Fix it in the (re)MIX
Today, I had my first encounter with the “remix” function of the sound devices MixPre 10ii.
Shooting an interview backstage at the Pokemon European championships, we had a Spanish language contributor, accompanied by a translator, who was the go-between for our interviewer & the competitor after they’re matches. The environment we shot in was extremely loud, being directly behind the competition stages, of which there were 4 of them, each with their own separate PA’s firing out live commentary from the livestream of the event, along with a capacity crowd for each stage, all cheering along to the Pokemon action. The content we were shooting, was being pressed straight to camera (with Timecode), to which the footage is being edited over night & played out the next day online, so the editors need to use a functional mix of the interviewee’s answers for the cut (no time to do a proper dialogue edit).
My initial setup had both the contributor & the translator on radion mics going through MixAssist, allowing for a smooth way of mixing between the two for answers to be quickly transcribed & cut down. That automix track was then processed with NoiseAssist on one of the main mix channels feeding camera, allowing me to separate the contributors program sound from the incredibly loud background sound. The issue though being that the interviewee was incredibly soft-spoken, so much so that the Lavalier positioned right at the neck line struggled to pull his voice out from the background noise captured by the translators lavalier mic, so the automix wasn’t switching the mix over to his mic as he began to speak, meaning the first part of his answers was missed on camera while I manually shifted the mix over to them.
After a few questions I managed to optimise the mix so it focused entirely on the contributor, getting the most out of his quiet voice where it could be heard over the background. The kicker being those first few questions weren’t properly covered in the mix, & while the ISO tracks were available, if the editors had to switch to only the interviewee channel, they’d miss out on the noise reduction I’d applied, which helped tip the balance of the recording away from the considerable background noise & toward the desired program audio, & so I pondered if I could find a solution to this that didn’t require the questions being asked again on camera, thus saving data, time to transfer, edit, and most importantly, the performance that the segment producer was happy with… What if I could re-process the take through noise assist, with the optimised mix settings?
The answer being the remix function!
Once the interview was over, I checked all the settings over in the mix to make sure the unwanted elements were taken out, & then triggered the remix function of the MixPre, & re-recorded the entire interview, this time focused entirely on the contributor. The new mix recording kept the same Timecode stamp from the original file (without deleting or over-writing the original), while only creating 2 additional tracks of the new mix (not duplicating all the ISO’s). The process was done in realtime, which I left running while I had a meal break. Once it was done, I offloaded to DIT along with the original audio file & clear instructions for the editors, to simply use the timecode metadata to re-sync the camera rushes (a simple process) & use the new refined mix, with the original available if any additional details were wanted.
& so I finally had the chance to use the remix function in a scenario where it’s functionality would save the overnight online editors a bunch of unnecessary extra audio work with the time-critical task to create finished deliverables for the next days show.
The tools we have available for us now are really rather incredible, & can truly save many hours in the field when used creatively to solve problems, while giving producers & directors more leeway in the moment to keep things authentic.
