Input Gain Levels Vs. Virtual Mixer's Level Sliders/Meters: Please Help

[Mike Freze]

Hello. I've asked a few questions before about input levels and the responses were really helpful. Now I am confused about another area with level balances in the signal chain.

I understand that you need to get as high an input level as possible when recording without going over that 0dB threshhold (to avoid clipping, distortion, to get the best signal-to-noise ratio, etc.) No problem. That means adjustments from the guitar volume knob to the amp volume knobs then to the input levels on the audio interface. OK.

I even understand that sometimes you might want to record those input levels a bit below what you could max out at if you plan to have dynamic variation in a song (raising an instrument or lead vocal during the chorus, for instance, but still not distorting the raised signal); then you can move the volume up or down through the ebb and flow of your song for one or all instruments.

Here's my question: oftentimes I have my levels through the signal chain set just right: no distortion, no clipping, then I get to my virtual mixer and play back what I recorded (or record on the spot) and the meters sometimes jump to the red when no red appeared on my interface after adjusting the earlier stages in the chain.

Why would it do so in the mixer?? Shouldn't the mixer reflect exactly what I set before it reached the computer recording software? If I turn doen the level sliders when red appears in the meters to get it doen in the green, then of course what I record will be weaker (quieter) tracks than I wanted.

Do you have to turn these level sliders down anyway? Is it a compromise you make between strong input signals earlier in the chain and the end of the chain (your mixer level readings??) Or does the level sliders/level meter just reflect what you will hear when you play back what you already recorded without affecting the levels you set for that recording? Seems to me that if I left the level sliders alone, let them go in the red, it stays that when when I save my prject: the red pops out as a part of the record; again, if I lower them down, then it defeats the purpose of having my original signals boosted up because now I'm forced to turn the signals down.

By the way: is the "channel level" I see in my mixer (I use Cubase LE) the same thing as my adjustments with the level sliders?

Mike Freze

 

[SouthSIDE Glen]

I understand that you need to get as high an input level as possible when recording without going over that 0dB threshhold (to avoid clipping, distortion, to get the best signal-to-noise ratio, etc.) No problem.

Except that that is not true. It one of those "common wisdoms" bouncing around the echo chamber of the Internet repeated ad nauseum by those that don't really know what they are talking about.

The "keep it below clipping part is true", but the "hot as possible to keep the S/N high" is not. The S/N range will be determined by the analog side of your signal chain (from microphone or instrument to A/D converter) because that will be a much narrower range than anything happening on the digital side. As long as your signal is anywhere near what anyone would consider "normal", the digital recording level won't have any effect on your S/N or noise floor level whatsoever.

What is "normal"? As long as your *peaks* are anywhere between -30dBFS and -6dBFS (give or take), you're doing just fine. Yes, it's that forgiving.

So how do you set the levels? If you want the whole story, you need to study up on the topic o "Gain structure". I have a good way to start with that with the "Metering and Gain Structure" applet resource on my website, but the key point for most folks is to leave the input faders on your recording software and driver software at unity gain and use the input levels on your interface to set the levels. Exact numbers depend upon your A/D converter calibration, but typically if you're recording that way so that your levels are peaking somewhere around -10 to -6dBFS *at most* with an RMS reading somewhere around -20 to -16 dBFS, you're cooking with gas.

If you're going into clipping on playback, after that it's either because you have something in the playback set too hot, or because you're playing back more than one track at once and the mix or the sum of them is pushing the total levels too high, in which case you need to drop the individual track volumes a dB or two each until the mix volume is within limits. BTW, if you record with the more conservative levels like I explain above, that summing going into the red will be less of a problem.

G.

 

[Chibi Nappa]

I understand that you need to get as high an input level as possible when recording without going over that 0dB threshhold (to avoid clipping, distortion, to get the best signal-to-noise ratio, etc.)

This is a common myth, and it's a bit of a problem. Don't record that hot. Record with the peaks around -18/-12.

Here's my question: oftentimes I have my levels through the signal chain set just right: no distortion, no clipping, then I get to my virtual mixer and play back what I recorded (or record on the spot) and the meters sometimes jump to the red when no red appeared on my interface after adjusting the earlier stages in the chain.

