acceptable noise level when recording

where's your noise level when you record

  • above -56db

    Votes: 8 16.0%

  • above -66db

    Votes: 7 14.0%

  • below -66db

    Votes: 21 42.0%

  • below -84db

    Votes: 14 28.0%
  • Total voters
    50
crosstudio

crosstudio

New member
how low is your noise level when recording?

i've been trying to keep the noise level down to -84db (24bit) or so, but that often means that i don't record quite as hot as i would like to. i'm using a sansamp bass di connected to my mackie1604 preamp and my peak level is smacking at around -18db or so with the noise level at -84db. if i crank the mackie pre up a bit, the noise level gets up around -70db, and the peak level is around -6. plenty hot enough for me, but i'm worried about the noise floor.
 
Well it makes no sense to compare noise floor readings (for me anyway) with mics plugged in because I don't have a really quiet recording space, but a fairly close comparison would be recording my synth direct thru a pre.
I get a -73 dB noise floor (occasionally as quiet as -84 dB but "chatter" from the Canvas brings it down to -73 dB) with levels that can get up to ~-2dB.
Sound Canvas analog out into a dbx386, then out at 16 bits via coax S/PDIF into a GINA card.
 
OK, so where is everybody else on this?

Are you looking at your meters at all?

Are you eating some nachos?

Put down the Pepsi and look at your screen and tell me what you usually get when there is dead quiet in your studio or the closest you can get direct or non-direct.

Don't be afraid to name names.

I'm not ashamed to admit that I can't get better than -30 dB in my studio at the quietest rare moment I've ever experienced here with a mic connected and the metering on.

But S/PDIF in from a quiet device to Gina can go below -90 dB and she just saves 16 bits. Supposedly some sort of 20 bit stuff in there but how can that apply when I fly the shit in via S/PDIF at 16 bits?
 
C'mon Fess Up. We'll all benefit.

This is an important aspect of the composite of components that make up your audio chain. When you multitrack- either one at a time or all at once the noise adds up to an effect that can range from none to mild to distracting to ridiculous.

One of the main advantages of having this site available is to compare the results you're getting with those using the same or different chains of equipment. I think Ed is onto a nominal crossover point with -75 dB, but again that's HIS point of view.

And does the background noise picked up by mics that raise the noise floor to obscene levels cover up a poor noise floor or add to it.
 
Im still confused at what we are measuring, but with a sm57 and pre hooked up to my Delta Audiophile, im getting -85 with the pre turned all the way down, but with a nice hot setting that will clip with a whisper, it sets around -55...of course the air conditioner is on and the tv and a 4 year old is only a room away.....

I think it would be a good test of cards if we just measured the noise of the card, with nothin plugged in...measuring like we are doing is measuring the noise level of our mics, pres, board, etc...
 
now i tried my getting into the preamp and set it at a nice hot setting and im getting -65
 
my setup and noise floor

i've got my PC in an adjacent closet with holes drilled into the baseboard to pass the wires through, so PC noise is very much minimized. I've got plush carpeting so foot noise is reduced, and i either put padding (glued to a magnet sheet) over the vent or turn the heat/air off while recording (much to the shagrin of my wife).

i'm using the mic-pre in my mackie 1604vlz, and sending that signal (using the inserts) directly into a frontier design tango which is connected to a frontier design wavecenter pci card through lightpipe.

vocals:
i'm not a strong (loud) singer, but using an AT4033 and standing about 6" to 1' away from the mic, i get strong enough levels (-6 to -12) with a noise floor of -84db or so.

mic'd guitar:
when recording guitar (fender american strat, steinberger paddle ore), through a fender - deville using an sm57, and raising the trim till i get a -84db noise floor, i get levels around -18db or so. i could stand for it to be a little hotter, but i don't want to crank up the amp any higher, because then i don't get the sweet skank i'm looking for.

bass DI:
when recording bass, through a tech 21 sans amp, and raising the trim till i get a -84db noise floor, i get levels down in the -20's. i've got to raise the trim to about -70 to get the kind of signal strength i'm looking for (-6 to -12).

my keyboard and drum machine are no problem.

i'm just trying to get the most signal for the least noise, and wanted to know if anyone with close to my level of gear was getting a better signal ratio from their gear.
 
cross,

if i were you, id bring up the level of the preamp to get a hotter signal, probably up around -1, even if it means the noise floor comes up...i think the advantages would outweigh the disadvantages...you are leaving alot of headroom on the table....i could be wrong, but i had to throw my opinion out at ya.....
 
I'm set up for scratch vocal tracking in the downstairs room now for a tune I'm working on. With the NT2 up, and the mic pre set for ~50dB gain, the noise floor runs at about -68dBFS. When recording my wife (who can actually sing), it's about 4-6dB better- I can run significantly less gain at the pre for her. Monitoring the residual cruft soloed through headphones, you can tell that the largest portion is fan noise from the DAW and the D1624. If the HVAC fires up, that drops to about -60dB. The mods I did on the 1624 and the rackmount DAW were good for about 8dB of attenuation of their contribution to the noise. With the top off the DAW case, just to see the effect (because that makes it sound a bit like it did in stock condition), the floor rises to -58dB. And that is *really* irritating noise...

