Remove all use of stringstream in an attempt to fix
[dcpomatic.git] / src / lib / ffmpeg_decoder.cc
index e90c33c80a848bca3c3faecbd7032c1df72d0526..ace54f7194a3eaffd0cd13f984e80af97e0be12e 100644 (file)
@@ -1,19 +1,20 @@
 /*
-    Copyright (C) 2012-2014 Carl Hetherington <cth@carlh.net>
+    Copyright (C) 2012-2016 Carl Hetherington <cth@carlh.net>
 
-    This program is free software; you can redistribute it and/or modify
+    This file is part of DCP-o-matic.
+
+    DCP-o-matic is free software; you can redistribute it and/or modify
     it under the terms of the GNU General Public License as published by
     the Free Software Foundation; either version 2 of the License, or
     (at your option) any later version.
 
-    This program is distributed in the hope that it will be useful,
+    DCP-o-matic is distributed in the hope that it will be useful,
     but WITHOUT ANY WARRANTY; without even the implied warranty of
     MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
     GNU General Public License for more details.
 
     You should have received a copy of the GNU General Public License
-    along with this program; if not, write to the Free Software
-    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+    along with DCP-o-matic.  If not, see <http://www.gnu.org/licenses/>.
 
 */
 
  *  @brief A decoder using FFmpeg to decode content.
  */
 
-#include <stdexcept>
-#include <vector>
-#include <iomanip>
-#include <iostream>
-#include <stdint.h>
-#include <sndfile.h>
-extern "C" {
-#include <libavcodec/avcodec.h>
-#include <libavformat/avformat.h>
-}
 #include "filter.h"
 #include "exceptions.h"
 #include "image.h"
 #include "util.h"
 #include "log.h"
 #include "ffmpeg_decoder.h"
+#include "subtitle_decoder.h"
 #include "ffmpeg_audio_stream.h"
 #include "ffmpeg_subtitle_stream.h"
-#include "filter_graph.h"
+#include "video_filter_graph.h"
 #include "audio_buffers.h"
 #include "ffmpeg_content.h"
 #include "raw_image_proxy.h"
+#include "video_decoder.h"
 #include "film.h"
-#include "timer.h"
+#include "audio_decoder.h"
+#include "compose.hpp"
+#include "subtitle_content.h"
+#include "audio_content.h"
+#include <dcp/subtitle_string.h>
+#include <sub/ssa_reader.h>
+#include <sub/subtitle.h>
+#include <sub/collect.h>
+extern "C" {
+#include <libavcodec/avcodec.h>
+#include <libavformat/avformat.h>
+}
+#include <boost/foreach.hpp>
+#include <boost/algorithm/string.hpp>
+#include <vector>
+#include <iomanip>
+#include <iostream>
+#include <stdint.h>
 
 #include "i18n.h"
 
-#define LOG_GENERAL(...) _video_content->film()->log()->log (String::compose (__VA_ARGS__), Log::TYPE_GENERAL);
-#define LOG_ERROR(...) _video_content->film()->log()->log (String::compose (__VA_ARGS__), Log::TYPE_ERROR);
-#define LOG_WARNING_NC(...) _video_content->film()->log()->log (__VA_ARGS__, Log::TYPE_WARNING);
-#define LOG_WARNING(...) _video_content->film()->log()->log (String::compose (__VA_ARGS__), Log::TYPE_WARNING);
+#define LOG_GENERAL(...) _log->log (String::compose (__VA_ARGS__), LogEntry::TYPE_GENERAL);
+#define LOG_ERROR(...) _log->log (String::compose (__VA_ARGS__), LogEntry::TYPE_ERROR);
+#define LOG_WARNING_NC(...) _log->log (__VA_ARGS__, LogEntry::TYPE_WARNING);
+#define LOG_WARNING(...) _log->log (String::compose (__VA_ARGS__), LogEntry::TYPE_WARNING);
 
 using std::cout;
 using std::string;
@@ -59,49 +69,39 @@ using std::vector;
 using std::list;
 using std::min;
 using std::pair;
-using std::make_pair;
+using std::max;
+using std::map;
 using boost::shared_ptr;
+using boost::is_any_of;
+using boost::split;
 using boost::optional;
 using boost::dynamic_pointer_cast;
 using dcp::Size;
 