The mixer on your input shows just the signal you are recording. The mixing in your DAW shows the recorded signal + everything you are doing to it (faders, plugins, EQ, etc). That's one reason of many to not record anywhere near zero.

 

[Mike Freze]

Thanks, guys, for your valuable advice. I'm beginning to get a lot of what you're saying but I'm still a bit confused on one thing, I guess. Sorry to try your patience.

I still don't understand why one can have his amp up at a certain level he desires on guitar (say, to drive the speakers with a tube amp), adjust the input level on the interface (probably way down) to avoid it going in the red on the input signal. but then when you look at the sound being recorded on your virtual mixer (the channels), often one or several of the channels (even the guitar channel I talked about) might be going in the red ONLY on the mixer. Why would it go in the red there but NOT on the interface? I can play in real time, look at my interface, see it's staying in the green, then look up at my mixer channel and the same playing is showing up a bit in the red (even if I didn't jack the mixer level up at all).

Maybe the interface level and the mixer level are two different things. Is that the case? Is the recorded signal (software program) different? Is it more sensitive than my interface to going in the red?

Mike Freze

 

[SouthSIDE Glen]

can play in real time, look at my interface, see it's staying in the green, then look up at my mixer channel and the same playing is showing up a bit in the red (even if I didn't jack the mixer level up at all).

It could be a number of things:

The green on your interface may not have the fastest "ballistics", and may be more of a VU-style average-reading indicator. If that's the case, the fastest peaks may not b registering.

It could also be that the driver software (the software that acts as the bridge between your USB or Firewire interface, or your soundcard if you're coming in through there is "turned up" and not passing the signal through to Cubase unaltered.

There are other possibilities, but check those out first.

G.

 

[Chibi Nappa]

Like I said, the meter in your DAW is the instrument's signal + everything else being done by the DAW. If a fader is raised in your DAW, that will show up on the DAW meter. If you have a compressor or EQ on the track, that will show up on the DAW meter, etc.

The meter on your input shows only the instrument's signal.

 

[Farview]

Tell us what interface you have and which part of the mixer in cubase is clipping.

In cubase, there are the input channels (which show you the recording level) and the mixer channels (which show the level in the mix) and the output channels (which show the level going back out to the interface)

If you interface only has a clip light, you can't trust it.
If the interface has a VU meter, it's reading something completely different than the peak meters in Cubase.

Sometimes it's just a clip sensitivity issue. Technically, it is impossible to go over 0dbfs, so the only way to know if it clips is by counting the number of full scale samples you have in a row. Different companies set the clip light to trigger after a different number of full scale samples. In other words, your interface might be looking for 10 consecutive full scale samples and cubase could be looking for 3.

The bottom line is, at no time should you be within 5db of clipping. So the problem takes care of itself once you get proper levels.

Also, don't confuse recording level with mix volume.

 

[Mike Freze]

OK, Farview. So you're saying that an interface and a software recordi8ng program coud have different readings on "going into the red." If I'm understanding that correctly, then I just have to make minor adjustments along the signal chain.

You asked which interface I use: it's a TASCAM US-122L audio/midi combo. It's a 2.0, 24 bit, 96 kHz. Cubase LE cam bundled with it.

You differentiated between input channels and mixer channels (recording level vs. level of the mix).

Here's where I'm confused: what's the difference?? If I record at a certain level (input), fine. But if I make adjustments in my virtual mixer because it reads in the red (mixer channels/faders), like say I need to bring the levels down in the mixer, then re-save my recording, all my tracks are lower in volume: I lose some of the punch I wanted for all of the tracks when I play it back (or burn it to a CD).

Is that normal? Just the way it is?? If you want a final signal that's hotter than after you brought the mixer trim levels down to avoid clipping, do you have to resort to compression to get a louder signal back up again like you originally recorded before you brought the mixer levels down?

I DO understand your point about the output levels to hear what you recorded. I'm lost in the other two stages BEFORE you just want to listen to the output.