With the mics set up upstairs in the living room for serious vocal or piano work, I can get -73dB, but only after 7pm when there's no traffic in the neighborhood. Solo that up and listen to it, and you can hear the computer fans in my office and the refrigerator compressor in the kitchen- which are 3 and 4 rooms away, respectively. Shoot, you can hear the cats walking on the carpet, and the wind in the trees outside, the BNSF coal train grinding its way south on the tracks 3 miles away, and the neighbors watching Oprah. Seriously. I'm lucky to live in a very, very quiet neighborhood, and I ain't _never_ gonna get no -75dB!

I'm with Ed: if you can do as well as -75dBFS for a noise floor when you're rigged for mic'd tracking, you are doing *damned* well. And you probably have concrete walls a foot thick! I'll take the ~-70 I can get and be happy as a little pig in slop, and feel very lucky indeed for it. I really feel for you folks who are trying to track acoustic music in urban areas, with noise floors in the -50dBfs range: that's just gotta drive you _nuts_. Been there, done that.

This thread got me to instrument that wheezy, whiny noise floor on my Audiophile 2496: it's -74 dBFS. No wonder it pisses me off so badly.

Getting to -70dBFS for your mic'd acoustic noise floor doesn't happen by accident: it is very hard work indeed. It's all well and good to talk about sound cards with theoretical noise floors of -100dBFS, but doing a reality check like this is *very* useful sometimes- this is where the flux meets the tape, so to speak.

I like this thread a lot. Bruce- seriously, what do you get in your room? Inquiring minds want to know...
 
c7sus. The consultant I used long ago suggested a Line Voltage Regulator for my digital based studio. He went on some voltage techy babble thing about ADC's and DAC's needing very stable power to do their job effectively, blahblahblah...I just didn't doubt him (like I didn't on most things) and bought one. My gear has stayed in tip top shape and nothing has fried out.

One point I wanted to make on noise, and this should not be discounted when thinking about noise.

Noise is almost never broad band!

"Why is that so damn important"? you may ask.

Simply that just because noise exist's doesn't mean that it is in a range of your hearing that you will detect very well. Obviously, achieving an excellent signal to noise ratio is important for improving the appearent "depth" in your recording, but just because you have noise doesn't mean that you cannot HEAR through to noise to usable musical content that is lower volume than the noise, because the musical content is not always in the same frequency band as the noise is, thus, the noise does not mask the musical content.

Anyway, skippy is much better at explain that vodoo magical electricity and obscure spec stuff than I am, and could probably throw a little more light on this subject.

All that I am saying is that the "noise" floor you see has little to do with the noise floor you hear in most cases.

If you understand that Fletcher/Munson Relative Loudness Curves, you will get an idea of just what band of noise would make a difference in your recordings and not.

Noise is not always a good test of how much dynamic range you actually have available, and will be audible in your recordings.

Ed
 
is it correct that every 6 decibels are equal to one bit, and therefore recording at -18 at 24 bit, one is actually only using 21 bits?

THe noise in my room is 55dbs, when the mic is in this room, at acceptable gain, the noise level is -57 dbs, in the other room, its about 6-10 dbs less. with everything turned off.
 
Fletcher-Munson equal loudness curves: here's just the first hit I got in searching on them.

http://www.sfu.ca/sonic-studio/Equal_Loudness_Contours.html

It's a roadmap to your auditory apparatus, showing how sensitive your hearing is to energy in different frequency bands, and also how that sensitivity _changes_ with the absolute sound pressure level.

Power. It's a heck of a complicated issue. First, on powerline conditioning: good powerline conditioning will probably help, and may help a lot. It probably will _not_ reduce your line-noise problems to zero, but every little bit can help. If you have reasonably quiet power, it probably won't help that much. But if you're in an area with a 50KW AM transmitter 4 blocks away, then just having a big hunk of iron with a well-grounded electrostatic shield between primary and secondary between you and the outside world could be a huge win.

There are powerline conditioners, and then there are powerline conditioners. I'm probably the patron saint of surge suppression: here in lightning country, I actually have 4 levels of surge suppression between the power company and my gear: entry point surge suppression, the UPS that powers my DAW, the Furman power sequencer box, and then the multi outlet strips plugged into the Furman to distribute the power. All of those have MOV surge suppressors. They will hopefully give their lives to help keep the lightning out, but don't really do much for the quality of the power per se, in terms of addressing noise. If what you're picking up is a fairly small multi-outlet power strip that _says_ it is a power conditioner, that's probably all you're getting.

