Lesson 1.2: Copy Metadata
In this lesson, you will learn how to copy metadata and chapter information
from the input to the output. You will also learn about what types of metadata
are supported in different file formats and whether they support chapters or
not. As with the previous lesson, there are test videos in the videos/inputs
folder and depending on whether the format supports it they may have file
metadata, stream metadata, or chapter data. Commands for remuxing
these files can be found in the run.sh file with comments about what data
is able to be successfully copied. All the code for this tutorial can be found
here.
File Metadata
// in main
const char *in_filename, *out_filename, *title;
...
title = argv[3];
...
if ((ret = av_dict_copy(&out_fmt_ctx->metadata,
in_fmt_ctx->metadata, AV_DICT_DONT_OVERWRITE)) < 0)
{
fprintf(stderr, "Failed to copy file metadata.\n");
goto end;
}
if ((ret =
av_dict_set(&out_fmt_ctx->metadata, "title", title, 0)) < 0)
{
fprintf(stderr,
"Failed to set title for output format context.\n");
goto end;
}First we will take in a third argument on the command line that will be used to
set the value of the metadata title for the output file. Next we will use the
av_dict_copy function to copy all metadata from the input file to the output
context. This data will be written to the file when we call
avformat_write_header. Finally, we will use av_dict_set to set the metadata
title on the output context to the value passed in on the command line.
Chapters
// in main
if ((ret = copy_chapters(out_fmt_ctx, in_fmt_ctx)) < 0)
{
fprintf(stderr, "Failed to copy chapters.\n");
goto end;
}int copy_chapters(AVFormatContext *out_fmt_ctx,
AVFormatContext *in_fmt_ctx)
{
AVChapter *in_chapter, *out_chapter;
int ret = 0;
if (!(out_fmt_ctx->chapters =
av_calloc(in_fmt_ctx->nb_chapters, sizeof(AVChapter *))))
{
fprintf(stderr,
"Failed to allocate output format chapters array.\n");
ret = AVERROR(ENOMEM);
return ret;
}
for (int i = 0; i < in_fmt_ctx->nb_chapters; i++)
{
in_chapter = in_fmt_ctx->chapters[i];
if (!(out_chapter = av_mallocz(sizeof(AVChapter)))) {
fprintf(stderr,
"Failed to allocate out_chapter for chapter %d", i);
ret = AVERROR(ENOMEM);
return ret;
}
out_chapter->id = in_chapter->id;
out_chapter->time_base = in_chapter->time_base;
out_chapter->start = in_chapter->start;
out_chapter->end = in_chapter->end;
if ((ret = av_dict_copy(&out_chapter->metadata,
in_chapter->metadata, 0)) < 0)
{
fprintf(stderr, "Failed to copy chapter metadata.\n");
av_freep(&out_chapter);
return ret;
}
out_fmt_ctx->chapters[i] = out_chapter;
out_fmt_ctx->nb_chapters++;
}
return ret;
}To copy chapters, we must first allocate the chapters array on the output
context. The size of the array will depend on the value of the nb_chapters
field on the input context. Then we’ll loop over the chapters. For each one,
we’ll get a reference to it from the chapters array field on the input
context. Then we’ll allocate a new AVChapter struct and copy all data from
the input chapter to the new chapter. Then we’ll use av_dict_copy to copy
any metadata from the input chapter to the new chapter. Finally, we’ll asign
the new chapter to the chapters array field of the output context and
increment the nb_chapters field.
Stream Metadata
// in initialze_stream
if ((ret = av_dict_copy(&stream->metadata,
in_fmt_ctx->streams[*stream_idx]->metadata,
AV_DICT_DONT_OVERWRITE)) < 0)
{
fprintf(stderr, "Failed to copy video metadata.\n");
return ret;
}Considering how similar the code is for initializing the video stream and the
audio stream are, I broke this code out of the main function into it’s own
initialize_stream function. The function takes in an AVMediaType value
which determines whether to initialize the audio stream or the video stream.
This value is given to av_find_best_stream and the stream index returned
is used to copy the codec parameters to the new stream just like before. In
addition, I’ve added the above code which again uses av_dict_copy to copy
the metadata from the input stream to the new stream.
With these new additions to the code, we will now produce files that have all
metadata and chapter information copied from the input file, so long as the
format container supports it. We can also set metadata values if we like. This
example only sets the file metadata title value but you can experiment to see
what other values you can set using av_dict_set.
In the next lesson we will learn how to add chapter markers to a file. We will
input a raw text file into avformat_open_input using the ffmetadata format.
The chapter data will parsed and the values will be set on metadata field of
the AVFormatContext struct which will then be copied to the output context
using av_dict_copy.
Go To Next Lesson - 1.3: Add Chapters