Transcribing recorded speech

Dragon can transcribe spoken audio recordings of yourself or someone else.


Record yourself

When you're away from your computer, dictate into a recording device or mobile app, then transcribe the audio with Dragon later. Record things like notes, observations, or memos.

Record someone else

Record someone else with your recording device or mobile app, then transcribe the audio with Dragon. Record things like lectures, presentations, or interviews and use the text in documents, reports, or reference notes.

Technical requirements for audio files

Supported audio file formats .wav, .mp3, .wma, .dss, .ds2, and .m4a
Minimum audio file sampling rate 16.000 kHz

When the recorded voice is not yours

If you are transcribing an audio file and the voice speaking is not yours, consider the following options and constraints.

Which user profile to use

Use your user profile:
You can transcribe the file with your own user profile. Speech data from the recording won't affect the speech data in your profile, as long as you choose Someone else in the Select the speaker field.

Use a transcription user profile:
If you will have several recordings of a person to transcribe, you will get the most accurate results using a transcription user profile for that person. To create this profile, from the DragonBar, select Settings>Profile>New User Profile and select the appropriate check box when prompted.

You cannot create a transcription-only user profile when Nuance Management Center/NMS mode is enabled.

How to make corrections

When you transcribe a recording of someone else's voice, you will only be able to correct the output manually, using mouse and keyboard.

Tutorial: How can I transcribe a recording?

To transcribe a recording

  1. If necessary, open the correct user profile:
    • Open your own profile to transcribe a recording of yourself, or of someone else for whom you don't have a transcription profile.
    • Open a transcription profile to transcribe a recording of someone else, when you have a transcription profile for that person.

  2. From the DragonBar, select Tools>Transcribe Audio>Transcribe Recording.
  3. Click Select the speaker and select who the voice in the recording belongs to – Me or Someone else.

    If you are working in a transcription user profile, Dragon does not display this field.

  4. In the Input audio file field, enter the file name of the recording and the directory path where it's located, or click Browse to navigate to it.
  5. In the Output text file field, enter a file name for the transcribed output file and enter the directory path where you want Dragon to save it, or click Browse to navigate to the directory.
  6. Optionally deselect Automatically add commas and periods if you do not want Dragon to add this punctuation to the transcription.

    Audio quality and dictation manner affect the punctuation results. Also, when transcribing multiple audio files in one session, accuracy may degrade when this option is selected, so you should deselect it.

  7. Optionally click Options to configure how you want Dragon to handle spoken commands it detects in your recording.
  8. Click Transcribe and if the file is encrypted (a .DSS or .DS2 file), specify the password.

    Dragon transcribes your recording and then prompts you to choose what you want to do next.

  9. Select an option and click Done.


Related procedures

Additional information

About microphones and other dictation sources

7/13/2017 12:32 PM