Hey Mr S, you talking to me?
About audio caching.
If the
audio cache amount of memory is
big enough to hold the entire decoded track, Cross will decode and cache it as fast as possible so that you can reach any position in the file asap.
This operation of course require a lot of CPU power (making one core usage reach 100%) and often lasts a few seconds.
If the
audio cache amount of memory is
too small, Cross will directly stream from the file by decoding small chunks in a background task. Of course, less CPU usage peak at load, but a bit more CPU usage all along the playback and more risk to have "cache miss" issues (decoded audio unavailable when seeking far from the current position) resulting in small "blanks".
Audio cache at 0% is the extreme case for which all the files are streamed (like in Cross 1.2.5).
Moreover, keep in mind that track analysis is a separate task and then, if you load an unanalyzed file, Cross will need to decode it twice (one for caching or streaming, one for analysis) !
Hope it's a bit clearer now.
Of course you shouldn't have drop-out issues by a loading a file, and something seems to be messing up the CPU usage repartition...