Audio level different between podcast episodes

[FangNet]

Mac M2 – Sonoma 14.1 - Audition CC - multitrack mode – wav input - mp3 output.

Refer to the two attachments. In the first session, audio output volume is as expected at earphones or other listening method. In the second session, audio output is considerably lower. When playing back audio within Adobe Audition, volume in the audio level panel is visibly and plainly different between these two sessions.

Both clips were recorded with the same audio settings (with Rodecaster 2).

Adobe Audition capability for volume matching was utilized for both sessions and set to:

                  ITU-R BS.1770-3 Loudness

                  Target Loudness: -19 LUFS

                  Tolerance:          1.5 LU

I do notice that individual segments for both clips are displayed with audio levels ranging from 12.3 db to 13.9 db. Perhaps this is due to speaking style at the microphone. I haven’t encountered a method for adjusting these with Audition.

Can the community comment on why the two sessions might have such different audio levels? Thanks.

[SteveG_AudioMasters_]

One thing I can see is that it looks as though the positive peaks in your waveforms somewhat exceed the negative ones, and that sort of assymetry is quite sufficient to slew the measured values. There are several things that can exacerbate this, but it's not that unusual. Male human voices and brass instruments often suffer from this.

It used to be relatively easy to effect a cure for this in Audition, but the tools as currently configured don't appear to allow you to do this. iZotope's RX suite does, though, using an all-pass filter. One thing you could try (although some sources report that this can make things worse) is to use one of Audition's peak limiters with the threshold set very carefully so that it just knocks off about 3dB of the peaks. That would hit the positive peaks and not the negative ones, without really altering the sound.