feat(encoder): copy input streams' metadata to output streams (#1284)

* feat(encoder): copy language tag for audio and subtitle streams

Containers with audio streams for different languages use a tag
to signal which track contains which language. This information is
saved in the metadata object of a stream and needs to be copied
in addition to the codec properties.

* feat(encoder): copy input streams' metadata to output streams

---------

Signed-off-by: k4yt3x <i@k4yt3x.com>
Co-authored-by: k4yt3x <i@k4yt3x.com>

CHANGELOG.md

@@ -10,6 +10,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ### Added
 
 - Multi-versioning to critical functions to enhance performance in generic architecture builds.
+- The feature to copy input streams' metadata to the output streams (#1282).
 
 ### Changed
 

src/encoder.cpp

@@ -225,14 +225,21 @@ int Encoder::init(
                 return AVERROR_UNKNOWN;
             }
 
-            // Copy codec parameters from input to output
+            // Copy codec parameters from the input stream to the output stream
             ret = avcodec_parameters_copy(out_stream->codecpar, in_codecpar);
             if (ret < 0) {
                 logger()->error("Failed to copy codec parameters");
                 return ret;
             }
             out_stream->codecpar->codec_tag = 0;
 
+            // Copy all metadata from the input stream to the output stream
+            AVDictionaryEntry* tag = nullptr;
+            while ((tag = av_dict_get(in_stream->metadata, "", tag, AV_DICT_IGNORE_SUFFIX)) !=
+                   nullptr) {
+                av_dict_set(&out_stream->metadata, tag->key, tag->value, 0);
+            }
+
             // Copy time base
             out_stream->time_base = in_stream->time_base;