 FFmpegDecoder::FFmpegDecoder (shared_ptr<const FFmpegContent> c, shared_ptr<Log> log)
-       : VideoDecoder (c)
-       , AudioDecoder (c)
-       , SubtitleDecoder (c)
-       , FFmpeg (c)
+       : FFmpeg (c)
        , _log (log)
 {
-       /* Audio and video frame PTS values may not start with 0.  We want
-          to fiddle them so that:
-
-          1.  One of them starts at time 0.
-          2.  The first video PTS value ends up on a frame boundary.
-
-          Then we remove big initial gaps in PTS and we allow our
-          insertion of black frames to work.
-
-          We will do pts_to_use = pts_from_ffmpeg + pts_offset;
-       */
-
-       bool const have_video = c->first_video();
-       bool const have_audio = c->audio_stream () && c->audio_stream()->first_audio;
-
-       /* First, make one of them start at 0 */
+       if (c->video) {
+               video.reset (new VideoDecoder (this, c, log));
+               _pts_offset = pts_offset (c->ffmpeg_audio_streams(), c->first_video(), c->active_video_frame_rate());
+       } else {
+               _pts_offset = ContentTime ();
+       }
 
-       if (have_audio && have_video) {
-               _pts_offset = - min (c->first_video().get(), c->audio_stream()->first_audio.get());
-       } else if (have_video) {
-               _pts_offset = - c->first_video().get();
-       } else if (have_audio) {
-               _pts_offset = - c->audio_stream()->first_audio.get();
+       if (c->audio) {
+               audio.reset (new AudioDecoder (this, c->audio, log));
        }
 
-       /* Now adjust both so that the video pts starts on a frame */
-       if (have_video && have_audio) {
-               ContentTime first_video = c->first_video().get() + _pts_offset;
-               ContentTime const old_first_video = first_video;
-               _pts_offset += first_video.round_up (c->video_frame_rate ()) - old_first_video;
+       if (c->subtitle) {
+               subtitle.reset (
+                       new SubtitleDecoder (
+                               this,
+                               c->subtitle,
+                               bind (&FFmpegDecoder::image_subtitles_during, this, _1, _2),
+                               bind (&FFmpegDecoder::text_subtitles_during, this, _1, _2)
+                               )
+                       );
        }
 }
 
@@ -109,31 +109,35 @@ void
 FFmpegDecoder::flush ()
 {
        /* Get any remaining frames */
-       
+
        _packet.data = 0;
        _packet.size = 0;
-       
+
        /* XXX: should we reset _packet.data and size after each *_decode_* call? */
-       
-       while (decode_video_packet ()) {}
-       
-       if (_ffmpeg_content->audio_stream()) {
+
+       while (video && decode_video_packet ()) {}
+
+       if (audio) {
                decode_audio_packet ();
-               AudioDecoder::flush ();
+               audio->flush ();
        }
 }
 
 bool
-FFmpegDecoder::pass ()
+FFmpegDecoder::pass (PassReason reason, bool accurate)
 {
        int r = av_read_frame (_format_context, &_packet);
 
-       if (r < 0) {
+       /* AVERROR_INVALIDDATA can apparently be returned sometimes even when av_read_frame
+          has pretty-much succeeded (and hence generated data which should be processed).
+          Hence it makes sense to continue here in that case.
+       */
+       if (r < 0 && r != AVERROR_INVALIDDATA) {
                if (r != AVERROR_EOF) {
                        /* Maybe we should fail here, but for now we'll just finish off instead */
                        char buf[256];
                        av_strerror (r, buf, sizeof(buf));
-                       LOG_ERROR (N_("error on av_read_frame (%1) (%2)"), buf, r);
+                       LOG_ERROR (N_("error on av_read_frame (%1) (%2)"), &buf[0], r);
                }
 
                flush ();
@@ -141,16 +145,17 @@ FFmpegDecoder::pass ()
        }
 
        int const si = _packet.stream_index;
+       shared_ptr<const FFmpegContent> fc = _ffmpeg_content;
 
