Topic: Help with removing a buzz from a recording

[pwig]

I use Audacity and run Mac OS X - I have a few recordings that have a buzz in them. What tools will remove the buzz and preserve the music. I can post a sample if needed.

[page]

I think the best thing you can do in Audacity is a very very tight notch EQ of about 10db or so for the offending frequency.

[pwig]

I think the best thing you can do in Audacity is a very very tight notch EQ of about 10db or so for the offending frequency.

Thank you - could you explain where this is found and how it works please?

Here is a sample of what I am attempting to fix: http://www.mrbryant.com/jtesample/jtesample.html

[page]

I think the best thing you can do in Audacity is a very very tight notch EQ of about 10db or so for the offending frequency.

Thank you - could you explain where this is found and how it works please?

http://taperssection.com/index.php?topic=149003.0 (Notch filter discussion)

and

http://taperssection.com/index.php?topic=148725.0 (removing narrow band noise)

[pwig]

I think the best thing you can do in Audacity is a very very tight notch EQ of about 10db or so for the offending frequency.

Thank you - could you explain where this is found and how it works please?

http://taperssection.com/index.php?topic=149003.0 (Notch filter discussion)

and

http://taperssection.com/index.php?topic=148725.0 (removing narrow band noise)

Thank you - still needing help. Attached is a screen shot - where is my hum and where do I remove in Izotope please?

[ScoobieKW]

IF you look at the graph there is a line going across around 500Hz, it's consistent through the whole file. I'd use a notch filter around this freq and see if it helps.

[pwig]

IF you look at the graph there is a line going across around 500Hz, it's consistent through the whole file. I'd use a notch filter around this freq and see if it helps.

Thank you. I have no experience with notch filters. I have izotope rx and Audacity for Mac OS X.

[vanark]

If you google audacity notch filter you should get some good instructions.  I've used it with pretty good results in the past.

[page]

Drink machine? I'm trying to think of what else would cause that mid-range constant...

In RX, use the vertical selector, zoom in, and select just slightly more than the hum itself, and then just do gain reduction. It's sooo much simpler than Audacity in that regard.

As for figuring out where it is, there is a isolation method in Izotope's products (it varys product to product). In RX, select a 200hz band (using the box selector, not the vertical or horizontal ones) and make sure it's plenty long so your audio doesn't stop while sampling. A couple minutes worth should be fine. Now as the audio is playing, you can move that around and hear just certain frequencies. Now hit the preview play and move it around until you clearly hear the hum and that should give you the vicinity. After that, it's just as has been previously described.

[pwig]

Drink machine? I'm trying to think of what else would cause that mid-range constant...

In RX, use the vertical selector, zoom in, and select just slightly more than the hum itself, and then just do gain reduction. It's sooo much simpler than Audacity in that regard.

As for figuring out where it is, there is a isolation method in Izotope's products (it varys product to product). In RX, select a 200hz band (using the box selector, not the vertical or horizontal ones) and make sure it's plenty long so your audio doesn't stop while sampling. A couple minutes worth should be fine. Now as the audio is playing, you can move that around and hear just certain frequencies. Now hit the preview play and move it around until you clearly hear the hum and that should give you the vicinity. After that, it's just as has been previously described.

I have isolated the bar that appears to be the issue and have had no luck. Anyone want to take a run at the sample and see if they have better luck?

[pwig]

Drink machine? I'm trying to think of what else would cause that mid-range constant...

In RX, use the vertical selector, zoom in, and select just slightly more than the hum itself, and then just do gain reduction. It's sooo much simpler than Audacity in that regard.

As for figuring out where it is, there is a isolation method in Izotope's products (it varys product to product). In RX, select a 200hz band (using the box selector, not the vertical or horizontal ones) and make sure it's plenty long so your audio doesn't stop while sampling. A couple minutes worth should be fine. Now as the audio is playing, you can move that around and hear just certain frequencies. Now hit the preview play and move it around until you clearly hear the hum and that should give you the vicinity. After that, it's just as has been previously described.

I have isolated the bar that appears to be the issue and have had no luck. Anyone want to take a run at the sample and see if they have better luck?

I am thinking my mics or patch cord may have gone bad.

[vanark]

Do you have a section of the recording with just the noise (no music).  I've had decent success using a noise reduction if I have a clean sample of the noise.  I cannot isolate the hum enough to get a notch filter to work.  I suspect it is more broad spectrum than a single frequency.

[pwig]

Do you have a section of the recording with just the noise (no music).  I've had decent success using a noise reduction if I have a clean sample of the noise.  I cannot isolate the hum enough to get a notch filter to work.  I suspect it is more broad spectrum than a single frequency.

