2

I want to build an application with a Textbox that captures the voice transcription create by Dragon Medical Practice Edition 4 (DMPE4) in a Windows environment. By default DMPE4 will place the text of whatever is transcribed into the current location of the cursor. This means that if a user clicks somewhere else other than where they want their dictation then the speech will be transcribed into a different window (or worse if no text box is available what they dictate won't appear anywhere).

I want a WPF application with a textbox that will capture the text regardless of where the current cursor location is, it should instead place the transcribed text into the textbox where the cursor was last. As long as the application is open and the window with the textbox is active it should not be possible to transcribe text anywhere else. I still want to use the DMPE4 application and all of the other features of it I just want the transcribed text to go into a specific textbox.

The few tutorials I found were based on System.Speech which is Microsoft's built in speech recognition, I want to use DMPE4 and I can't seem to find a good starting point.

DMPE4 has something similar called the Dictation Box, but in order to get the text from the Dictation Box to the location of where the text needs to be there are additional steps, I want the text to go directly into the Textbox of the new application.

Pryach
  • 391
  • 2
  • 8
  • 18
  • I don't kinow anything about DMPE, but I came across this: http://www.rwilke.de/dragoncapture/ it seems to do what you're looking for already, and there is a trial version you could check out. – Shrimperator Oct 19 '21 at 17:54
  • If the Dragon app can output the text to any textbox that has focus, I'm thinking there might be some 'Standard ouput' stream that you can hook into and then capture it. It's a longshot, but have you tried running the Dragon app using 'Process.Start' and then reading the standard output stream? https://learn.microsoft.com/en-us/dotnet/api/system.diagnostics.process.standardoutput?view=net-5.0 – mnemonic Oct 22 '21 at 11:53

1 Answers1

0

Since DMPE is a closed application and most probably simulates key strokes, you would need to somehow wrap it, capture its output and use it. My first idea to do so would be to run it in virtual machine (or maybe some container like docker would be sufficient), where it will just write the characters to your listener app and you can than transmit and use it somewhere else.

However, if you just need to create a solution based on Dragon Medical's speech recognition, I'd start with their SDK, which can give speech recognition capabilities to any software: https://www.nuance.com/dragon/for-developers/dragon-software-developer-kit.html

karolgro
  • 181
  • 1
  • 8