diff options
author | Justin Ruggles <justin.ruggles@gmail.com> | 2012-02-28 18:51:04 -0500 |
---|---|---|
committer | Justin Ruggles <justin.ruggles@gmail.com> | 2012-02-29 14:44:15 -0500 |
commit | eb35ef2932e42d9b203a8bf9e5dba6f1c666ce1e (patch) | |
tree | 8ead4d5a6fb201501c7d32cc983d7d569f0adf56 /libavcodec/libvorbis.c | |
parent | 4e99501f629f6baebac0414d92d841b64ead30fe (diff) | |
download | ffmpeg-eb35ef2932e42d9b203a8bf9e5dba6f1c666ce1e.tar.gz |
libvorbis: cosmetics: renaming/pretty-printing/comments/unused code
Diffstat (limited to 'libavcodec/libvorbis.c')
-rw-r--r-- | libavcodec/libvorbis.c | 224 |
1 files changed, 115 insertions, 109 deletions
diff --git a/libavcodec/libvorbis.c b/libavcodec/libvorbis.c index 3226f86486..e519a1dbb8 100644 --- a/libavcodec/libvorbis.c +++ b/libavcodec/libvorbis.c @@ -20,7 +20,7 @@ /** * @file - * Ogg Vorbis codec support via libvorbisenc. + * Vorbis encoding support via libvorbisenc. * @author Mark Hills <mark@pogo.org.uk> */ @@ -35,24 +35,26 @@ #undef NDEBUG #include <assert.h> +/* Number of samples the user should send in each call. + * This value is used because it is the LCD of all possible frame sizes, so + * an output packet will always start at the same point as one of the input + * packets. + */ #define OGGVORBIS_FRAME_SIZE 64 #define BUFFER_SIZE (1024 * 64) typedef struct OggVorbisContext { - AVClass *av_class; - vorbis_info vi; - vorbis_dsp_state vd; - vorbis_block vb; - uint8_t buffer[BUFFER_SIZE]; - int buffer_index; - int eof; - - /* decoder */ - vorbis_comment vc; - ogg_packet op; - - double iblock; + AVClass *av_class; /**< class for AVOptions */ + vorbis_info vi; /**< vorbis_info used during init */ + vorbis_dsp_state vd; /**< DSP state used for analysis */ + vorbis_block vb; /**< vorbis_block used for analysis */ + uint8_t buffer[BUFFER_SIZE]; /**< output packet buffer */ + int buffer_index; /**< current buffer position */ + int eof; /**< end-of-file flag */ + vorbis_comment vc; /**< VorbisComment info */ + ogg_packet op; /**< ogg packet */ + double iblock; /**< impulse block bias option */ } OggVorbisContext; static const AVOption options[] = { @@ -61,6 +63,7 @@ static const AVOption options[] = { }; static const AVClass class = { "libvorbis", av_default_item_name, options, LIBAVUTIL_VERSION_INT }; + static int vorbis_error_to_averror(int ov_err) { switch (ov_err) { @@ -71,27 +74,31 @@ static int vorbis_error_to_averror(int ov_err) } } -static av_cold int oggvorbis_init_encoder(vorbis_info *vi, AVCodecContext *avccontext) +static av_cold int oggvorbis_init_encoder(vorbis_info *vi, + AVCodecContext *avctx) { - OggVorbisContext *context = avccontext->priv_data; + OggVorbisContext *s = avctx->priv_data; double cfreq; int ret; - if (avccontext->flags & CODEC_FLAG_QSCALE) { - /* variable bitrate */ - float q = avccontext->global_quality / (float)FF_QP2LAMBDA; - if ((ret = vorbis_encode_setup_vbr(vi, avccontext->channels, - avccontext->sample_rate, + if (avctx->flags & CODEC_FLAG_QSCALE) { + /* variable bitrate + * NOTE: we use the oggenc range of -1 to 10 for global_quality for + * user convenience, but libvorbis uses -0.1 to 1.0 + */ + float q = avctx->global_quality / (float)FF_QP2LAMBDA; + if ((ret = vorbis_encode_setup_vbr(vi, avctx->channels, + avctx->sample_rate, q / 10.0))) goto error; } else { - int minrate = avccontext->rc_min_rate > 0 ? avccontext->rc_min_rate : -1; - int maxrate = avccontext->rc_min_rate > 0 ? avccontext->rc_max_rate : -1; + int minrate = avctx->rc_min_rate > 0 ? avctx->rc_min_rate : -1; + int maxrate = avctx->rc_min_rate > 0 ? avctx->rc_max_rate : -1; - /* constant bitrate */ - if ((ret = vorbis_encode_setup_managed(vi, avccontext->channels, - avccontext->sample_rate, minrate, - avccontext->bit_rate, maxrate))) + /* average bitrate */ + if ((ret = vorbis_encode_setup_managed(vi, avctx->channels, + avctx->sample_rate, minrate, + avctx->bit_rate, maxrate))) goto error; /* variable bitrate by estimate, disable slow rate management */ @@ -101,14 +108,15 @@ static av_cold int oggvorbis_init_encoder(vorbis_info *vi, AVCodecContext *avcco } /* cutoff frequency */ - if (avccontext->cutoff > 0) { - cfreq = avccontext->cutoff / 1000.0; + if (avctx->cutoff > 0) { + cfreq = avctx->cutoff / 1000.0; if ((ret = vorbis_encode_ctl(vi, OV_ECTL_LOWPASS_SET, &cfreq))) goto error; } - if (context->iblock) { - if ((ret = vorbis_encode_ctl(vi, OV_ECTL_IBLOCK_SET, &context->iblock))) + /* impulse block bias */ + if (s->iblock) { + if ((ret = vorbis_encode_ctl(vi, OV_ECTL_IBLOCK_SET, &s->iblock))) goto error; } @@ -126,59 +134,59 @@ static int xiph_len(int l) return 1 + l / 255 + l; } -static av_cold int oggvorbis_encode_close(AVCodecContext *avccontext) +static av_cold int oggvorbis_encode_close(AVCodecContext *avctx) { - OggVorbisContext *context = avccontext->priv_data; -/* ogg_packet op ; */ + OggVorbisContext *s = avctx->priv_data; - vorbis_analysis_wrote(&context->vd, 0); /* notify vorbisenc this is EOF */ + /* notify vorbisenc this is EOF */ + vorbis_analysis_wrote(&s->vd, 0); - vorbis_block_clear(&context->vb); - vorbis_dsp_clear(&context->vd); - vorbis_info_clear(&context->vi); + vorbis_block_clear(&s->vb); + vorbis_dsp_clear(&s->vd); + vorbis_info_clear(&s->vi); - av_freep(&avccontext->coded_frame); - av_freep(&avccontext->extradata); + av_freep(&avctx->coded_frame); + av_freep(&avctx->extradata); return 0; } -static av_cold int oggvorbis_encode_init(AVCodecContext *avccontext) +static av_cold int oggvorbis_encode_init(AVCodecContext *avctx) { - OggVorbisContext *context = avccontext->priv_data; + OggVorbisContext *s = avctx->priv_data; ogg_packet header, header_comm, header_code; uint8_t *p; unsigned int offset; int ret; - vorbis_info_init(&context->vi); - if ((ret = oggvorbis_init_encoder(&context->vi, avccontext))) { - av_log(avccontext, AV_LOG_ERROR, "oggvorbis_encode_init: init_encoder failed\n"); + vorbis_info_init(&s->vi); + if ((ret = oggvorbis_init_encoder(&s->vi, avctx))) { + av_log(avctx, AV_LOG_ERROR, "oggvorbis_encode_init: init_encoder failed\n"); goto error; } - if ((ret = vorbis_analysis_init(&context->vd, &context->vi))) { + if ((ret = vorbis_analysis_init(&s->vd, &s->vi))) { ret = vorbis_error_to_averror(ret); goto error; } - if ((ret = vorbis_block_init(&context->vd, &context->vb))) { + if ((ret = vorbis_block_init(&s->vd, &s->vb))) { ret = vorbis_error_to_averror(ret); goto error; } - vorbis_comment_init(&context->vc); - vorbis_comment_add_tag(&context->vc, "encoder", LIBAVCODEC_IDENT); + vorbis_comment_init(&s->vc); + vorbis_comment_add_tag(&s->vc, "encoder", LIBAVCODEC_IDENT); - if ((ret = vorbis_analysis_headerout(&context->vd, &context->vc, &header, - &header_comm, &header_code))) { + if ((ret = vorbis_analysis_headerout(&s->vd, &s->vc, &header, &header_comm, + &header_code))) { ret = vorbis_error_to_averror(ret); goto error; } - avccontext->extradata_size = - 1 + xiph_len(header.bytes) + xiph_len(header_comm.bytes) + - header_code.bytes; - p = avccontext->extradata = - av_malloc(avccontext->extradata_size + FF_INPUT_BUFFER_PADDING_SIZE); + avctx->extradata_size = 1 + xiph_len(header.bytes) + + xiph_len(header_comm.bytes) + + header_code.bytes; + p = avctx->extradata = av_malloc(avctx->extradata_size + + FF_INPUT_BUFFER_PADDING_SIZE); if (!p) { ret = AVERROR(ENOMEM); goto error; @@ -193,100 +201,97 @@ static av_cold int oggvorbis_encode_init(AVCodecContext *avccontext) offset += header_comm.bytes; memcpy(&p[offset], header_code.packet, header_code.bytes); offset += header_code.bytes; - assert(offset == avccontext->extradata_size); + assert(offset == avctx->extradata_size); -#if 0 - vorbis_block_clear(&context->vb); - vorbis_dsp_clear(&context->vd); - vorbis_info_clear(&context->vi); -#endif - vorbis_comment_clear(&context->vc); + vorbis_comment_clear(&s->vc); - avccontext->frame_size = OGGVORBIS_FRAME_SIZE; + avctx->frame_size = OGGVORBIS_FRAME_SIZE; - avccontext->coded_frame = avcodec_alloc_frame(); - if (!avccontext->coded_frame) { + avctx->coded_frame = avcodec_alloc_frame(); + if (!avctx->coded_frame) { ret = AVERROR(ENOMEM); goto error; } return 0; error: - oggvorbis_encode_close(avccontext); + oggvorbis_encode_close(avctx); return ret; } -static int oggvorbis_encode_frame(AVCodecContext *avccontext, - unsigned char *packets, +static int oggvorbis_encode_frame(AVCodecContext *avctx, unsigned char *packets, int buf_size, void *data) { - OggVorbisContext *context = avccontext->priv_data; + OggVorbisContext *s = avctx->priv_data; ogg_packet op; signed short *audio = data; - int l; + int pkt_size; + /* send samples to libvorbis */ if (data) { - const int samples = avccontext->frame_size; + const int samples = avctx->frame_size; float **buffer; - int c, channels = context->vi.channels; + int c, channels = s->vi.channels; - buffer = vorbis_analysis_buffer(&context->vd, samples); + buffer = vorbis_analysis_buffer(&s->vd, samples); for (c = 0; c < channels; c++) { + int i; int co = (channels > 8) ? c : ff_vorbis_encoding_channel_layout_offsets[channels - 1][c]; - for (l = 0; l < samples; l++) - buffer[c][l] = audio[l * channels + co] / 32768.f; + for (i = 0; i < samples; i++) + buffer[c][i] = audio[i * channels + co] / 32768.f; } - vorbis_analysis_wrote(&context->vd, samples); + vorbis_analysis_wrote(&s->vd, samples); } else { - if (!context->eof) - vorbis_analysis_wrote(&context->vd, 0); - context->eof = 1; + if (!s->eof) + vorbis_analysis_wrote(&s->vd, 0); + s->eof = 1; } - while (vorbis_analysis_blockout(&context->vd, &context->vb) == 