Lowering the input gain to remove background hiss

[TBman]

I thought I would start another thread to highlight this for discussion. Zoom H5 is the equipment used.

This is exactly what I was doing, using an input gain of 5 on the Zoom and then bringing up the gain in post eq and fighting the hiss with noise/hiss removal tools. This guy says that using normal input gain levels results in less hiss to remove in post eq.

[ljguitar]

Originally Posted by TBman

I thought I would start another thread to highlight this for discussion. Zoom H5 is the equipment used.

Hi Barry
My 'ultimate' test is when my final-mix is played through ear-buds (bluetooth or wired) can it be heard. Also can it be heard on the average car stereo system?

That's where my audio is listened to by friends and/or family.

[Doug Young]

Gain staging is always tricky.

You have a mic, which has a certain amount of self noise, and also any noise from your room that the mic picks up. *any* gain that you apply to that, whether its from the gain on your Zoom, or gain you add when you mix, or even the volume control as I listen to your track will bring up the guitar and that noise equally.

Then there's noise added by Zoom's electronics. If you assume it's a constant (maybe, maybe not), then adding gain later will also amplify that.

So the trick is to find the setting on the Zoom that optimizes that last part, and gets you more signal relative to the Zoom's internal noise (which is pretty low). Nothing you do there will help with noise from your room or the mic, tho. One thing you can do is close-mic (and play louder), to bring up the guitar relative to all these other sources of noise.

I generally set the gain on my H6 at about 5 or 6 - it's the level that matters, not the numerical setting, so that I have a solid signal, but am sure not to overdrive.

What the heck, as long as I'm answering, I just did a quick demo. I recorded a few chords in my video room - untreated, just a spare bedroom. My wife's doing laundry downstairs, so it's not really very quiet right now, definitely some ambient noise. I recorded using an AudioTechnica mic into the H6. Here's the raw track



Here's what it looks like in Adobe Audition:

Screen Shot 2020-05-17 at 11.33.06 AM.jpg

And here it is in RX:

Screen Shot 2020-05-17 at 11.33.38 AM.jpg

My noise floor looks to be -64db (that's total of Zoom internal noise, mic internal noise, room internal noise). but we have to take into account my volume, which is a peak of -7.81 db. 'll be adding that when I bring the track up to full level. So subtracting, I have a actual noise floor of around -57db. Not horrible, especially since the washing machine is running downstairs, but not studio quality.

Screen Shot 2020-05-17 at 11.34.05 AM.jpg

But I can improve this a bit. I generally add a high pass filter to knock out low mud below the guitar's range. I can do that with an EQ, but since I'm in RX, I'll do it there, so I can measure the impact. I just select everything below 40Hz, and hit delete. That takes my noise floor down to -67db, so I gained 3db, probably some low rumble from the washing machine. NOTE: I'm not doing any noise reduction here, I'm just using RX to do a high pass filter. I could easily knock down the noise using the noise reduction feature, but let's not.

OK, so now, let's "mix" this, and bring up the levels.

Here's the track after running thru Ozone with a target of -14LKFS. My peak volume is now -1db, and my noise floor is -59db, about as expected from the previous measurements. Not bad, but not fantastic either.



Then finally, I'll do a bit of manually EQing, add some reverb, and most importantly, trim off the beginning "silence" (aka "noise") and fade out the tail. This gives me this, without using any noise reduction. I could do better if I wanted to use RX to reduce the noise, but I don't think it's needed.




How does this seem? Is the Zoom/mic/room noise an issue, or is this clean enough, as Larry says, for playing in typical listening environments? I do hear noise on the tail - probably the washing machine :-)

[Rudy4]

I'm curious as to the benefit of recording at a -10 or -12 level (to gretly reduce the chance of clipping or digital overs) and normalizing the track afterward. In theory that should remove any additional electronics noise by using actual circuitry to boost levels for a given track.

The use of compression or limiting on the input can also give you some wacky hiss levels if not used properly. I sometimes wonder if this is something that is experienced by home recordists.

[TBman]

Originally Posted by Doug Young

Gain staging is always tricky.

