How to improve your transcription quality?

The accuracy of transcription can vary greatly depending on the quality of the audio data and the recording environment. By implementing the following measures, you can improve the quality of your audio data and enhance recognition accuracy.
 
  1. Check the Transcription Language

    If the transcription result is completely different from the recorded content, or if it is not transcribed in the desired language, there is a possibility that the correct language was not selected before starting the transcription.

    Please select the language spoken in the recording or video as the "Transcription Language" before starting the transcription, and try again.

    After transcription, the Transcription Language can be confirmed at the top of the record details page in the Notta Web version, to the right of the creation date and time.

     

  2. Record while being mindful of the following points
    • Reduce ambient noise
    • Speak directly into the microphone
    • Speak slowly and clearly
    • Minimize dialects and accents, and avoid using slang
    • Avoid two or more people speaking simultaneously
    • Use a directional microphone whenever possible, and record with the microphone positioned close to the speaker.
       
  3. Upload a file with good audio quality

    Please prepare audio files of good quality, meaning they should not have noise or be too quiet.

    Avoid compressed formats such as MP3. When recording, selecting the uncompressed WAV format is generally best.
     

  4. Register Vocabulary

    During meetings, company-specific terminology and specialized terms are often spoken, which can be challenging for speech recognition engines. You can improve transcription accuracy by registering these terms and phrases in the My Vocabulary beforehand.
     

  5. Check the Recommended Operating Environment

    If you are not using the recommended operating environment, transcription may not work correctly. Please review the recommended operating environment in the help article Supported Environments.

    If you are using the app version, please update to the latest version.

 

If you have tried the methods above and transcription accuracy remains low, please share your audio/video file and the transcription result with Notta Support. Our support team will conduct a more detailed investigation.
If it is difficult to share the original file, we ask that you re-record and transcribe content that can be shared under the same environment and conditions, and send that to Notta Support as investigation material.

Articles in this section

Was this article helpful?
6 out of 15 found this helpful
Share