-       if (si == _video_stream) {
+       if (_video_stream && si == _video_stream.get() && !video->ignore() && (accurate || reason != PASS_REASON_SUBTITLE)) {
                decode_video_packet ();
-       } else if (_ffmpeg_content->audio_stream() && _ffmpeg_content->audio_stream()->uses_index (_format_context, si)) {
-               decode_audio_packet ();
-       } else if (_ffmpeg_content->subtitle_stream() && _ffmpeg_content->subtitle_stream()->uses_index (_format_context, si)) {
+       } else if (fc->subtitle_stream() && fc->subtitle_stream()->uses_index (_format_context, si)) {
                decode_subtitle_packet ();
+       } else if (accurate || reason != PASS_REASON_SUBTITLE) {
+               decode_audio_packet ();
        }
 
-       av_free_packet (&_packet);
+       av_packet_unref (&_packet);
        return false;
 }
 
@@ -158,47 +163,51 @@ FFmpegDecoder::pass ()
  *  Only the first buffer will be used for non-planar data, otherwise there will be one per channel.
  */
 shared_ptr<AudioBuffers>
-FFmpegDecoder::deinterleave_audio (uint8_t** data, int size)
+FFmpegDecoder::deinterleave_audio (shared_ptr<FFmpegAudioStream> stream) const
 {
-       assert (_ffmpeg_content->audio_channels());
-       assert (bytes_per_audio_sample());
+       DCPOMATIC_ASSERT (bytes_per_audio_sample (stream));
 
-       /* Deinterleave and convert to float */
+       int const size = av_samples_get_buffer_size (
+               0, stream->stream(_format_context)->codec->channels, _frame->nb_samples, audio_sample_format (stream), 1
+               );
 
-       assert ((size % (bytes_per_audio_sample() * _ffmpeg_content->audio_channels())) == 0);
+       /* Deinterleave and convert to float */
 
-       int const total_samples = size / bytes_per_audio_sample();
-       int const frames = total_samples / _ffmpeg_content->audio_channels();
-       shared_ptr<AudioBuffers> audio (new AudioBuffers (_ffmpeg_content->audio_channels(), frames));
+       /* total_samples and frames will be rounded down here, so if there are stray samples at the end
+          of the block that do not form a complete sample or frame they will be dropped.
+       */
+       int const total_samples = size / bytes_per_audio_sample (stream);
+       int const frames = total_samples / stream->channels();
+       shared_ptr<AudioBuffers> audio (new AudioBuffers (stream->channels(), frames));
 
-       switch (audio_sample_format()) {
+       switch (audio_sample_format (stream)) {
        case AV_SAMPLE_FMT_U8:
        {
-               uint8_t* p = reinterpret_cast<uint8_t *> (data[0]);
+               uint8_t* p = reinterpret_cast<uint8_t *> (_frame->data[0]);
                int sample = 0;
                int channel = 0;
                for (int i = 0; i < total_samples; ++i) {
                        audio->data(channel)[sample] = float(*p++) / (1 << 23);
 
                        ++channel;
-                       if (channel == _ffmpeg_content->audio_channels()) {
+                       if (channel == stream->channels()) {
                                channel = 0;
                                ++sample;
                        }
                }
        }
        break;
-       
+
        case AV_SAMPLE_FMT_S16:
        {
-               int16_t* p = reinterpret_cast<int16_t *> (data[0]);
+               int16_t* p = reinterpret_cast<int16_t *> (_frame->data[0]);
                int sample = 0;
                int channel = 0;
                for (int i = 0; i < total_samples; ++i) {
                        audio->data(channel)[sample] = float(*p++) / (1 << 15);
 