Mike Freze

 

[Mike Freze]

Guess I just don't undertand the difference between "recording levels" and "levels in a mix." Why would they be different? Is it just because some tracks you recorded end up unbalanced when you mix (maybe you accidentally recorded the guitar track too loud and the bass track too soft, for example) so you need to balance the overalll track levels later (the mix)?? Again, if you lower the levels on any track in the mixer, re-save it, it will end up a softer track then you intended. I've savede those settings, played them back, and some of the drive gets lost.

Mike Freze

 

[SouthSIDE Glen]

Guess I just don't undertand the difference between "recording levels" and "levels in a mix." Why would they be different?

Try asking it the other way around: Why would they be the same?

When one records, they are (or should be, anyway) recorded in a way that optimizes the quality of the recording. When it comes to the levels, that's all about "gain structure" - i.e. choosing the proper levels at each point in the recording chain to give you the best sound you want by keeping distortion low (unless you purposely want it from somewhere) and nose low.

None of which has anything to do with how loud you want that particular track to wind up sounding relative to the other tracks in the final mix. That's the major part of what mixing is all about - why it's called "mixing" and not just "adding" or "combining". Let's say you want your drums to be "behind" the vocals (just as an example.) That doesn't mean you want to record your drums quieter; you still want to record them at the level which will give you the best quality recording, and then adjust the level in the mix to make it quieter *after* you record it.

Additionally - and an important point that many newbs aren't ready for just yet, but eventually will make the difference between a passable mix and a good mix - is that you shouldn't count on just setting a stationary mix level for a track and just forgetting it. You will probably be wanting to take track X and raise it's level here to accentuate a cool little fill or riff, drop it there when it's less important, perhaps even mute it for that measure over there to make room for something else, that after listening to and working with the mix for a bit would sound better than what you originally planned. As you get better, MIX those tracks; don't just lay them on top of each other like dead pieces of plywood.

G.

 

[miroslav]

Guess I just don't undertand the difference between "recording levels" and "levels in a mix." Why would they be different?

Recording level is the individual level of your source(s).
Mix levels are the balance between tracks.
You can still keep your level the same as you recorded it for a given track....BUT...all the other tracks will then have to be adjusted to that reference. When you mix, you have to think of the MIX...not just "how hot is my guitar level".

Lets think about recording ALL of our individual tracks at a nice hot/healthy level. OK...they sound great individually.
But now you get to the mix and you have them all at that same hot/healthy level as they were during recording...and all of a sudden, everything is fighting for the same sonic space, so you have to balance the mix levels as to how you want the mix to sound...which is not the same as how you wanted just an individual track to sound.

When you mix, if your guitar isn't as "hot" sounding...first, set you speakers to a good listening level...then set the guitar track to where you like the level...now, adjust all the other tracks to blend in.
If the total mix level is making the stereo meters clip in the DAW...then bring all the faders down equally/proportionally and then raise the level of your speakers to compensate so it still sounds as loud to you, but it's no not clipping the DAW meters.

You can record every track at -0.1 dBFS going in...and it won't clip...but try playing them ALL back at that same level and your MIX will clip.
Basically something's gotta' give...so you need to balance the mix...which changes the levels from what they were individually during recording.

 

[Mike Freze]

Ah!! Miroslav. So you CAn record tracks at -0.1 dBFS that won't clip BUT they might in the mixer so you have to adjust that after. Got it!! One final question (PLEASE): if it doesn't clip in the recording tracks but might in the mix levels, WHY?? Does the software recording program give different responses to what it hears from your input levels on your interface that makes the tracks in the mixer more sensitive to clipping?? Is that why you need to re-adjust volume levels with the faders in the mixer??

I DO understand the need to adjust volume levels in the mixer to balance levels between all separate tracks (like you said). Fine. BUT that's different than having the separate tracks (channels) in the mixer vary according to going in the red on some, not on others EVEN before you adjust different track levels for an overall balance for your project.

Mike Freze

 

[Farview]

If you EQ any of the individual tracks, you are raising the level of the track (oddly enough, even if you are using subtractive EQ) Anything you do to a track that peaks at -0.1dbfs is going to push it over the edge.

