The image above shows our decisions on the numerical levels for the sound that both the added white noise and soundtrack should rise and fall to when people speak. We wanted to make sure that the soundtrack fell so low that it was barely even noticeable, however we decided to keep it slightly audio in order to ‘carry’ the visuals and the content.
The screenshot above captures the way I used keyframes on the sound levels to fade the interview audio and white noise in and out.
This above image shows how I cut the soundtrack to match when the interviewer talks and stops talking in order to fade the sound down to a low level and back up when cutaway shots are featured.
This final screenshot illustrates how we manually adjusted the levels of the interview sound in order to balance the audio and keep the levels at a comfortable level (between -18db and -12db) making sure that they didn’t exceed -6.