I've got a rough cut of a short I'm producing independently, doing all of my own editing etc (cant afford a production team yet). There's some excess background noise and white noise I'd like to get rid of while retaining as much of the speakers' vocals as possible. Any suggestions for best settings or even program to use for this? Thanks
-
Wondering which camera, gear, computer, or software to buy? Ask in our Gear Guide.
Best Way to Removie Excess Noise?
- Thread starter kinosight
- Start date
It depends on the nature of the noise.. if it's other people talking and moving around and whatnot, there's probably not a whole lot you can do. Hums, buzzing, static, and the like call all be reduced somewhat with filters and some EQing.
What software do you have available to work with? You're going to want something that's specifically for working with audio.
What software do you have available to work with? You're going to want something that's specifically for working with audio.
Two simple methods: using your NLE's graphic equalizer, find which frequency the background noise primarily resides and drop it as much as you can without disrupting the overall sound picture.
You can also ADD enough appropriate room tone/aural atmosphere to cover up the white noise. Where does the scene take place?
You can also ADD enough appropriate room tone/aural atmosphere to cover up the white noise. Where does the scene take place?
See if you can get into a one room cafe before it opens and do some listening...is the coffee maker dripping? Dishwasher or refrigerator running? Talk radio on softly in the background? You can create a believable space aurally and cover up that white noise at the same time. Let us know how it goes...we'd love to see (hear) the results.
Get yourself a (free) copy of audacity and play around with the noise removal tools, here's info about noise removal in audacity: http://wiki.audacityteam.org/index.php?title=Noise_Removal
Or, you could try playing around with REAPER. It's not free software, but the demo is freely available and totally uncrippled.
Or, you could try playing around with REAPER. It's not free software, but the demo is freely available and totally uncrippled.
Last edited:
Waves has all kinds of plug-ins including noise reduction software. It can only be used to a degree; After a while artifacts get added. You get a "phasey" sound.
You could always try ADR'ing it. Lotta work though. Gotta use the same microphone at the same distance from talent as other scenes in the movie.
You could always try ADR'ing it. Lotta work though. Gotta use the same microphone at the same distance from talent as other scenes in the movie.
The lesson learned here is to record it correctly in the first place. Read my blog here on IndieTalk to get yourself up to speed.
To reduce hiss you would use a Low-Pass Filter. Start at about 12kHz and slowly work your way down. Be careful you do not "dull" the tone of the dialog.
Hums tend to be 60 Cycle Hums; most heavy electronic machinery such as refrigerators work at 60 cycles. Use a Hi-Pass Filter to reduce low frequency hums, which can include wind noise and traffic. Start at 60Hz and slowly work your way up. You can go up to about 180Hz without "thinning out" the voices - unless they sound like James Earl Jones.
Other frequencies can be reduced using a Parametric Equalizer. You identify the offending frequency and use a narrow Q (band-width).
Using EQ is a balancing act of reducing noise and retaining voice clarity.
Pops and clicks can be "ducked"; that is reducing the volume for the millisecond or so that they last using volume automation (As an experienced audio guy I also do "patching"; I remove the pop or click and time stretch to fill the gap). You can also duck the noise between lines of dialog, again using volume automation.
There are numerous Noise Reduction plug-ins out there such as Izotope RX (fantastic at the price - about $300) all the way up to the Waves Restoration Bundle ($1,350) and the Cedar Cambridge System (a full-blown system can be as much or more than $25,000).
None of these options - except for maybe the Cedar - is a cure-all; it only mitigates the problem.
To reduce hiss you would use a Low-Pass Filter. Start at about 12kHz and slowly work your way down. Be careful you do not "dull" the tone of the dialog.
Hums tend to be 60 Cycle Hums; most heavy electronic machinery such as refrigerators work at 60 cycles. Use a Hi-Pass Filter to reduce low frequency hums, which can include wind noise and traffic. Start at 60Hz and slowly work your way up. You can go up to about 180Hz without "thinning out" the voices - unless they sound like James Earl Jones.
Other frequencies can be reduced using a Parametric Equalizer. You identify the offending frequency and use a narrow Q (band-width).
Using EQ is a balancing act of reducing noise and retaining voice clarity.
Pops and clicks can be "ducked"; that is reducing the volume for the millisecond or so that they last using volume automation (As an experienced audio guy I also do "patching"; I remove the pop or click and time stretch to fill the gap). You can also duck the noise between lines of dialog, again using volume automation.
There are numerous Noise Reduction plug-ins out there such as Izotope RX (fantastic at the price - about $300) all the way up to the Waves Restoration Bundle ($1,350) and the Cedar Cambridge System (a full-blown system can be as much or more than $25,000).
None of these options - except for maybe the Cedar - is a cure-all; it only mitigates the problem.
Last edited:
It would be helpful if you could post a sound clip so we could give you a more specific set of strategies rather than the general directions we're giving you now.
What software do you have to work with, Mac/Linux/Windows, Final Cut, Premeire, Vegas, Avid, WMM?
We have very little information to give you as we have very little to work with from you up front.
What software do you have to work with, Mac/Linux/Windows, Final Cut, Premeire, Vegas, Avid, WMM?
We have very little information to give you as we have very little to work with from you up front.
mute the whole scene , re record clean voices and build up the background noise from audio libraries,
Hey Alcove,
I'm getting ready to start on sound post for my feature and am planning to use Soundtrack Pro. My audio is fairly clean, but is mostly outdoor location so I'd like to isolate it even more. Do you ever use STP and what do you think of it? Just curious.
Sorry for the hijack.
I'm getting ready to start on sound post for my feature and am planning to use Soundtrack Pro. My audio is fairly clean, but is mostly outdoor location so I'd like to isolate it even more. Do you ever use STP and what do you think of it? Just curious.