                        ++channel;
-                       if (channel == _ffmpeg_content->audio_channels()) {
+                       if (channel == stream->channels()) {
                                channel = 0;
                                ++sample;
                        }
@@ -208,25 +217,25 @@ FFmpegDecoder::deinterleave_audio (uint8_t** data, int size)
 
        case AV_SAMPLE_FMT_S16P:
        {
-               int16_t** p = reinterpret_cast<int16_t **> (data);
-               for (int i = 0; i < _ffmpeg_content->audio_channels(); ++i) {
+               int16_t** p = reinterpret_cast<int16_t **> (_frame->data);
+               for (int i = 0; i < stream->channels(); ++i) {
                        for (int j = 0; j < frames; ++j) {
                                audio->data(i)[j] = static_cast<float>(p[i][j]) / (1 << 15);
                        }
                }
        }
        break;
-       
+
        case AV_SAMPLE_FMT_S32:
        {
-               int32_t* p = reinterpret_cast<int32_t *> (data[0]);
+               int32_t* p = reinterpret_cast<int32_t *> (_frame->data[0]);
                int sample = 0;
                int channel = 0;
                for (int i = 0; i < total_samples; ++i) {
-                       audio->data(channel)[sample] = static_cast<float>(*p++) / (1 << 31);
+                       audio->data(channel)[sample] = static_cast<float>(*p++) / 2147483648;
 
                        ++channel;
-                       if (channel == _ffmpeg_content->audio_channels()) {
+                       if (channel == stream->channels()) {
                                channel = 0;
                                ++sample;
                        }
@@ -234,61 +243,81 @@ FFmpegDecoder::deinterleave_audio (uint8_t** data, int size)
        }
        break;
 
+       case AV_SAMPLE_FMT_S32P:
+       {
+               int32_t** p = reinterpret_cast<int32_t **> (_frame->data);
+               for (int i = 0; i < stream->channels(); ++i) {
+                       for (int j = 0; j < frames; ++j) {
+                               audio->data(i)[j] = static_cast<float>(p[i][j]) / 2147483648;
+                       }
+               }
+       }
+       break;
+
        case AV_SAMPLE_FMT_FLT:
        {
-               float* p = reinterpret_cast<float*> (data[0]);
+               float* p = reinterpret_cast<float*> (_frame->data[0]);
                int sample = 0;
                int channel = 0;
                for (int i = 0; i < total_samples; ++i) {
                        audio->data(channel)[sample] = *p++;
 
                        ++channel;
-                       if (channel == _ffmpeg_content->audio_channels()) {
+                       if (channel == stream->channels()) {
                                channel = 0;
                                ++sample;
                        }
                }
        }
        break;
-               
+
        case AV_SAMPLE_FMT_FLTP:
        {
-               float** p = reinterpret_cast<float**> (data);
-               for (int i = 0; i < _ffmpeg_content->audio_channels(); ++i) {
+               float** p = reinterpret_cast<float**> (_frame->data);
+               /* Sometimes there aren't as many channels in the _frame as in the stream */
+               for (int i = 0; i < _frame->channels; ++i) {
                        memcpy (audio->data(i), p[i], frames * sizeof(float));
                }
+               for (int i = _frame->channels; i < stream->channels(); ++i) {
+                       audio->make_silent (i);
+               }
        }
        break;
 
        default:
-               throw DecodeError (String::compose (_("Unrecognised audio sample format (%1)"), static_cast<int> (audio_sample_format())));
+               throw DecodeError (String::compose (_("Unrecognised audio sample format (%1)"), static_cast<int> (audio_sample_format (stream))));
        }
 
        return audio;
 }
 
 AVSampleFormat
-FFmpegDecoder::audio_sample_format () const
+FFmpegDecoder::audio_sample_format (shared_ptr<FFmpegAudioStream> stream) const
 {
-       if (!_ffmpeg_content->audio_stream()) {
-               return (AVSampleFormat) 0;
-       }
-       
-       return audio_codec_context()->sample_fmt;
+       return stream->stream (_format_context)->codec->sample_fmt;
 }
 
 int
-FFmpegDecoder::bytes_per_audio_sample () const
+FFmpegDecoder::bytes_per_audio_sample (shared_ptr<FFmpegAudioStream> stream) const
 {
-       return av_get_bytes_per_sample (audio_sample_format ());
+       return av_get_bytes_per_sample (audio_sample_format (stream));
 }
 
