Help

Recordings with audio complexities

DoReveal provides highly accurate transcription by default. However, in cases with poor audio quality, multiple languages, or complex recordings, you can use Advanced Transcription to improve accuracy.

Enable Advanced Transcription


Note:
Advanced transcription is not enabled by default.
Contact support@doreveal.com to enable it for your account (no additional cost).





When to Use Advanced Transcription


Use advanced transcription if your recordings include:

  • Background noise or echo
  • Speakers far from the microphone
  • Multiple speakers with similar voices
  • Mixed languages in the same conversation
  • Less common or unsupported languages


Default vs Advanced Transcription

  • Default Upload
    • Automatically detects language and speakers
    • Best for clear audio and common languages
  • Advanced Transcription
    • Lets you control audio complexity and language settings
    • Improves accuracy in challenging scenarios


Upload Files with Advanced Options

  1. Add an Interview
    • Click the + (Add Interview) button in your study.
  2. Select Upload with Options
    • Choose Upload with Options instead of the default upload.


Set Audio Complexity


Audio complexity determines how the system processes your recording.

Option 1: Simple


Use when:
  • Audio quality is good (e.g., Zoom recordings)
  • Minimal noise or echo
  • You need support for additional languages

Steps:
  1. Select Simple
  2. Choose the language for transcription


Option 2: Complex


Use when:

  • Audio has noise, echo, or quality issues
  • Multiple languages are used in the same recording

Steps:

  1. Select Complex
  2. Choose one or more languages used in the recording
    (e.g., English + Hindi)

Please note:
  • Complex mode supports up to 60 languages (Standard transcription supports up to 100 languages)
  • Complex mode does not support redaction of sensitive information like PHI or PII


Set Number of Voices


Providing the number of speakers improves transcription accuracy.

  1. Disable Auto Detect
    • Uncheck the auto-detect option.
  2. Enter Number of Voices
    • Count all unique voices, including:
      • Moderator(s)
      • Participant(s)
      • Voices) in audio/video stimuli

No simply select the recording to upload. 
© Synthefai Inc.