InnerTune
InnerTune copied to clipboard
z-huang
Replay Gain Normalization
In a queue, some songs are louder than others. To solve this an option of Replay Gain Normalization can be introduced in settings ( disabled by default).
Will this help? https://github.com/videolan/vlc/search?q=replaygain&type=
On Thu 14 Apr, 2022, 9:53 AM Zion Huang, @.***> wrote:
I'm still finding how to achieve this in ExoPlayer.
— Reply to this email directly, view it on GitHub https://github.com/z-huang/music/issues/24#issuecomment-1098695872, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALIG6ORUCYLU2WM36YCF52TVE6MS3ANCNFSM5TLM4AIQ . You are receiving this because you authored the thread.Message ID: @.***>
I'm not sure if YouTube use ReplayGain, because some formats are webm. However, YouTube provides "loudnessDb" and "perceptualLoudnessDb", which may be the way to implement normalized volume.
I found that YouTube uses loadnessDb to convert the volume to a normalized value to make every song under a specific loadness.
However, NewPipe Extractor doesn't extracct "loudnessDb" and "perceptualLoudnessDb". Also, NewPipeExtractor doesn't provide the type and year of albums, and the song thumbnail is always rectangle due to fetching data from YouTube, not YouTube Music. I'll make a modified version of NewPipe Extractor to support these things.
According to my roadmap, I'll first fix the UI issues, and then getting started to remake NewPipe Extractor.
By remaking the extractor, you wont be able to just merge the extractor from Newpipe when Youtube breaks somethin in the site right? The "total shutdown" days in Newpipe always happen once in while. Those will be needed to fixed separately. So maybe you can leave it as it is? This is not that necessary feature I think.
Or maybe just make a PR in the main Newpipe extractor.
On Thu, 21 Apr, 2022, 2:41 pm Zion Huang, @.***> wrote:
I found that YouTube uses loadnessDb to convert the volume to a normalized value to make every song under a specific loadness. [image: image] https://user-images.githubusercontent.com/39427017/164409259-6d1ac580-0f38-47d3-a8a1-2c83049f17d9.png
However, NewPipe Extractor doesn't extracct "loudnessDb" and "perceptualLoudnessDb". Also, NewPipeExtractor doesn't provide the type and year of albums, and the song thumbnail is always rectangle due to fetching data from YouTube, not YouTube Music. I'll make a modified version of NewPipe Extractor to support these things.
According to my roadmap, I'll first fix the UI issues, and then getting started to remake NewPipe Extractor.
— Reply to this email directly, view it on GitHub https://github.com/z-huang/music/issues/24#issuecomment-1104925554, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALIG6OR65PLWSLVIONSFRTTVGELUVANCNFSM5TLM4AIQ . You are receiving this because you authored the thread.Message ID: @.***>
I know that a quick hotfix is important for user experience. But I have to do that not only for the audio normalization feature, but also for more information about albums, artists and square song thumbnails. If the modification doesn't cause many conflicts with the upstream, a hotfix won't be too hard, I think. A PR may be feasible. No matter what, I'll try to achieve that. 😊
Relevant https://github.com/vfsfitvnm/ViMusic/commit/8a94c7e714536ef7ba6b925332b208f590ff500e