setting levels, tracking, and mixing the "right" way

S

skiz

New member
Hey guys, just wanted to ask a few questions that might seem a bit noob but i dont want to learn bad habits when recording. Instead i want to learn to do it the right way.

i record with a firepod directly into my computer which runs cubase SX3 with firewire.

when i set levels i turn the gain knob up till the red peak light comes on then i back off a few clicks.

When mixing if i see my track is clipping i usually just pull down on the master fader.

Now ive been reading a couple threads on this forum about having a quiet mix which i definitely agree with. Then people speak of having your mix at some sort of volume like -24dbRMS and embarrasingly... i dont know what that means.. i dont see that level in cubase and there's nothin like that on the firepod

Also ive read people like massive master say that its better to leave the master fader at 0 and then mix the levels in the track to be at about -24dbRMS, why is that?

then once the track is mixed and mixed down is a limiter used to reach a level of around -14dbRMS? or is it just that the fader is pushed up on the whole track?

anyway, id like to start getting these things right in future projects and develop better recording habits, so any advice is appreciated!
 
The most important thing here is the input. You're probably tracking too hot (depending on what "a few clicks" means anyway) and there is no way to repair that damage. Do yourself a favor as long as you're trying to avoid bad habits -- Manuals are good - But a lot of what is said in manuals ("set your levels as hot as possible without clipping" is a fine example) is complete and utter dung.

-24dBRMS is an overall level over time (while -24dBFS is a peak level at a given point in time) although you might find people referring to dBFS as an overall level also - and in your manual if they bothered to put in where the converters are calibrated (probably somewhere between -18 and -15dBFS, which would equal line level or 0dBVU).

I have a fairly extensive rant here: http://www.massivemastering.com/blog/html/blog_files/Proper_Audio_Recording_Levels.html

The master fader... IMO, it should stay at unity, period. With some programs, there's no particular difference if it's lowered vs. the individual faders being lowered. But it's just "really bad form" IMO and should be avoided. If your master fader is clipping consistently with all the individual faders at unity, you probably tracked too hot. Not that it won't clip here and there - Even with normal tracking levels, you're probably going to have to attenuate most tracks somewhat. But if your mix is even approaching full scale, then grab all the faders and back them down.

To "dumb it down" considerably -- Individual tracks should hit *maybe* -10dBFS if they're heavy transients (kick, snare, etc.) and considerably lower (maybe -18dBFS) for most everything else. One thing you can bet on - Almost anything you do for the sake of volume with insure that you'll never get it.*** Worry about clean, clear, focus, performance, core sounds while tracking at levels where you won't even be withing a cannon-shot of clipping the input.


*** Not that "sheer volume" should ever be the goal in the first place - But it's the best-sounding (clearest, cleanest, most focused) mixes with hoards of headroom at every possible stage that have the potential for that sheer volume. Less is more in almost every single case.
 
Read your blog and hoping i understood it!

Alright so basically there's no way, with my gear, to set where each individual track is peaking. (cause i dont have a scale on my gear)

So best solution to this is when tracking, turn up the gain knobs of each channel till it clips, then back off quite a substantial ammount instead of just 2-3 clicks?

In doing so ill leave myself enough usable headroom to create a clean airy better sounding recording that wont clip when all faders are on unity, and that will then stand up against limiting etc better later on?

sorry if i totally misunderstood your post, tryin to catch on tho!

thanks a lot tho massive master :)
 
skiz said:
Alright so basically there's no way, with my gear, to set where each individual track is peaking. (cause i dont have a scale on my gear)
Click to expand...
Actually there is. What you wan to do is set the trim and the input/recording faders in Cubase to unity gain - that's the 0 level about 2-3rds-3/4s of the way up the fader scale. This means that your Cubase software is not boosting or cutting the level of the signal coming in from your Firepod.

When set this way - which is how I do ALL my software tracking - your meters in Cubase will now be showing the unadulterated signal level coming in from the Firepod, in effect becoming meters monitoring the signal in the Firepod. You can then watch those as your adjust the gain on the Firepod.

Chances are two or three click below clipping like you're doing now is probably not a horrible level, but using the above method for metering it could help you fine-tune that. Maybe on one track you might only want to go 2 clicks while you 4 clicks on another, etc.

G.
 
right i understand what you mean

so then in doing this where abouts would i want the "meters" to peak?

about half way down the bar? a little more than that?
 
a little more than that, I'd say.

