Hello. I have a Son SX712 and have been trying to create a profile specifically for it. But, no matter what I do, Dragon will not let me get to the point of activating the profile. Following the instructions, I read the training text into the recorder, transfer it to the computer, point Dragon to the file and let it read it...at which point, it always says (just after a few seconds) "This profile has been insufficiently trained...please rerecord the training text, being careful to ...(speak clearly, position microphone, etc). Now the recording is done in a very quiet environment and the audio file is crisp and clear. Yet I have tried this several times with the same result. I do note it is an MP3 (if the DVR can output wav, I haven't yet discovered how).

Interestingly, it will transcribe from the file (though with a greater than average number of errors) when in an already established profile. But that doesn't seem like a good way to do it, as that profile was not set up for the Sony recorder. Does anyone have any idea what the problem might be?

Hi there, Ive done alot of DVR work with Dragon and here I can say the following:

1) I have found no difference in accuracy between a trained voice rec profile and the profile I use for usual dictation through a headset/microphone.

2) If not using already get a good quality headset mic or mic stubb to go into the recorder and use that. The mics built in to the voice recorder are just not up to it. Also a headset mic can stop you waving the recorder mic all over the place as you dictate.

3) Expect slightly less accuracy than when going straight into the PC with your headset. But not much less. 

Im sure you will hear that the Sony's arent the best for DVR work with Dragon, but before you go changing you should still be able to get pleasing results with the above approach. I have an old Sony (circa 2002) as a back up to my Olympus and its not to shabby at all.

R 

Click KB 2012 REV D to Download a 30 Day Evaluation of KnowBrainer 2012 

I am confused by this reply. The Sony ICD SX712 is listed under Nuance's own hardware compatibility list with "six dragons" (the top rating available) here:

http://support.nuance.com/compatibility/Search.asp?PRO=DNS&CID=7

(it is the second one from the top of the list).

Forgetting about any controversy about the Nuance hardware recommendation list for a moment, ..., and going back to your original question, why is it you cannot get past the initial training with the recorder? - What audio format are your recording the training text in? How long is reading of the training text? Which version of Dragon are you currently using actually?

Rüdiger

Well, it's past the point where we can make any changes in the code, but we can still make changes to the Easter Egg!

Hi. I do not know why I am unable to get past that training. I recorded the training in mp3 320kbps (the recommendation by Nuance). However, I have also converted that to a .wav. Still no luck. The length of the file I have varied between about 5 mins to 30 mins (the entirety of the "Arthur C Clarke" text). Dragon 11.5 premium.

I have also tried to "add a new dictation source to an existing profile." The help instructions within dragon for that menu appear to suggest that "digital audio recorder" will appear as one of the options in that menu. However, it DOES NOT! The options that appear are *microphone plugged into mic-in jack* *microphone plugged into line-in jack* *bluetooth microphone* *dragon remote microphone application* *handheld or smartphone with recording application" end of list.

What particular WAV format have you been using? You should make it one of these:

PCM 11.025 Hz, 16 bit, mono
PCM 22.050 Hz, 16 bit, mono

If "digital recorder" is not available from the list, this usually indicates that the profile you are currently using has been set up as digital recorder already.

Rüdiger

Well, it's past the point where we can make any changes in the code, but we can still make changes to the Easter Egg!

Hello. In the software I am using to convert, I do not see an option that corresponds exactly to what you posted above. I chose the closest possible from the list, which was "PCM, signed, 16 bit, mono". Same result as before.

You are right about the item on the list. It seems that once you "try" to add it, it takes it as "added" even if the process does not successfully complete itself (which it does not with me).

Then try with a different converter, such as Audacity which is free. I am bringing this up just because recently there was a user coming onto the German forum that I run describing exactly the same issue. However, using the appropriate WAV audio format solved it.

Rüdiger

Well, it's past the point where we can make any changes in the code, but we can still make changes to the Easter Egg!

Hello. I am trying to comply with your suggestion, as I also have Audacity on my system as well as Goldwave. I just don't use it often. Is there a way to specify that output exactly? Again, in the "export" option, I simply get one .wav possibility: "WAV (Microsoft) signed 16 bit PCM".

EDIT: I see, that choosing "uncompressed other" gives some more .wav options...but nothing closer than the above.

I am not certain, but it may be because of the original format you are converting from:

http://forum.audacityteam.org/viewtopic.php?f=20&t=34101#p91660

Can you try recording in a different format? - Other than that, just go on using a profile set up as microphone source for transcription, it won't be so much different.

Rüdiger

Well, it's past the point where we can make any changes in the code, but we can still make changes to the Easter Egg!

I will send you a training recording of mine in DS2 format via e-mail that you can use for testing purposes, just to determine whether or not the bug is with the software or with the audio format.

Rüdiger

Well, it's past the point where we can make any changes in the code, but we can still make changes to the Easter Egg!

Click KB 2012 REV D to Download a 30 Day Evaluation of KnowBrainer 2012