You have a mic, which has a certain amount of self noise, and also any noise from your room that the mic picks up. *any* gain that you apply to that, whether its from the gain on your Zoom, or gain you add when you mix, or even the volume control as I listen to your track will bring up the guitar and that noise equally.

Then there's noise added by Zoom's electronics. If you assume it's a constant (maybe, maybe not), then adding gain later will also amplify that.

So the trick is to find the setting on the Zoom that optimizes that last part, and gets you more signal relative to the Zoom's internal noise (which is pretty low). Nothing you do there will help with noise from your room or the mic, tho. One thing you can do is close-mic (and play louder), to bring up the guitar relative to all these other sources of noise.

I generally set the gain on my H6 at about 5 or 6 - it's the level that matters, not the numerical setting, so that I have a solid signal, but am sure not to overdrive.

What the heck, as long as I'm answering, I just did a quick demo. I recorded a few chords in my video room - untreated, just a spare bedroom. My wife's doing laundry downstairs, so it's not really very quiet right now, definitely some ambient noise. I recorded using an AudioTechnica mic into the H6. Here's the raw track



Here's what it looks like in Adobe Audition:

Attachment 37893

And here it is in RX:

Attachment 37894

My noise floor looks to be -64db (that's total of Zoom internal noise, mic internal noise, room internal noise). but we have to take into account my volume, which is a peak of -7.81 db. 'll be adding that when I bring the track up to full level. So subtracting, I have a actual noise floor of around -57db. Not horrible, especially since the washing machine is running downstairs, but not studio quality.

Attachment 37895

But I can improve this a bit. I generally add a high pass filter to knock out low mud below the guitar's range. I can do that with an EQ, but since I'm in RX, I'll do it there, so I can measure the impact. I just select everything below 40Hz, and hit delete. That takes my noise floor down to -67db, so I gained 3db, probably some low rumble from the washing machine. NOTE: I'm not doing any noise reduction here, I'm just using RX to do a high pass filter. I could easily knock down the noise using the noise reduction feature, but let's not.

OK, so now, let's "mix" this, and bring up the levels.

Here's the track after running thru Ozone with a target of -14LKFS. My peak volume is now -1db, and my noise floor is -59db, about as expected from the previous measurements. Not bad, but not fantastic either.



Then finally, I'll do a bit of manually EQing, add some reverb, and most importantly, trim off the beginning "silence" (aka "noise") and fade out the tail. This gives me this, without using any noise reduction. I could do better if I wanted to use RX to reduce the noise, but I don't think it's needed.




How does this seem? Is the Zoom/mic/room noise an issue, or is this clean enough, as Larry says, for playing in typical listening environments? I do hear noise on the tail - probably the washing machine :-)

Doug thanks. What do you use for input gain with just the Zoom mics?

[TBman]

Here's what I'm dealing with -

Zoom H5, just internal mics, gain at about 5, 5.5. - on the gain dial, not the db meter. I should look at that tonight and see what it is. It looks like Doug has almost no hum at the beginning of the unedited file.

I bumped the gain up in Rx by 12



Notice all that hum at the beginning. Here's what it sounds like:



I can remove that hum in Audition. Here's the after:



And the sound file



But then I get that slightly "hollow" sound for lack of a better word. I tried a couple of times to reduce the noise removal but don't get an acceptable result. Maybe I need to learn to adjust the tool better, not sure.

Here's a cleaned up "final" version:

[Doug Young]

Originally Posted by TBman

Doug thanks. What do you use for input gain with just the Zoom mics?

I'll have to try it. But I never use the Zoom mics for anything I'd care about. They sound "OK", but an external mic is easier to place, and typically sounds better

[Doug Young]

Originally Posted by Rudy4

I'm curious as to the benefit of recording at a -10 or -12 level (to gretly reduce the chance of clipping or digital overs) and normalizing the track afterward. In theory that should remove any additional electronics noise by using actual circuitry to boost levels for a given track.

The use of compression or limiting on the input can also give you some wacky hiss levels if not used properly. I sometimes wonder if this is something that is experienced by home recordists.

The reason for leaving a bit of headroom is to be sure you don't go "over" by accident. Also it leaves room for processing and EQ. With 24 bits, you have more dynamic range available than a guitar is capable of, you won't miss the top few DB. I generally shoot for leaving 6 or so, tho, not 10-12. but 10-12 is still more than fine.

