I stumbled across this a while back and I am not sure how seriously to take it:
"Some keys are just plain difficult to play in.
A and E concert are strange keys to the orchestra. Colourful pedal notes allude you when you use these keys. The keys are also out of favour with transposing instruments. E concert transposes Eb instruments to six sharps and Bb instruments to seven sharps. Just a semitone higher than E, Eb concert is a wonderful key for colour, transposes Eb instruments to C and Bb instruments to F."
- The best keys are: Eb, Bb and F concert.
- Secondary keys are: G, C and D concert.
- Notably bad keys are: A and E concert.
[ More at http://www.musicarrangers.com/star-theory/p21.htm ]
I am in the process of orchestrating a musical, using a fairly traditional pit orchestra (4 reeds, 4 brass, small string section, 2 keyboard, bass, drums and percussion). I have used the guidelines above when I have gotten into trouble (when the ranges of wind instruments get too high or don't sound right). I've seen a lot of numbers in Eb and Bb (concert key) and less often F or C or G (concert key). So I am assuming there is some validity to this theory. I'm just wondering how far to trust it.
I'm talking about real world composing experience, not academic theory. I usually compose on the piano and pick the key by the way it sounds (and change it only to accomodate the vocal ranges of the characters). They pretty much all sound good on keyboards. I especially liked a song that I had written in C# (which, truthfully, was not all that hard to play on the piano) but I had to change it to Bb because the brass section didn't sound as good as I expected in that key. As ever, when using sample libraries, I wonder if what I am hearing accurately represents how it will sound with real instruments (and real musicians having to play those keys.)