For tracking, I was told to...

Set my master fader at unity
Set my track faders at unity
and adjust the trim for the incoming signal to where the meters were about 2/3rds.

Hope it helps man....
Kel
 
skiz said:
right i understand what you mean

so then in doing this where abouts would i want the "meters" to peak?

about half way down the bar? a little more than that?
Click to expand...
If you really want to calibrate everything correctly and optimize your tracking levels, you shouldn't judge by the peak levels (other than to make sure they are not clipping, of course.) Rather, you want your signal "averaging" somewhere around -18dBFS or so.

If your meters do not have a number scale on them, then what I recommend is to get a test tone. For example, this site has a downloadable set of WAV files that includes a 1kHz sine wave recorded at -20dBFS. Download that file and play it back in Cubase with all your trim controls and faders set at unity gain, and it should show you where -20dBFS is located on your meters.

Then when you record your music, each track should be bouncing around above and below that line so that you can kind of eyeball that it's averaging right about that line.

Alternately, you can get RMS (average) reading meters as VST plug-ins that you can use to check levels without having to guess. Roger Nichols Digital has a free version of their "inspector" plugin that has all sorts of analysis metering on it, but mainly has both peak and RMS meters on it. Also PSP Ware has for free what they call their Vintage Meter, which simulates an old-style VU meter reading more averaged levels than a peak meter.

After a short while you'll get used to reading these meters and seeing how your Cubase metering corresponds to them, and setting the right levels based upon average levels instead of peak levels will become as second-nature as breathing, and you'll guarantee getting "sweet spot" recording levls for getting the most out of your tracking.

G.
 
SouthSIDE Glen said:
Actually there is. What you wan to do is set the trim and the input/recording faders in Cubase to unity gain - that's the 0 level about 2-3rds-3/4s of the way up the fader scale. This means that your Cubase software is not boosting or cutting the level of the signal coming in from your Firepod.

When set this way - which is how I do ALL my software tracking - your meters in Cubase will now be showing the unadulterated signal level coming in from the Firepod, in effect becoming meters monitoring the signal in the Firepod. You can then watch those as your adjust the gain on the Firepod.


G.
Click to expand...

+1 This is how I set my firepod up as well. Like Glen said if you have numbers on your meters just get your average peak around -18 and your set. This will insure plenty headroom for later mixing.
 
Massive Master said:
The most important thing here is the input. You're probably tracking too hot (depending on what "a few clicks" means anyway) and there is no way to repair that damage. Do yourself a favor as long as you're trying to avoid bad habits -- Manuals are good - But a lot of what is said in manuals ("set your levels as hot as possible without clipping" is a fine example) is complete and utter dung.

-24dBRMS is an overall level over time (while -24dBFS is a peak level at a given point in time) although you might find people referring to dBFS as an overall level also - and in your manual if they bothered to put in where the converters are calibrated (probably somewhere between -18 and -15dBFS, which would equal line level or 0dBVU).

I have a fairly extensive rant here: http://www.massivemastering.com/blog/html/blog_files/Proper_Audio_Recording_Levels.html

The master fader... IMO, it should stay at unity, period. With some programs, there's no particular difference if it's lowered vs. the individual faders being lowered. But it's just "really bad form" IMO and should be avoided. If your master fader is clipping consistently with all the individual faders at unity, you probably tracked too hot. Not that it won't clip here and there - Even with normal tracking levels, you're probably going to have to attenuate most tracks somewhat. But if your mix is even approaching full scale, then grab all the faders and back them down.

To "dumb it down" considerably -- Individual tracks should hit *maybe* -10dBFS if they're heavy transients (kick, snare, etc.) and considerably lower (maybe -18dBFS) for most everything else. One thing you can bet on - Almost anything you do for the sake of volume with insure that you'll never get it.*** Worry about clean, clear, focus, performance, core sounds while tracking at levels where you won't even be withing a cannon-shot of clipping the input.


*** Not that "sheer volume" should ever be the goal in the first place - But it's the best-sounding (clearest, cleanest, most focused) mixes with hoards of headroom at every possible stage that have the potential for that sheer volume. Less is more in almost every single case.
Click to expand...

Great post, dude - I'm not as guilty as some, probably, but I'd been following the old rule of thumb about trying to get your signal to clip no higher than -6db, to keep the noise floor as low as possible while not distorting the input signal. I'm thinking this is something I should re-evaluate, so I'll be grabbing those Roger Nichols Digital plug-ins, for sure.

