_Practical Mixing ?

[lawless]

I recently read "Practical Mixing" by Hugh Robjohns. Anyway, I have a question reguarding this excerpt:

"Whether you're using an analogue or digital mixer, it's
designed to operate most effectively with its faders at or near the 0dB or 'unity gain' position
(sometimes simply marked with a 'U'). This provides the best compromise between a useful
amount of headroom and adequately low mix-buss noise. So, when setting up your mix,
you should open each fader to the unity mark and then adjust that channel's input gain
control so that the signal is at roughly the right level for the final mix. Clearly, you won't
know what that is for sure, and EQ'ing the signal will affect its level too, but you can usually
get it pretty close with a little thought and some experience."


I am mixing inside Cubase sx 3... does this apply to me as well, or just to those with a hardware mixer ?

[ikon]

i beleive that when using digital software you want to be around the home of -18 and peakin around -9 because its not the same as analog where you would record close to 0...im not positive but i think this is correct

[Ironklad Audio]

http://www.digido.com/

there's mad amounts of info there about peak levels, and the differences between analog/digital peaking...and it's from one of the world's highest regarded ME's

i guess i'm somewhat fortunate in the sense that i have an analog mixer with digital sends/returns, so i don't have to pay all that much attention to digital levels - i set levels when tracking according the meters on my mixer, then calibrate the master fader in my DAW to match that of the mixer...this way the only digital levels i need to watch are the individual channels, and with this method, they rarely ever clip or get anywhere near 0db

[NYMorningstar]

Originally Posted by lawless

I am mixing inside Cubase sx 3... does this apply to me as well, or just to those with a hardware mixer ?

This doesn't apply to the mixer in Cubase because the channel strips don't have a separate gain for each channel. I believe what he is referring to is an outboard digital mixer.

[lawless]

Originally Posted by NYMorningstar

This doesn't apply to the mixer in Cubase because the channel strips don't have a separate gain for each channel. I believe what he is referring to is an outboard digital mixer.

Actually, sx 3 does have gain knobs for each channel; above the fader. I did, however, get the impression that Hugh was talking about hardware mixers but I was not sure...

[NYMorningstar]

My apologies, Sx does have input gain above each channel. I guess my head is stuck in Cubasis which didn't show input busses. Anyhow then yes, this would apply to what he is saying even though he's talking about outboard equipment. You want to activate the Input VU to check the level of your input bus and have the signal as loud as possible without exceeding 0 db.

[Raw-Tracks]

Whether on an analog or a digital hardware console, or a DAW based software console, I think the advice in the original post is pure rubbish.

This is based on my personal experience, your experience may be different.

When I first started learning to mix on an analog console, I approached the mix in a fashion similar to what that author is suggesting. Every single time, I would wind up with a mix where the 2-bus was just getting hammered. I quickly learned that I would have much better results if I pulled all the faders up to about -15db from unity as a starting point.

Also, there is confusion in this thread between "0dB or 'unity gain' position" on the fader and 0dbfs, or full scale in digital systems. Every DAW I've ever seen has had a "0dB or 'unity gain' position" on it's faders. You generally have about 6db of gain on the fader above that position, ProTools now gives you 12db.

[TuoKaerf]

Totally depends on the console. Some of the Yamaha digital boards I've used did not like to have the stereo buss hit very hard, while an SSL G series board sounds better with the meters in the red some.

[SouthSIDE Glen]

Originally Posted by Raw-Tracks

Every single time, I would wind up with a mix where the 2-bus was just getting hammered. I quickly learned that I would have much better results if I pulled all the faders up to about -15db from unity as a starting point.

Where were the imput trims on your strips set? What you are describing is entirely contrary to my experience and sounds like it might be rooted in improper technique or gain staging.

One of the biggest misconceptions in the use of mixers, IMHO, is the belief that the channel faders should be used to set actual track levels. I take the approach - as I think most should - that the faders should be used pretty much for levels RIDING only, and not for levels SETTING.

As was described in the original post, set the faders for unity, and then adjust the input/trim to get the right track level. If the levels wind up being too hot, then turn the trim down. If the trim is all the way down and the levels are still too hot, then there might be either a gain staging or a mis-wiring problem somewhere before the mixer.

G.

[Raw-Tracks]

Originally Posted by SouthSIDE Glen

Where were the imput trims on your strips set? What you are describing is entirely contrary to my experience and sounds like it might be rooted in improper technique or gain staging.

Well, your experience must just be different then. My above statements are based on years of experience working in pro studios. Many of which were spent watching mixing engineers who had far more experience and credentials than I'll ever have.

[SouthSIDE Glen]

Originally Posted by Raw-Tracks

Well, your experience must just be different then. My above statements are based on years of experience working in pro studios. Many of which were spent watching mixing engineers who had far more experience and credentials than I'll ever have.

That still doesn't answer the question of where were the trim pots in all this.

Regardless of who you were watching, the gain staging within mixers is designed such where what you describe should not be happening unless the signals going into the faders were too hot to begin with.

