Quantcast
Audio cleaning - Gearspace.com
The No.1 Website for Pro Audio
Audio cleaning
Old 31st January 2009
  #1
Here for the gear
 
🎧 10 years
Audio cleaning

Recently I was doing location sound for Video and unfortunately the location we had to use was by some Large AC units. (60hz ~)

The Editor wants to work predominately in Final Cut; does anyone know a good method to EQ the low hertz out of the track, with say a final cut plugin (or Soundtrack program).. or is the best method w/ final cut its EQ filter tool?

If he can't find a good solution, or I can't recommend one, perhaps we should go through PT and will have more success... its been a prob. 18 months since I've gotten to play w/ PT so can some of ya'll refresh me w/ some good plugins for this situation.. I wanna say X-noise, X-hum, 7-band EQ... what about in Waves.? I may have access to the plugins in the Diamond bundle.

Thanks!
Old 31st January 2009
  #2
Lives for gear
 
🎧 10 years
If the editor just wants to hear dialog more clearly while editing, Final Cut has some audio filters that should work, just make sure you use real time filters and nothing that has to be rendered so that when you send it to be mixed, the re-recording mixer can have the original files to clean up in a proper mix environment. If you need to render the audio, duplicate the tracks and save the original noisy track to be cleaned up later with the right tools and go to town on the duplicate track for editing.

Izotope RX, Cedar, Waves all have good noise reduction plugins for ProTools.
Old 31st January 2009 | Show parent
  #3
Lives for gear
 
TVPostSound's Avatar
 
🎧 10 years
Quote:
when you send it to be mixed, the re-recording mixer can have the original files to clean up in a proper mix environment.
I believe reading between the lines of the OP's post, the editor will mix it in FCP.
This becoming the norm on ultra low budgets.
Old 31st January 2009 | Show parent
  #4
Lives for gear
 
🎧 10 years
Quote:
Originally Posted by TVPostSound ➑️
I believe reading between the lines of the OP's post, the editor will mix it in FCP.
This becoming the norm on ultra low budgets.
That's called a hobby.
Old 31st January 2009 | Show parent
  #5
Lives for gear
 
georgia's Avatar
 
🎧 10 years
you can use FCPs multiband EQ or Hi-pass filter... but it still ain't gonna work.. you will get the low end out, but AC systems are wideband noise issues and hum issues....


cheers
geo
Old 31st January 2009 | Show parent
  #6
Here for the gear
 
🎧 10 years
hmm... I think the main problem is their is no "budget" to have it mastered in post.. It will be mastered, but mastered in Final Cut that is.. their is no plan for them to export the audio into any audio mastering program (eeeek); this annoys me personally because location sound is important to me but noise reduction was beyond my control at this location; I may volunteer to personally clean up the audio (in PT) if he is unable to do it himself in Final Cut.. Is their a program related w/ Soundtrack, that if you you a part of "room tone" you can essentially eliminate those freq's?
Old 31st January 2009 | Show parent
  #7
Lives for gear
 
TVPostSound's Avatar
 
🎧 10 years
Quote:
Is their a program related w/ Soundtrack, that if you you a part of "room tone" you can essentially eliminate those freq's?
Izotope RX besides being stand alone, and a PT plugin, will work in FCP, and STP.
Old 31st January 2009 | Show parent
  #8
Lives for gear
 
🎧 15 years
Quote:
Originally Posted by f1sound ➑️
hmm... I think the main problem is their is no "budget" to have it mastered in post.. It will be mastered, but mastered in Final Cut that is.. their is no plan for them to export the audio into any audio mastering program (eeeek); this annoys me personally because location sound is important to me but noise reduction was beyond my control at this location; I may volunteer to personally clean up the audio (in PT) if he is unable to do it himself in Final Cut.. Is their a program related w/ Soundtrack, that if you you a part of "room tone" you can essentially eliminate those freq's?
This is a common audio post myth. Yes, there are several "fingerprint" type NR apps out there, but they universally follow "Berger's Law": they work the best when you need them the least. With a very wide band noise source like HVAC, attempts to use such apps to do more than a slight reduction in BG noise start bringing up artifacts right away--the dreaded "head in a toilet" sound. With loud BG like that you have to chip away at it with various tools a little at a time, but short of really high-end tools like Cedar Cambridge etc the amount of improvement that can be made without damaging the speech audio is actually fairly small.