1) { - vorbis_analysis(&context->vb, NULL); - vorbis_bitrate_addblock(&context->vb); + /* retrieve available packets from libvorbis */ + while (vorbis_analysis_blockout(&s->vd, &s->vb) == 1) { + vorbis_analysis(&s->vb, NULL); + vorbis_bitrate_addblock(&s->vb); - while (vorbis_bitrate_flushpacket(&context->vd, &op)) { + /* add any available packets to the output packet buffer */ + while (vorbis_bitrate_flushpacket(&s->vd, &op)) { /* i'd love to say the following line is a hack, but sadly it's * not, apparently the end of stream decision is in libogg. */ if (op.bytes == 1 && op.e_o_s) continue; - if (context->buffer_index + sizeof(ogg_packet) + op.bytes > BUFFER_SIZE) { - av_log(avccontext, AV_LOG_ERROR, "libvorbis: buffer overflow."); + if (s->buffer_index + sizeof(ogg_packet) + op.bytes > BUFFER_SIZE) { + av_log(avctx, AV_LOG_ERROR, "libvorbis: buffer overflow."); return -1; } - memcpy(context->buffer + context->buffer_index, &op, sizeof(ogg_packet)); - context->buffer_index += sizeof(ogg_packet); - memcpy(context->buffer + context->buffer_index, op.packet, op.bytes); - context->buffer_index += op.bytes; -// av_log(avccontext, AV_LOG_DEBUG, "e%d / %d\n", context->buffer_index, op.bytes); + memcpy(s->buffer + s->buffer_index, &op, sizeof(ogg_packet)); + s->buffer_index += sizeof(ogg_packet); + memcpy(s->buffer + s->buffer_index, op.packet, op.bytes); + s->buffer_index += op.bytes; } } - l = 0; - if (context->buffer_index) { - ogg_packet *op2 = (ogg_packet *)context->buffer; - op2->packet = context->buffer + sizeof(ogg_packet); - - l = op2->bytes; - avccontext->coded_frame->pts = ff_samples_to_time_base(avccontext, - op2->granulepos); - //FIXME we should reorder the user supplied pts and not assume that they are spaced by 1/sample_rate - - if (l > buf_size) { - av_log(avccontext, AV_LOG_ERROR, "libvorbis: buffer overflow."); + /* output then next packet from the output buffer, if available */ + pkt_size = 0; + if (s->buffer_index) { + ogg_packet *op2 = (ogg_packet *)s->buffer; + op2->packet = s->buffer + sizeof(ogg_packet); + + pkt_size = op2->bytes; + // FIXME: we should use the user-supplied pts and duration + avctx->coded_frame->pts = ff_samples_to_time_base(avctx, + op2->granulepos); + if (pkt_size > buf_size) { + av_log(avctx, AV_LOG_ERROR, "libvorbis: buffer overflow."); return -1; } - memcpy(packets, op2->packet, l); - context->buffer_index -= l + sizeof(ogg_packet); - memmove(context->buffer, context->buffer + l + sizeof(ogg_packet), context->buffer_index); -// av_log(avccontext, AV_LOG_DEBUG, "E%d\n", l); + memcpy(packets, op2->packet, pkt_size); + s->buffer_index -= pkt_size + sizeof(ogg_packet); + memmove(s->buffer, s->buffer + pkt_size + sizeof(ogg_packet), + s->buffer_index); } - return l; + return pkt_size; } AVCodec ff_libvorbis_encoder = { @@ -298,7 +303,8 @@ AVCodec ff_libvorbis_encoder = { .encode = oggvorbis_encode_frame, .close = oggvorbis_encode_close, .capabilities = CODEC_CAP_DELAY, - .sample_fmts = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_NONE }, + .sample_fmts = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_S16, + AV_SAMPLE_FMT_NONE }, .long_name = NULL_IF_CONFIG_SMALL("libvorbis Vorbis"), .priv_class = &class, }; |