Recorded a ground I can't get rid of

Rokket

Trailing Behind Again
I goofed. I was video taping the wife for some project she has going on. She has to do a 10 minute presentation on video. The mic on the cam wasn't cutting it, so I tried a shotgun mic I have, only to hear the hum of a ground in the damn soundtrack. We tried it again with a different mic setup, but I fear the problem is in the jack on the camera, and not in my mics or cables. She refuses to redo the presentation, so I am stuck trying to edit out the offending noise.

I was thinking notching it out with a parametric eq. How would you do it?
 

gecko zzed

Grumpy Mod
I goofed. I was video taping the wife for some project she has going on. She has to do a 10 minute presentation on video. The mic on the cam wasn't cutting it, so I tried a shotgun mic I have, only to hear the hum of a ground in the damn soundtrack. We tried it again with a different mic setup, but I fear the problem is in the jack on the camera, and not in my mics or cables. She refuses to redo the presentation, so I am stuck trying to edit out the offending noise.

I was thinking notching it out with a parametric eq. How would you do it?

You have an audio track that consists of her vocals with a peristent hum going right through it?

One thing you can try is this:

1 Find a space on the audio track where there is just the hum, say, between a phrase.

2 Copy this chunk, invert it, then loop it on a new track that's the length of the video audio.

3 Mix the two together. With luck the inverted phase of these hum chunks will cancel the univerted hum on the audio track.

4 The caveat is that it could actually reinforce it unless you get these chunks to line up appropriately.

Or: Record a video length of nothing but hum, invert this and mix with audio hum. That way you only have to make one lining up adjustment.

Or: do as you say, i.e. notch it out.
 

Waffleness

New member
A mixture of what gecko said and EQ should do an ok job.

Electricity isn't a perfect sine wave, so harmonics will vary across the length of the audio track. Phase cancelling should remove the fundamental, and eq should just about deal with the rest.
 

Rokket

Trailing Behind Again
You have an audio track that consists of her vocals with a peristent hum going right through it?

One thing you can try is this:

1 Find a space on the audio track where there is just the hum, say, between a phrase.

2 Copy this chunk, invert it, then loop it on a new track that's the length of the video audio.

3 Mix the two together. With luck the inverted phase of these hum chunks will cancel the univerted hum on the audio track.

4 The caveat is that it could actually reinforce it unless you get these chunks to line up appropriately.

Or: Record a video length of nothing but hum, invert this and mix with audio hum. That way you only have to make one lining up adjustment.

Or: do as you say, i.e. notch it out.
I'll give this a shot. I haven't had much luck with the parametric eq. I am struggling with identifying the freq the hum is on. When I get close it cuts too much into her vocal.
 

gecko zzed

Grumpy Mod
If it's a ground hum, it should be around 50hz ~ 60hz, which should be below the vocal range.

If it is not in that frequency range, then I'm not sure (a) what the problem is, and (b) what other steps you can take.
 

noisewreck

New member
If it's a ground hum, it should be around 50hz ~ 60hz, which should be below the vocal range.

If it is not in that frequency range, then I'm not sure (a) what the problem is, and (b) what other steps you can take.

The fundamental is, however the harmonics are not. You can try sharp notches at multiples of 50Hz or 60Hz depending on whether you're in Europe or the US. However, I like the polarity reversal idea. It's definitely worth a shot.
 

Rokket

Trailing Behind Again
If it's a ground hum, it should be around 50hz ~ 60hz, which should be below the vocal range.

If it is not in that frequency range, then I'm not sure (a) what the problem is, and (b) what other steps you can take.

The fundamental is, however the harmonics are not. You can try sharp notches at multiples of 50Hz or 60Hz depending on whether you're in Europe or the US. However, I like the polarity reversal idea. It's definitely worth a shot.
Thanks again. I am going to give it a shot tonight. I have been on airplanes and trains for the last few days. My ass is killing me and my body still doesn't know what time zone I am in! :eek:

But I'll post the results as soon as I get it figured out. I am excited to try the polarity reversal thing too!
 

