Akshara,
as usually happens with these matters, it depends. The most common method I have seen in the field, as well as at my current institute, is a combination of loudspeakers, and conducting experiments from within a sound attenuated chamber, which should satisfy
the noise requirements for most set-ups. If you are especially paranoid about EM interference, then you would also invest in making sure your SA chamber is also a faraday cage.
Hope this helps!
- Andrew