Ok, so I've just got my new RH10 and decide to do some comparisons between ATRAC and MP3. I take one of my favourite test tracks ("Fonkalishus" from the Bluesiana II album, Windham Hill Records, great, natural recording with lots of low level detail), encode it with ATRAC3+ @ 64 and 256, MP3-CBR @ 128 and MP3-VBR at different qualities (I use Creative Media Source where for VBR you specify a quality percentage, 50% giving approx. the same file size as CBR @ 128), download the lot on the RH10 together with a WAV file for reference, plug in my reference Beyerdynamic DT-880 headphones. I do a sort of blind test (set it to random, listen to part of the track, make notes, then look to see which one is it, go on to the next track, etc.) and on a first quick pass listening to just a few seconds and not thinking much I get this: VBR 40% - no treble, splashy CBR 128 - no treble, splashy VBR 100% - no treble but good WAV - perfect VBR 80% - no treble but good ATRAC3+ 64 - lots of artifacts, useless VBR 10% - no treble, some artifacts, splashy ATRAC3+ 256 - almost perfect, some very subtle artifacts VBR 60% - no treble, slightly splashy So, I think, "boy, that MP3 thing just throws away all the highs!", double-check in my computer with Windows Media Player, and nope, the highs are there, with more or less quality, but the quantity is practically the same as in the WAV file. What are they on about? Is this just to promote ATRAC over MP3? I mean, even with the very recessed treble, I prefer MP3 CBR 128 over ATRAC3+ 64 (which is a joke), and when you compare the "good" MP3 from Media Player it's just no contest, but I can see where in listening tests people may prefer the livelier presentation of ATRAC3+ 64 if the test was carried out through something like an RH10 or a similar software implementation of the decoder. That's naughty, Sony