Philip Perkins
Old 31st January 2009 | Show parent
  #9
Gear Guru
 
Thomas W. Bethe's Avatar
 
1 Review written
🎧 15 years
Philip,

About 6 years ago I had a videographer bring me some footage that he had shot in front of an industrial air conditioner. He wanted me to take the noise out of the audio but leave the interviewer's voice intact. I tried a lot of things including the Sony NR package and some filtering but could not do a good job of getting out the sound without making the announcer sound like the man from Mars.

Fast forward to a year ago when I got RX from Izotope. The first thing I thought of when I got the program was to try it on that piece of video and low and behold it took out the background noise and left the announcer pretty much intact.

If you have not tried it RX is an excellent program and I would highly recommend it to any one working with sound. You have to do some tweaking with the sliders but it can do things no other program I have used can do. It is almost and I say almost as good as Cedar and for somethings maybe even better...

Best of luck!
Old 31st January 2009 | Show parent
  #10
Lives for gear
 
Jesse Peterson's Avatar
 
🎧 10 years
+1 for RX.

also, stacking NR plugs can work sometimes too. I did a bit once where I stacked 8 Z-Noise's, each one cuts out just a little, and the next one sampled noise from what was left.. (lots of tweaking) so on and so on. The same technique with RX works too. I got the idea from something I read on this board I think, talking about cicadas at the grave scene from forest gump. Anyone ever read that one?
Old 31st January 2009 | Show parent
  #11
Lives for gear
 
🎧 15 years
Quote:
Originally Posted by Thomas W. Bethe ➑️
Philip,

About 6 years ago I had a videographer bring me some footage that he had shot in front of an industrial air conditioner. He wanted me to take the noise out of the audio but leave the interviewer's voice intact. I tried a lot of things including the Sony NR package and some filtering but could not do a good job of getting out the sound without making the announcer sound like the man from Mars.

Fast forward to a year ago when I got RX from Izotope. The first thing I thought of when I got the program was to try it on that piece of video and low and behold it took out the background noise and left the announcer pretty much intact.

If you have not tried it RX is an excellent program and I would highly recommend it to any one working with sound. You have to do some tweaking with the sliders but it can do things no other program I have used can do. It is almost and I say almost as good as Cedar and for somethings maybe even better...

Best of luck!
I have RX (since it's release) and several other NR apps. I'm very glad that RX worked for you on the files you tried it on, but for very loud very broadband noise it is difficult to make a major dent in the BG without affecting the voice in unpleasant ways. I work on docs mostly, and this is an issue for me every day. My clients are not technical people but they are acutely sensitized to the sound and emotion contained in the voices in their films, and when we discuss and demo various NR techniques to deal with their more problematic scenes we almost always come to the conclusion that less is more. Yes, w/ RX and etc you can reduce really heavy BG noise down to zero, nearly, but when you A/B it with the original a lot of the life has gone out of the voice, esp if the voice recording was lo-fi to begin with. What we are finding is that chipping away at a scene with EQ, expander, a multiband limiter and then, finally, maybe something like RX seems to yield the most natural sounding audio--far more so than just RX or its competitors alone. We also find that our clients are usually willing to live with a certain amount of residual noise rather than kill all the room ambiance and make the voices sound in any way odd (or different).

Philip Perkins
Old 7th March 2009 | Show parent
  #12
Lives for gear
 
🎧 10 years
Great advice Philip. I totally agree. I spend enough time with my head down a toilet as it is , without NR programs (as fantastic as they are, especially RX from what I've seen).

I do 95% of my NR using EQ - automated to buggery - and drawn-in volume automation, riding the volume up over the words. And when things are really bad, I tend to embrace subtle multiband expansion more often than noise reduction.

If an engineer's not got the time to EQ and automate, then please go easy on the NR plugs, it's gonna screw it up (realism and emotion and all)! With no budget, I bet there are still tonnes of starting-out engineers out there who'd love to help and have a proper play with it for free.

good luck,
Oli
Old 14th March 2009 | Show parent
  #13
Lives for gear
 
Jfriah's Avatar
 
🎧 10 years
Quote:
Originally Posted by TVPostSound
I believe reading between the lines of the OP's post, the editor will mix it in FCP.
This becoming the norm on ultra low budgets.


Quote:
Originally Posted by Rick Sanchez ➑️
That's called a hobby.
-----



Well...that's called recent and future budgetary reality for those that don't care about ears and those that get paid to use them. (and, thus, why I'm boning up on my "FC" [a user who refuses to add the P] skills to prep for that near future).
Old 14th March 2009 | Show parent
  #14
Lives for gear
 
Jfriah's Avatar
 
🎧 10 years
Quote:
Originally Posted by Thomas W. Bethe ➑️
If you have not tried it RX is an excellent program and I would highly recommend it to any one working with sound. You have to do some tweaking with the sliders but it can do things no other program I have used can do. It is almost and I say almost as good as Cedar and for somethings maybe even better...

