Factorized some container/codec handling code + removed the no longer needed transcode output formats mka and webm

This commit is contained in:
emeric
2025-11-30 18:48:04 +01:00
parent c3cd30f248
commit d834f34683
65 changed files with 1424 additions and 1003 deletions
@@ -26,8 +26,9 @@
#include "core/ILogger.hpp"
#include "core/IResourceHandler.hpp"
#include "core/String.hpp"
#include "core/media/MimeType.hpp"
#include "audio/AudioTypes.hpp"
#include "audio/AudioProperties.hpp"
#include "audio/Exception.hpp"
#include "audio/IAudioFileInfo.hpp"
#include "audio/IAudioFileInfoParser.hpp"
@@ -37,6 +38,7 @@
#include "database/objects/PodcastEpisodeId.hpp"
#include "database/objects/Track.hpp"
#include "database/objects/TrackLyrics.hpp"
#include "database/objects/Types.hpp"
#include "database/objects/User.hpp"
#include "services/artwork/IArtworkService.hpp"
@@ -54,84 +56,73 @@ namespace lms::api::subsonic
{
namespace
{
std::optional<audio::OutputFormat> subsonicStreamFormatToTranscodingOutputFormat(std::string_view format)
struct OutputFormat
{
for (const auto& [str, avFormat] : std::initializer_list<std::pair<std::string_view, audio::OutputFormat>>{
{ "mp3", audio::OutputFormat::MP3 },
{ "opus", audio::OutputFormat::OGG_OPUS },
{ "vorbis", audio::OutputFormat::OGG_VORBIS },
})
{
if (core::stringUtils::stringCaseInsensitiveEqual(str, format))
return avFormat;
}
return std::nullopt;
std::string name;
core::media::ContainerType container;
core::media::CodecType codec;
};
constexpr std::array outputFormats{
OutputFormat{ "mp3", core::media::ContainerType::MPEG, core::media::CodecType::MP3 },
OutputFormat{ "opus", core::media::ContainerType::Ogg, core::media::CodecType::Opus },
OutputFormat{ "vorbis", core::media::ContainerType::Ogg, core::media::CodecType::Vorbis },
};
std::optional<OutputFormat> getOutputFormatByName(std::string_view format)
{
std::optional<OutputFormat> res;
const auto itFormat{ std::find_if(outputFormats.begin(), outputFormats.end(), [&format, &res](const OutputFormat& outputFormat) {
if (core::stringUtils::stringCaseInsensitiveEqual(format, outputFormat.name))
{
res = outputFormat;
return true;
}
return false;
}) };
if (itFormat != outputFormats.end())
res = *itFormat;
return res;
}
audio::OutputFormat userTranscodeFormatToTranscodingFormat(db::TranscodingOutputFormat format)
std::optional<OutputFormat> userTranscodeFormatToOutputFormat(db::TranscodingOutputFormat format)
{
std::optional<OutputFormat> res;
switch (format)
{
case db::TranscodingOutputFormat::MP3:
return audio::OutputFormat::MP3;
res = getOutputFormatByName("mp3");
break;
case db::TranscodingOutputFormat::OGG_OPUS:
return audio::OutputFormat::OGG_OPUS;
case db::TranscodingOutputFormat::MATROSKA_OPUS:
return audio::OutputFormat::MATROSKA_OPUS;
res = getOutputFormatByName("opus");
break;
case db::TranscodingOutputFormat::OGG_VORBIS:
return audio::OutputFormat::OGG_VORBIS;
case db::TranscodingOutputFormat::WEBM_VORBIS:
return audio::OutputFormat::WEBM_VORBIS;
res = getOutputFormatByName("vorbis");
break;
}
return audio::OutputFormat::OGG_OPUS;
return res;
}
bool isCodecCompatibleWithOutputFormat(audio::CodecType codec, audio::OutputFormat outputFormat)
audio::AudioProperties getAudioProperties(const std::filesystem::path& trackPath)
{
switch (outputFormat)
{
case audio::OutputFormat::MP3:
return codec == audio::CodecType::MP3;
case audio::OutputFormat::OGG_OPUS:
case audio::OutputFormat::MATROSKA_OPUS:
return codec == audio::CodecType::Opus;
case audio::OutputFormat::OGG_VORBIS:
case audio::OutputFormat::WEBM_VORBIS:
return codec == audio::CodecType::Vorbis;
}
return true;
}
struct StreamParameters
{
std::filesystem::path filePath;
std::string fileMimeType; // set if known
std::optional<audio::TranscodeParameters> transcodeParameters;
