Looking For Advice On Using A Noise Gate

  • Thread starter Thread starter Doctor Varney
  • Start date Start date
D

Doctor Varney

Cave dwelling Luddite
I'm a bit lost with noise gating at the moment. Wonder if I could get some advice? I think I know enough to know how little I actually know on this subject, so any beginner advice would be appreciated.

Purpose: Speech for audio narration. Bits of music are involved, but we're concentrating on vocals here.

What I've Done So Far:
Removed DC offset from samples; recorded at 44100 Hz, 32bit stereo. Added a noise gate to the vocal track on the mixer and given the volume about -10dB overhead - for now...

Environment: Computer. VST. Digitally recorded voice samples.

Results I'm Getting: The samples have some hiss. It seems to occupy a wide range of bands, not just in the uppermost registers. ( Before anyone says it - this is the LOWEST amount of noise I can achieve. Mic positioning, etc is the best it can be at this point. I've done everything I can in that dept. )

So, with the gate on, I'm stuck in the inevitable balance between hiss trails and losing the ends of words. A lot of these noise gates I'm trying happen to be features on compressors so it occurred to me to compress the signal a little to bring up the quieter bits that seem to be suffering the most from the effect of the gate closing too quickly down on it, before it's fully through. I reckoned a slight boost there would help push them through and my pronunciation is generally good, so t's and d's wouldn't get lost at least. But it's not really working. The results are still ropey and messy. So I don't know where to go from here.

Some of them don't always have a threshold setting and I thought that was important for gating, so you can set the noise threshold to what the gate should start to trigger...?

What I can't seem to suss with these noise gates though, is the compression side of things. I get the 'gating' idea:- release/ attack/ amount and all that... I also get the basic principle of compression (I know what it is) but I don't really know how to do it properly. I have no idea what ratio to use, for a start.

With all that said, these are only my guesses, going by what I've learnt so far. I have no idea if I'm doing this right.

And About The Order Of Things: Also I'm not sure where in the chain my noise gate should go. Makes sense it should go on the vocal track itself. But, also, I've experimented with it on a send insert, so I can route more than one source into it if I choose. I figured the EQ should come before the noise gate and the reverb should go last, so that the reverb tails weren't cut off by the gate. Does that seem about right?

Any help and advice would be most appreciated. I know you guys know your stuff and really look forward to hearing from you.

Dr. V
 
Hi mate
Just a query on why there's so much hiss in your recorded files.
Going right back to basics, are you running your mic through a preamp or just plugging straight into some kind of audio interface/soundcard on your PC?
A decent preamp can boost your recording level without adding any circuit noise which makes for very, very clean vocals and will allow a noise gate to perform much more effectively (should you still need it)

Also, just as a technical tip, a single voice doesn't need to be tracked in stereo.

When I need a noise gate I usually put it as the first plugin directly on the instrument/vocal channel and set the key input to focus on the frequency range I want to let through.
An alternative option, If you're using Protools, is to use the Expander plugin which allows you to set a rate of decay/closing of the gate which brings down the unwanted noise by degrees instead of slamming it shut once the signal drops below threshold.
This may be a bit more pleasing to the ear.

But if you can eliminate the noise in the first place, maybe through using a decent preamp if you're not already using one, then the rest is a doddle!

Dags
 
Hi, Dags.

A decent preamp can boost your recording level without adding any circuit noise which makes for very, very clean vocals and will allow a noise gate to perform much more effectively (should you still need it)

I'm using the pre-amp stage of my physical mixer. Behringer 1204fx-Pro. It's possible the hiss is coming from there but I don't know how to get rid of it. Strangely, I found that by turning the input gain all the way up, this lessened the noise coming in. So I think possibly these samples are the best they can be, given the situation.

Also, just as a technical tip, a single voice doesn't need to be tracked in stereo.

Would that make a difference?

An alternative option, If you're using Protools, is to use the Expander plugin which allows you to set a rate of decay/closing of the gate which brings down the unwanted noise by degrees instead of slamming it shut once the signal drops below threshold.
This may be a bit more pleasing to the ear.

I don't have Protools, I'm using FL Studio but I've added to the native plugins with third party VSTs. They're free and some of them are very good but I'm still on the hunt for a noise gate with a softer touch. Any suggestions? That is, if a noise gate is the right way to go...

But if you can eliminate the noise in the first place, maybe through using a decent preamp if you're not already using one, then the rest is a doddle!

I don't doubt for a second you're right. When funds allow, I'll be looking at getting at a better mic/ pre-amp. For now, I want to see how much I can get done with what I have.

Thanks for your help, mate. :guitar:
 
I can understand why the Pt plugs were recommended. They have fantastic GUIs that really simplify how these tools work.

Seriously. Check out a youtube tutorial for the digi expander/gate, even though you don't have it.
 
Hi again!
Hmmm....having to crank the gain on the desk all the way up is going to introduce loads of circuitry noise to your signal. I'd hazard a guess and say that this is where the majority of your noise problems are coming from. But you mention that the circuitry noise is *reduced* by fully turning the knob all the way up? weird :confused:

Maybe check your gain structure through the desk. (You probably have already done this but I'm just trying to help pinpoint where the problem is occurring)
As a test, back off the gain all the way anticlockwise and lower the channel fader/aux send all the way down.
Leave the master faders at unity (the zero mark about 3/4 of the way up the fader rail)
Press the PFL button and start talking into the mic whilst turning up the gain dial and see if the level meter shows that the input level is reaching the upper green or yellow area of the LEDs.
If this is all good, then your preamp is allowing you to get sufficient gain with the mic you're using.

Oh, I'd recommend using an XLR cable if you don't already as this can affect the level of the incoming signal from the mic. The jack input is for line level devices so this would definitely cause a signal drop causing you to have to crank the gain. In fact if you're using a jack plug cable with the mic this is probably the main culprit as to why you can't get sufficient mic level within the mixer.

I don't know how you have the mixer connected to the soundcard, but if it is via the control room outs, arm the track in Fruity Loops and start raising the fader until the level reaches the 'goldilocks' range in the FL channel. If it is via an Aux or Buss 3-4 out, then raise that pot/fader.
With any luck your signal should be sufficiently loud without the addition of too much unwanted circuitry noise.

Hope that something in here can fix the problem

Dags

Oh: re: mono vocal file
Would that make a difference?

Nope - not to the noise issue, but it will reduce the file size on disk and may give slight gains to your processing with only having to render a mono audio file, especially if you have loads of vocal files to work with.
More of a housekeeping and software efficiency tip than anything particularly useful :)
 
What microphone are you using?

All mics have their own self noise--it's generally low enough not to cause a problem but there are some noisy mics out there. If you're using a dynamic mic, you may wish to try one of the many inexpensive condenser mics which have a much higher output, because...

...Behringer pre-amps frequently get noisy when you turn them up beyond half to three quarters of their range. There are quite big variations...it might be worth trying each input to your mixer to see if any channels have less noise than others.

It's also worth seeing if you can increase the level coming out of your mic by getting closer to it. The way the physics of sound works, if you can halve the distance from sound source to mic, you get four times as much level, allowing you to turn things down and get less noise. Obviously, your set up will determine how much of this is possible.

Finally, I've never had much luck using a noise gate as a noise reducer. No matter how careful I am with setting the threshold and release, either I get some splashes of noise or some clipped words. I don't know if FL has a noise reduction effect but it's worth experimenting with this if it does (or searching for a plug in--sorry, can't help because Audition--DAW I use--has an excellent one built in).

However, it's always best to try and eliminate the problem at source rather than fix it later--a better mic, change of gain structure or similar might help because, in normal working, you should never have to put up with electronic hiss.
 
Trying what you said, Dags. I'll post back later on the results. Have to turn this computer off, because of the hum. My DAW computer doesn't make anywhere near as much noise.

Thanks again

Dr. V
 
Well, hi again, Dags. I did the following, as you suggested:

PFL on.

Gain cranked full, I'm afraid. But, with input track fader about half way, I think this did the trick of lowering the noise. Is that what you mean by 'unity'?

Alt 3/4 is my output to the computer. Had to leave that on full.

(When I say 'full' I mean: +0, not the physical limit of the slider).

Input strip panned full left. Alt 3 up full; while 4 is right down. Audio logger set to 'mono'.

Just pipping the bottom of the 'Goldilocks' range on the FL meter now...

Low cut on.

XLR cables all the way! (I've tried jack plugs in the past and had to jury rig an old guitar amp to get any gain!)

Got up close to the mic! As Bobbsy suggested. I had made myself a 'pop' filter out of an old radio-cassette speaker grill and flexi-rod. Since it moves about, I decided to utilise this to set, and fix, mouth-mic distance. So, chin against the grill, I can't go wrong. Works well! draped a pretty cloth over it to attenuate my b's and p's.

Normalised the waveform (0dB). Removed the DC offset (for good measure); not that it makes any difference to my hearing.

Result: LOADS BETTER! Nice big, even waveform without the obvious hiss. I still had to remove residual background noise from the gaps, by applying a volume envelope in the editor.

So thank you for the tips! You've definitely helped me out loads!

Cheers!

Dr. V :listeningmusic:
 
Last edited:
What microphone are you using?

KAM KDM400 Dynamic Microphone kit. Came with both XLR and jack lead. I use the XLR.

Someone told me these mics are directional and won't pick up BG noise if they're pointed away from equipment, like computer fans, etc. It cost £15 quid from a discount store. I aim to get a better one in future.

I get the impression Behringer are looked down upon by a lot of people and I can understand why but I get on alright with this little mixer, for what it cost me, second hand. I'm assuming that with a better mic, (like a condenser?) I can set the gain control much lower.

I took your suggestion by getting in closer, as said in my last post. Thanks for the tip!

I'm still using the noise gate to pick out residual hiss but with the settings MUCH LESS harsh on the noise gate, I'm finding it much more effective without needing such sharp attack and release times. I can't hear the effects of the gate now but it is definitely taking out that last bit of hiss quite seamlessly.

It could be better, but I'm on the look out for a better mic when I can afford it.

Thanks again!

Dr. V
 
Keep in mind, when you get your condensor mic, it will pick up every little noise in the room. So, if you're having problems with background noise now, the condensor mic will be much worse. In a music recording, it's easy to hide background noise, but for spoken word, you get no such luxury.

I suggest you really figure out the source of your hiss and eliminate it. Hopefully, it is the mic, but I'm doubtful.
 
I cant remember the last time I used a noise gate. I never had luck with it being easier than just editing out the 'in betweens'.

As Chili said, find the source of the noise, and fix the problem.
 
Keep in mind, when you get your condensor mic, it will pick up every little noise in the room. So, if you're having problems with background noise now, the condensor mic will be much worse. In a music recording, it's easy to hide background noise, but for spoken word, you get no such luxury.

Thanks for the knowledge, Chili. The background noise in the room at present could present an even bigger problem, in that case. I'm planning a sound proofed booth at one end of my small studio, which I'll pad out with insulation. Since you've said that, it seems I won't be wanting a condenser until I've finished building the booth. I've built an effective isolation cabinet for the DAW computer, as I just couldn't situate it outside of the room. Despite how some iso-cabinets can cause more problems (so I've heard) it actually does reduce the noise significantly, but not totally. Then there's a slight hum from the amplifier - and of course, occasional road noise, as I'm in the front room. Luckily, I don't live on a busy street and I tend to record at night anyway. If I have other actors over for readings, it will be during the day, so I will really need that booth.

Though, the mic isn't really sensitive enough for BG noise to be the issue. It's more hiss, which is something different, I think. But, like I said, I've minimized that to something I can now cope with.
 