Compression is best avoided for acoustic guitar. I use some, usually, but very little. Probably no more than 0.5 db of compression, nowhere near enough to aggravate noise or hiss, just enough to slightly smooth things.

I don't think you can assume that you're removing noise by boosting levels, it all depends on where the noise is coming from, what the sweet spot is in your gear, etc.

[Bob Womack]

You know, a lot of that B/G noise sounds like "room tone," an aggregate of A/C flow noise and home sounds that typically becomes loud from not having the mics near enough to the source. I do audio post production for video and had to remove this every working day. You can deal with this noise to a degree by learning it and removing it with iZotope Spectral Noise Reducer but honestly, the best way to start is by working with a combination of mic placement and turning off the A/C during takes to reduce it.

Bob

[Rudy4]

Originally Posted by Doug Young

The reason for leaving a bit of headroom is to be sure you don't go "over" by accident. Also it leaves room for processing and EQ. With 24 bits, you have more dynamic range available than a guitar is capable of, you won't miss the top few DB. I generally shoot for leaving 6 or so, tho, not 10-12. but 10-12 is still more than fine.

Compression is best avoided for acoustic guitar. I use some, usually, but very little. Probably no more than 0.5 db of compression, nowhere near enough to aggravate noise or hiss, just enough to slightly smooth things.

I don't think you can assume that you're removing noise by boosting levels, it all depends on where the noise is coming from, what the sweet spot is in your gear, etc.

Thanks, Doug.

I always record at 24 bits, specifically for that reason. I know you don't get rid of the noise when you boost the levels, but you're simply masking what's present in the signal as background noise.

[Doug Young]

Originally Posted by Bob Womack

You know, a lot of that B/G noise sounds like "room tone," an aggregate of A/C flow noise and home sounds that typically becomes loud from not having the mics near enough to the source.

My impression every time this comes up, is that the real culprit is room noise, not mic-self noise, or preamp. In home recording, that's almost always the issue.

[Doug Young]

OK, I tried the Zoom internal mics. They do need more gain, I ended up setting them at 7, resulting in a recording with peak level of -12db. So more gain on the preamp, and less final level. They may also be slightly noisier (not surprising, the cost of those internal mics has to be measured in pennies, not dollars). With a signal peak of -12 and noise level of about -67, that gives me a noise floor of -55 db, slightly worse than with the AT mics.

On the other hand, my wife's still doing laundry, but has switched from washer to drier, so maybe the noise difference is just due to that. Close enough in any case. Here's the raw zoom track:



and here it is brought up to level and mixed (no noise reduction). I left the silence at the beginning and end so you can hear the noise level.

[Chipotle]

It always surprises people what a microphone picks up. Is there a way to monitor what the Zoom is hearing through good headphones? Often background noise you don't even notice suddenly becomes apparent.

Another test to try, though probably less practical: go throw the main circuit breaker and shut off all power to your house. See how quiet it is then.

[Doug Young]

Originally Posted by Chipotle

Another test to try, though probably less practical: go throw the main circuit breaker and shut off all power to your house. See how quiet it is then.

I sometimes do that when I'm doing "serious" recording, especially if my wife is out of the house. It's astonishing. You think things are quiet, and then you you throw that switch, and it goes down a whole other level. One of the nice things about the Zoom, is you can use batteries, and really record with everything else off.

[TBman]

Originally Posted by Doug Young

OK, I tried the Zoom internal mics. They do need more gain, I ended up setting them at 7, resulting in a recording with peak level of -12db. So more gain on the preamp, and less final level. They may also be slightly noisier (not surprising, the cost of those internal mics has to be measured in pennies, not dollars). With a signal peak of -12 and noise level of about -67, that gives me a noise floor of -55 db, slightly worse than with the AT mics.

On the other hand, my wife's still doing laundry, but has switched from washer to drier, so maybe the noise difference is just due to that. Close enough in any case. Here's the raw zoom track:



and here it is brought up to level and mixed (no noise reduction). I left the silence at the beginning and end so you can hear the noise level.

The level of hum in both of these recordings is way lower than what I'm getting. Odd.