SouthSIDE Glen

independentrecording.net
Note that since you're talking only a small section of the hum and replicating it for the length of the recording, that there's two important details you need to watch for it to work properly:

- First any loop you create should be edited to star and stop at the same point in the fundamental wave cycle. This can be anywhere in the wave, but typically one picks the positive zero crossing (the point where the wave "starts" by climbing up from the zero centerline and "ends" by reaching the zero centerline from below.) Zoom in tight and make your cuts precise on the centerline. This will help ensure that there are no pops or clicks at the end of the sample and that the overall waveform remains in the same phase.

- you will probably need to make sure that you phase align the original and inverted signals by sliding them fore and back until they appear to be images of each other. If they are mis-aligned, this will lessen the effectiveness of the exercise.

Finally, unless you have very good power conditioning in your studio, there is a real chance that the frequency of the hum may drift by a cycle or two per second over the length of the recording; the power directly from the utilities is rarely exactly 120V/60Hz (in the US) and tends to have a fairly wide margin of error over time.

If this happens, it will affect how well the phase inversion/cancellation will hold over the 10 minutes of the recording. The negative effect will probably be minimal, but if you notice that there's a certain section of your recording that doesn't cancel the way the rest of the recording does, you might want to try taking a second sample from that part of the recording, if a quiet space is available there, and crossfade a punch-in of the inverted new sample into just that trouble section.

G.
 

Rokket

Trailing Behind Again
Note that since you're talking only a small section of the hum and replicating it for the length of the recording, that there's two important details you need to watch for it to work properly:

- First any loop you create should be edited to star and stop at the same point in the fundamental wave cycle. This can be anywhere in the wave, but typically one picks the positive zero crossing (the point where the wave "starts" by climbing up from the zero centerline and "ends" by reaching the zero centerline from below.) Zoom in tight and make your cuts precise on the centerline. This will help ensure that there are no pops or clicks at the end of the sample and that the overall waveform remains in the same phase.

- you will probably need to make sure that you phase align the original and inverted signals by sliding them fore and back until they appear to be images of each other. If they are mis-aligned, this will lessen the effectiveness of the exercise.

Finally, unless you have very good power conditioning in your studio, there is a real chance that the frequency of the hum may drift by a cycle or two per second over the length of the recording; the power directly from the utilities is rarely exactly 120V/60Hz (in the US) and tends to have a fairly wide margin of error over time.

If this happens, it will affect how well the phase inversion/cancellation will hold over the 10 minutes of the recording. The negative effect will probably be minimal, but if you notice that there's a certain section of your recording that doesn't cancel the way the rest of the recording does, you might want to try taking a second sample from that part of the recording, if a quiet space is available there, and crossfade a punch-in of the inverted new sample into just that trouble section.

G.
I am working a bit on it and had to take a break. It's a lot tougher than I figured because it was recorded off a small camcorder with a really cheap mic (the only one I had with an 1/8 inch T/S connector). I am having trouble finding the frequency this thing is eating, and it changes from not being there at all to a full on hum. I thought about trying to notch it out as best I can and adding a music soundtrack to mask it some more. I recorded this in our bedroom because she didn't want to go into the studio (too sterile???), and it's her baby, not mine so she won that battle. I have good power conditioners in there and could have run a good mic through the board and edited it onto the video later, but now I am stuck because she doesn't want to redo it, and I am now 4,500 miles from home... :(


EDIT: After listening to this thing, it sounds more like the hum you get from flourescent lights in an amp than a ground thing. Now I am really stuck. Notching it with a parametric eq in the 1000hz range seems to quiet it some, but nothing I've done takes it away without messing up the rest of the recording (i.e. her quieter speaking parts). I tried a noise gate to no avail. I can't seem to get the loop right to try phase reversal.

Any other ideas???
 