If all your tracks peak so close to clipping, when they are mixed together, they will add up and clip the master buss.

None of this is a problem if you don't record so hot. Turning it down does not change the sound, that's one of the cool things about digital recording. If you are hearing a difference, it is really just your perception. Turn down the tracks and turn up the volume of your monitors.

 

[Mike Freze]

Thanks, Farview, for your great advice!

Hey, what if I DON'T EQ or do anything to my original tracks and some still go in the red in the mixer? How do you explain that?

OK: so the imput level of all instruments should be moderatley "hot" when recording to get a good signal-to-noise ratio. Then, if some tracks still"clip" in the virtual monitor, you just lower them down a bit. But when saving that version in your mixer that you lowered down, then what?? Have to compress everything to get all tracks to boost where they were at the interface level where they didn't clip??


Maybe I'm confusing what happens when you record a track. I can record a track, say, of my guitar (guitar through the amp through a mic to my interface and then into my computer). But I can do this without even bothering with the mixer: it will still record and not distort if I do my signal chain right. It's only after I look at my channel in my mixer that I might see the meter going into the red. I haven't played with EQ settings, effects, nothing. Just looking at what I recorded. No adjustments to the mixer whatsoever. Why would some tracks go "red" in this case??

Even if they do, here's a real big confusion: does your adjustment in your mixer for channels that are too hot really have anything to do with the good signal you recorded from the first stage through the interface when you had the adjustment right on that interface to not clip?? Sorry, still don't understand how a virtual mixer in your software program would ever go in the red if it stayed in the green on your interface (THE LAST STEP of your signal before it hits your computer).

I'm not talking about balancing volume levels amongst all the tracks in a project during the mixing process. That I understand: you can't record every instrument/vocal and exspect them to be perfectly balanced volume-wise when you play them back and look at them in the mixer. BUT that's a different issue. I'm talking about previously recorded tracks that go "red" in the mixer when they didn't throughout the earlier signal chains. See what I'm getting at??

Final question then I'll leave you alone: Does raising or lowring the fader volumes for any channels (tracks) in a virtual mixer affect the recording you made even before you looked at the mixer? Seems like it would. Try it: resave your recording by bringing down the faders on some clipped tracks and there you have it; lower (weaker) tracks you end up with compared to the other tracks that didn't clip in the mixere that you left alone. After that, now you not only have the problem of balancing all tracks to fit together volume-wise, but you have lower recordings to balance for those channels that were clipping. Raise them up won't don't anything to match those tracks that didn't clip, right? Try it: raise one of those tracks back up that clipped in your final mix, and guess what? It will clip again like it did to begin with. It doesn't solve the problem unless you lower ALL the other tracks doen to those that clipped.

Mike Freze

 

[Mike Freze]

One last thought I forgot: If you need to adjust volume levels in your mixer for individual channels (tracks), then what's the point of the gain stages right up to the interface if you have to re-adust all your volume settings anyway in the mixer? Seems a waste of time if anything you do wrong for the signal-to-noise ratio through the interface stage doesn't really matter because you have to re-adjust volume levels anyway when you mix. I once thought that was what the signal chain was all about: get all your signals to record at a decent, solid volume WITHOUT having to change them all when you see them light up in the "red" in your mixer.

Mike Freze

 

[bdenton]

Methinks you may be overthinking, and overworrying...

Don't get me wrong, signal-to-noise is always important. But, assuming you have reasonably decent gear, the noise present through your A/D converters and into your computer is insignificant. But, quite a bit of noise existed between the inputs of an analog tape recorder and the tape (and the tape itself introduced noise). That was just the nature of the beast.

Assuming the rest of your signal chain (prior to the A/D converters and including your room) has low noise levels, it's just not that much of an issue any more.

There are specifications, just like there are speed limits on the road. Which means that if I'm building interfaces, I can set zero anywhere I want it to be. And if I'm building a DAW, I can set zero anywhere I want it to be. And sometimes, between the engineers and when it gets off the boat from China, tolerances can slip.