bool estimateContentLength{};
};
bool isOutputFormatCompatible(const std::filesystem::path& trackPath, audio::OutputFormat outputFormat)
{
// TODO: put this information in db during scan
// It is in base only for tracks, not yet for podcasts
try
{
const auto parser{ audio::createAudioFileInfoParser(audio::AudioFileInfoParserBackend::FFmpeg) };
audio::AudioFileInfoParseOptions parseOptions;
parseOptions.audioPropertiesReadStyle = audio::AudioFileInfoParseOptions::AudioPropertiesReadStyle::Fast; // only coded needed
parseOptions.audioPropertiesReadStyle = audio::AudioFileInfoParseOptions::AudioPropertiesReadStyle::Average;
parseOptions.readImages = false;
parseOptions.readTags = false;
const auto audioFile{ parser->parse(trackPath, parseOptions) };
if (!audioFile->getAudioProperties())
const audio::AudioProperties* properties{ audioFile->getAudioProperties() };
if (!properties)
throw RequestedDataNotFoundError{};
return isCodecCompatibleWithOutputFormat(audioFile->getAudioProperties()->codec, outputFormat);
return *properties;
}
catch (const audio::Exception& e)
{
@@ -143,9 +134,7 @@ namespace lms::api::subsonic
struct AudioFileInfo
{
std::filesystem::path path;
std::chrono::milliseconds duration{};
std::size_t bitrate{};
std::string mimeType; // set if known
audio::AudioProperties audioProperties;
};
AudioFileInfo getAudioFileInfo(db::Session& session, AudioFileId audioFileId)
@@ -161,8 +150,20 @@ namespace lms::api::subsonic
throw RequestedDataNotFoundError{};
res.path = track->getAbsoluteFilePath();
res.duration = track->getDuration();
res.bitrate = track->getBitrate();
if (track->getContainer() && track->getCodec())
{
res.audioProperties.container = *track->getContainer();
res.audioProperties.codec = *track->getCodec();
res.audioProperties.duration = track->getDuration();
res.audioProperties.bitrate = track->getBitrate();
res.audioProperties.channelCount = track->getChannelCount();
res.audioProperties.sampleRate = track->getSampleRate();
res.audioProperties.bitsPerSample = track->getBitsPerSample();
}
else
{
res.audioProperties = getAudioProperties(res.path);
}
}
else if (const db::PodcastEpisodeId * episodeId{ std::get_if<db::PodcastEpisodeId>(&audioFileId) })
{
@@ -173,14 +174,33 @@ namespace lms::api::subsonic
std::filesystem::path podcastCachePath{ core::Service<podcast::IPodcastService>::get()->getCachePath() };
res.path = podcastCachePath / episode->getAudioRelativeFilePath();
res.duration = episode->getDuration();
res.bitrate = episode->getEnclosureLength() / std::chrono::duration_cast<std::chrono::seconds>(episode->getDuration()).count() * 8;
res.mimeType = episode->getEnclosureContentType();
if (episode->getContainer() && episode->getCodec())
{
res.audioProperties.container = *episode->getContainer();
res.audioProperties.codec = *episode->getCodec();
res.audioProperties.duration = episode->getDuration();
res.audioProperties.bitrate = episode->getBitrate();
res.audioProperties.channelCount = episode->getChannelCount();
res.audioProperties.sampleRate = episode->getSampleRate();
res.audioProperties.bitsPerSample = episode->getBitsPerSample();
}
else
{
res.audioProperties = getAudioProperties(res.path);
}
}
return res;
}
struct StreamParameters
{
std::filesystem::path filePath;
audio::AudioProperties audioProperties;
std::optional<audio::TranscodeParameters> transcodeParameters;
bool estimateContentLength{};
};
StreamParameters getStreamParameters(RequestContext& context)
{
// Mandatory params
@@ -201,32 +221,31 @@ namespace lms::api::subsonic
StreamParameters parameters;
parameters.filePath = audioFileInfo.path;
parameters.fileMimeType = audioFileInfo.mimeType;
parameters.estimateContentLength = estimateContentLength;
if (format == "raw") // raw => no transcoding
return parameters; // TODO: what if offset is not 0?
