Headroom needed for "perfect mix"?

[crabtwins]

I never like hearing all this "we need headroom" talk. (just kiddintutt ) thats not the middle finger is it?....anyway...


Assuming optimum perfect mix (which is subjective anyway), how much headroom would you like?


Also i know that might depend on how much you plan on squashing it?

Anyway. i think i answered my own question again .not sure

[TheArchitect]

Originally Posted by crabtwins

I never like hearing all this "we need headroom" talk. (just kiddintutt ) thats not the middle finger is it?....anyway...


Assuming optimum perfect mix (which is subjective anyway), how much headroom would you like?


Also i know that might depend on how much you plan on squashing it?

Anyway. i think i answered my own question again .not sure

Why would I squash it if I were concerned about headroom?

[crabtwins]

Originally Posted by TheArchitect

Why would I squash it if I were concerned about headroom?

maybe i mean the range of peaks and valleys like i say often, im not even sure if im explaining it right.

maybe I mean WhyTF do mastering e's complain there isnt enough headroom then?
even though my words can't speak it i think i have answered my own question up there in the matrix. But I tried to make sens e of your riddle Architect

[Ged Leitch]

Originally Posted by crabtwins

maybe i mean the range of peaks and valleys like i say often, im not even sure if im explaining it right.

maybe I mean WhyTF do mastering e's complain there isnt enough headroom then?
even though my words can't speak it i think i have answered my own question up there in the matrix. But I tried to make sens e of your riddle Architect

Generally with digital mixes, the more headroom the better!

I find it much much easier to maintain the sound of the mix IF theres plenty of headroom to start with.

I need Spare DB's for EQ'ing!!! not much though...

it's when novice mixers strap their plugin limiter on the 2buss to make it louder that gives me headaches!......some compression is fine as long as they think it makes the "Sound" better, but anything done for level at the mixing stage really isnt needed.

Im sure the tech guys can give a more detailed explanation though...

[heathen]

Leaving a few db of headroom is very important, a lot of mix engineers really want to be mastering engineers too, me included. Mixing with too much compression or poor eq choice on the stereo buss will almost guarantee a worse off final result, compression more so as it is almost totally impossible to fix. The problem being listening to the same tracks so many times it is easy to lose perspective on a final master, thus you will either most likely overcompress or under compress or over eq something. Mix into your favourite comp and turn it off for the actual mixdown or do 2 mixes 1 with and 1 without. Then even take you fav comp too the mastering session if you cant live without it.

Now the headroom part, trying to use up all the available headroom is not needed in a digital medium as the noise floor is so low. Using all the available headroom is most likely going to result in the ME turning your mix down so it wont distort the inputs on thier analog gear thus losing resolution as the bit rate is truncated as the digital faders are brought down to a workable level. Most good mastering engineers have rooms, monitors, comps and eq's which are extremely expensive and of utmost quality. So thier work most likely will result in a better and cleaner volume and loudness boost.

I'd say good levels to send a mastering engineer depends on the music but a good guide is , do not go over -3 db peak and try to stay around - 18 db rms. It will result in a superior, louder, cleaner, more precise master. Also try to adopt the k system reference monitoring level, if you do you will be unlikely to ever have a problem with headroom.

[Ged Leitch]

...What he said

[davehome]

I generally try to make all my mixes peak around -6dB. That way when I bounce it, EQing or other post-proc. doesn't push it over the top.

However, some people don't like to do that, cause then they have to turn up their monitors during the mix... so my solution to them is to use a "mixing" limiter.

Find a limiter that preserves as much transients as possible (like Voxengo's Elephant2 - Settings: EL-2, Max Speed, 4x oversampling) and use it while mixing, then before you bounce, disable it. However, it may not be a good idea because after you EQ your mix, the same or other limiters will make it sound different and your target volume after limiting will likely be different due to extra peaking.

maybe i mean the range of peaks and valleys like i say often, im not even sure if im explaining it right.

Thats dynamic range. The difference between the loudest parts in your mix and the quietest parts. If you want to preserve dynamic range, First: turn up your monitors. Second: turn off any "mixing" limiter, Third: compress less.

Most of the time, compression is only needed mostly on the bass (0-140hz), so a multi-band compressor works wonders here. Sometimes it will be needed on mids (500hz-2.5khz) because of loud transients in snares and such. Compress very little on the highs.

Always EQ your mix after applying multi-band compression, because it changes the spectrum information as you compress a range and adjust it's gain.

Anyways... that's pretty much all the advice I can give you aside from simply using your ears as much as possible and that less is more.

[Timecode]

Another thing to bear in mind is that, depending on how the matrix is weighted, the mid channel in an M/S setup is generally at a much higher level than either the left or the right channel. So if the ME is going to go M/S without turning the original down he's going to need more headroom than usual (for the M/S plug I use I generally pull the peaks down to -8dbFS or so before hitting the plug to avoid clipping). At 24 bit or above leaving this much headroom is no problem at all.

