Sound Editing - Audacity and removing voice hiss

Status
Not open for further replies.

jetjackson

Wannabie Member
Feb 17, 2012
29
5
8
Europe atm
www.thegreatgallivant.com
I ride a
F650 GS Twin '08
Hi Guys,

I am currenly using Audacity to try and clean up an audio file. I recorded a bit of footage for Stahlkoffer panniers and I am putting together some Youtube clips for them to promote their gear. Anyway, the recording situation was not ideal. In a workshop with (it was -2 degrees outside at the time) banging and clanging going on in the background.

So far using Audacity - A great free tool btw. I have used the Noise reduction effect to pretty much eliminate the white noise in the background. This did dullen the voice a bit though. I have then used the envelope tool to cut out the isolated clangs and bangs in the background. I am left with a couple that I cannot get rid of because they overlap the things that the subject is saying and if I cut those clangs I cut the voice as well - I am going to edit these parts out most likely. After all this the sound of the voice was still too low so I used the normalise effect (on default settings, using all effects on default settings atm) to increase the volume of the voice. This amplication has resulted in the s sounds on the start and end of words being a very high hiss - sss. If you dl the attached file you will be able to hear it in the background - pariculary in the last 5 or so seconds with the thing the subject says then.

I am trying to get rid of this sound. Have played around with the EQ but dont know enough about sound to know which frequencies to cut and I am having no luck. Anyone know how to do this?

The end result I am going for is loud voice - doesnt matter if itis too rough as going to layer it over background music in order to distract viewer/listener from it's 'roughness'.

I have been on the sound engineering forums but everything there is directed at higher end sound production and not a basic level so I thought maybe here someone might be able to tell me in laymans terms :)
 

Yarek

Wannabie Member
Jan 29, 2012
457
3
0
I ride a
Street Triple 'R' - 2010 Matt Orange
I'd be really interested in understanding more about this - I'm getting stupid interference with sound through my crappy mic and while I have Audacity I've never ventured into it yet. It does sound like it'll be a decent tool to clean up some of my audio, although I don't think there's a 'fix Scottish accent' filter :p
 

Gadgetsandwheels

Wannabie Member
Mar 10, 2012
147
0
0
Huddersfield
I ride a
BMW R1200GS
If the audio is that bat with the background noise etc, could you mute the sound on the video altogether and re-record the sound fresh again separately? Or can you see your face in the video while you are talking?
 

Gadgetsandwheels

Wannabie Member
Mar 10, 2012
147
0
0
Huddersfield
I ride a
BMW R1200GS
Ignore my previous message - I just re-read your note and it says you filmed somebody else.

There are ways of using EQ, compression and noise filters in most audio software to try and remove unwanted noise. Even the best ones however leave the remaining sounding mechanical and often alien. It depends how bad it is in the first place really. Do you have a short sample?
 

xXxDarkKnightxXx24

Wannabie Member
Mar 23, 2012
117
0
0
New York
www.youtube.com
I ride a
Kawasaki Ninja 650r
Im looking for the same thing to get rid of the wind noise from riding but not lose my voice in the process and i have tried multiple noise reducers and they get rid of the wind but also make my voice sound funny and muffled so im in the same boat as you
 

lauriejennifer

Wannabie Member
Jul 23, 2011
214
0
16
I ride a
Kawasaki KLX330sm * Keen BadBoy CX50 * Kawasaki ER-6n
The "sss" sound in the human voice is generally right around 7k-8k hertz. Making a slight EQ cut at around 7k, with the Q band ranging from about 6k to 8k should do the trick. And/or, you could use a de-esser, if there is dynamics available for that. A de-esser is essentially a type of compressor that is meant to specifically and solely compress that "s" range.

Be careful about overdoing either the cut or the de-esser. This can end up making the voice sound like it has a lisp. Moderation is key here.



...I'm a total sound nerd. :-/
 

jetjackson

Wannabie Member
Feb 17, 2012
29
5
8
Europe atm
www.thegreatgallivant.com
I ride a
F650 GS Twin '08
lauriejennifer said:
The "sss" sound in the human voice is generally right around 7k-8k hertz. Making a slight EQ cut at around 7k, with the Q band ranging from about 6k to 8k should do the trick. And/or, you could use a de-esser, if there is dynamics available for that. A de-esser is essentially a type of compressor that is meant to specifically and solely compress that "s" range.

Be careful about overdoing either the cut or the de-esser. This can end up making the voice sound like it has a lisp. Moderation is key here.



...I'm a total sound nerd. :-/

Thanks heaps, going to have another crack at the audio over the next couple of days.. as soon as I finish the next video blog :)
 

ariderslife

Wannabie Member
Jul 26, 2011
879
79
28
37
Melbourne - Australia
www.youtube.com
I ride a
Currently: 2004 CB 919 - Previous rides - Honda Hornet 600 aka CB599 - Previous: 1999 Honda VTR 250: Previous Honda f4i - Previous: CBR 929 Previous: DRZ400sm: Previous 2006 gsxr 600: Previous 1989 Honda Spada VT250
all i know is i go -6 on the audio setting son the 2nd last and 3rd last line in fcpx. lol
 

Gorguruga

Wannabie Member
Jun 23, 2012
3
0
0
I ride a
bike
The best way to get rid of the sss sounds is literally to zoom in to each offending sss sound and reduce the volume to an acceptable level. Obviously it takes more time but if you do it this way you don't need to apply a hiss reduction effect to the whole recording so you retain the sound quality. When you do an automatic hiss reduction it can make the recorded vocals sound muffled and you lose quality.
 

lauriejennifer

Wannabie Member
Jul 23, 2011
214
0
16
I ride a
Kawasaki KLX330sm * Keen BadBoy CX50 * Kawasaki ER-6n
Gorguruga said:
The best way to get rid of the sss sounds is literally to zoom in to each offending sss sound and reduce the volume to an acceptable level. Obviously it takes more time but if you do it this way you don't need to apply a hiss reduction effect to the whole recording so you retain the sound quality. When you do an automatic hiss reduction it can make the recorded vocals sound muffled and you lose quality.

This is like a manual de-esser. I'd rather just apply a de-esser and be conservative on the settings. Set the threshold a bit higher. Beats spending the better part of an hour isolating and manually cutting each "s." Also, it will only cut the "s" part of the offending "s," rather than the entire level. It will only cut that specific frequency.

But, you're right about over-using it. As I said, overzealous use of a de-esser will make the speaker sound like they have a lisp. Moderation is key.
 
Status
Not open for further replies.

Winners Video

Website Supported by Ipswich SEO

Latest posts