X-Git-Url: https://git.sesse.net/?a=blobdiff_plain;f=ffmpeg.c;h=ac428d5d1f0c120ed7913a66fb3052e75d4dab96;hb=aca0f78114a4f74e1b59d706c11c1bd5ff6f92d9;hp=cc1b32fc7995e0962d43270578003878cb06e68e;hpb=53a3748ed23136615e488dc463b91aa57c0e9ec6;p=ffmpeg

diff --git a/ffmpeg.c b/ffmpeg.c
index cc1b32fc799..ac428d5d1f0 100644
--- a/ffmpeg.c
+++ b/ffmpeg.c
@@ -249,6 +249,12 @@ typedef struct InputStream {
     int      resample_channels;
     uint64_t resample_channel_layout;
 
+    struct sub2video {
+        int64_t last_pts;
+        AVFilterBufferRef *ref;
+        int w, h;
+    } sub2video;
+
     /* a pool of free buffers for decoded data */
     FrameBuffer *buffer_pool;
     int dr1;
@@ -504,6 +510,155 @@ static void update_benchmark(const char *fmt, ...)
     }
 }
 
+/* sub2video hack:
+   Convert subtitles to video with alpha to insert them in filter graphs.
+   This is a temporary solution until libavfilter gets real subtitles support.
+ */
+
+
+static int sub2video_prepare(InputStream *ist)
+{
+    AVFormatContext *avf = input_files[ist->file_index]->ctx;
+    int i, ret, w, h;
+    uint8_t *image[4];
+    int linesize[4];
+
+    /* Compute the size of the canvas for the subtitles stream.
+       If the subtitles codec has set a size, use it. Otherwise use the
+       maximum dimensions of the video streams in the same file. */
+    w = ist->st->codec->width;
+    h = ist->st->codec->height;
+    if (!(w && h)) {
+        for (i = 0; i < avf->nb_streams; i++) {
+            if (avf->streams[i]->codec->codec_type == AVMEDIA_TYPE_VIDEO) {
+                w = FFMAX(w, avf->streams[i]->codec->width);
+                h = FFMAX(h, avf->streams[i]->codec->height);
+            }
+        }
+        if (!(w && h)) {
+            w = FFMAX(w, 720);
+            h = FFMAX(h, 576);
+        }
+        av_log(avf, AV_LOG_INFO, "sub2video: using %dx%d canvas\n", w, h);
+    }
+    ist->sub2video.w = ist->st->codec->width  = w;
+    ist->sub2video.h = ist->st->codec->height = h;
+
+    /* rectangles are PIX_FMT_PAL8, but we have no guarantee that the
+       palettes for all rectangles are identical or compatible */
+    ist->st->codec->pix_fmt = PIX_FMT_RGB32;
+
+    ret = av_image_alloc(image, linesize, w, h, PIX_FMT_RGB32, 32);
+    if (ret < 0)
+        return ret;
+    memset(image[0], 0, h * linesize[0]);
+    ist->sub2video.ref = avfilter_get_video_buffer_ref_from_arrays(
+            image, linesize, AV_PERM_READ | AV_PERM_PRESERVE,
+            w, h, PIX_FMT_RGB32);
+    if (!ist->sub2video.ref) {
+        av_free(image[0]);
+        return AVERROR(ENOMEM);
+    }
+    return 0;
+}
+
+static void sub2video_copy_rect(uint8_t *dst, int dst_linesize, int w, int h,
+                                AVSubtitleRect *r)
+{
+    uint32_t *pal, *dst2;
+    uint8_t *src, *src2;
+    int x, y;
+
+    if (r->type != SUBTITLE_BITMAP) {
+        av_log(NULL, AV_LOG_WARNING, "sub2video: non-bitmap subtitle\n");
+        return;
+    }
+    if (r->x < 0 || r->x + r->w > w || r->y < 0 || r->y + r->h > h) {
+        av_log(NULL, AV_LOG_WARNING, "sub2video: rectangle overflowing\n");
+        return;
+    }
+
+    dst += r->y * dst_linesize + r->x * 4;
+    src = r->pict.data[0];
+    pal = (uint32_t *)r->pict.data[1];
+    for (y = 0; y < r->h; y++) {
+        dst2 = (uint32_t *)dst;
+        src2 = src;
+        for (x = 0; x < r->w; x++)
+            *(dst2++) = pal[*(src2++)];
+        dst += dst_linesize;
+        src += r->pict.linesize[0];
+    }
+}
+
