I don't understand, why isn't simply navigating the media tree taken into consideration?
UPnP ContentDirectory has two ways for Control Points to navigate the content -- namely the
Browse() method and the
Search() method.
Simple Control Points, which do not have memory for a local database, (like most media renderers), will use the Browse() method to navigate item by item up and down the branches of the server's content directory folder tree. This means that such CPs are constrained to only display the tree in whatever fixed folder structure has been hard wired into the server.
Sophisticated Control Points (like most software applications, including MC), which do have memory for a local database, will use the Search() method to download the whole content folder tree from the server. This means that they can then display the library dynamically to the user according to flexible views and layouts which are defined locally by the user in the Control Point software.
In other words, MC uses the
Search() method to download all the library items from the Twonky server library. During this download process, MC will only display whatever subset of the library information that has been downloaded so far. Depending on the network speed and the processor speeds of the Twonky PC and the MC PC it might take many hours before MC has fully downloaded a library of 70k tracks. (e.g. my experience is that MC can download about 30k tracks per hour from Whitebear Media server, but Twonky may be faster or slower..)
Connect MC to the Twonky library server and let it run overnight; in MC select Audio (say) and its status bar displays the number of tracks loaded; that figure will increase during the load process...