My experience:
Used music/Equipment: Asus WinXP laptop (circa 2006)-> Foobar -> ASIO4ALL -> generic USB cord -> Maverick DAC -> Completely stock Ultrasone HFI 780s.
Music: FLAC: Rush Moving Pictures / Sarah McLachlan Closer: The Best of.. / Jay-Z Empire State of Mind / Them Crooked Vultures
320mp3: Once OST / Jay-Z Empire State of Mind / Eric Clapton Unplugged
256VBR mp3: The Clipse / Popular Demand
Some other, lower quality mixtapes were also used -- Normally at 192
I played around with it some more and had a nagging feeling that it was 'colouring' the music, especially female vocals. I decieded to get a second opinion with my girlfriend, who is a trained vocalist. I loaded up a song I knew she was very familiar with - Building a Mystery from Sarah McLachlan's 'Best of CD' - and asked her to listen. I began with the Dolby headphone as set up in the OP (with 1.25 centre) and she right away took it off the headphones and gave me a look. I asked her to try it again with nothing and then switched to bs2b on the cmoy setting. She could not tell the difference between the vanilla sound and the crosfeed sound but noted that the vocals on the DH were unnatural.
Personally, I prefer the stock sound but become fatigued quite quickly. I use the bs2b setting on the default or the cmoy depending on listening times. After a couple days use with multiple 4 + hour sessions, I find the the dolby headphone to be the least fatiguing setting, the cmoy bs2b to be the best balance, and no DSP to be the best sounding.
The cmoy bs2b was also the most generous with lossy files, I am assuming this is because some of the highs are rolled off. Dolby Headphone completely exposed poor files; however, 320 and FLAC were mostly indistinguishable.