 l = 0;
00266 if (context->buffer_index) {
00267 ogg_packet *op2 = (ogg_packet *)context->buffer;
00268 op2->packet = context->buffer + sizeof(ogg_packet);
00269
00270 l = op2->bytes;
00271 avccontext->coded_frame->pts = av_rescale_q(op2->granulepos, (AVRational) { 1, avccontext->sample_rate }, avccontext->time_base);
00272
00273
00274 if (l > buf_size) {
00275 av_log(avccontext, AV_LOG_ERROR, "libvorbis: buffer overflow.\n");
00276 return AVERROR(EINVAL);
00277 }
00278
00279 memcpy(packets, op2->packet, l);
00280 context->buffer_index -= l + sizeof(ogg_packet);
00281 memmove(context->buffer, context->buffer + l + sizeof(ogg_packet), context->buffer_index);
00282
00283 }
00284
00285 return l;
00286 }
00287
00288 static av_cold int oggvorbis_encode_close(AVCodecContext *avccontext)
00289 {
00290 OggVorbisContext *context = avccontext->priv_data;
00291
00292
00293 vorbis_analysis_wrote(&context->vd, 0);
00294
00295 vorbis_block_clear(&context->vb);
00296 vorbis_dsp_clear(&context->vd);
00297 vorbis_info_clear(&context->vi);
00298
00299 av_freep(&avccontext->coded_frame);
00300 av_freep(&avccontext->extradata);
00301
00302 return 0;
00303 }
00304
00305 AVCodec ff_libvorbis_encoder = {
00306 .name = "libvorbis",
00307 .type = AVMEDIA_TYPE_AUDIO,
00308 .id = CODEC_ID_VORBIS,
00309 .priv_data_size = sizeof(OggVorbisContext),
00310 .init = oggvorbis_encode_init,
00311 .encode = oggvorbis_encode_frame,
00312 .close = oggvorbis_encode_close,
00313 .capabilities = CODEC_CAP_DELAY,
00314 .sample_fmts = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_NONE },
00315 .long_name = NULL_IF_CONFIG_SMALL("libvorbis Vorbis"),
00316 .priv_class = &class,
00317 };