Cheers
Matthew

[davehome]

Very good point.

[MASSIVE Master]

Adding to Heathen's post -

If there's a "natural" (non-limited) peak almost anywhere below -0, I'm happy. It could be -1, it could be -3, -6, -12... One engineer that sends in projects frequently rarely *ever* has a peak above -12dBFS. And as you'd expect, when he wants "sheer volume, he gets it.

But he doesn't get it because of where he mixes - He gets it because of how he tracks. He (as most of the better engineers I know) rarely goes above 0dBVU on the input chain. So his individual track levels are hitting around -18dBFS during tracking.

He has headroom *from the start* - He isn't overdriving his preamps, he isn't overdriving his outboard during tracking or mixing - He's using his gear as it was designed to be used.

It absolutely shocks me how few "noobs" understand that - They're always tracking "as hot as I can without clipping" - ramming preamps up to +12, maybe +18 -- And then wondering why their mixes don't sound "big" -

I know I sound like a broken record with this - Maybe not here, but at several forums. I don't mean to - I just don't get it. I suppose it's because I was brought up on "0dBVU" instead of "-0dBFS."

[heathen]

Yep, spot on John, ahhh I'm glad I learnt to track on analog tape all those years ago as well. A hell of a lot of mixes I do for young people these days I find exactly that, clipped tracks and not from deliberate clipping either, just very poor tracking levels, way too hot, add 24 together and there is little which can be done to make it sound good. It's like shooting yourself in the foot.

[MASSIVE Master]

I actually use "It's like shooting yourself in the foot before you even get a chance to load the gun."

[sergioelectro]

hello ,

I am quite a newbie in the mastering and recording world , but I was wondering what is the advantage of having a bigger headroom:

For example I read an article where a guy tells that the classic Mackie clipps at +16db, whereas other higher professionnal desks can go up to +26db or maybe more .

Does it means your mixes can be +10 db louder on the higher quality desks ?

If someone could explain me that I would appreciate .

What is the available headroom in a software like cubase SX ? ?

Thanks for your help

Sergio

[Bob Olhsson]

Everything we do distorts and degrades the sound in addition to whatever subject benefit there might be. At some point it turns to mush. The idea of maintaining headroom is to try and push that breaking point just as far downstream in the overall process as possible.

In the case of peak limiting, the precise amount needed is unknown outside the context of the final album sequence. Different methods sound better for different amounts of limiting. Until you know the precise amount needed, you can't choose the best sounding method.

[MASSIVE Master]

Originally Posted by sergioelectro

I am quite a newbie in the mastering and recording world , but I was wondering what is the advantage of having a bigger headroom:

For example I read an article where a guy tells that the classic Mackie clipps at +16db, whereas other higher professionnal desks can go up to +26db or maybe more .

Does it means your mixes can be +10 db louder on the higher quality desks?

No, it means those desks (probably the preamps in this case) have more headroom. You aren't supposed to use up all the headroom - It's there so you don't have to.

Originally Posted by sergioelectro

What is the available headroom in a software like cubase SX ? ?

Exactly what you put into it. No more. Digital has no headroom in itself... The gear before it and after it does.

For instance - If you record at "normal" levels (around 0dBVU) then it's probably going to equate to a signal dancing around -18dBFS on your DAW. There's your headroom.

If you do what a lot of the "noobs" have been doing lately (getting the signal "hot without clipping") then you're already using up all of your headroom in the analog AND digital stages of your front end. Shooting yourself in the foot before you even loaded the gun.

[mahasandi]

good point Bob thanks!

[Watersound]

So someone clarify something for me...I tend to drive my ANALOG preamps pretty hard (1073/Wunder PEQ-1.) Am I making a mistake, should I be well below that point where saturation occurs unless I'm "going" for that sound? I never bring that signal into my DAW more than -12 DBFS, so am I cool, or will driving my pre's hard cause problems down the road in mastering?

[MASSIVE Master]

If your digital levels are around -12dBFS, you're probably averaging somewhat lower (maybe -18dBFS).

If your converters are calibrated to -18dBFS, you're probably driving an ideal signal.

But in any case, it (overdriving preamps) will cause problems long before the mastering stage...

[Masterer]

Originally Posted by Watersound

So someone clarify something for me...I tend to drive my ANALOG preamps pretty hard (1073/Wunder PEQ-1.) Am I making a mistake, should I be well below that point where saturation occurs unless I'm "going" for that sound? I never bring that signal into my DAW more than -12 DBFS, so am I cool, or will driving my pre's hard cause problems down the road in mastering?

