External sound devices for laptops

[Andrew Faulkner <andyf@xxxxxxxxxxxxxxxxxxx>]

Stuart Rosen and othere here have been using a Dell laptop for masking
experiments, 
but he is not  completely happy with the sound quality (occasional
glitches; a bit too much 
noise in the background -- in particular a 200 Hz periodic complex), so
is
looking into an external box. The SoundBlaster Extigy looks good (96 kHz
sampling rate, 24 bits) -- but does anyone have any experience with this
or 
similar external sound boxes? I guess it should plug in to a USB port to
be 
readily used, as does the Extigy. I'd also like to have ready access to
its 
workings from Windows 2000 and Matlab, perhaps through an ActiveX 
controller. At about 149 dollars or pounds, this seems very good value to
me.

Thank you all in advance --
P.S. If anyone is interested in running psychacoustic tasks through
Matlab, 
I am having reasonable success using an ActiveX sound card controller
that 
Rhodri Cusack developed to control Windows sound cards (thank you, thank
you 
Rhodri! And to Bob Carlyon who let me know about it).
http://www.mrc-cbu.cam.ac.uk/personal/rhodri.cusack/vb/download.htm
I only manipulate the master volume control (keeping the
digital signal as 
close to maximum as possible) as the background noise level from the
laptop 
is determined by its setting -- the lower the better. The wave volume
control has no effect on this background noise. You need to measure the
behaviour of your volume control but the attenuations on my laptop were
in 
1.5 dB steps for small attenuations (up to 19.5 dB) and in 3 dB steps
after 
that to about 42.5 dB.
I also have been using a Logitech serial 3-button mouse to get responses,
as 
Matlab has good access to serial devices implemented explicitly. (Dontcha
just hate having to select a box on the screen, as well as pressing a
button!) A student of mine has also cannibalised one of these to make a
more 
typical response pad simply by soldering other buttons in parallel to the
switches in the mouse. You need to be a little careful about what you
look 
for on the serial port as different serial mice send different numbers of
bytes, but it is a relatively simple matter to make it work right. I
currently have this set up working during PET, with the listener flat on
his 
back and resting his hand on an inclined response box with two
buttons.

Dr Andrew Faulkner
Principal Research Fellow
Dept Phonetics and Linguistics
UCL (University College London)
Wolfson House
4 Stephenson Way
LONDON NW1 2HE
tel 44 (0)20 7679 7408 (direct)
fax 44 (0)20 7383 0752
e-mail andyf@phon.ucl.ac.uk
WWW: http://www.phon.ucl.ac.uk/home/andyf/