Don't Assume Dialogue Consistency in Multi-Cam Shoots

I learned the hard way that assuming consistent dialogue quality across multi-camera setups is a recipe for a tough mix. The problem was a documentary where we had an AMIRA for A-cam interviews and a LUMIX S1H on a wide, static shot, both recording scratch audio for sync. During the shoot, the director loved the spontaneity of certain moments captured by the S1H's wide angle, so some lines from that camera's scratch track made it into the edit, despite my Mixer using an 888 feeding MKH 416s for the primary audio.

What went wrong was my initial approach to dialogue editing. I focused heavily on cleaning the primary lav and boom tracks, trusting the scratch audio would only be for reference. I didn't treat the S1H's scratch track with the same critical ear, thinking it wouldn't be used for anything beyond a temp sync. When I got to the final mix and realized several key lines were sourced from the S1H's internal mic (which, predictably, sounded thin, roomy, and completely different from the professional recordings) I had to scramble. I spent days trying to match the spectral profile, adding subtle reverb, and EQing just to get it palatable enough to cut in without a jarring shift.

The solution is now a non-negotiable step in my workflow: always perform a quick pass on all scratch audio from every camera used in the edit, even if it's just a preliminary listen for quality and consistency. If a camera's scratch track is used for even one word, it needs to be processed to match the primary audio as closely as possible. It's extra time up front, but far less painful than trying to salvage mismatched audio in the final mix. Does anyone else have a similar experience with unexpected scratch track usage?