 void
 FFmpegDecoder::seek (ContentTime time, bool accurate)
 {
-       VideoDecoder::seek (time, accurate);
-       AudioDecoder::seek (time, accurate);
-       
+       if (video) {
+               video->seek (time, accurate);
+       }
+
+       if (audio) {
+               audio->seek (time, accurate);
+       }
+
+       if (subtitle) {
+               subtitle->seek (time, accurate);
+       }
+
        /* If we are doing an `accurate' seek, we need to use pre-roll, as
           we don't really know what the seek will give us.
        */
@@ -299,22 +328,37 @@ FFmpegDecoder::seek (ContentTime time, bool accurate)
        /* XXX: it seems debatable whether PTS should be used here...
           http://www.mjbshaw.com/2012/04/seeking-in-ffmpeg-know-your-timestamp.html
        */
-       
-       ContentTime const u = time - _pts_offset;
-       int64_t s = u.seconds() / av_q2d (_format_context->streams[_video_stream]->time_base);
 
-       if (_ffmpeg_content->audio_stream ()) {
-               s = min (
-                       s, int64_t (u.seconds() / av_q2d (_ffmpeg_content->audio_stream()->stream(_format_context)->time_base))
-                       );
+       optional<int> stream;
+
+       if (_video_stream) {
+               stream = _video_stream;
+       } else {
+               shared_ptr<FFmpegAudioStream> s = dynamic_pointer_cast<FFmpegAudioStream> (_ffmpeg_content->audio->stream ());
+               if (s) {
+                       stream = s->index (_format_context);
+               }
        }
 
-       av_seek_frame (_format_context, _video_stream, s, 0);
+       DCPOMATIC_ASSERT (stream);
+
+       ContentTime u = time - _pts_offset;
+       if (u < ContentTime ()) {
+               u = ContentTime ();
+       }
+       av_seek_frame (
+               _format_context,
+               stream.get(),
+               u.seconds() / av_q2d (_format_context->streams[stream.get()]->time_base),
+               AVSEEK_FLAG_BACKWARD
+               );
 
-       avcodec_flush_buffers (video_codec_context());
-       if (audio_codec_context ()) {
-               avcodec_flush_buffers (audio_codec_context ());
+       if (video_codec_context ()) {
+               avcodec_flush_buffers (video_codec_context());
        }
+
+       /* XXX: should be flushing audio buffers? */
+
        if (subtitle_codec_context ()) {
                avcodec_flush_buffers (subtitle_codec_context ());
        }
@@ -326,17 +370,29 @@ FFmpegDecoder::decode_audio_packet ()
        /* Audio packets can contain multiple frames, so we may have to call avcodec_decode_audio4
           several times.
        */
-       
+
        AVPacket copy_packet = _packet;
-       
+
+       /* XXX: inefficient */
+       vector<shared_ptr<FFmpegAudioStream> > streams = ffmpeg_content()->ffmpeg_audio_streams ();
+       vector<shared_ptr<FFmpegAudioStream> >::const_iterator stream = streams.begin ();
+       while (stream != streams.end () && !(*stream)->uses_index (_format_context, copy_packet.stream_index)) {
+               ++stream;
+       }
+
+       if (stream == streams.end ()) {
+               /* The packet's stream may not be an audio one; just ignore it in this method if so */
+               return;
+       }
+
        while (copy_packet.size > 0) {
 
                int frame_finished;
-               int decode_result = avcodec_decode_audio4 (audio_codec_context(), _frame, &frame_finished, &copy_packet);
+               int decode_result = avcodec_decode_audio4 ((*stream)->stream (_format_context)->codec, _frame, &frame_finished, &copy_packet);
                if (decode_result < 0) {
                        /* avcodec_decode_audio4 can sometimes return an error even though it has decoded
                           some valid data; for example dca_subframe_footer can return AVERROR_INVALIDDATA
-                          if it overreads the auxiliary data.  ffplay carries on if frame_finished is true,
+                          if it overreads the auxiliary data.  ffplay carries on if frame_finished is true,
                           even in the face of such an error, so I think we should too.
 
