head room, input gain and 0db fader level. whats the relation?

Fetusborg

New member
hello all
i've got a question thats been bothering me for some time.
i home record, but am pretty amatuerish. just doing things with not much theory backing.

i know that a mixing aim is to try and get your fader levels (the actual signal, and not the knob) as close to 0 db as possible. of course, considering that a high hat usually should be at a lower volume level than a vox track, for example.

but when i record a guitar track for example... i try and record at maximum gain before clipping. this means this particular tracks wave form is very full, almost box like, but of course, not clipping.

when this track is in the digital sequencer (i use kristal audio engine), the softwares mixer shows it peaking at like -(that infinity symbol) db, and the actual fader knob set at -21db. any louder, the guitar track will be too loud in the mix. and when i test out the fader knob, raising the db level to see where i begin to clip in this particular track... it clips at -6 db. but of course, the guitar is WAY too loud in the mix.


am i doing something wrong?
what is " head room"? does that have anything to do with what im doing?


ok, thanks for all the help.
bye
 
Fetusborg said:
am i doing something wrong?
what is " head room"? does that have anything to do with what im doing?

Probably. Headroom is an analog concept, because there is none in digital. It's the amount of gain above nominal level (0 VU) before clipping. If your nominal level is +4 dBu, then if maximum level in that particular piece of gear is say +24 dBu, then you have 20 dB of headroom. The theory is you would want the average level to be around +4dBu, which would be 0 on your analog meter, with plenty of headroom for peaks.

Digital, unlike most analog gear, sounds terrible if you clip, so you have to avoid it. Maximum level for digital is 0 dBFS, and that typically corresponds with about +15dBu at the converter. Therefore your converters will clip long before your analog gear might. Knowing this, your software seems to indicate 0 VU at around -15dBFS (a lot of software is 0 VU = 0 dBFS though, so be sure of what your meters are telling you!)

If your peak level really are -22dBFS, you'll want to track a bit louder than that, nothing too crazy, but -6 to -9 dBFS is good. Don't worry about maximizing levels until you master. If you've tracked a bunch of stuff way too quiet, then raise the level a bit in mixing, but again don't bother with normalizing everything as that won't help at this stage.
 
It's pretty simple.

Unity gain on a fader will represent no change to what the output of the preamp is, as far as using a mixer. Let's say you're setting up a live sound console: for each instrument, you adjust the gain or trim pot to get a nominally good signal. The sound person gets the musician to play while slowly turning up the gain knob. There's usually a clip light or something on the channel or preamp. Once this light starts freaking out, you're running with too much gain and things might start to distort, so you back it off a little until it's generally not lit. If you get the "occasional" blip from a stray note, that's usually normal and ok - it means you're in the right ballpark, but it's not essential as long as there's a good, strong sound signal coming from the preamp. If you have a properly calibrated meter, it should be reading around 0 dB from an analog console.

Headroom is the amount of space, or a cushion or safety zone before you start running into severe distortion problems. A decent preamp might have 12 to 18 dB of headroom. The optimum signal level is 0. Unlike digital, you can go over the 0 level and things won't instantly start distorting uncontrollably. The headroom is built it. If the signal is too hot, bad distortion could happen. If it's too cold, you get a wimpy sound and noise from op amps and other circuitry within the chain, adding mud.

Instruments with transients (fast percussion sounds) and high frequency content (like cymbals) can benefit from extra headroom.

These are generalizations. There's guys that obsess over this stuff, and things like calibrating their meters properly so that they know EXACTLY where 0 dB is. They want the maximum signal with absolutely no distortion added. A 1 kHz sine wave outputting 1 volt is a good test signal for calibrating your meters.

The process of setting this stuff up properly is called Gain Staging.

Once the band does its sound check, the balance might be a bit off. The fader is used to adjust this a bit - you can boost a little signal, and cut a lot, just to get a balance between the sound of all of the instruments playing together. The sound characteristics of the preamps remain where you set them.

In digital recording, the scale of the meters is different. It's dBFS as opposed to dBv or dBVU. They're all different, but dBFS (a digital signal) means that 0 dB is the very, very maximum signal level possible without bad things happening. There's absolutely no headroom beyond that point. Also, it's common for a properly set-up signal from an analog console or preamp running at 0 dBv or so to roughly correspond with -12 dBFS once it hits your computer or digital recorder. It means roughly 1 volt of signal. If your levels peak at -12, you have 12 dB of headroom.

Headroom is good.

This might be a bit confusing, because there are too many digital recording guides that tell you to pin your levels. For 16 bit recording, the advantage is that you're using "more bits" as the levels go up, so the resolution gets better. The down side is that if your signal goes over, it sounds bad and there's nothing you can do to correct it short of re-recording the track at a lower level.

For 24 bit recording, there's absolutely no excuse not to leave ample headroom.

So let's say you record 24 tracks, and none of them go over -12 dB in digital. Once you go to mix the song, if you have all of the faders set to unity, the added output of the tracks will still be likely to slam way over 0. You'll still have to turn things down to keep your levels reasonable. There's nothing wrong with this. If you're recording 2 tracks, this might not be the case.

