Skip to main content
unauthorizedrob
unauthorizedrob
Inspiring
January 25, 2019
Answered

Any experts in L.A.? [Need help with the various stages of pre-processing optimization before loudness targeting]

  • January 25, 2019
  • 1 reply
  • 1010 views

i have been struggling with an audiobook, trying to get consistent voice quality -- running files through Match Loudness - but not getting a consistent resonance, timbre, tone.... I really am asking someone to help me finish this interminable job, submit it and get paid so we can pay the rent.  I'm in Studio City.

    This topic has been closed for replies.
    Correct answer paulfigg

    It sounds like you need help with the various stages of pre-processing optimization before loudness targeting. Keep in mind at it's core Match Loudness is nothing more than Loudness Normalization with supplimental tolerance margin options. It basicily applies a gain offset based on the measured Integrated Loudness relative to the user defined target. E.g. If the audio checks in at -20 LUFS, and the user defined target is -16 LUFS, +4 dB (LU) of gain is applied to meet the target. The user also defines a True Peak ceiling. If the gain offset results in overshoots - limiting will be applied at the defined ceiling.

    The key is the audio needs to be "optimized" before you apply Loudness Normalization. Yes, it's a generalization that represents subjective processing, such as leveling, compression, EQ, noise reduction, etc. Of course this all varies based on what you are working with (spoken word, music, etc.). In most cases you want to avoid the necessity for excessive limiting, especially when lossy (low bit rate) coding is in the pipeline. Ultimately if the audio was well recorded and all stages of processing were optimal, your processed (spoken word) audio will exhibit optimized intelligibility.

    -paul

    @produceNewMedia

    1 reply

    P
    paulfiggCorrect answer
    Inspiring
    January 25, 2019

    It sounds like you need help with the various stages of pre-processing optimization before loudness targeting. Keep in mind at it's core Match Loudness is nothing more than Loudness Normalization with supplimental tolerance margin options. It basicily applies a gain offset based on the measured Integrated Loudness relative to the user defined target. E.g. If the audio checks in at -20 LUFS, and the user defined target is -16 LUFS, +4 dB (LU) of gain is applied to meet the target. The user also defines a True Peak ceiling. If the gain offset results in overshoots - limiting will be applied at the defined ceiling.

    The key is the audio needs to be "optimized" before you apply Loudness Normalization. Yes, it's a generalization that represents subjective processing, such as leveling, compression, EQ, noise reduction, etc. Of course this all varies based on what you are working with (spoken word, music, etc.). In most cases you want to avoid the necessity for excessive limiting, especially when lossy (low bit rate) coding is in the pipeline. Ultimately if the audio was well recorded and all stages of processing were optimal, your processed (spoken word) audio will exhibit optimized intelligibility.

    -paul

    @produceNewMedia

      unauthorizedrob
      unauthorizedrobAuthor
      Inspiring
      January 25, 2019

      Thank you, Paul.

      I think that's where the problems are, you're right.  But you have really just stated more expertly the problem -- I am a novice, not a pro -- and finding a standardized approach to these files (52 chapters) is a nightmare for me.  I have tried and tried and in several cases gone all the way back to my original, unedited session files and cleaned them of all noises, adjusted levels manually again, and -- without applying ANY eq or processing -- mixed down and exported those WAVE files, and STILL they do not match up, before or after I run them through the Match Loudness process.  I have set those parameters to meet ACX requirements of a max peak of -3db and a total RMS between -18 and -23db.  I can achieve those numbers easily, but the QUALITY of the sound for some reason differs from the rich, intimate voice we started with, and I am frankly at my wit's end.   I am sure that a friendly expert in L.A. would whiz through this in an hour, but all the audio guys I know work in ProTools.

      We are dead broke now, so this would have to be an act of kindness or some kind of barter deal.  Professionally, I am a writer/director and only did this myself, ironically, to save money.

      Thanks very much for your response.  Your answer is no doubt correct, but the result I seem is an actual human angel to come in and take the sweaty reins away from me.

      Thanks,

      Rob

      P
      paulfigg
      Inspiring
      January 25, 2019

      You're welcome, Rob. And good luck. Have a look at something that I posted a while back. It displays a RT workflow as opposed to using an offline option such as Match Loudness. The documentation doesn't touch on all aspects of "optimization" (e.g. EQ, noise reduction, etc.). I hope you find it useful ...

      Intelligibility Optimization

      -paul.

      @produceNewMedia

      Terms & ConditionsAccessibility statement
      Using the community Terms of use Privacy Cookie preferences Do not sell or share my personal information ad choicesAdChoices