                           Returning from the method here caused mantis #352.
@@ -350,18 +406,26 @@ FFmpegDecoder::decode_audio_packet ()
                }
 
                if (frame_finished) {
-                       ContentTime const ct = ContentTime::from_seconds (
+                       ContentTime ct = ContentTime::from_seconds (
                                av_frame_get_best_effort_timestamp (_frame) *
-                               av_q2d (_ffmpeg_content->audio_stream()->stream (_format_context)->time_base))
+                               av_q2d ((*stream)->stream (_format_context)->time_base))
                                + _pts_offset;
-                       
-                       int const data_size = av_samples_get_buffer_size (
-                               0, audio_codec_context()->channels, _frame->nb_samples, audio_sample_format (), 1
-                               );
 
-                       audio (deinterleave_audio (_frame->data, data_size), ct);
+                       shared_ptr<AudioBuffers> data = deinterleave_audio (*stream);
+
+                       if (ct < ContentTime ()) {
+                               /* Discard audio data that comes before time 0 */
+                               Frame const remove = min (int64_t (data->frames()), (-ct).frames_ceil(double((*stream)->frame_rate ())));
+                               data->move (remove, 0, data->frames() - remove);
+                               data->set_frames (data->frames() - remove);
+                               ct += ContentTime::from_frames (remove, (*stream)->frame_rate ());
+                       }
+
+                       if (data->frames() > 0) {
+                               audio->give (*stream, data, ct);
+                       }
                }
-                       
+
                copy_packet.data += decode_result;
                copy_packet.size -= decode_result;
        }
@@ -370,6 +434,8 @@ FFmpegDecoder::decode_audio_packet ()
 bool
 FFmpegDecoder::decode_video_packet ()
 {
+       DCPOMATIC_ASSERT (_video_stream);
+
        int frame_finished;
        if (avcodec_decode_video2 (video_codec_context(), _frame, &frame_finished, &_packet) < 0 || !frame_finished) {
                return false;
@@ -377,15 +443,16 @@ FFmpegDecoder::decode_video_packet ()
 
        boost::mutex::scoped_lock lm (_filter_graphs_mutex);
 
-       shared_ptr<FilterGraph> graph;
-       
-       list<shared_ptr<FilterGraph> >::iterator i = _filter_graphs.begin();
+       shared_ptr<VideoFilterGraph> graph;
+
+       list<shared_ptr<VideoFilterGraph> >::iterator i = _filter_graphs.begin();
        while (i != _filter_graphs.end() && !(*i)->can_process (dcp::Size (_frame->width, _frame->height), (AVPixelFormat) _frame->format)) {
                ++i;
        }
 
        if (i == _filter_graphs.end ()) {
-               graph.reset (new FilterGraph (_ffmpeg_content, dcp::Size (_frame->width, _frame->height), (AVPixelFormat) _frame->format));
+               graph.reset (new VideoFilterGraph (dcp::Size (_frame->width, _frame->height), (AVPixelFormat) _frame->format));
+               graph->setup (_ffmpeg_content->filters ());
                _filter_graphs.push_back (graph);
                LOG_GENERAL (N_("New graph for %1x%2, pixel format %3"), _frame->width, _frame->height, _frame->format);
        } else {
@@ -397,12 +464,12 @@ FFmpegDecoder::decode_video_packet ()
        for (list<pair<shared_ptr<Image>, int64_t> >::iterator i = images.begin(); i != images.end(); ++i) {
 
                shared_ptr<Image> image = i->first;
-               
+
                if (i->second != AV_NOPTS_VALUE) {
-                       double const pts = i->second * av_q2d (_format_context->streams[_video_stream]->time_base) + _pts_offset.seconds ();
-                       video (
+                       double const pts = i->second * av_q2d (_format_context->streams[_video_stream.get()]->time_base) + _pts_offset.seconds ();
+                       video->give (
                                shared_ptr<ImageProxy> (new RawImageProxy (image)),
-                               rint (pts * _ffmpeg_content->video_frame_rate ())
+                               llrint (pts * _ffmpeg_content->active_video_frame_rate ())
                                );
                } else {
                        LOG_WARNING_NC ("Dropping frame without PTS");
@@ -411,7 +478,7 @@ FFmpegDecoder::decode_video_packet ()
 
        return true;
 }
-       
+
 void
 FFmpegDecoder::decode_subtitle_packet ()
 {
@@ -421,40 +488,102 @@ FFmpegDecoder::decode_subtitle_packet ()
                return;
        }
 
