does digital voice recorder need to be trained

I'm new to WSR Toolkit, and am having trouble usable transcriptions from my Olympus voice recorder. I first have to convert the Olympus WMA files to WAV, and I'm not sure that the result is recognized as "me" by WSR. I speak into the recorder with the same microphone that I use for direct WSR. My question is, do I have to train WSR to recognize the recorder output before I use WSR Toolkit? Do I have to set up a separate speech profile for the voice recorder output? I hope these questions make sense! Thanks, ed

mmarkoe wrote: ed wrote: I'm new to WSR Toolkit, and am having trouble usable transcriptions from my Olympus voice recorder. I first have to convert the Olympus WMA files to WAV, and I'm not sure that the result is recognized as "me" by WSR. I speak into the recorder with the same microphone that I use for direct WSR. My question is, do I have to train WSR to recognize the recorder output before I use WSR Toolkit? Do I have to set up a separate speech profile for the voice recorder output? I hope these questions make sense! Thanks, ed What Olympus recorder were you using? What software were you using to convert the WMA file to a WAV file?

Olympus VN-6200PC. This is much less expensive than the Philips & Zoom mics you mention, do I guess the sound quality can't be expected to be as good. For WMA to Wav I use WinFF. I'm running Vista 64 bit.

It does seem that the Olympus is the problem, though after a few experiments I see that I could also improve my dictation style. Using Vista's native sound recorder and also sndrec32 imported from my old system (no difference in output sound quality), I recorded the Rainbow text, also some paragraphs similar to those I was trying to capture from the Olympus. I found that these wav files gave much better results than those files taken from the Olympus. (I haven't had time to create a file in WMR, convert to WAV, and transcribe that.)

We have tested the transcription function with the Phillips 9600 recorder and the Zoom H2 recorder from Samson. The Phillips recorder used files converted to WAV I believe with the Philips 9600 recorder software. The Zoom H2 files were native MP3 files. We did not have to create a new user profile. As long as we enunciated clearly and kept the recorder in a consistent position (we used the built in microphones set to directional on both) accuracy was just as good as dictating using the WSR Software. Marty

It's very interesting that you can pull this off by using the built-in mics. I got the best results (such as they were) with the Olympus by using my external mic - a Zalman ZM-MIC1 glued onto an otherwise unuseable headset. This jury-rigged setup works pretty well for direct dictation.

Thanks much for your reply. And the WSRToolbox is a trip! - ed

mmarkoe wrote: ed wrote: Olympus VN-6200PC. This is much less expensive than the Philips & Zoom mics you mention, do I guess the sound quality can't be expected to be as good. For WMA to Wav I use WinFF. I'm running Vista 64 bit. The recording format of the Zoom H2, Philips 9600 and Olympus D-5000 recorders is at least CD quality (44 kHz). This high quality sound is what makes their dictation recordings usable in speech recognition software.