Sorry for the hijack.
I've used STP a couple of times and found it rather buggy and limited - although the limitations were due more to the fact that I have a lot of third party plug-ins for my personal Pro Tools system.
And that is a large part of the equation, the tools available to you. Yes, STP will do the job. The EQs are okay as are the reverbs, although the 'verbs tend to be CPU hogs (not that some of the plug-ins I use are any more frugal). What is equally important is your listening environment and your speakers. You need to be in a fairly quiet space, and it needs to have at least some sound treatment; otherwise you will not hear the audio accurately. And the speakers need to be relatively flat and accurate as well. And the system needs to be CLEAN; in other words you don't want to be correcting for noise created by the audio monitoring system.
Isolating the dialog from the background/ambient noise takes lots of time and patience. The whole job is to eliminate the noise without compromising the tonal quality of the dialog. Having copious, well recorded room tones (indoors) and ambient tones (outdoors) makes the job a lot easier.
First, make sure that you work non-destructively, in other words you can always undo what you have done.
Start by making an original production sound session and then never use it again except as a reference.
Do a "Save As" of the original session. I generally do 8 to 15 minute sessions.
In each scene you will checkerboard the dialog - each character gets his/her own track.
Do the noise reduction.
Eliminate all of the ambient sound between lines of dialog.
Pull in your room tone.
Mix to taste.
It's that simple and a lot more complicated. Yes, it's hard and sometime boring, but you'll get better with experience.
And that is a large part of the equation, the tools available to you. Yes, STP will do the job. The EQs are okay as are the reverbs, although the 'verbs tend to be CPU hogs (not that some of the plug-ins I use are any more frugal). What is equally important is your listening environment and your speakers. You need to be in a fairly quiet space, and it needs to have at least some sound treatment; otherwise you will not hear the audio accurately. And the speakers need to be relatively flat and accurate as well. And the system needs to be CLEAN; in other words you don't want to be correcting for noise created by the audio monitoring system.
Isolating the dialog from the background/ambient noise takes lots of time and patience. The whole job is to eliminate the noise without compromising the tonal quality of the dialog. Having copious, well recorded room tones (indoors) and ambient tones (outdoors) makes the job a lot easier.
First, make sure that you work non-destructively, in other words you can always undo what you have done.
Start by making an original production sound session and then never use it again except as a reference.
Do a "Save As" of the original session. I generally do 8 to 15 minute sessions.
In each scene you will checkerboard the dialog - each character gets his/her own track.
Do the noise reduction.
Eliminate all of the ambient sound between lines of dialog.
Pull in your room tone.
Mix to taste.
It's that simple and a lot more complicated. Yes, it's hard and sometime boring, but you'll get better with experience.
Thanks for the tips, AA!
I have some decent noise-canceling Senny headphones I'm using during the sound edit. Not sure yet how/where I'll do the final mix. My studio is not the environment for that, at present.
I've actually already checkerboarded the dialog tracks -- did it during the pix edit to save me time later (I've done a lot of editing - this is my third feature).
STP has that bg noise suppression feature that I'm anxious to try out to minimize generator noise and white noise. Just wondered if you'd used it.
Thanks again!
I have some decent noise-canceling Senny headphones I'm using during the sound edit. Not sure yet how/where I'll do the final mix. My studio is not the environment for that, at present.
I've actually already checkerboarded the dialog tracks -- did it during the pix edit to save me time later (I've done a lot of editing - this is my third feature).
STP has that bg noise suppression feature that I'm anxious to try out to minimize generator noise and white noise. Just wondered if you'd used it.
Thanks again!
I believe my composer used a program called "sound soap" that cleaned everything up real nice.
Go back to my previous post in this thread (#8, I believe) for getting rid of hums and hiss. I've never used the NR in STP; I use SoundSoap Pro (not too bad) and Digidesigns DINR (also decent) as well as lots of EQ which works better in a lot of cases. On my "to get" list is Izotope RX which has been raved about by many of my peers.
You are better off doing multiple passes of mild noise reduction rather than trying to do radical passes all in one shot.
Just watched the feature highlight video on their website.. I wasn't terribly impressed, since most of that functionality was in Adobe Audition a few years ago...
Then, last feature was demoed and it blew me away.
"iZotope RX has the ability to regenerate missing audio..."
Obviously that wouldn't work on dialogue and the like, but in the example shown.. WOW!
The NR in STP is pretty good, but suffers from the same problems as all the others, you can only push audio so far before it's painful to listen to.
I just saw a tutorial on how to use the spectrum analyzer/ frequency graph within STP to remove specific frequencies directly from the graph!
http://www.youtube.com/user/AppleShakeGuru#p/u/15/_s7pVw-WpUU
I just saw a tutorial on how to use the spectrum analyzer/ frequency graph within STP to remove specific frequencies directly from the graph!
http://www.youtube.com/user/AppleShakeGuru#p/u/15/_s7pVw-WpUU
Thanks knightly, I've already seen that tutorial a couple of times -- I'm a fan of Eric Meek on YouTube. His tuts have helped me immensely!
Looking forward to starting on audio for this project. It's a supernatural thriller so sound is gonna be 75% of the experience, IMO.
BTW, that homemade highlight softening filter you described to me in another post is AWESOME. I love how it smooths out kicks and how you can adjust it to soften detail in unimportant areas. I use it in combination with Eric's complexion softening Color filter technique on nearly every shot.
Looking forward to starting on audio for this project. It's a supernatural thriller so sound is gonna be 75% of the experience, IMO.
BTW, that homemade highlight softening filter you described to me in another post is AWESOME. I love how it smooths out kicks and how you can adjust it to soften detail in unimportant areas. I use it in combination with Eric's complexion softening Color filter technique on nearly every shot.