Best of luck!
Izo is a great program. Personally, I'm now finding I have 'better' luck/response going back to Waves X-Noise. I find "it depends" on the source/situation. If a user has iZo Rx, Waves restoration, CEDAR, one of those will work. But, I'd say, bang for the buck? iZotope Rx wins hands-down.

-Jeff
Old 16th March 2009 | Show parent
  #15
Lives for gear
 
BIGBANGBUZZ's Avatar
 
🎧 10 years
Quote:
Originally Posted by philper ➑️
I have RX (since it's release) and several other NR apps. I'm very glad that RX worked for you on the files you tried it on, but for very loud very broadband noise it is difficult to make a major dent in the BG without affecting the voice in unpleasant ways. I work on docs mostly, and this is an issue for me every day. My clients are not technical people but they are acutely sensitized to the sound and emotion contained in the voices in their films, and when we discuss and demo various NR techniques to deal with their more problematic scenes we almost always come to the conclusion that less is more. Yes, w/ RX and etc you can reduce really heavy BG noise down to zero, nearly, but when you A/B it with the original a lot of the life has gone out of the voice, esp if the voice recording was lo-fi to begin with. What we are finding is that chipping away at a scene with EQ, expander, a multiband limiter and then, finally, maybe something like RX seems to yield the most natural sounding audio--far more so than just RX or its competitors alone. We also find that our clients are usually willing to live with a certain amount of residual noise rather than kill all the room ambiance and make the voices sound in any way odd (or different).

Philip Perkins
Totally agree :-)

I work on allot of Interviews shot by VJ's who really don't know what they are doing the backgrounds are so noisy that RX can do little to repair it in most cases.
I have more success notching out the freq.
What expanders and multiband limiters do you use?
Old 19th March 2009
  #16
Gear Maniac
 
roughly's Avatar
 
🎧 15 years
well of course that is true, we all want clients that have a budget and demand high quality work we can be proud of --- that said out of the pro-sumer grade soft out there for noise redux, i have to say that izotope's RX is pretty impressive and lets hope that the high end EG cedar, finds it way down to a larger middle end market as tech improves.

OT: So what does the cedar NR cost anyway? Just wondering if it is that good and there seams to be a endless supply of projects that are feeling their skimp once post starts perhaps this [audio resteration] a business to look into.
Old 20th March 2009 | Show parent
  #17
Gear Head
 
OJCamero's Avatar
 
🎧 10 years
Quote:
Originally Posted by f1sound ➑️
Is their a program related w/ Soundtrack, that if you you a part of "room tone" you can essentially eliminate those freq's?
I thought Soundtrack had a NR filter that came in the software. I swear I remember taking a noise print in there before. Either way, it's not going to come out as easily as you might think. I remember getting artifacts a lot with the supplied NR filter.
Old 20th March 2009 | Show parent
  #18
Gear Head
 
OJCamero's Avatar
 
🎧 10 years
Quote:
Originally Posted by roughly ➑️
OT: So what does the cedar NR cost anyway? Just wondering if it is that good and there seams to be a endless supply of projects that are feeling their skimp once post starts perhaps this [audio resteration] a business to look into.
Last I checked the DNS2000 goes for anywhere in the neighborhood of 8-9k. Really pricey but if you work on a lot of dialog, it is worth it.
Old 20th March 2009 | Show parent
  #19
Lives for gear
 
minister's Avatar
 
🎧 15 years
Quote:
Originally Posted by roughly ➑️
OT: So what does the cedar NR cost anyway
Give or take 8 Grand USD.

Not a box for the underresourced.
Old 24th March 2009 | Show parent
  #20
Lives for gear
 
Jfriah's Avatar
 
🎧 10 years
Yes, at the time, our DNS 2000 was over $10k CDN. It really 'can' do wonders from time to time (again, all tools are dependant on situation).
Old 26th March 2009 | Show parent
  #21
Gear Head
 
OJCamero's Avatar
 
🎧 10 years
Quote:
Originally Posted by Jfriah ➑️
Yes, at the time, our DNS 2000 was over $10k CDN. It really 'can' do wonders from time to time (again, all tools are dependant on situation).
People also need to realize that you don't just throw a track through a noise filter and expect it to come out clean on the other end. There are things you need to do BEFORE you put it through the filter to get the best results. Always try and EQ as much of the problem without affecting the DX track first. After you feel that you have done the best that you can, THEN run it through noise reduction. You will get much better results.
Old 28th March 2009 | Show parent
  #22
Here for the gear
 
🎧 10 years
Applying the NR plug-ins in series IMHO is the best approach. The key is choosing good noise samples for training and not using too much reduction at any stage as to damage the good audio.

