Speaker Suggestions for Acoustic-Model Building

[rif <rif@xxxxxxx>]

I'm trying to build an acoustic model for speech recognition over a
noisy channel (a specific microphone, used in a specific reverberant
room, under far-field conditions).  I'm planning to generate channel
specific recordings by playing clean high-quality recordings of known
speech through a speaker and recording them through the microphone.
Does anyone have any suggestions as to what sort of speaker setup
might be good for this kind of task?

Cheers,

rif