+static void sub2video_push_ref(InputStream *ist, int64_t pts)
+{
+    AVFilterBufferRef *ref = ist->sub2video.ref;
+    int i;
+
+    ist->sub2video.last_pts = ref->pts = pts;
+    for (i = 0; i < ist->nb_filters; i++)
+        av_buffersrc_add_ref(ist->filters[i]->filter,
+                             avfilter_ref_buffer(ref, ~0),
+                             AV_BUFFERSRC_FLAG_NO_CHECK_FORMAT |
+                             AV_BUFFERSRC_FLAG_NO_COPY);
+}
+
+static void sub2video_update(InputStream *ist, AVSubtitle *sub, int64_t pts)
+{
+    int w = ist->sub2video.w, h = ist->sub2video.h;
+    AVFilterBufferRef *ref = ist->sub2video.ref;
+    int8_t *dst;
+    int     dst_linesize;
+    int i;
+
+    if (!ref)
+        return;
+    dst          = ref->data    [0];
+    dst_linesize = ref->linesize[0];
+    memset(dst, 0, h * dst_linesize);
+    for (i = 0; i < sub->num_rects; i++)
+        sub2video_copy_rect(dst, dst_linesize, w, h, sub->rects[i]);
+    sub2video_push_ref(ist, pts);
+}
+
+static void sub2video_heartbeat(InputStream *ist, int64_t pts)
+{
+    InputFile *infile = input_files[ist->file_index];
+    int i, j, nb_reqs;
+    int64_t pts2;
+
+    /* When a frame is read from a file, examine all sub2video streams in
+       the same file and send the sub2video frame again. Otherwise, decoded
+       video frames could be accumulating in the filter graph while a filter
+       (possibly overlay) is desperately waiting for a subtitle frame. */
+    for (i = 0; i < infile->nb_streams; i++) {
+        InputStream *ist2 = input_streams[infile->ist_index + i];
+        if (!ist2->sub2video.ref)
+            continue;
+        /* subtitles seem to be usually muxed ahead of other streams;
+           if not, substracting a larger time here is necessary */
+        pts2 = av_rescale_q(pts, ist->st->time_base, ist2->st->time_base) - 1;
+        /* do not send the heartbeat frame if the subtitle is already ahead */
+        if (pts2 <= ist2->sub2video.last_pts)
+            continue;
+        for (j = 0, nb_reqs = 0; j < ist2->nb_filters; j++)
+            nb_reqs += av_buffersrc_get_nb_failed_requests(ist2->filters[j]->filter);
+        if (nb_reqs)
+            sub2video_push_ref(ist2, pts2);
+    }
+}
+
+static void sub2video_flush(InputStream *ist)
+{
+    int i;
+
+    for (i = 0; i < ist->nb_filters; i++)
+        av_buffersrc_add_ref(ist->filters[i]->filter, NULL, 0);
+}
+
+/* end of sub2video hack */
+
 static void reset_options(OptionsContext *o, int is_input)
 {
     const OptionDef *po = options;
@@ -560,9 +715,9 @@ static enum PixelFormat choose_pixel_fmt(AVStream *st, AVCodec *codec, enum Pixe
         int has_alpha= av_pix_fmt_descriptors[target].nb_components % 2 == 0;
         enum PixelFormat best= PIX_FMT_NONE;
         if (st->codec->strict_std_compliance <= FF_COMPLIANCE_UNOFFICIAL) {
-            if (st->codec->codec_id == CODEC_ID_MJPEG) {
+            if (st->codec->codec_id == AV_CODEC_ID_MJPEG) {
                 p = (const enum PixelFormat[]) { PIX_FMT_YUVJ420P, PIX_FMT_YUVJ422P, PIX_FMT_YUV420P, PIX_FMT_YUV422P, PIX_FMT_NONE };
-            } else if (st->codec->codec_id == CODEC_ID_LJPEG) {
+            } else if (st->codec->codec_id == AV_CODEC_ID_LJPEG) {
                 p = (const enum PixelFormat[]) { PIX_FMT_YUVJ420P, PIX_FMT_YUVJ422P, PIX_FMT_YUVJ444P, PIX_FMT_YUV420P,
                                                  PIX_FMT_YUV422P, PIX_FMT_YUV444P, PIX_FMT_BGRA, PIX_FMT_NONE };
             }
@@ -608,9 +763,9 @@ static char *choose_pix_fmts(OutputStream *ost)
 
