How to simulate distance with lavalier mic in post-processing

Question

I am recording a vlog (Basically a travel video while e.g. hiking) where me and the other person are both speaking into their own lavalier mic (DJI Mic; seperated on the left and right channel). One person is behind the camera, the other one is walking in front of it, although not always facing towards the camera (Basically I am filming the back with some distance, e.g. 3 meters or so). The movement is relatively dynamic and changes quite a lot (e.g. suddenly facing the camera, then again into the other direction etc).

I'm now having the "problem" that the two audio voices are basically sounding the same "perfectly", even though in reality, the other person would sound of course much different.

Is there a way to simulate/edit this? Or which effects happen in reality e.g. when the distance and the direction change? I just turned down the volume of the other microphone, which had a bit of an effect, but it still sounds a bit unnatural. I guess there are other effects as well (I think some frequencies getting more muted than the other or something like that?).

(I'm using Davinci Resolve/Fairlight for editing)

Answer 1

Contrary to some of the other answers, I wouldn't start with removing high frequencies with EQ. I would instead start with removing low frequencies.

If the lav mics are cardioid, then they will emphasize low frequencies in a way that a typical outdoor location shotgun mic does not. Even omni lav mics can end up with some low frequency emphasis if they are pinned to the chest of the speaker. Removing low end also helps manage rustle and rumble that may get into lav mics.

The other big difference in the lav mic sound is what the mic doesn't pick up as well, which are the ambient sounds. For example, if there is a road or the ocean or a forest nearby, it will help to be able to hear (at a low level) the sounds typical for these environments.

Finally, as some other have suggested, reverb can help - but for outdoor scenes it really has to be the right reverb. A typical algorithmic reverb will not easily sound like a forest, hill, cliff, beach, roadway, city, etc. Instead, careful use of a convolution reverb (the kind that uses IRs) with the right IR can really make a difference. I have found and used in the past IRs taken in woods and other outdoor spaces that have some reflective surfaces (like the ground) and this kind of reverb can really make the sound more natural. This is assuming you can find the right IR and take the time to adjust the amount and processing of the reverb.

I know I said "finally" above, but I do want to circle back to reducing high frequencies. On a very dry day (clear skies, bright sun, cooler weather, no haze), high frequency attenuation can be audible at closer distances. If the on-camera speaker is within 3 meters/10 feet of the camera, it's not worth it decreasing the high frequencies. At 10 meters/30 feet at 50% relative humidity, you may see 3-6 dB of attenuation above 10 kHz. So only when a sound seems to be pretty far away does high frequency attenuation help with the illusion of distance. Think of a horn from far off or someone shouting from the other side of a sports field/pitch.

When a person is facing away from the camera, their head and body also block highs, so you could automate rolling off some highs when they are facing away and bringing them back when they turn towards the camera.

I do want to say that since it's a vlog, there may not be much expectation by viewers of a natural sound.

Answer 2

Wild guesses:

EQ to remove some of higher frequencys. Add reverb to taste.

Answer 3

You can add a bit of reverb, take a bit of the high frequencies. You want the "remote" speaker to end up with a similar volume and comprehensibility as the "local" speaker, but you can "emulate" turning up the volume for the remote speaker by defaulting them to a lower volume, adding a bit of white noise, and "gating" the result (essentially turning up the speaker as well as their "personal" noise with a very slight delay and possibly ducking the "local" speaker's sound). Essentially you emulate a less than optimal remote sound quality in combination with a switch of speakers. That makes for an audible "explanation" of why you hear both speakers at similar volume when they are talking.

The problem is that this is the kind of setup you'd use for split-screen scenarios. Outside in nature, you tend to have very little reverb either way. Talking towards you will yield higher volume, but that kind of directional characteristic is much more pronounced with higher frequencies. You'd use a shelving highcut when turned away from the camera.

Answer 4

I'll add my two cents to these answers... When I think of sound over distance, I think of the smudging of transients. The smudging of the whole sound, actually. This can be achieved with some dry, diffuse, very short decay reverb. Don't use any of the original signal, pure reverb. We want phase interactions and diffuse signal grains to smudge the sound slightly. (This may not work properly with all reverb effect processors)

As others have said, the reverb is a big part, as the sound will reflect off surfaces before it reaches the receiver. With larger distances, only a small amount will be the original soundwave, and even that will be temporally smudgy. SMUDGY IS THE W0RD! A big part is knowing how your environment should affect the sound, too. But as the other mic is unaffected, I think you want a really subtle effect.

As for low or high shelving, both are necessary, as sound tends to "thin" with distance (unless the environment is a factor; standing waves etc). A static flanging effect would also help with the thinness (and yes, the smudginess).

I could go on about things like how you can't hear whispers at a relatively large distance, so there's a psychological factor to consider, but I'll leave it there. All you need is a slight smudging. You can automate this effect, too, so it changes correctly with distance.

Answer 5

One additional info (partly answer I guess) which I learned in the meantime, and would have probably kept me from asking the initial question:

The key to do everything is (or seems to be?) automation to change the parameters over time.

Even just changing the volume in response to the movement in the video helps a lot. Watching the video and then lowering the volume as the person is moving away or facing another direction gives immediate feedback on how it can sound.