thesoundguy
Carbonite Member
I was on set today, and, like my usual sound guy self, was complaining that we needed to close up a few things to get better sound. The place we were filming at is on a major street with the front door facing the road. Electric had routed many cables through this front door and were unable to close it. I mentioned this issue when I first got wind that this was electrics plan.
The response I got was "We can flip the room tone sound 180 degrees in post and that will get rid of the traffic"
Being a post guy as well, i know this is not as easy as they think. This post is to inform you of how noise removal actually works, so listen closely my videographer friends.
It is true that an inverse phase (flipping the sound 180 degrees) will in fact destroy background noise. Sound waves work just like regular waves in water in that when they collide they add and subtract in size. two identical waves of the same size traveling in opposite directions will cancel each other out. two IDENTICAL waves. in an uncontrolled, uncomputerized environment, you will never record two identical audio wave forms. flipping them 180 degrees does nothing, in fact it will make it worse, regardless of changing the phase.
Again, think of how waves in the water act during... lets say... a storm. when two turbulent waves collide you get what... a new turbulent wave. When you translate this to audio, your background is one turbulent wave and your room tone is another. these two waves are their own sounds that are easy recognizable, traffic. when these two waves crash together you get a jumble of frequencies, some added from the collision, some removed because of the collision. The result is a sound that is often unrecognizable and worse than what you started with.
Now lets put all that mumbo jumbo aside for a minute (or if you don't understand just move on to the next point). What if for some crazy mysterious reason, you did acquire a recording of your room tone that you could make exactly 180 degrees out of phase, what happens when you try and place it over your location audio with dialogue??? it comes out sounding like your actors were inside a dryer.
This is because you have frequencies of noise (in my case traffic) that are in the same frequencies as your actors voice. The dialog and the unwanted noise blend together in the same wave. When you apply this 180 degree out of phase "noise destroyer", it cannot differentiate between what frequencies are required for the dialogue and what frequencies are not. It simply destroys all parts of the wave that it blends with, effectively leaving only parts of your dialogue wave in tact often leaving your actors voice sounding muddy and dirty. (if you have tried noise reduction you know what I am talking about)
The way noise reduction processors work is by using a lot of heavy math to generate that magic signal I mentioned that will destroy your noise. the problem is that you still have the issue with destroying your dialog data if the noise overlaps in frequency. low pitched hums and high pitched sounds are simple to remove, but you cannot easy remove frequencies that interferes directly with the dialog.
I know that a lot of this is complicated and a lot of what I just said will go over a lot of peoples heads. I'm not the best writer, but if there are questions, ask them and I will try to explain a little better. Also it is way past my bed time and I am falling asleep as I write this. it is entirely possible i left out some details, but hopefully this gives you an idea as to what actually goes into getting rid of all that noisy stuff in your dialogue tracks, and why your sound guy gets frustrated with you.
The response I got was "We can flip the room tone sound 180 degrees in post and that will get rid of the traffic"
Being a post guy as well, i know this is not as easy as they think. This post is to inform you of how noise removal actually works, so listen closely my videographer friends.
It is true that an inverse phase (flipping the sound 180 degrees) will in fact destroy background noise. Sound waves work just like regular waves in water in that when they collide they add and subtract in size. two identical waves of the same size traveling in opposite directions will cancel each other out. two IDENTICAL waves. in an uncontrolled, uncomputerized environment, you will never record two identical audio wave forms. flipping them 180 degrees does nothing, in fact it will make it worse, regardless of changing the phase.
Again, think of how waves in the water act during... lets say... a storm. when two turbulent waves collide you get what... a new turbulent wave. When you translate this to audio, your background is one turbulent wave and your room tone is another. these two waves are their own sounds that are easy recognizable, traffic. when these two waves crash together you get a jumble of frequencies, some added from the collision, some removed because of the collision. The result is a sound that is often unrecognizable and worse than what you started with.
Now lets put all that mumbo jumbo aside for a minute (or if you don't understand just move on to the next point). What if for some crazy mysterious reason, you did acquire a recording of your room tone that you could make exactly 180 degrees out of phase, what happens when you try and place it over your location audio with dialogue??? it comes out sounding like your actors were inside a dryer.
This is because you have frequencies of noise (in my case traffic) that are in the same frequencies as your actors voice. The dialog and the unwanted noise blend together in the same wave. When you apply this 180 degree out of phase "noise destroyer", it cannot differentiate between what frequencies are required for the dialogue and what frequencies are not. It simply destroys all parts of the wave that it blends with, effectively leaving only parts of your dialogue wave in tact often leaving your actors voice sounding muddy and dirty. (if you have tried noise reduction you know what I am talking about)
The way noise reduction processors work is by using a lot of heavy math to generate that magic signal I mentioned that will destroy your noise. the problem is that you still have the issue with destroying your dialog data if the noise overlaps in frequency. low pitched hums and high pitched sounds are simple to remove, but you cannot easy remove frequencies that interferes directly with the dialog.
I know that a lot of this is complicated and a lot of what I just said will go over a lot of peoples heads. I'm not the best writer, but if there are questions, ask them and I will try to explain a little better. Also it is way past my bed time and I am falling asleep as I write this. it is entirely possible i left out some details, but hopefully this gives you an idea as to what actually goes into getting rid of all that noisy stuff in your dialogue tracks, and why your sound guy gets frustrated with you.
