You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This api lets you create a virtual file where you can save the data and then obtain it through the demux by reading the file.
I have already made progress in demux webm, ffmpeg and webCodecs by reducing memory.
I would only be missing the mp4 demuxer.
I don't know exactly where the memory expenditure is, it could be when reading the file that everything is stored in memory and then reading it to process it or it could be when creating the chunks that it stores them in memory and that is why it uses so much memory.
In either case you could use this api to read directly from the file.
Library uses too much memory for large files
In this case the file weighs 500 MB and uses 700 MB, I think the memory could be lowered to 0 using this api.
https://developer.mozilla.org/en-US/docs/Web/API/File_System_API
This api lets you create a virtual file where you can save the data and then obtain it through the demux by reading the file.
I have already made progress in demux webm, ffmpeg and webCodecs by reducing memory.
I would only be missing the mp4 demuxer.
I don't know exactly where the memory expenditure is, it could be when reading the file that everything is stored in memory and then reading it to process it or it could be when creating the chunks that it stores them in memory and that is why it uses so much memory.
In either case you could use this api to read directly from the file.
If you use it, I recommend using this one, which is compatible with webview
https://developer.mozilla.org/en-US/docs/Web/API/FileSystemSyncAccessHandle
You only need to save the file's offsets and then read them.
Someone who knows the library well could implement it quickly, greetings.
The text was updated successfully, but these errors were encountered: