USB mic hiss/white noise

Audacity 3 Help Forum Windows
1

Audacity 2.0.5 from exe
Laptop: Win7 Ultimate 32 bit PC: Win7 home premium 64 bit


ok, so the ultimate answer here is very likely: get a USB preamp and a proper 1/4’’ jack microphone, instead of using a 16 bit USB mic. But for now, a USB mic is what I have. It’s a condenser with a shock cage which I have inside a foam lined box on a microphone stand for recording, on the end of a 5m shielded cable.

The problem is hiss. Quite a lot of hiss. I understand electrical interference is usually more of a hum, and this is definitely a hiss - it’s the very definition of white noise. Above an input level of 2.0 in the audacity interface the hiss and general noise level goes up accordingly, but what’s interesting is that if I lower the mic input below 2.0, the hiss does NOT get quieter. It can go to ZERO, and record… there will be hiss. If I go into the windows recording devices interface and check the box to listen to the device and apply, then go into levels and put it to zero… there’s that hiss again. There are no advanced features (e.g. noise cancellation etc) to choose from so there’s nothing to tweak. So something is not right here, and I would suspect it’s the mic or the USB interfaces (but that would be at least 3 entirely separate USB bus devices as there’s two in the main PC, for front and back panel).

I’ve tested the mic on a different computer (one with a louder fan), same issue (the hiss was exactly the same level, this is how I know it isn’t background noise), I’ve got the mic isolated enough that even the fan in my full size PC isn’t disturbing things in terms of actual ambient noise. I’ve changed it’s cable - it was already using a shielded cable that it came with, I bought another 2m shielded cable and tried that. Marginal improvement, but I could be imagining it. Ran the mic through a powered USB hub, connected to the PC with the shielded cable - again, marginal improvement but I think I’m imagining less of that.

Only thing I haven’t tried is a different USB mic to make sure it’s not the USB ports or sound card etc. That’s on it’s way because I’ve not recorded in a few months and I used to record at 0.60 input absolutely fine, so I’m fairly confident a large part of the issue is that the mic has somehow started to deteriorate. If it doesn’t sort it then I guess it’s save up a few hundred quid to retool with a USB pre-amp and a 1/4’’ condenser because frankly how else am I going to record to the computer?

In the meantime whilst I establish that, I want to know what i can do best to minimise the impact of the hiss at the point of recording. Are there any plug ins or settings i can tweak at the point of capture (e.g. dither)? I’ve been scouring the internet all day and trying things out (like noise gates, noise removal etc - which with the level of hiss involved just leave too many artefacts behind, again they didn’t used to), finally gave up and just decided to ask - post processing is only for polish anyway in my mind, this is something that needs to be addressed earlier than that.

Also see the attached spectrum analysis of some ‘silence’ (input to 0.2, no background noise etc), in case it’s any use. I’ve tried to use notch filtering from the tutorials but it doesn’t seem to make any difference. Wav file also attached.

Any help gratefully appreciated.

static spectrum.jpg

2

You can’t patch white noise in post production. You have to get the microphone right. Noise removal works by sampling the noise by itself and then trying to generate filters and notches to get rid of it in the show. White noise contains all audible tones, so Noise Removal tries to remove the whole show. It’s entertaining when it tries to do that, but usually not useful.

You have a different problem. I have microphone systems that are noisier than that. You are very probably using the mic wrong — or it’s broken. Do you have a good, robust speaking voice and are you about a foot away from the microphone? Can you overload it? This seems counterintuitive, but knowing where the vocal overload point is can be useful for testing. Speak louder and louder until the Audacity red recording meters go all the way up. Can you get there in your speaking voice? Do you have to increase the volume controls? The volume controls should be wherever your voice sounds good. Volume control is only a problem at the extremes. If you routinely run the controls all the way up or all the way down, then we need to find out why.

All microphones have noise; it comes from the preamplifier electronics and you can’t stop it. The goal is to make your voice louder than the noise.

What’s the make and model number?

I know of two different microphones that look similar, but one is a side-fire and one an end-fire. If you mount them and use them wrong, your voice will be suppressed, exhibit bad levels or distorted — and there’s nothing wrong with the microphone.

If you’ve tried all the variations and two different computers, then yes, the microphone could just be bad.

We don’t yet know what you have now, but what were you planning on buying? Generally a good microphone choice is an XLR type with a USB preamp. Computer microphones with 1/8" connector or rock band 1/4" connector can have cable problems. They may work OK in a dance club or a Skype connection at home, but the next step for both of those people is an XLR microphone, shielded XLR mic cable and USB preamp.

Our official position is to use a good USB microphone, but those can have cable and computer USB problems and you may be trading one problem for another.

Koz

3

The noise is NOT that bad! I wonder if you have hiss on the playback side? Try Generate → Silence.

That will create a digitally dead-silent file. I you hear hiss when you play back that file, the noise is a playback/monitoring issue.

What are you recording? Faint sounds, or solo acoustic instruments or solo vocals are probably the most difficult because they are dynamic with quiet parts that don’t mask the noise very well.

What are you recording levels like? …You have to consider both sides of the signal-to-noise ratio. If you have a quiet recording, you have to boost the gain to get the proper playback level, and you have to boost the signal and noise together.

Noise is analog. The noise on the recording has nothing to do with the computer (or the USB cable). There is always some electrical and acoustic noise. With a USB mic, ALL of the electrical noise is generated inside the mic.

4

Thank you for the replies.

I actually have no idea of the mic make/model - it is unbranded in terms of markings on it, and I no longer have the leaflet it came with. I suspect a chinese knock off frankly although I got it from a uk online store, but it really has been rock solid up to now. I can’t find a picture of the exact one now. It looks pretty much exactly like a Samson CUC1 (right down to the shock mount and tripod) and came in a fairly well presented and robust foam padded box. I paid about £70 for it and I’ve certainly never had cause to regret the purchase up to now.

Perhaps I’m using it incorrectly - I am assuming given the metal strip that divides the head that I must be singing into one side of the microphone rather than end on, and that’s what I’m doing.

In terms of volume - quiet is not what I’m known for frankly, i don’t think that’s the problem. I tried turning up the mic to get a decent wave form and belting something out to see if it’s any better and I guess it is, slightly, as long as I then turn the gain clider down to about -12 the hiss becomes bearable and the track reasonably clean before mixing it… but then of course the resultant track is very quiet compared to other recordings or music - full volume on the headphones was probably equivalent to 75% volume on a “professional” track. Not a complete disaster, sure, and probably livable with for the ‘demo’ level recordings I’m really putting out at the moment but far from ideal because I’d like to be able to do better than demo quality now and again. Note I don’t want ‘studio quality’ as such - I don’t care if the recording is not perfect and there’s odd ambient noise or a bit of echo from poor acoustics in the room (there are neither of those things as it happens, I’m just saying by way of demonstration) - but the hiss is really overbearing, at least to me - and I am not known for the quality of my ears and choosiness when it comes to perfect pitch and what not - I am one of those folks who listens to supposedly ‘hilariously bad’ singing on american sit coms and thinks: “sounded passable to me”. But this really is quite bad, perhaps the wav didn’t do it justice, I’ll try to do another with some actual sound on it to demonstrate.

Perhaps I’m being a perfectionist or perhaps I really am too quiet (never been an issue recording anywhere before mind you). So let’s rule that out, by asking a question: what level range would you normally expect your mic to be in for a reasonably loud singer or an acoustic guitar about 12 inches from the microphone?

If mine is set higher than 0.20 the hiss is really loud. If I set to 0.60 on the input slider, and record ‘silence’ the attached is what I get. Where would you normally put your input slider? I realise it will depend on the circumstances but ideally where? Bottom half or top half even?
frequency 0.6 input.jpg

5

Also - I tried to generate silence and I got… silence. Better mic it is. The new one might be an improvement though its the same ‘class’ of mic I guess, bottom end prosumer. If not, then I guess its save up for that preamp kit and an xlr. Hopefully my guitar>usb lead will arrive and will be decent enough that I get at least good guitar. Vocal tracks can be helped by muting in between sections at least.

6

The noise is NOT that bad!

See? I’m not crazy — that you know of.

Make a mono recording with your normal voice. “Testing, One Two, etc.” Then shut up and stop moving. Leave it just sit there with “Room Tone” for a couple of seconds. Stop it, Export as a 128 Quality MP3 and post it here on the forum. We’re too far into this while flying blind — or deaf.

https://forum.audacityteam.org/t/how-to-post-an-audio-sample/29851/1

Do you see those eight or nine upward peaks on your blue analysis pattern? Those are not supposed to be there. That could be a computer problem or cable problem. That kind of error I call “frying mosquitoes.” We come to a dead stop quickly because you changed both the computer and the cable.

???

Koz

7

I’ve said multiple times you don’t need thousands of dollars of microphone and sound equipment to get a reasonable recording, but you do have to start with a quiet room, pay attention, and sometimes get lucky. Koz

8

Kos - I will post a recording as soon as I can, probably tomorrow, but may have to be day after.

That spectrum analysis follows me from machine to machine - I can’t confirm whether it does that for every combination of the two cables and/or powered USB hub (e.g cable 1 no hub, cable 2 no hub, mic cable 1>hub>cable 2 PC etc), but the hiss certainly seems to. I think it’s fairly inescapably mic barring issues such as power interference (PC uses a different socket to the laptop, but same room, same house), I was just hoping there was some way to filter it before it became part of the recording but evidently not. It still might not be as bad as I think and maybe its a question of mixing things correctly and getting gain etc right instead, so I’ll get that mp3 made asap.

Oh another question folks have asked that i haven’t actually answered directly:
I am trying to record acoustic guitar and vocals (with occasionally bazouki and very very occasionally basic drum beat from maybe a bhoran or single floor tom or similar - at the moment I’m sticking to guitar and vocals because of my mic issues till I get the basics sorted). However, I am a strummer not a finger picker, on either instrument, so volume ‘in the real world’ is really not an issue, its just about what I can get into the mic over and above the hissing.

Thanks again for all the replies

9

Missed a question someone asked:

What mic have I got on the way: a Samson Meteor

http://www.amazon.co.uk/SAMSON-METEOR-Samson-Meteor-Microphone/dp/B004MF39YS

If that fails, I’m going to get something like this later in the year:

http://www.gear4music.com/Recording-and-Computers/Focusrite-Scarlett-Studio-Recording-Package/P6G?origin=product-ads&utm_source=google&utm_medium=cpc&utm_term=&utm_campaign=PLA+Focusrite&utm_content=uSq6CK2F|dc_pcrid_42682430239_plid__kword__match__&gclid=CImt2diyz70CFWbLtAodLCoAXA

10

Yes, but as a general observation, the RF characteristics of an operating USB device and its power demands can leak into the analogue stages and contaminate the analogue signal during the sampling process, before it gets converted to digital. The noise may not only come from the microphone.

Double-shielded USB cable or ferrite beads wrapped around the cable can sometimes help with RF noise.


Gale

11

Double-shielded USB cable of ferrite beads wrapped around the cable can sometimes help with RF noise.

Terrific idea. Part numbers and sources?

Noise will manifest inside the microphone, that’s the place your voice gets converted into electrical and then on to digital, that’s the place the system is most vulnerable, but it can come from anywhere. Have you heard your cellphone calling the tower for a quality check? It does that periodically, even if it’s sitting on its charger. I can hear that on my big sound system and haven’t bothered to separate the two sets of cables.

“This week on BBC-America, the new…[bee bee boop boop zap zap]”

One guy at work had his charger sitting right next to his computer and his sound system would announce the cell negotiation periodically — no matter how many ear plugs I used to tune it out.

I have a theory where Frying Mosquitoes is coming from. The digital sound signal goes down the USB cable from the microphone and on to the computer processing, but while inside the cable, it’s traveling right next to the USB power lines. Those are analog. I have no doubt that absent good shielding, your voice’s digital “fuzz” is picked up by the power lines and injected right into the arteries of the microphone.

We already have evidence in one case of the noise decreasing with cable length so that’s one point in favor of this theory.

Since everybody is always trying to save every penny, the computer makers are leaving out regulators and filters on the USB lines. Those cost money. Who cares if the battery signal to your mouse or keyboard is a little “dirty?”

But the microphone makers cheap out by removing protective shielding, power processing, filtering and regulation.

What’s wrong with that picture?

It used to be a safe bet that you could plug a USB microphone in and you were good to go as long as you conformed to the general USB microphone constraints. Now it’s a crap shoot.

“OK,” he said unwrapping his new USB microphone, “What’s the chances this is going to work the first time?”

Not as good as it used to be.

Koz

12

ok, so attached are files for the following:

80% mic input speaking slightly louder than natural

80% mic input speaking naturally

70% mic input speaking naturally

60% mic input speaking naturally

13

Right away there’s something very wrong. The first clip has no audio frequencies — pitch tones past about 16KHz. That’s slightly better than FM radio. Not anything you would want to make a studio recording with.

Audacity > Edit > Preferences > Quality > Sampling: 44100, 32-floating or 16-bit.

What does yours say?

Koz

14

44100hz, 32 bit float,

real time conversion medium (not sure why it’s usually set to high, not that it seems to make a difference to the hiss) with no dither

High quality conversion best and shaped dither (not sure why its usually none… not that it seems to make a difference to the hiss).

15

Right new mic new world. All is well. Nothing to see here…

With the mic to 0.2 still I easily can get the mic to clip at about 18 inches with voice only. Plenty of room to drop the gain on the final mix and remove what hiss is left without even bothering with noise removal I think. I’ll get some thing done and post it later and you can give me polishing tips instead. £30 quid off on a Samson meteor at Amazon folks get em while the getting good cos its a really good cheap option at least fresh out the box. Cable is a bit thin and only about 2m so I have to be closer to the laptop and its fan… And its fine.

Edit: I’m not naive enough to claim the war is won. I just wanted that out in the open :slight_smile:

16

Ferrite Rings:
http://www.maplin.co.uk/p/ferrite-ring-hem3016-qt26d
http://www.radioshack.com/product/index.jsp?productId=2103222

Double-shielded USB cable:
https://www.google.com/search?q=double-shielded++USB+cable


Gale

17

OMG I’m sorry - I forgot to say thank you for the help so far: THANK YOU.

I still learned stuff, so it was worth it for me. (still to find time to record a sample file of the new mic, but I will).

18

So you’re making these. They don’t come with the ferrite chokes.

http://www.radioshack.com/product/index.jsp?productId=2103222
I was going to say, I have clip-ons.

How would you know to add the bead/chokes — and where? Radio interference can be seriously magic.

Koz

19

As I understand it, you wind a few turns of the USB cable around the ring, or bead/choke close to both connections.


Gale

20

That makes sense. You’re trying to prevent radio energy from dancing down the cable’s shield.

If you have a jump rope connected to the barn door and you swing it fast enough, it will eventually break into two swings instead of one (Illustration).

Radio waves can do that, too and if you put a bead or choke right in the middle where it’s quiet, it won’t have any effect at all. This is a special case, but but it illustrates why suppressing radio trash is so darn much fun.

Koz
Screen shot 2014-04-12 at 7.30.14 PM.png