Thank you!
 
I'd been following the old rule of thumb about trying to get your signal to clip no higher than -6db, to keep the noise floor as low as possible while not distorting the input signal.
Click to expand...
There's a rub in there -- A lot (a LOT) of preamps get considerably noisier approaching and passing line level than they otherwise are below. Although the SNR is measured at line level, it can be even better below line level.
 
interesting enough...ive always heard the -18FS = 0db (analog) rule.......but it never seems to happen that way for me.

what i do is.... (please correct me if im doing it wrong)...

i use an analog mix board/console for inputs and preamps, and run that into my presonus Firebox (for additional gain trim/preamp).

Ill set my mix console gain trim so the meters on the analog console are pumping almost at 0db (not quite clipping into the red). then ill check my Presounus box, and set that gain trim so its blinking into the red zone, and then click back on the gain so its not longer in the red.

Ill keep my master fader and recording track (in the DAW) set to (0) zero....and hit record.

......yet ive noticed the levels in the DAW fall around -8 to -6FS......and peak at about -3FS.......... Am i still to hot?? if i am, which gain trim should i be lowering? the analog mix console, or the presonus? or am i OK where im at?
 
bkkornaker said:
interesting enough...ive always heard the -18FS = 0db (analog) rule.......but it never seems to happen that way for me.
Click to expand...
Well, to be totally accurate, the 0VU = -18dBFS "rule" is a "safe" average of how the A/D converters in prosumer/pro gear are calibrated. There is no industry-wide standard, unfortunately. A lot of (especially older) European gear used a -14dBFS ADC calibration "standard", DAT and ADAT machines use(d) -15dBFS. Many current interfaces use -18dBFS - though some are at -16, others at -20. modern DAW desks can run from -16 on some of the cheaper models to as low as -24 on many of the newest, better-quality digital desks. There seems to be a trend to quieter conversion rates as ADCs and digital technology in general gets better.

Rather than having to explain that every time someone asks (and people are asking all thetime ;) ) the short answer is easier to just say "around -18dBFS, give or take". The best answer, IMHO, is to actually find out the specs on your converters and calibrate your software metering accordingly. On the Presonus Firepod - last time I checked, anyway - it was indeed -18dBFS.
bkkornaker said:
......yet ive noticed the levels in the DAW fall around -8 to -6FS......and peak at about -3FS.......... Am i still to hot?? if i am, which gain trim should i be lowering? the analog mix console, or the presonus? or am i OK where im at?
Click to expand...
Based upon your description, unless your meters on your analog mixer are off (doubtful, but not impossible), it sounds as though your extra gain is coming in from the gain control on the Firepod.

As described earlier, when everything on your computer (including any driver level controls, if available) is set for unity gain, the metering will reflect what is coming out of the Firepod, and what is coming out is a direct result of what is going into it. And what is going into it is determined by the input gain knobs.

It just goes to show you the amount of headroom (and possibly S/N ratio) that can be wasted. On your setup even though you are dialing back a couple of clicks from clipping the converter, you're still pumping the input by some 10dB or more than you really need to be. That's 10dB *per channel* that's just being burned for no really worthwhile reason.

Of course the absolute best answer would be inf the manufacturers put calibrated metering on their gear so we could actually see what was up. But god forbid that they actually have to raise the price another hundred bucks or so to do so. ;)

G.
 
while we're on the subject of less is more etc, i wanted to ask something about the makeup gain on something like a compressor.

when i use my VST compressors and i've set the attack release and ratio to what i want, i usually crank the makeup gain till just before it clips. Im taking it, from what massive master said about less is more, that this is definitely incorrect and will just chew up my headroom?

should the makeup gain just be used to match the volume of the track before the compression was engaged?

hah its starting to all make sense :P
 
skiz said:
should the makeup gain just be used to match the volume of the track before the compression was engaged?
Click to expand...
Pretty much, yes. Hence the name "make-up" gain :); it makes up for the loss in peak level brought about by the compression. Or, on an EQ, depending on just how you have the EQ set, it could be forcing a net gain or a net loss in overall signal level, and you'd use the output gain to bring the overall level back in line again.

It can get a bit more complicated than that, especially on the analog side of things where one might choose to set the amount of make-up gain lower or louder than just returning to nominal for the purpose of optimizing signal-to-noise level or for feeding a custom-level signal to the next box in the chain in order to take advantage of a certain type of distortion (or lack of it). But the short answer in general - especially on the digital plug-in side - is that output gain, or makeup gain, is mostly used to bring levels back to something nominal after they have been affected by the processing that plug just did.

