TDAC;Help needed salvaging comedic dialogue

[person]

After wasting a lot of time trying things to improve the sound of these voices, I have come to the conclusion that I am never gonna get it done to my satisfaction alone.

The track consists of two speakers with simple music to be mixed into the background.

The voices were recorded about 6 years ago, and I am not sure how I managed to record them so poorly. The act was largely improvised and we didn't want to re-record it. Pitch shifting was part of the act and it appears I did not save the unshifted tracks. I'm not sure why, because I usually save everything.

The background music track I uploaded was redone this year.

The voices have a lo-fi sound and span a significant dynamic range. What is causing the weird sound on lines like Tommy's "what was the noise" and Shawn's "clothes that don't fit and stuff", among others?! ..and is there
anything to do about it?

On the voices, I have experimented with boosting the highs and lows and cutting mids, applying some compression and/or riding to reduce the large dynamic range, using some noise reduction to reduce the noise brought up from the dynamics processing, eqing the plosives out, and rolling off the extreme top and bottom(the mics only pick up to about 15k and after shifting Shawn down there probably isn't much useful in his track above 12.7k or so.). I even experimented starting by eqing down the frequncies that were really peaking in the loudest spots.

All of this seemed to help some, but not enough. Compression never sounds good to me, and riding seems to artificially bring the ambience up and down too much.

Maybe someone here can mess with these files and have better luck and share with me what they have done, so I can learn from all of this. I want to finally have this project(among others) completed

I was gonna place the voice files to begin at about 260585 samples into the music.

Thanks in advance to anyone who tries to 'mix' this.

[person]

Nobody even listened?!

[Geert van den Berg]

As discussed per our PM, here's an example, offcourse quality is very low, going from MP3 to WAV and then back to MP3...

[Geert van den Berg]

Here's the screenshot with faderrides