Depending upon how you are doing it: destructively/bouncing/bussing the approach is the same.

1) Apply a HPF to get rid of LF noise/rumble/DC offset. Apply notch filtering, as appropriate.
2) Look for the noise. When you find a section, listen to a normalized version of the selection to make sure that it is just steady noise. Listen for any clicks, pops, taps or what not. If you hear them, either trim the sample from the beginning/end or choose another noise sample.
3) Train the noise reduction with the non-normalized noise sample.
4) When setting the plug-in NR parameters, solo what it is being removed. Be sure that you cannot hear _any_ of the good audio. Turn up your monitors to really hear it clearly. It should sound like steady noise that may be ducked a bit by the good audio.
5) Reset the monitor level and listen to the processed (noise-reduced) audio. If you hear any degradation, back off some of the setting and go back to #4 . If you can't find a setting that is removing noise without damaging the good audio, then choose a new noise sample.
6) "Apply" the NR. If you need to get rid of more noise, repeat/add another NR plug-in.
7) Go back to #2 using the NR processed sound for training. If possible, select it from a different part of the track.

Keep referencing back to the original source to make sure the processed audio sounds the same, only cleaner.

Think of this kind of noise reduction as a multi-band gate: a bunch of parallel gates each operating on a small slice of the audio spectrum. The noise profile is used to determine the thresholds. If you use settings that are too aggressive or inappropriate samples for training, you begin to loose bits of the good signal's spectrum.

--jim
Old 28th March 2009 | Show parent
  #23
Lives for gear
 
Jfriah's Avatar
 
🎧 10 years
That's a good chronology, Jim.
Old 31st March 2009 | Show parent
  #24
Here for the gear
 
🎧 10 years
Dont Forget the most important step..

Save....

How many times has that screwed me...
Old 3rd December 2016
  #25
Here for the gear
 
Would it be near impossible to seriously clean a quiet conversation buried under the noise of running water AND if the people speaking were talking using voice distortes over some RF type device? Heh, just curious!
Old 3rd December 2016 | Show parent
  #26
Lives for gear
 
🎧 15 years
Quote:
Originally Posted by Madtrax ➑️
Would it be near impossible to seriously clean a quiet conversation buried under the noise of running water AND if the people speaking were talking using voice distortes over some RF type device? Heh, just curious!
That only works in movies and on TV. In real life, no. This is why ADR was invented.
Old 3rd December 2016
  #27
Here for the gear
 
ADR?
Old 5th December 2016 | Show parent
  #28
Gear Guru
 
Thomas W. Bethe's Avatar
 
1 Review written
🎧 15 years
Quote:
Originally Posted by Madtrax ➑️
ADR?
]See this[ https://en.wikipedia.org/wiki/Dubbing_(filmmaking)
Old 7th December 2016
  #29
Lives for gear
 
huub's Avatar
 
🎧 15 years
Soundtrack has a noisereduction function with learn/noiseprint.
This works quite well actually.
Take a noiseprint of the ac Sound and see how far you can take it.
Old 7th December 2016 | Show parent
  #30
Lives for gear
 
lemix's Avatar
 
🎧 15 years
Quote:
Originally Posted by huub ➑️
Soundtrack has a noisereduction function with learn/noiseprint.
This works quite well actually.
Take a noiseprint of the ac Sound and see how far you can take it.
Yes.. but let's just hope the OP has finished the project by now..




6 years later
πŸ“ Reply

Similar Threads

Thread / Thread Starter Replies / Views Last Post
replies: 16 views: 6617
Avatar for Jerry Tubb
Jerry Tubb 24th December 2020
replies: 60 views: 8013
Avatar for Jim@SSL
[email protected] 10th September 2020
replies: 217 views: 51634
Avatar for Leverson
Leverson 7th June 2014
replies: 129 views: 20418
Avatar for bam_sound
bam_sound 30th May 2021
Post Reply

Welcome to the Gearspace Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…

Forum Jump
Forum Jump