-       /* Sometimes we get an empty AVSubtitle, which is used by some codecs to
-          indicate that the previous subtitle should stop.
-       */
        if (sub.num_rects <= 0) {
-               image_subtitle (ContentTimePeriod (), shared_ptr<Image> (), dcpomatic::Rect<double> ());
+               /* Sometimes we get an empty AVSubtitle, which is used by some codecs to
+                  indicate that the previous subtitle should stop.  We can ignore it here.
+               */
                return;
-       } else if (sub.num_rects > 1) {
-               throw DecodeError (_("multi-part subtitles not yet supported"));
        }
-               
+
        /* Subtitle PTS (within the source, not taking into account any of the
-          source that we may have chopped off for the DCP)
+          source that we may have chopped off for the DCP).
        */
-       ContentTimePeriod period = subtitle_period (sub) + _pts_offset;
-
-       AVSubtitleRect const * rect = sub.rects[0];
+       FFmpegSubtitlePeriod sub_period = subtitle_period (sub);
+       ContentTimePeriod period;
+       period.from = sub_period.from + _pts_offset;
+       if (sub_period.to) {
+               /* We already know the subtitle period `to' time */
+               period.to = sub_period.to.get() + _pts_offset;
+       } else {
+               /* We have to look up the `to' time in the stream's records */
+               period.to = ffmpeg_content()->subtitle_stream()->find_subtitle_to (subtitle_id (sub));
+       }
 
-       if (rect->type != SUBTITLE_BITMAP) {
-               /* XXX */
-               // throw DecodeError (_("non-bitmap subtitles not yet supported"));
-               return;
+       for (unsigned int i = 0; i < sub.num_rects; ++i) {
+               AVSubtitleRect const * rect = sub.rects[i];
+
+               switch (rect->type) {
+               case SUBTITLE_NONE:
+                       break;
+               case SUBTITLE_BITMAP:
+                       decode_bitmap_subtitle (rect, period);
+                       break;
+               case SUBTITLE_TEXT:
+                       cout << "XXX: SUBTITLE_TEXT " << rect->text << "\n";
+                       break;
+               case SUBTITLE_ASS:
+                       decode_ass_subtitle (rect->ass, period);
+                       break;
+               }
        }
 
+       avsubtitle_free (&sub);
+}
+
+list<ContentTimePeriod>
+FFmpegDecoder::image_subtitles_during (ContentTimePeriod p, bool starting) const
+{
+       return _ffmpeg_content->image_subtitles_during (p, starting);
+}
+
+list<ContentTimePeriod>
+FFmpegDecoder::text_subtitles_during (ContentTimePeriod p, bool starting) const
+{
+       return _ffmpeg_content->text_subtitles_during (p, starting);
+}
+
+void
+FFmpegDecoder::decode_bitmap_subtitle (AVSubtitleRect const * rect, ContentTimePeriod period)
+{
        /* Note RGBA is expressed little-endian, so the first byte in the word is R, second
           G, third B, fourth A.
        */
-       shared_ptr<Image> image (new Image (PIX_FMT_RGBA, dcp::Size (rect->w, rect->h), true));
+       shared_ptr<Image> image (new Image (AV_PIX_FMT_RGBA, dcp::Size (rect->w, rect->h), true));
 