A real power line conditioner would be something on the order of a set of coupled inductors and capacitors (like the Emcor inline noise filters). Those are good at eating common-mode noise impulses, and bucking RF that is riding in on the power. They are also large, have only one or two outlets, and aren't very common.

Better yet would be a ferroresonant isolation transformer. These help take care of single-cycle transients (like blower motors or compressors kicking on), and actually filter and reshape the output waveform, by storing magnetic energy in their cores and smoothing successive AC cycles with it. The bad news is that they tend to be heavy, expensive, acoustically noisy, and get very unpleasantly hot. However, in the AM radio case above, or if you live next door to a welding shop, something like that may be the only way to get quiet power. Some vendors take this one step further, and produce "balanced power", where the neutral is floated with respect to ground, completely decoupling any possible moise from riding in on the neutral. There is different hardware to address different problems.

Ed was talking about noise almost never being broadband. True, in this day and age. The flipside of that statement is also informative: "broadband noise is typically much less offensive than noise with spiky energy concentrations". What the hell does that mean? The human brain is _unbelieveably_ capable of pulling signal out of noise. It can detect and comprehend signals (voices or music) that are 30-40dB _lower in level_ than masking noise, if the noise is broadband in nature. Ever hear someone calling you over the sound of a windstorm or a heavy rain, or fish a shortwave station out of the hash while tuning the reciever by ear? Your brain sees correlation, and energy concentractions in the audio band, as signals: and it focuses right in on them. Maybe it's a sign of primordial danger, or the call to the next meal back in the cave...

Which is why pitched noises are so damned offensive. We have evolved to be able to fish that stuff out of the soup, and there's no way to turn off that part of our perceptual machines. I can cope with tape noise or semiconductor junction noise, both of which are beautifully broadband. Digital wheezes (signals with local correlation that come across as pitched) drive me *nuts*, and stick out like a sore thumb, or a snake that's fixin' to bite me....
 
And in response to CyanJaguar: yes, one bit is essentially 6dB. If you are recording at -18dBFS, you are really only using 21 of your 24 bits, or thereabouts. However, you are probably still getting a couple more bits of usable resolution than you would if you were recording at 16bits, going absolutely full-scale, and sweating over the faders trying to avoid overs. Which is one big reason I love 24-bit recording: being able to print tracks between -6dBFS and -12dBFS, and not sweat as much...

If you're recording down at -18, your'e probably giving away a bit more than you really need to, and you probably ought to consider bringing the levels up at the preamp stage to something closer to 0dBFS. I try to shoot for -6, but your mileage may vary. It's always best to try to get the levels right up there _with the mic pre_, because your signal will never be treated any better by any other gain stage, all of which will add their own noise. At 24 bits, the quantization noise is well down there, but that doesn't mean that minimizing it is no longer necessary. Every step in the chain brings more noise, so you have to stay on top of it all the way through.

Crosstudio was using his mic pres to bring the resultant noise floor up to -84dB with no signal present, and then looking at where the peaks end up (at -18dBfs). That's one hell of a room you've got there, sir! Still, I think you might do better to do it the other way around, and run the gains on the mic pres according to the program material, and then let the noise floor end up where it ends up. After, it's the *peaks* you want to keep: you want them as far up there as you can comfortably keep them without overs, since you will be doing more processing on them down the pipe.

I understand that you want to run the guitar amp at exactly the level you have now for the tone, but your mic pre gives you the ability to then fit that signal exactly into your gain structure, without leaving any headroom on the table. There is no "right way" to do this, but my finding has *always* been that attentuating a hot signal (running the fader below zero) yields better results than trying to pour more gain onto it later to bring it up out of the soup (running the fader above zero). That's a great killer of overall noise floors, because the mix amps are never as quiet as the mic pres...

Your mileage may vary, though. Don't let me try to convince you to change your working style, if you're comfortable with it and the results you get!
 
I lied

sorry,

the best noise I could get even in the quiet room was -57 db. The vocal I tested with was clipped in places though, so I am pretty sure that I can get a -61 or -62 best case scenario.
 
the difference between comfort and acceptance

thanks skip. i'm comfortable with how i do things, but i can't accept that this is the best i can do. i've been toying with raising the gain on the pre's and so far i'm happy.

as far as letting the instrument dictate the mic-pre gain, what a brain fart. why didn't i think of that before?

i think that the question i really should have been trying to solve is:

once i get the sound i want from my gear (tone, and level), how hard can i push my mic-pre gain before the gain in sound level and the gain in noise/distortion from the mic-pre cease to be proportiionate.
 
i have two recording situations so far.

When im in my apartment with friends goofind around, especially in the summer the AC will be on so i can expect -48 with the AC on and -56 when its off dependin on day/night/business.

When i go to my parents house to record, this place is a fortress and "my" room is above the garage which gives it silence underneath. but i can get -60 WITH the AC on, and -68 without. I think i could do better if i tried harder. but i like just getting down to playing.
 
You must log in or register to reply here.
Back
Top