My advice would be to try to track things at lower levels and compare to see if you get cleaner recordings. Use the instruments (level meters) to set up the gain staging to where you think it should be initially. Let your ears be the final judge.


sl
 
Last edited:
Another concept for you: Look at many analog mixers you will notice that on the fader strip there is a boxed section or shaded section that extends just above and just below the 0dB mark (Unity Gain). This is used as measure for setting levels while tracking. When tracking/mixing just one input you can crank the fader to the top of that line with minimal risk of using your headroom up. If you are tracking/mixing a few tracks than it is probably best to stay right at 0dB (or less if the track warrants), but if you are tracking/mixing a whole bunch of tracks (i.e. the majority of the available inputs), run the faders at the bottom of the boxed area to ensure enough headroom. You can of course go outside of this area in either direction if you so choose, but generally this region is designated by the manufacturer, based on their analysis of the best signal-to-noise ratio.

This in principal can be practiced in the digital domain. By running your digital gear a couple of dB below 0dBFS, you are creating your own headroom. Many people on this board have stated in the past that they always mix digitally at about -6dB, and I have found that mixing this way, helps give me much clearer mixes that don't sound quite as crowded.

Be sure also that the operating level (+4/-10) matches the settings with your soundcard, this will insure compatible mixer settings between your computer, and your analog gear.
 
i know that a mixing aim is to try and get your fader levels (the actual signal, and not the knob) as close to 0 db as possible. of course, considering that a high hat usually should be at a lower volume level than a vox track, for example.

but when i record a guitar track for example... i try and record at maximum gain before clipping. this means this particular tracks wave form is very full, almost box like, but of course, not clipping.


The intent is no sot much of getting things to zero as it is having a good balance of sound on the way in. In digital recording, the concept of "hotter is better" is not nessessarly true.

It might seem natural to boost things just before clipping, but the best thing you can do in digital is give yourself headroom. Yes, it does exsist in the digital realm, too. Although it's not the same as analog.

Good practice could be setting a maximum setting of no more than -3 db. Usually at around -6 is fine.

Let the track's breathe a little. If it's clean guitar, then it should look like a series of dips and peaks. If it's constant distortion guitar mayhem, then it would probably look more square-like, still with dips and peaks. Almost like a double edge saw.

It's obvious stuff, but the main emphasis being studying the shape of the waveform.

Bob Katz has some good articles on the best ways of achieving an "analog" sound in the digital realm. He explains the importance of leaving sufficent headroom, 24-bit recording versus 16-bit and a whole lot more theory that will guarantee mental collapse if you're not prepared. :D

Digital Domain (Bob Katz)
 
ahhhhhhh
wow, that cleared up many things
thanks for all the answers.

i did make a mistake though, and didnt specify my guerrila recording methods.

after the replies, i understood the difference between recording digital and analog.

i am recording everything digital. and when i say guerila, i mean i am not using racks or outboards. NOTHING. i record guitar and bass lining out from my amp straight to soundcard on pc. i record directly into my sequencer, kristal audio engine. some songs i mic my amp. it gets better results, but im still a bit hesitant as i have a band amp.
voice i do direct as well. but take out the amp. sometimes i pass it through a guitar pedal, as a means of a pre amp. but not sure if that works , haha.

when im about to record a guitar track, for example, i test signals in kristal. i play my lowest and my high note/chord and make sure that i dont hit the top of the meter. (it is a vertical light meter, but with no number markings, nothing. it doesnt even show up on the built in mixer in kristal. only AFTER i record the track, of course, i can monitor where i peaked).

as for head room... i believe i am recording wrong than. because when im signal testing... i try and get as high up on the vertial meter/bar as i can before any peak reaches the top, signifiying clipping. i control this using the sound card volume knob or the amp or guitar volume. in any case, after recording, i get a full, nice looking wave form.

but when i go to mix, my fader level can only go to AT MOST -6 db before clipping that track. and even at -6db it is FAR TOO LOUD in comparison with my drum tracks for example. (which i do in fruity loops, then render and import to kirstal)


what am i doing thats dumb?
i figure im recording pretty blindly if i dont even have a means of knowing how many dbs my signal is when going in.

i think maybe i should record at a lower level, having my meter go up only maybe half way, or less. this will mean my wave form will be smaller, yes, but i will have more head room on my fader in the mixer.


i think i went crazy now.

thanks, and sorry for the long ass post.
 
Fetusborg said:
but when i go to mix, my fader level can only go to AT MOST -6 db before clipping that track. and even at -6db it is FAR TOO LOUD in comparison with my drum tracks for example. (which i do in fruity loops, then render and import to kirstal)


what am i doing thats dumb?

Nothing really. If your guitars need to be at -6dB in the mix, OK.

i think maybe i should record at a lower level, having my meter go up only maybe half way, or less. this will mean my wave form will be smaller, yes, but i will have more head room on my fader in the mixer.

Really depends on the meter--how much is halfway? See, this is the problem with digital meters--there are no standards. Halfway might be -24dB or it might be -6dB, and if it ain't labeled . . . If you want to drop 6dB during tracking, you'll have to calibrate that meter so you know where that is.

Post a printscreen of your mix as a link if you can, that might be helpful.
 
Back
Top