If you are beyond the "saturation" point and you are not "going for that sound" then you are getting a "saturated" sound unintentionally. Also known as a mistake. Do not do that. Getting an intentionally distorted sound is fine and dandy. Getting a distorted sound by mistake - not so fine or dandy.

What you send a mastering engineer has more to do with the level of the mix in total rather than the tone of an individual track or group of tracks [afterall you could distort the hell out of your preamps while still delivering a mix with very low overall level].


...sorry, I have a cold.

[SANDS]

I realize this is an older post and I have read many posts on here about this topic, with that said my question is with noise floor.

If we track where we sit around -18 inside the daw are we bringing up more of a noise floor when compressing and mixing in such down the road, compared to tracking hot but not clipping?

not sure if that makes sense, I have tracked at levels and found that in the mixing stage my noise floor sneaks up out of nowhere sometimes.

I assume this has alot to do with crap gear as well.

Since I have gotten better gear but I still haven't tracked as low as people have stated on here. I'm usually around -8 when coming inisde the daw. I also tend to track parts that are going to be louder in the mix louder when tracking,

example: say a pop style rock song: let's say guitars are tracked in the daw around -10, but the vocal takes are usually tracked in louder say -6

does that approach work or should I track everything around -18 inside the daw?

[Kayo]

Originally Posted by crabtwins

I never like hearing all this "we need headroom" talk. (just kiddintutt ) thats not the middle finger is it?....anyway...


Assuming optimum perfect mix (which is subjective anyway), how much headroom would you like?


Also i know that might depend on how much you plan on squashing it?

Anyway. i think i answered my own question again .not sure

I was just talking about the same, on another thread...

"Much prefer a spotless stage by stage transition to top form.
My personal suggestions are, try bring the mixes in at -14db RMS or maybe -12db RMS [raw mixes].
If I’m mixing, I aim for -14db RMS. Nice, clear and pristine and allows for mastering to achieve it's best. No matter what the genre of music it is..."


Ciao,
FD

[MASSIVE Master]

If we track where we sit around -18 inside the daw are we bringing up more of a noise floor when compressing and mixing in such down the road, compared to tracking hot but not clipping?

You're increasing the noise floor by adding gain to get the signal "hot but not clipping" as much, if not more than adding clean gain digitally (if you even need to).

For that matter, if your converters are calibrated to -18dBFS (plenty are - not low enough for my tastes, but it's not unreasonable), then "around 18dBFS" during tracking is exactly where the gear is designed to run. It's right where it's all spec'd at.

I don't track much anymore - But I tend to have individual elements *peaking* at maybe around -20dBFS (some extreme transients - Kick, snare, etc., maybe a little hotter, but I'm not going to sweat it - I'd rather have the extra headroom in most cases). That's where they're going to be in the mix anyway. If I track hotter, I need to turn it down anyway. I might was well capture it as clean (and focused, and clear) as possible. Then when it's time to start mixing, just throwing all the faders up to unity is 75% of the work.

[Jesse Graffam]

Originally Posted by karyoky****

tracking unnecessarily low is not good practice, and gets you nothing but files that aren't even using the available bit width.

Oh god, here we go again. Another topic that's going to spin way off into people arguing about bit depth (not width).

Please, can we NOT go there? Cos there will end up being 100 posts of people arguing with each other, and maybe 1 post that shows (for the thousandth time) how recording with -18dBfs peaks into 24bit resolution is not a big deal at all, and still ends up beating the noise floor (+ at least another 20db) of the capabilities of 24bit INT PCM.

So let's not go there again. I like that we finally have one conversation about ANALOG headroom going on for once. Albeit brought up in relation to tracking, it applies to mixing as well... if any outboard gear is being used, as well as for monitoring the master outputs. Let's talk about those ramifications.

I'll start by saying I'm in agreement with everyone who suggests not overloading your analog signal paths.

[Jesse Graffam]

You contradicted yourself there. Ignoring that; if you're beating the noise floor, so that the only thing left to resolve is noise with no audible (non random) signal within it... what else is left to resolve but more noise?

That's at -144dB mind you. -18dB and an additional -20dB for perceivable signal within the noise, still leaves you with 106dB of completely noiseless dynamic range (assuming your ADC is even that good) if all you're recording is signal that has the same RMS as peak. In a more real-world tracking environment you'll see more like 20dB sample-value peaks above RMS peaks. So a more realworld dynamic range is somewhere between 116-124dB, not including the assumed 20dB of noise above -144dB.

As far as overdriving, if a piece of gear is calibrated for -18dB RMS after ADC being the same thing as 0dBu on it's analog path, then yes!! Going above that is overdriving the gear BY DEFINITION.

Also, if a piece of gear is calibrated for -18dB RMS = 0dBu, then operating it as those levels is NOT "soft" but in actuality as loud as you should go before the manufacture decided that you will start running into situations where your audio quality becomes compromised.

