I'm writing a simple text-to-speech application using the SAPI SDK and part of it involves being able to save the spoken text to a wav file. I've got the saving part figured out, that's simple, but I'm running into problems setting the format of the wav file. When the voice is spoken normally (not saved) it uses 44KHz 16 bit stereo, but when I try to save it, it always uses 22KHz 8 bit mono. I've tried setting the voice object's speechaudioform at to the different format, but it doesn't take. I also tried setting the SpFileTream object's format type but it gives me an error every time. Does anyone know how I might accomplish this? I don't like how low quality the 22KHz file sounds. Ultimately this is going into a monitoring program that will keep track of things like disk space and CPU temp and give an audible voice notification of the event. Events will be recorded and people will be able to play them back from the recorded audio files. I also just want to learn how to do it.