         p = ost->enc->pix_fmts;
         if (ost->st->codec->strict_std_compliance <= FF_COMPLIANCE_UNOFFICIAL) {
-            if (ost->st->codec->codec_id == CODEC_ID_MJPEG) {
+            if (ost->st->codec->codec_id == AV_CODEC_ID_MJPEG) {
                 p = (const enum PixelFormat[]) { PIX_FMT_YUVJ420P, PIX_FMT_YUVJ422P, PIX_FMT_YUV420P, PIX_FMT_YUV422P, PIX_FMT_NONE };
-            } else if (ost->st->codec->codec_id == CODEC_ID_LJPEG) {
+            } else if (ost->st->codec->codec_id == AV_CODEC_ID_LJPEG) {
                 p = (const enum PixelFormat[]) { PIX_FMT_YUVJ420P, PIX_FMT_YUVJ422P, PIX_FMT_YUVJ444P, PIX_FMT_YUV420P,
                                                     PIX_FMT_YUV422P, PIX_FMT_YUV444P, PIX_FMT_BGRA, PIX_FMT_NONE };
             }
@@ -745,7 +900,10 @@ static void init_input_filter(FilterGraph *fg, AVFilterInOut *in)
         s = input_files[file_idx]->ctx;
 
         for (i = 0; i < s->nb_streams; i++) {
-            if (s->streams[i]->codec->codec_type != type)
+            enum AVMediaType stream_type = s->streams[i]->codec->codec_type;
+            if (stream_type != type &&
+                !(stream_type == AVMEDIA_TYPE_SUBTITLE &&
+                  type == AVMEDIA_TYPE_VIDEO /* sub2video hack */))
                 continue;
             if (check_stream_specifier(s, s->streams[i], *p == ':' ? p + 1 : p) == 1) {
                 st = s->streams[i];
@@ -1014,8 +1172,7 @@ static int configure_input_video_filter(FilterGraph *fg, InputFilter *ifilter,
     AVFilterContext *first_filter = in->filter_ctx;
     AVFilter *filter = avfilter_get_by_name("buffer");
     InputStream *ist = ifilter->ist;
-    AVRational tb = ist->framerate.num ? (AVRational){ist->framerate.den,
-                                                      ist->framerate.num} :
+    AVRational tb = ist->framerate.num ? av_inv_q(ist->framerate) :
                                          ist->st->time_base;
     AVRational fr = ist->framerate.num ? ist->framerate :
                                          ist->st->r_frame_rate;
@@ -1025,6 +1182,12 @@ static int configure_input_video_filter(FilterGraph *fg, InputFilter *ifilter,
     int pad_idx = in->pad_idx;
     int ret;
 
+    if (ist->st->codec->codec_type == AVMEDIA_TYPE_SUBTITLE) {
+        ret = sub2video_prepare(ist);
+        if (ret < 0)
+            return ret;
+    }
+
     sar = ist->st->sample_aspect_ratio.num ?
           ist->st->sample_aspect_ratio :
           ist->st->codec->sample_aspect_ratio;
@@ -1413,6 +1576,7 @@ void av_noreturn exit_program(int ret)
         av_freep(&input_streams[i]->decoded_frame);
         av_dict_free(&input_streams[i]->opts);
         free_buffer_pool(&input_streams[i]->buffer_pool);
+        avfilter_unref_bufferp(&input_streams[i]->sub2video.ref);
         av_freep(&input_streams[i]->filters);
         av_freep(&input_streams[i]);
     }
@@ -1678,7 +1842,7 @@ static void do_subtitle_out(AVFormatContext *s,
     /* Note: DVB subtitle need one packet to draw them and one other
        packet to clear them */
     /* XXX: signal it in the codec context ? */
-    if (enc->codec_id == CODEC_ID_DVB_SUBTITLE)
+    if (enc->codec_id == AV_CODEC_ID_DVB_SUBTITLE)
         nb = 2;
     else
         nb = 1;
@@ -1708,7 +1872,7 @@ static void do_subtitle_out(AVFormatContext *s,
         pkt.size = subtitle_out_size;
         pkt.pts  = av_rescale_q(sub->pts, AV_TIME_BASE_Q, ost->st->time_base);
         pkt.duration = av_rescale_q(sub->end_display_time, (AVRational){ 1, 1000 }, ost->st->time_base);
-        if (enc->codec_id == CODEC_ID_DVB_SUBTITLE) {
+        if (enc->codec_id == AV_CODEC_ID_DVB_SUBTITLE) {
             /* XXX: the pts correction is handled here. Maybe handling
                it in the codec would be better */
             if (i == 0)
@@ -1800,7 +1964,7 @@ static void do_video_out(AVFormatContext *s,
         return;
 
