I find amarok is evolving nicely, but there\'s one thing that prevents me from using it : I have doubts on the quality of audio it produces.

Delegating most of the decoding, and sound processing, to third-party engines is a flexible and powerful solution, but neither xine nor gstreamer seem to focus much on internal sound quality - their prime concern is not whether a few bits of sound precision is lost between source and sink, but rather that all possible video/sound formats playback is somewhat satisfactory.

That\'s comprehensible, but then since neither amarok nor the engines it uses focuses on audio quality, it\'s rather hard from the end-user point of view to get any detail about it - let alone choose his own performance/quality balance..


so here\'s my wish/rant :
1. could amarok provides details about audio quality to the end-user ? e.g.,
whether sound samples travel thru those engines as 32bit values, or 24, or only as many as original material - or as many as sink bitdepth, when/how is resampling taking place, ..
e.g., give those details in the FAQ
2. could amarok provide user-friendly settings on the quality of the processing inside the engines ? for instance, allow the user to make the engines process sound with 32bits, so that the numerous processing steps won\'t kill quality [maybe this is already the case] ; or choose between several resampling algorithms - in case there\'s any need to resample [I don\'t know whether the engines provide several resampling algorithms at this point though..] ; choose whether to dither where it makes sense to (at every reduction of bitdepth. e.g., when converting high precision samples decoded from MP3 to integer values, and right before the sink if it has less bits than internal resolution) ; etc ...