Last edited:

Bristol Posse

Okey Dokey
Leave thu hum in the presentation wrap it up and give it to your wife as a fait accoplis, If she doesn't even notice it then hey nothing lost. If she hates the hum tell her the only way to fix is to record the vocal in the "Sterile" studio on a propper mic and then highlight the work you already did to try and eliminate it and show her this thread.
 

SouthSIDE Glen

independentrecording.net
I like Bristol's response ;).

But if that is just not good enough for your wife to let you back into the house when you get home (;)), there's a couple of more ideas:

- Note what Noisewreck had to say about harmonics (he beat me to the punch on that point.) If it sounds more like a floescent light, that probably indicates some higher frequency stuff - FL's tend to be "sharper" sounding than a simple 60 cycle hum. If the hum's fundamental is at (for example) exactly 60 Hz, try some additional (but probably smaller) filtering at 120, 180 and 240hz, as needed.

If you have a decent FFT/spectrum analyzer display, take a snapshot of the hum during an otherwise quiet period and see if you can identify a harmonic series like that. Or it may even point out another non-harmonic bump that needs addressing with a tight EQ.

- If the hum is significantly lower level than the important components of the signal you want to keep, you might want to try using an expander (the opposite of a compressor) to push the hum down to inaudible levels. Set the threshold to just above the hum level and expand everything below that threshold down.

G.
 

Rokket

Trailing Behind Again
I like Bristol's response ;).

But if that is just not good enough for your wife to let you back into the house when you get home (;)), there's a couple of more ideas:

- Note what Noisewreck had to say about harmonics (he beat me to the punch on that point.) If it sounds more like a floescent light, that probably indicates some higher frequency stuff - FL's tend to be "sharper" sounding than a simple 60 cycle hum. If the hum's fundamental is at (for example) exactly 60 Hz, try some additional (but probably smaller) filtering at 120, 180 and 240hz, as needed.

If you have a decent FFT/spectrum analyzer display, take a snapshot of the hum during an otherwise quiet period and see if you can identify a harmonic series like that. Or it may even point out another non-harmonic bump that needs addressing with a tight EQ.

- If the hum is significantly lower level than the important components of the signal you want to keep, you might want to try using an expander (the opposite of a compressor) to push the hum down to inaudible levels. Set the threshold to just above the hum level and expand everything below that threshold down.

G.
A couple cool ideas, Glen. I am going to give them a shot. I use Reaper and it does have an FTT/analyzer, but I haven't used it, so I don't know how good it is. I'll try that first, then go for the expander.
Thanks, man!
 
gotta sample of it? I'll throw it through some of my noise reduction plugins and see what I can come up with for you. I love being able to use those things when I can (I mean, I paid for 'em...so might as well use 'em)
;)

hums and high pitched whines are the easiest to get rid of since they don't reside too much around the voice. I get those kind of things all the time.
 

gecko zzed

Grumpy Mod
hums and high pitched whines are the easiest to get rid of since they don't reside too much around the voice. I get those kind of things all the time.

I get those things all the time too. But they usually come from people, not electronics, and they are a bit harder to get rid of.
 

Rokket

Trailing Behind Again
gotta sample of it? I'll throw it through some of my noise reduction plugins and see what I can come up with for you. I love being able to use those things when I can (I mean, I paid for 'em...so might as well use 'em)
;)

hums and high pitched whines are the easiest to get rid of since they don't reside too much around the voice. I get those kind of things all the time.
I'll see what I can do...
 

SouthSIDE Glen

independentrecording.net
What do you think???
Unfortunately I think it's difficult to make any specific diagnosis from that screen shot because I can't see the frequency scale which I assume is along either the bottom or the top of the graph. Without being able to verify the actual frequencies of the various peaks, it's not a complete help.

Is that what it looks like when there's nothing but hum? Because that's what we really need to see - along with the frequency scale. It looks like *maybe* there's at least some regular repetition there, but I can't say for sure yet.

G.
 
Top