Here's how I do it...

When setting levels, I look at the meters on the DAW while I adjust the level to -10 to -6dBFS. Then I look at the interface to make sure I'm not getting clipping there. If I am, I turn the level down until it's not clipping on the interface. If the level is too low there, I turn up the source (preamp, instrument, whatever).

That process may not have the USDA seal of approval, but I get pretty good results with it...

 

[miroslav]

I'm talking about previously recorded tracks that go "red" in the mixer when they didn't throughout the earlier signal chains. See what I'm getting at??

If you record at -6dBFS at your digital interface...it will/should still be -6dBFS in your DAW.
If it's somehow hotter and even clipping WITHOUT you doing anything to the track once it's in the DAW...then something is not right.

Final question then I'll leave you alone: Does raising or lowering the fader volumes for any channels (tracks) in a virtual mixer affect the recording you made even before you looked at the mixer? Seems like it would. Try it: resave your recording by bringing down the faders on some clipped tracks and there you have it; lower (weaker) tracks you end up with compared to the other tracks that didn't clip in the mixere that you left alone. After that, now you not only have the problem of balancing all tracks to fit together volume-wise, but you have lower recordings to balance for those channels that were clipping. Raise them up won't don't anything to match those tracks that didn't clip, right? Try it: raise one of those tracks back up that clipped in your final mix, and guess what? It will clip again like it did to begin with. It doesn't solve the problem unless you lower ALL the other tracks doen to those that clipped.

You're getting yourself all twisted up with this.
If you change levels and save...of course the levels will be different.
If you push them up they clip...if you bring them down they don't.
What's the real confusion here....????

Just forget the "red" while you are mixing/balancing. Most DAWs run at 32bit float, so you can literally max everything so it looks red...but it still doesn't sound like it's clipping. The DAW's math takes care of it for you.

So...ingore the "red" for the moment at the stereo master. Get your mix balanced how you like it...then if the stereo master is clipping, just lower the stereo master level until it's not. Of course that will also make it sound lower in level at your speakers...so, turn up the speakers. The key is to get the mix balanced!...then worry about final stereo master level.
If you were mixing in analog, you wouldn't want to just drop the stereo master if it was clipping, you would actually readjust your individual levels and leave the stereo master up at "0"...and it's good practice to do it that way also with digital...BUT...you don't have to. Digital allows you to just drop the stereo master level at the final stage without any detrimental effect to the audio.

 

[SouthSIDE Glen]

Digital recordng and editing software uses a technique called "non-destructive editing" in editing and mixing. What this means is that when you record, the software records the sound information to a file on your computer's hard drive. That file will have it's own sound properties such as tone and volume.

When you put that track into the virtual mixer, and make volume changes to that track and clip off a few seconds here or there, ad so forth, unless and until you specifically tell the software to "Render as.." or "Export as..", those changes are non-destructive. That means the original file that you recorded remains untouched and is still sitting there on your hard drive in it's original condition, retrievable as such any time you need to. It's only how it's played back within your mix project that's affected.

And even then, when you "Render as" or Export as", unless you pick the identical file name and location as your original recording, you will be rendering or exporting an brand new file with the new attributes; the original file will stil remain sitting untouched on your hard drive if you ever need to go backto it again.

G.

 

[Mike Freze]

Thanks again, guys! I think I have it all understood now. Bdenton, your last few sentences on explaining how you check and adjust volume levels for inputs and recording was very helpful. You kind of go backwards in the chain til you see no clipping: check DAW first, then check the interface, then go to the instruments (if need). That makes sense to me.

Mike Freze

 

[Mike Freze]

Miroslav, thanks. Guess I was assuming that when one lowered the faders in the mixer, like you said (of course the volume goes down). I just thought that when the volume went down there, so did your input signal strength (gain) from the instruments you tried so hard to keep as high as possible for a good, strong signal. I guess the strong input signal is a different thing than the volume of the recording levels in the mixer (channels). One can keep those input signals strong but still lower what level it's being recorded at and there will still be a strong signal-to-noise ratio.

Mike Freze