I cant remember the last time I used a noise gate. I never had luck with it being easier than just editing out the 'in betweens'.

That's exactly what I was doing, Jimmy, before I got wind of the noise gate... editing out the noisy gaps, manually. That still left me with a slight hiss during the speech. So it sounded weird, going from hissy to silent, in between words. Now I've attenuated the gain a touch, by lowering the input channel's slider, it's greatly reduced and a very... I mean VERY slight... noise gate applied, I think this is the best I can hope for, with this set up.
 
. editing out the noisy gaps, manually. That still left me with a slight hiss during the speech. So it sounded weird, going from hissy to silent, in between words. .

But that's exactly what a noise gate will do too, even if it's applied perfectly. A noise gate doesn't do anything to the noise while you're talking. It only cuts it out when you stop talking and the level goes below the threshold. Editing by cutting out the silences does the same thing.
 
But that's exactly what a noise gate will do too, even if it's applied perfectly. A noise gate doesn't do anything to the noise while you're talking. It only cuts it out when you stop talking and the level goes below the threshold. Editing by cutting out the silences does the same thing.

Exactly what Rami said... That's why it's called a gate. Open it up and it lets everything through; signal, noise, neighbor's dog.

I think the small vocal booth might be a suitable solution in this particular case. As long as it really dead. Usually a small booth is not a good idea because it emphasizes the wrong freq band, but maybe for spoken word, it won't matter.
 
But that's exactly what a noise gate will do too, even if it's applied perfectly. A noise gate doesn't do anything to the noise while you're talking. It only cuts it out when you stop talking and the level goes below the threshold. Editing by cutting out the silences does the same thing.

Except, applying a noise gate is much less time consuming than picking through hundreds of .wavs with the mouse... and the same setting works across all of my recordings, as long as they're recorded at roughly the same over all level. The amount of time and labour I've managed to save by using the noise gate... Otherwise, this project would have gone into next year.
 
Exactly what Rami said... That's why it's called a gate. Open it up and it lets everything through; signal, noise, neighbor's dog.

Well, I thought that's why it has an attack and release curve. For adjusting the sensitivity. Actually, it seems to work on words as well, now that the over all level of hiss is much lower.

I think the small vocal booth might be a suitable solution in this particular case. As long as it really dead. Usually a small booth is not a good idea because it emphasizes the wrong freq band, but maybe for spoken word, it won't matter.

I think you're right, Chili. But it's designed to be big enough to seat two, thereabouts. I reckon I can get it anechoic and fairly well proofed. For spoken word, I think you can get away with less exacting conditions. The only music I ever make, is cooked up completely inside the computer.
 
Last edited:
Except, applying a noise gate is much less time consuming than picking through hundreds of .wavs with the mouse... and the same setting works across all of my recordings, as long as they're recorded at roughly the same over all level. The amount of time and labour I've managed to save by using the noise gate... Otherwise, this project would have gone into next year.
Well, that's a different reason than what you originally gave.
Well, I thought that's why it has an attack and release curve. For adjusting the sensitivity. Actually, it seems to work on words as well, now that the over all level of hiss is much lower.
No, the attack and release control how fast the gate opens and closes. When you're speaking, the gate is open and it's letting everything in, noise included. That's how they work. It's a gate. It's either open or closed.
 
Well, that's a different reason than what you originally gave.

But not the only reason I had...

No, the attack and release control how fast the gate opens and closes. When you're speaking, the gate is open and it's letting everything in, noise included. That's how they work. It's a gate. It's either open or closed.

Then perhaps it must be some other action of the compressor, which features the noise gate, that's reducing the hiss during speech?
 
But not the only reason I had...
Nothing wrong with that. :)



Then perhaps it must be some other action of the compressor, which features the noise gate, that's reducing the hiss during speech?
Probably not. It's simply because when you're speaking, you're masking the noise. The good news is that your voice sugnal is louder than the hiss. But the hiss is still there.
 
Back
Top