BTW the files sound somewhat better on the PC. There is a bunch of digital artifacts from the low bitrate ("Fixed rate Stereo 32kbits/s" per empeg info screen) and on the PC they seem to be filtered somewhat and come out sort of high & squeaky on the empeg.
Based on your description, I'm betting the WMA files where you're complaining about the audio quality are not CD rips, they are something else, perhaps sound bites downloaded from the web. Right?
If your WMA files are not 44.1khz 16-bit stereo (in other words, they are some kind of lower-sample-rate web sound byte instead of actual CD rips), then I would expect that your sound quality problems are due to the sample rate conversion algorithm not being optimized for that rate.
There is a whole complicated science to sample rate conversion. It's a lot like image scaling/resizing in a graphics application: Some do it better than others.
Imagine that your 11khz audio file were a bitmap image. Say, for example, it was a small 320x200 graphic.
If you resize a tiny 320x200 graphic to fill a high resolution monitor, say, 1152x864, you will get blocky jagged edges on the image, right? Unless you did some kind of bilinear resampling of the image, then it wouldn't have jaggies but it would be blurry.
Well, sampling an 11khz 8-bit mono audio file up to 44.1khz 16-bit stereo is very much the same thing. High frequency artifacts in an audio file are like the jaggies in that image. And a lowpass filter on an audio file is like the blurring feature in an image editor.
So. Your windows WMA-player applet is probably doing a better job of samplerate conversion than the empeg car player because it's designed to play crappy little soundbytes and it's already tweaked to hide the "jaggies" in the upconversion to your soundcard. The empeg-car player, on the other hand, was meant to play nothing but CD rips from the get-go, and they probably didn't bother with a fancy samplerate converter algorithm, figuring that no one would want to play web soundbites on their high-quality car player.