+#ifdef DCPOMATIC_HAVE_AVSUBTITLERECT_PICT
        /* Start of the first line in the subtitle */
        uint8_t* sub_p = rect->pict.data[0];
        /* sub_p looks up into a BGRA palette which is here
           (i.e. first byte B, second G, third R, fourth A)
        */
        uint32_t const * palette = (uint32_t *) rect->pict.data[1];
+#else
+       /* Start of the first line in the subtitle */
+       uint8_t* sub_p = rect->data[0];
+       /* sub_p looks up into a BGRA palette which is here
+          (i.e. first byte B, second G, third R, fourth A)
+       */
+       uint32_t const * palette = (uint32_t *) rect->data[1];
+#endif
+       /* And the stream has a map of those palette colours to colours
+          chosen by the user; created a `mapped' palette from those settings.
+       */
+       map<RGBA, RGBA> colour_map = ffmpeg_content()->subtitle_stream()->colours ();
+       vector<RGBA> mapped_palette (rect->nb_colors);
+       for (int i = 0; i < rect->nb_colors; ++i) {
+               RGBA c ((palette[i] & 0xff0000) >> 16, (palette[i] & 0xff00) >> 8, palette[i] & 0xff, (palette[i] & 0xff000000) >> 24);
+               map<RGBA, RGBA>::const_iterator j = colour_map.find (c);
+               if (j != colour_map.end ()) {
+                       mapped_palette[i] = j->second;
+               } else {
+                       /* This colour was not found in the FFmpegSubtitleStream's colour map; probably because
+                          it is from a project that was created before this stuff was added.  Just use the
+                          colour straight from the original palette.
+                       */
+                       mapped_palette[i] = c;
+               }
+       }
+
        /* Start of the output data */
        uint32_t* out_p = (uint32_t *) image->data()[0];
 
@@ -462,31 +591,46 @@ FFmpegDecoder::decode_subtitle_packet ()
                uint8_t* sub_line_p = sub_p;
                uint32_t* out_line_p = out_p;
                for (int x = 0; x < rect->w; ++x) {
-                       uint32_t const p = palette[*sub_line_p++];
-                       *out_line_p++ = ((p & 0xff) << 16) | (p & 0xff00) | ((p & 0xff0000) >> 16) | (p & 0xff000000);
+                       RGBA const p = mapped_palette[*sub_line_p++];
+                       /* XXX: this seems to be wrong to me (isn't the output image RGBA?) but it looks right on screen */
+                       *out_line_p++ = (p.a << 24) | (p.r << 16) | (p.g << 8) | p.b;
                }
+#ifdef DCPOMATIC_HAVE_AVSUBTITLERECT_PICT
                sub_p += rect->pict.linesize[0];
+#else
+               sub_p += rect->linesize[0];
+#endif
                out_p += image->stride()[0] / sizeof (uint32_t);
        }
 
-       dcp::Size const vs = _ffmpeg_content->video_size ();
-
-       image_subtitle (
-               period,
-               image,
-               dcpomatic::Rect<double> (
-                       static_cast<double> (rect->x) / vs.width,
-                       static_cast<double> (rect->y) / vs.height,
-                       static_cast<double> (rect->w) / vs.width,
-                       static_cast<double> (rect->h) / vs.height
-                       )
+       dcp::Size const vs = _ffmpeg_content->video->size ();
+       dcpomatic::Rect<double> const scaled_rect (
+               static_cast<double> (rect->x) / vs.width,
+               static_cast<double> (rect->y) / vs.height,
+               static_cast<double> (rect->w) / vs.width,
+               static_cast<double> (rect->h) / vs.height
                );
-       
-       avsubtitle_free (&sub);
+
+       subtitle->give_image (period, image, scaled_rect);
 }
 
-list<ContentTimePeriod>
-FFmpegDecoder::subtitles_during (ContentTimePeriod p, bool starting) const
+void
+FFmpegDecoder::decode_ass_subtitle (string ass, ContentTimePeriod period)
 {
-       return _ffmpeg_content->subtitles_during (p, starting);
+       /* We have no styles and no Format: line, so I'm assuming that FFmpeg
+          produces a single format of Dialogue: lines...
+       */
+
+       vector<string> bits;
+       split (bits, ass, is_any_of (","));
+       if (bits.size() < 10) {
+               return;
+       }
+
+       sub::RawSubtitle base;
+       list<sub::RawSubtitle> raw = sub::SSAReader::parse_line (base, bits[9]);
+
+       BOOST_FOREACH (sub::Subtitle const & i, sub::collect<list<sub::Subtitle> > (raw)) {
+               subtitle->give_text (period, i);
+       }
 }