std::optional<audio::OutputFormat> requestedFormat{ subsonicStreamFormatToTranscodingOutputFormat(format) };
std::optional<OutputFormat> requestedFormat{ getOutputFormatByName(format) };
if (!requestedFormat)
{
if (context.getUser()->getSubsonicEnableTranscodingByDefault())
requestedFormat = userTranscodeFormatToTranscodingFormat(context.getUser()->getSubsonicDefaultTranscodingOutputFormat());
requestedFormat = userTranscodeFormatToOutputFormat(context.getUser()->getSubsonicDefaultTranscodingOutputFormat());
}
if (!requestedFormat && (maxBitRate == 0 || audioFileInfo.bitrate <= maxBitRate))
if (!requestedFormat && (maxBitRate == 0 || audioFileInfo.audioProperties.bitrate <= maxBitRate))
{
LMS_LOG(API_SUBSONIC, DEBUG, "File's bitrate is compatible with parameters => no transcoding");
return parameters; // no transcoding needed
}
// scan the file to check if its format is compatible with the actual requested format
// Check if the input file is compatible with the actual requested format
// same codec => apply max bitrate
// otherwise => apply default bitrate (because we can't really compare bitrates between formats) + max bitrate)
std::size_t bitrate{};
if (requestedFormat && isOutputFormatCompatible(audioFileInfo.path, *requestedFormat))
if (requestedFormat && requestedFormat->container == audioFileInfo.audioProperties.container && requestedFormat->codec == audioFileInfo.audioProperties.codec)
{
if (maxBitRate == 0 || audioFileInfo.bitrate <= maxBitRate)
if (maxBitRate == 0 || audioFileInfo.audioProperties.bitrate <= maxBitRate)
{
LMS_LOG(API_SUBSONIC, DEBUG, "File's bitrate and format are compatible with parameters => no transcoding");
return parameters; // no transcoding needed
@@ -235,22 +254,24 @@ namespace lms::api::subsonic
}
// Need to transcode here
if (!requestedFormat)
requestedFormat = userTranscodeFormatToTranscodingFormat(context.getUser()->getSubsonicDefaultTranscodingOutputFormat());
if (!bitrate)
bitrate = context.getUser()->getSubsonicDefaultTranscodingOutputBitrate();
if (maxBitRate)
bitrate = std::min<std::size_t>(bitrate, maxBitRate);
audio::TranscodeParameters& transcodeParameters{ parameters.transcodeParameters.emplace() };
if (!requestedFormat) // no format provided => use user's default
requestedFormat = userTranscodeFormatToOutputFormat(context.getUser()->getSubsonicDefaultTranscodingOutputFormat());
if (!bitrate) // no bitrate provided => use user's default
bitrate = context.getUser()->getSubsonicDefaultTranscodingOutputBitrate();
if (maxBitRate) // but still honor provided maxBitRate, if any
bitrate = std::min<std::size_t>(bitrate, maxBitRate);
transcodeParameters.inputParameters.filePath = audioFileInfo.path;
transcodeParameters.inputParameters.duration = audioFileInfo.duration;
transcodeParameters.inputParameters.audioProperties = audioFileInfo.audioProperties;
transcodeParameters.inputParameters.offset = std::chrono::seconds{ timeOffset };
;
transcodeParameters.outputParameters.bitrate = bitrate;
transcodeParameters.outputParameters.format = *requestedFormat;
transcodeParameters.outputParameters.format.emplace();
transcodeParameters.outputParameters.format->container = requestedFormat->container;
transcodeParameters.outputParameters.format->codec = requestedFormat->codec;
transcodeParameters.outputParameters.stripMetadata = false; // We want clients to use metadata (offline use, replay gain, etc.)
return parameters;
@@ -371,7 +392,7 @@ namespace lms::api::subsonic
if (streamParameters.transcodeParameters)
resourceHandler = core::Service<transcoding::ITranscodeService>::get()->createTranscodeResourceHandler(*streamParameters.transcodeParameters, streamParameters.estimateContentLength);
else
resourceHandler = core::createFileResourceHandler(streamParameters.filePath, streamParameters.fileMimeType);
resourceHandler = core::createFileResourceHandler(streamParameters.filePath, core::media::getMimeType(streamParameters.audioProperties.container, streamParameters.audioProperties.codec).str());
}
else
{