I will find a section that may be between songs that might allow this work. Thank you!

[pwig]

Do you have a section of the recording with just the noise (no music).  I've had decent success using a noise reduction if I have a clean sample of the noise.  I cannot isolate the hum enough to get a notch filter to work.  I suspect it is more broad spectrum than a single frequency.

Here is a link to a page with a song sample as well as a sample of in-between songs that may allow this to be isolated. Thank you!

http://www.mrbryant.com/jtesample/jtesample.html

[vanark]

First, the noise in the "sample" is different in the left channel than in the "hum" version.  Also, you aren't making this very easy.  I could only isolate about 0.75 sec of noise only in the hum clip, but I think it does a decent job on addressing the right channel.  Whatever is happening in that left channel on the sample is very broad spectrum and it is going to be more difficult to work with.  I'm not sure how much of the set has it as bad as that.  The "hum" clip did not.  I think with the entire set as one file, you should be able to isolate the noise alone, get a good "sample" for the noise removal tool and remove it to make it at least listenable on the right channel.  You may decide to make it mono if you can only salvage one channel. Hope that helps.

Right channel improve with noise removal in audacity:
http://www.sendspace.com/file/k2c8dg

"Mono" version just using the right channel after noise removal:
http://www.sendspace.com/file/xxjh7k

[pwig]

First, the noise in the "sample" is different in the left channel than in the "hum" version.  Also, you aren't making this very easy.  I could only isolate about 0.75 sec of noise only in the hum clip, but I think it does a decent job on addressing the right channel.  Whatever is happening in that left channel on the sample is very broad spectrum and it is going to be more difficult to work with.  I'm not sure how much of the set has it as bad as that.  The "hum" clip did not.  I think with the entire set as one file, you should be able to isolate the noise alone, get a good "sample" for the noise removal tool and remove it to make it at least listenable on the right channel.  You may decide to make it mono if you can only salvage one channel. Hope that helps.

Right channel improve with noise removal in audacity:
http://www.sendspace.com/file/k2c8dg

"Mono" version just using the right channel after noise removal:
http://www.sendspace.com/file/xxjh7k

Ok, in Audacity, how do I do it - the right change; correction track sounds GREAT! Thank you!

[vanark]

I'm sure there are more detailed instructions on the use of the noise removal tool somewhere.  I've had good luck when googling them. Audacity is a widely used program so there are a good deal of discussions on it.

Basic outline (sorry, I don't have time to write it all out in detail).

1. If you are only going to try to salvage the right channel, split the stereo track into two separate channels.  Work on the right channel only.
2. Identify short period of ONLY noise and highlight.  And I mean only noise - no talking, no tuning, etc.  The better the sample is, the better it will do to identify and remove it.
3. Go the effect menu and choose Noise removal.
4. With the noise sample only highlighted, you can select "Get Noise Profile".
5. Go back to the effect menu and choose Noise removal.  With the noise profile identified, you can now run the Noise removal tool  There are 4 settings and I'm sure their effect is discussed somewhere - for some reason (whatever it is), I have mine set at Noise reduction: 24 dB, Sensitivity: 0.0 dB, Frequency smoothing: 150 Hz, Attack/decay time: 0.15 sec, Noise is set to remove.
6. Once the noise is removed from the right channel to your satisfaction, highlight the entire channel (only the right channel), Copy.  Go to the left channel, highlight the entire channel, Paste.  Then join the two tracks back together to make a stereo track.

Hope that helps get you started in the right direction.  You may need to tweak the Noise removal settings or the noise profile sample to find the right result.  There is still some noise on the de-noised result I gave you, but I didn't tweak it at all - just wanted to show how you could get a listenable result with some fairly standard tools.  Good luck.

[pwig]

First, the noise in the "sample" is different in the left channel than in the "hum" version.  Also, you aren't making this very easy.  I could only isolate about 0.75 sec of noise only in the hum clip, but I think it does a decent job on addressing the right channel.  Whatever is happening in that left channel on the sample is very broad spectrum and it is going to be more difficult to work with.  I'm not sure how much of the set has it as bad as that.  The "hum" clip did not.  I think with the entire set as one file, you should be able to isolate the noise alone, get a good "sample" for the noise removal tool and remove it to make it at least listenable on the right channel.  You may decide to make it mono if you can only salvage one channel. Hope that helps.

Right channel improve with noise removal in audacity:
http://www.sendspace.com/file/k2c8dg

"Mono" version just using the right channel after noise removal:
http://www.sendspace.com/file/xxjh7k

This worked beautifully! Thank you! I really appreciate the help - I am not great with this type of post work - most of my recordings need audio boosting and tracking - again, thank you!