Mark Edward Lewis is the owner of online education website Cinema Sound. He has over 30 years of experience writing and producing musical scores, re-recording, mixing, developing sound design, editing, writing, and directing. Through Cinema Sound and its production partner, MZed, Lewis has created over 85 hours of audio education videos for independent media creatives. He was the post-production supervisor on Marvel’s Avengers STATION interactive experience and, in 2016, he led the Sound Advice tour through North America and Australia where he taught over 1,200 filmmakers to improve their production value with better sound.
PS: What are some common mistakes that you see even experienced film audio professionals making when either recording on-set or in the field?
Mark Edward Lewis: I don’t know any sound recordists who consider sound delay when it comes to the on-camera reference microphone, whether it’s in-camera (please god no!) or on-camera. I’ve seen situations where they were very happy to have a Rode video mic on a camera and say, ‘This is going to be the reference,’ and the recordist has lavs on the actors and there are booms and they are 16 ft. or more away from the camera.
If we know that at 72 degrees at sea level, sound travels at 1.1308 milliseconds per foot, what we can very quickly realize – especially if we know that at a 24p [framerate], that’s about 42 milliseconds – is that at 16 ft. or so, that means your audio is hitting the camera reference half a frame late. At 36 ft., it’s a full frame late.
I’ve had people say, “But look, I’ve got PluralEyes or I use Adobe Audition,” and I say, “Yeah, but it’s syncing it to your reference, which is one frame late and by the time it comes to us in post, we’re like, ‘Why are all the lips rubbery? Who recorded this?’” What we find out later is that they just thought it was a reference and no big deal. They just put a mic on the camera or, worse, used the in-camera mic. Sound moves pretty slow and at 36.8 ft. you’re one frame back and there is no solution for it. The only thing that we can do is, if you absolutely cannot run your mixer to the camera for whatever reason or you’re too far away or the camera is moving or they don’t want it and the actors are far, then notate that in the log. We’ve got to know in post that we’re listening to audio that is probably half, three-quarters, or a full frame late or more because it was on-camera audio.
PS: Likewise, what is a common mistake that even professionals make in the studio?
MEL: It used to be that people would put an AKG C12 in front of a person to do ADR, but the whole purpose of ADR is to recreate what was done on-set. So, fortunately, now we’re starting to see people at studios everywhere who are putting lavs on them or they’re finding out what the original shotgun mic was and either pulling it out of their mic cabinet or renting it so that it’s the exact microphone that was used. That really helps us. But I think the primary thing that I see is mixing levels being wrong. I’ve been into several studios where they were doing a mix for the internet and the studio is tuned at 85dB dialogue, and that is great for cinema – though it’s actually too loud for me because I can’t mix for a 10-hour day at 85dB; I can do 80, but I can’t do 85. But on the internet, man, nobody listens to anything at 85dB. At worst, they listen to it on tiny little speakers, or even worse, headphones.
If you’re listening to your mix at 85dB, by the time it gets out of that little speaker, the dialogue is going to be so loud compared to the music, images, and foley because in a full-bandwidth environment at a strong decibel level, you’re going to have a lot more bass frequencies and a lot more high frequencies. But by the time we get to a tiny little speaker, obviously it’s stripped out all the low frequencies, maybe even from 200 Hz down, and it exacerbates the frequencies of dialogue. So if we were to mix at a lower level and take into account the roll-off of these speakers that most people are going to be listening on, we’re going to find that our dialogue mixes are way too loud and we need to mix them down or mix the music up or do really clever, what I call ‘notch EQing,’ which is layering the EQ into the music or sound effects or the images so that we can bring up the music – the emotional and immersive elements – without disturbing the dialogue, which might be not so great for the cinema but for the internet delivery, or even some DVD deliveries, we can accommodate that.
I hear so many online mixes on my laptop where I’m like, “Wow, nobody thought about this and it’s on YouTube.” No one is going to watch a YouTube video in the theatre. Why would we ever put a cinema mix or even a DVD mix on YouTube? You wouldn’t, hopefully. So, we shouldn’t be referencing our mixes at 85dB. They should be 75dB and some kind of roll-off.
In part two in the December 2018 issue, Mark discusses the common causes of poorly-recorded dialogue and, on the flip side, how to fix poorly-recorded dialogue in the studio.