     if (s->oformat->flags & AVFMT_RAWPICTURE &&
-        enc->codec->id == CODEC_ID_RAWVIDEO) {
+        enc->codec->id == AV_CODEC_ID_RAWVIDEO) {
         /* raw pictures are written as AVPicture structure to
            avoid any copies. We support temporarily the older
            method. */
@@ -2216,7 +2380,7 @@ static void flush_encoders(void)
 
         if (ost->st->codec->codec_type == AVMEDIA_TYPE_AUDIO && enc->frame_size <= 1)
             continue;
-        if (ost->st->codec->codec_type == AVMEDIA_TYPE_VIDEO && (os->oformat->flags & AVFMT_RAWPICTURE) && enc->codec->id == CODEC_ID_RAWVIDEO)
+        if (ost->st->codec->codec_type == AVMEDIA_TYPE_VIDEO && (os->oformat->flags & AVFMT_RAWPICTURE) && enc->codec->id == AV_CODEC_ID_RAWVIDEO)
             continue;
 
         for (;;) {
@@ -2335,10 +2499,10 @@ static void do_streamcopy(InputStream *ist, OutputStream *ost, const AVPacket *p
     opkt.flags    = pkt->flags;
 
     // FIXME remove the following 2 lines they shall be replaced by the bitstream filters
-    if (  ost->st->codec->codec_id != CODEC_ID_H264
-       && ost->st->codec->codec_id != CODEC_ID_MPEG1VIDEO
-       && ost->st->codec->codec_id != CODEC_ID_MPEG2VIDEO
-       && ost->st->codec->codec_id != CODEC_ID_VC1
+    if (  ost->st->codec->codec_id != AV_CODEC_ID_H264
+       && ost->st->codec->codec_id != AV_CODEC_ID_MPEG1VIDEO
+       && ost->st->codec->codec_id != AV_CODEC_ID_MPEG2VIDEO
+       && ost->st->codec->codec_id != AV_CODEC_ID_VC1
        ) {
         if (av_parser_change(ist->st->parser, ost->st->codec, &opkt.data, &opkt.size, pkt->data, pkt->size, pkt->flags & AV_PKT_FLAG_KEY))
             opkt.destruct = av_destruct_packet;
@@ -2636,13 +2800,16 @@ static int transcode_subtitles(InputStream *ist, AVPacket *pkt, int *got_output)
     AVSubtitle subtitle;
     int i, ret = avcodec_decode_subtitle2(ist->st->codec,
                                           &subtitle, got_output, pkt);
-    if (ret < 0)
-        return ret;
-    if (!*got_output)
+    if (ret < 0 || !*got_output) {
+        if (!pkt->size)
+            sub2video_flush(ist);
         return ret;
+    }
 
     rate_emu_sleep(ist);
 
+    sub2video_update(ist, &subtitle, pkt->pts);
+
     for (i = 0; i < nb_output_streams; i++) {
         OutputStream *ost = output_streams[i];
 
@@ -3021,7 +3188,7 @@ static int transcode_init(void)
             }
 
             if(ost->frame_rate.num)
-                codec->time_base = (AVRational){ost->frame_rate.den, ost->frame_rate.num};
+                codec->time_base = av_inv_q(ost->frame_rate);
 
             av_reduce(&codec->time_base.num, &codec->time_base.den,
                         codec->time_base.num, codec->time_base.den, INT_MAX);
@@ -3038,9 +3205,9 @@ static int transcode_init(void)
                 codec->frame_size         = icodec->frame_size;
                 codec->audio_service_type = icodec->audio_service_type;
                 codec->block_align        = icodec->block_align;
-                if((codec->block_align == 1 || codec->block_align == 1152) && codec->codec_id == CODEC_ID_MP3)
+                if((codec->block_align == 1 || codec->block_align == 1152) && codec->codec_id == AV_CODEC_ID_MP3)
                     codec->block_align= 0;
-                if(codec->codec_id == CODEC_ID_AC3)
+                if(codec->codec_id == AV_CODEC_ID_AC3)
                     codec->block_align= 0;
                 break;
             case AVMEDIA_TYPE_VIDEO:
@@ -3114,7 +3281,7 @@ static int transcode_init(void)
                 codec->time_base      = (AVRational){ 1, codec->sample_rate };
                 break;
             case AVMEDIA_TYPE_VIDEO:
-                codec->time_base = (AVRational){ost->frame_rate.den, ost->frame_rate.num};
+                codec->time_base = av_inv_q(ost->frame_rate);
                 if (ost->filter && !(codec->time_base.num && codec->time_base.den))
                     codec->time_base = ost->filter->filter->inputs[0]->time_base;
                 if (   av_q2d(codec->time_base) < 0.001 && video_sync_method != VSYNC_PASSTHROUGH
@@ -3198,7 +3365,8 @@ static int transcode_init(void)
             if ((ist = get_input_stream(ost)))
                 dec = ist->st->codec;
             if (dec && dec->subtitle_header) {
-                ost->st->codec->subtitle_header = av_malloc(dec->subtitle_header_size);
+                /* ASS code assumes this buffer is null terminated so add extra byte. */
+                ost->st->codec->subtitle_header = av_mallocz(dec->subtitle_header_size + 1);
                 if (!ost->st->codec->subtitle_header) {
                     ret = AVERROR(ENOMEM);
                     goto dump_format;
@@ -3847,6 +4015,8 @@ static int transcode(void)
             }
         }
 