[It'sJoeAgain]

back in the days when i used to do 8 trk pre-production work at a NY studio the owner used to demand i drive a hot signal to tape...we wuld be doin calibration tests to make sure we were *safe* but still it was a hot signal Ya...those who've worked with 4 and 8 trk multitrack know waht i m talkin bout im sure...then... hot out to 2 trck again! tis to get that SNR as low as u can with these *noisy* units...finally cassettes...no fifferent yo, just run the signal as hot as u can without distortion.... again due to snr issues...soooo runin a hot signal is not a bad thing with som old gear.....digital sh*t though, as ya'll know thats totally a different thing....

[MASSIVE Master]

here's a hint. when they came out with all the digital stuff they kind of screwed up as far as level standardization. so you have to learn to use input and output trims to get the best result.

No they didn't - Too many people don't bother to understand how those levels relate to the same voltages that have been the standard since the 40's (30's? - I'm old, but I'm not a historian).

converters do a better job of converting robust signals than they do converting less robust signals.

You should get new converters.

important components of musical tones are way below the fundamental. there is no reason to leave -18dBfs safety margin unless you are dealing with some totally unpredictable signal and you need to avoid the chance of ruining a recording with overs.

And there is also no reason to drive the system into distortion just so you can turn that distorted signal down later.

18dB of headroom is 24-bit audio is nothing.

Granted - We're generally talking about average levels as opposed to peak levels. That said, I tend to peak around there (or lower - Hey, it's 24-bit audio - I'd rather be 30dB too "low" than 3dB too hot).

[wado1942]

converters do a better job of converting robust signals than they do converting less robust signals.

Back in the early/mid 70s perhaps, when the S/N ratio of the converters were similar to cassette. In any converter, the distortions introduced in ADCs are every bit as high for loud signals as lower level signals. So if you feed your ADC a signal and it produces distortions 80dB below the fundamental, it will still be about 80dB below the signal even if you amp up the source signal before hand.

there is no reason to leave -18dBfs safety margin unless you are dealing with some totally unpredictable signal and you need to avoid the chance of ruining a recording with overs.

Most good listeners can hear up to a 20-bit range. You'd have to PEAK at -24dB in order to use the same resolution. You'd have to peak at -48dB to yield CD-quality sound, which most people still find acceptible. Averaging around -20dBfs RMS ensures you can set levels very quickly without getting any clipping. I just did a demonstration here locally and one guy was totally amazed at the fact that I never looked at my meters. He kept asking how I knew how loud it was and whether or not I was worried about distortion etc. I told him the monitors were calibrated so 85dB at the listening position was -20dBfs RMS on the mix medium so as long as the level was comfortable, there was no danger of distortion and I wouldn't be in danger of losing any resolution. So even on record, set the trim on the record channel to a comfortable listening level, you automatically have a good level in the DAW or multitrack without having to worry about it.

[Jesse Graffam]

Here's what I was going to post, before I got to page 2 and realized that karyoky**** is just here to troll the forums. fuuck

I'm not even going to bother to reply to his last message. He's obviously asking to get ban hammered.

-------------------------------------------

[edit]
removed my reply, it's not even worth it. this guy is obviously not going to listen to anything we have to say.
[/edit]

-------------------------------------------


John & Stephen made some very good points, which I originally had replied containing similar comments on before reading their replies, and for the people who are interested in improving the quality of their recordings... good advice to look in to.

Another thing I had commented on, yet removed, was that the CD Audio standard was created with these levels in mind. It's no secret that most early 80s CDs sound better than most CDs being released today. I'm not saying that the reduced levels on those early CDs is why, but I am saying that it doesn't get in the way of sounding great. Obviously. The recording is WHY they sound great, and most if not all were recorded with standardized levels for everything across the board.

If you're ignoring these standards now, and just tracking at a certain maximum dBfs peak, then it could only benefit you to research standardized signal levels (and monitoring levels for that matter) and how not adhering to them can cause a loss of clarity and transparency in the peaks of the waveform that FAR outweighs the loss of resolve in what is likely to be pure noise, at VERY low levels that won't even make it onto a CD. It's also worth noting that these distortions can effect other signals regardless of level - depending on the type of distortion, this can effect low level signals even more.

And @ tape levels... most tape machines are setup for around +24dBu before they really start to break down. This was done on purpose so that people could hit the tape with even up to +14dBu, if it sounded good on tape, so as to reduce noise. (which was already mentioned)

Most AD converters are not calibrated to have anywhere near that kind of headroom, +14dBu. For instance a converter calibrated to -18dB RMS in the digital world, you would be at -4dB RMS in the digital world by the time you got there. Obviously WAY too hot.



Now a question. At an -18dB (digital level) signal... how do you think this would related to +0dBu on a tape machine? I know it varies from model to model, brand to brand. But is there a general consensus on this, or is it just all over the board?