audio quality issue when recording voice narration



  • Paul Middlin

    You may have some success making edits in Audacity to account for differences. We also recommend these other tips:

    • Make sure the mic is in the exact same position in the exact same room each time you record
    • Make sure you are in the exact same position relative to the mic
    • Practice a ton; consistency will really only happen if you are recording and developing your voice. Voice over artists are great at this because they practice their voice so much, so they are able to sound consistent as long as the first two things are in place.

    It's tricky, and challenging, to be sure, to get really consistent results.

  • I have the same problem, Paul Middlin

    Following your advice, I have different results for the same mic - location, room, and settings. If I record twice you get different tones an also the noise removal audio effect produces different outcomes.

    I'm using a Samson UQ2 and Camtasia 2020.

    any clue would be very much appreciated.

    kind regards

    Carlos de la Barrera.

  • Erik R.

    I would try turning off the "Auto-normalize loudness" feature and see if that makes your narration sound more consistent.

    You can do this by right-clicking on the Canvas (where the video is previewed in the middle of Camtasia), choosing "Project Settings" and unchecking the box next to "Auto-normalize loudness":

  • bebecker

    Interesting. Any idea why turning off this setting would make the voice narration more consistent? 

  • Erik R.

    Auto-normalize loudness will try to detect the overall loudness (i.e. range and consistency of dB level) of your clips and try to bring them into a standard level.

    However, this feature works independently for each clip, meaning it may turn the volume up or down for each separate clip depending on how loud each one is. So a very loud clip may get turned down while a soft clip will get turned up. The ones that get turned up may have more noticeable background noise as the algorithm used does not distinguish between speaking volume and overall volume. 


Please sign in to leave a comment.