+        sub2video_heartbeat(ist, pkt.pts);
+
         // fprintf(stderr,"read #%d.%d size=%d\n", ist->file_index, ist->st->index, pkt.size);
         if ((ret = output_packet(ist, &pkt)) < 0 ||
             ((ret = poll_filters()) < 0 && ret != AVERROR_EOF)) {
@@ -4514,11 +4684,11 @@ static int opt_input_file(OptionsContext *o, const char *opt, const char *filena
         av_dict_set(&format_opts, "pixel_format", o->frame_pix_fmts[o->nb_frame_pix_fmts - 1].u.str, 0);
 
     ic->video_codec_id   = video_codec_name ?
-        find_codec_or_die(video_codec_name   , AVMEDIA_TYPE_VIDEO   , 0)->id : CODEC_ID_NONE;
+        find_codec_or_die(video_codec_name   , AVMEDIA_TYPE_VIDEO   , 0)->id : AV_CODEC_ID_NONE;
     ic->audio_codec_id   = audio_codec_name ?
-        find_codec_or_die(audio_codec_name   , AVMEDIA_TYPE_AUDIO   , 0)->id : CODEC_ID_NONE;
+        find_codec_or_die(audio_codec_name   , AVMEDIA_TYPE_AUDIO   , 0)->id : AV_CODEC_ID_NONE;
     ic->subtitle_codec_id= subtitle_codec_name ?
-        find_codec_or_die(subtitle_codec_name, AVMEDIA_TYPE_SUBTITLE, 0)->id : CODEC_ID_NONE;
+        find_codec_or_die(subtitle_codec_name, AVMEDIA_TYPE_SUBTITLE, 0)->id : AV_CODEC_ID_NONE;
     ic->flags |= AVFMT_FLAG_NONBLOCK;
     ic->interrupt_callback = int_cb;
 
@@ -5228,7 +5398,7 @@ static void opt_output_file(void *optctx, const char *filename)
         /* pick the "best" stream of each type */
 
         /* video: highest resolution */
-        if (!o->video_disable && oc->oformat->video_codec != CODEC_ID_NONE) {
+        if (!o->video_disable && oc->oformat->video_codec != AV_CODEC_ID_NONE) {
             int area = 0, idx = -1;
             for (i = 0; i < nb_input_streams; i++) {
                 ist = input_streams[i];
@@ -5243,7 +5413,7 @@ static void opt_output_file(void *optctx, const char *filename)
         }
 
         /* audio: most channels */
-        if (!o->audio_disable && oc->oformat->audio_codec != CODEC_ID_NONE) {
+        if (!o->audio_disable && oc->oformat->audio_codec != AV_CODEC_ID_NONE) {
             int channels = 0, idx = -1;
             for (i = 0; i < nb_input_streams; i++) {
                 ist = input_streams[i];
@@ -5258,7 +5428,7 @@ static void opt_output_file(void *optctx, const char *filename)
         }
 
         /* subtitles: pick first */
-        if (!o->subtitle_disable && (oc->oformat->subtitle_codec != CODEC_ID_NONE || subtitle_codec_name)) {
+        if (!o->subtitle_disable && (oc->oformat->subtitle_codec != AV_CODEC_ID_NONE || subtitle_codec_name)) {
             for (i = 0; i < nb_input_streams; i++)
                 if (input_streams[i]->st->codec->codec_type == AVMEDIA_TYPE_SUBTITLE) {
                     new_subtitle_stream(o, oc, i);