G.
 
SouthSIDE Glen said:
Many current interfaces use -18dBFS - though some are at -16, others at -20. modern DAW desks can run from -16 on some of the cheaper models to as low as -24 on many of the newest, better-quality digital desks. There seems to be a trend to quieter conversion rates as ADCs and digital technology in general gets better.
Click to expand...
I got a tc electronic interface and a dbx 286a mic pre/processor and i try to live by that -18 rule when tracking vocals sometime quieter. but tryin to mix that is stressin me cuz thru my mackies everything sounds good level wise but i take that out and my vocals sound quiet. should i be trackin drums, synth, fx and all that with the same rule about -18? and what do i do to raise up all the levels later after i mix without peaking? I DO HIP HOP / RAP
 
bezubz said:
I got a tc electronic interface and a dbx 286a mic pre/processor and i try to live by that -18 rule when tracking vocals sometime quieter. but tryin to mix that is stressin me cuz thru my mackies everything sounds good level wise but i take that out and my vocals sound quiet. should i be trackin drums, synth, fx and all that with the same rule about -18? and what do i do to raise up all the levels later after i mix without peaking? I DO HIP HOP / RAP
Click to expand...
The thing to remember is that tracking levels and final mix levels are not meant to be the same thing.

When you track/record your stuff, the idea is to try and get the best sounding recording you can. You do this by recording at levels that find the best balance between gain and noise level while getting the best "character" or desired color out of each piece of gear and leaving enough headroom for the mixing process. What levels you will eventually mix at is irrelevant at that point, you just want to record the best possible source material to work with at that point.

It's after you're done tracking that you start to mix the tracks together, and it's there where you set the relative levels to mix the tracks together at the proper relative levels. The thing to remember when mixing is to avoid the tendancy to adjust levels by just pushing everything up. If something is too soft, you can make it louder just as effectively by pulling back on the level of the louder track(s) as you can by boosting the quiet one(s). This is where you keep an eye on your master mix buss meters as you mix. If as you mix your mix levels start getting dangerously high, often the best solution is not to pull back on the master faders, but rather to bring down the gain on the individual tracks. (In fact, I almost *never" touch the master faders myself, I just leave them at unity gain and touch them only if I want to perform a master fade on the mix.)

G.
 
Anyone care to take a crack at how tracking at near clip (*without any clipping*) reduces headroom if the the track fader for monitoring/mixing is pulled down enough to about -18dB?

I have been tracking in 24 bit well below what I used to track at in 20bit (which was near clip), but the logic is still missing in terms of the "headroom". I always considered my "headroom" to be that usable space between enough signal to not be too low, and not enough signal to clip. Why it would matter if it is at the upper end is what I mean. If I track hot and just pull down the fader, should be the same (relatively minus any added noise from the equipment) as if I track a little less, but the difference coming from possibly boosting later, making that track a little noisier. I am not saying I haven't noticed a difference tracking at a more moderate level, I am just wondering why.

I am not in a studio. I am referring to home recording. Noise levels are of utmost concern...especially background.
 
Outlaws said:
Anyone care to take a crack at how tracking at near clip (*without any clipping*) reduces headroom if the the track fader for monitoring/mixing is pulled down enough to about -18dB?
Click to expand...
That's a very good question with a very important answer. :)

The issue there isn't so much final headroom, but rather signal-to-noise ratio. By pushing the preamp on the analog side that hot you're boosting the overall noise level (both cumulative and inherent) and decreasing S/N before you even get to the converter.

Additionally, many (not all, but many) converters add some extra distortion in the last few dB before clipping, so that even if you don't clip, if you come close enough you might be getting some peak distortion. A minor point in and of itself, perhaps, but still better to avoid it than to accept it.

Put it all together and it come down to signal quality going into the converter. Dropping the input level on the digital side may return the headroom, but the analog signal which it's recording will in most cases be noisier with less usable dynamic range and potentially higher distortion.

By keeping the input levels at unity, we are then using the digital side to monitor the levels in the final stage *on the analog side* and help ensure that we are recording the best *analog* signal to digital.

It's really just an extension of the whole idea of gain staging and caring for the input signal level and quality at every link in the signal chain.

G.
 
You must log in or register to reply here.
Back
Top