Hi @denisdawid , well AVFoundation is a great framework used for audio modeling, but as all Apple frameworks, is not so simple. AVAsset generally are container for single AVAssetTrack, each of this could be a different media type, for example audio, video, video/audio subtitles, and so on.
But, if I remember correctly, this objects abstraction are not correlated to media data they represent (that is, load in memory) until you execute an asynchronous data load.
Note that from iOS 16 syntax callbacks are more younger, look at this link https://developer.apple.com/documentation/avfoundation/media_assets/loading_media_data_asynchronously
Then, what you should try to manage, is this asynchronous loading of media that could optimise the media operations. Use the correct keys should avoid data not necessarily used for your app.
Bye
Rob