There are two main explanations for this happening in a pro studio where they should otherwise know what they are doing. Either A) they puropsely kept their trims open wider than textbook because they felt that the particular preamps on that particular board sounded better that way, or, B) they internally customized the board somehow (not uncommon in pro studios) with modifications to the strip preamps or custom aftermarket fader controls.

I was at a studio once that actually flipped their faders around backwards; "up" was low volume and "down" was high volume. The owner/head engineer liked it that way because it made more sense to him that raising volume brought stuff forward in the mix and lowering it pushed it back in the mix, so he turned his faders around to mirror that. Needless to say for anyone else going in (like me), adapting to that scheme was harder than learning to drive in England . Anyway, it would be obvious that on that board, fader unity would not be where the marking on the surface said they were. Similar things can happen when one installs new fader controls from another brand, or perhaps even when adding automation to a previously manual board.

But customization and engineer preferences aside, there should be no technical reason why one should have to throttle their channel strip faders in order to keep the buss levels cool.

G.

[Raw-Tracks]

The trims are generally set at unity.

I don't know if we are just not communicating here, or if our experience is really that different.

Let's say we are recording to a properly aligned 24-track where 0db on the VU meter is equal to a +4dBu. The tracks are all recorded at roughly 0dB on the tape machine's VU meters.

Now it's time to mix. If understand you correctly, you are basically suggesting that you set your faders at unity and mix with the trims. I don't see how that could, as a rule, be considered good gain staging practice. Take this example: It's a tambourine track. You want to have it relatively low in the mix and you also need to add a good bit of high end EQ to the track. You propose that I set the fader ad unity and bring down the input trim in order to keep the track low in the mix. However, the input trim is before EQ circuit. By bringing down the input trim, you are no longer sending a nominal signal level into the EQ circuit. It is a less than nominal signal. Now, when you boost with an 8K shelf, you are also really bringing up your noise floor.

Please tell me what is wrong with my thinking.

My experience is based on many different studios with a wide variety of consoles and a wide variety of engineers. Neve 80 series, Neve V Series, SSL's, Trident 80's and A-Ranges, Sphere's.....the list goes on. I will admit I have seen one or two engineers mix in a similar fashion, but the vast majority are using the faders and leaving the trims set to unity.

[SouthSIDE Glen]

Originally Posted by Raw-Tracks

I don't know if we are just not communicating here, or if our experience is really that different.

As is usually the case with things like this in this communication medium, I don't think we're all that far apart and that it's more of a communication thing.

No, I am not suggesting that mixing be done with the trim and the faders be left at unity. Frankly, after re-reading my post I thought that someone might have misuderstood me to be saying that. If that were the case, I'd recommend using sliders for the trim and knobs for the channel faders

What I'm describing is a method for determing the initial trim setting, instead of assuming unity on the trim, by setting the faders for unity and then adjusting trim for the "proper" (so to speak) input level for the strip. It sounds like you are describing just the opposite, to set the trim for unity and then adjust the fade for the right post-strip output.

I can see doing it either way, I suppose; I don't see much of a technical problem eitheer way off the top of my head. What makes sense to me about the way I describe, though, is that it is using gain staging to set the level going into the strip, rather than depending on the input signal from the outside world to be correct and not adjusting for it until it comes out the bottom side of the strip. If the trim were always just defaulted to unity, there'd be no need for the trim control at all. The trim, IMHumbleO is provided for the very reason to allow the engineer to insure the right levels going in.

Now, once the initial trim is set, then of course one adjusts and rides the faders as needed to actually mix to the buss, not only for the obvious dynamic riding of the levels, but also for relative adjustment of the levels between tracks. Again, the key here is that the trim is insuring proper staging control of the levels going into the strip and all it's various functions, and then the faders are used for the actual mix levels.

I don't know if your familiar with flying airplanes at all; I don't mean the big ones used by airlines and religious fanatics, I mean little single-engine private planes and the like. They have what's called a trim control in them as well, often in the form of a wheel...not unlike the trim knob on a mixer. The idea is to turn this wheel to "trim" the controls of the airplane so that it flies basically straight and narrow, even if you take your hands and feet off the wheel. It's not an autopilot, it's kind of more like steering alignment on a car. If you have to turn or go up or down, the pilot still does that manually; the trim is just setting the right "default settings" for all the steering controls on the plane.

I view the trim on a mixer in much the same way; it's setting the right default signal level for the channel strip operation and for passing through to the output fader at the bottom when it's set for straight and level, ignoring the insert and EQ processing in the strip. The output fader, OTOH, is like the wheel and pedals for the channel strip, which needs to be moved off of unity to actually fly the mixer, but not until after the trim is first set.

G.

[TuoKaerf]

With your EQ example, it is all dependant on how hot the signal is coming into the board. If the signal is hot, it might work well to drop down the trim some.

In the long run, use your ears. The workflow you use is what you know, and if you don't have any problems with it, keep at it.

Check out the book The Mixing Engineer's Handbook. The info on mixing is super simple, but the back of the book has interviews with 10-15 major mixing engineers. Almost all of them start with the faders at about 0 or -5.

[Raw-Tracks]

Originally Posted by SouthSIDE Glen

As is usually the case with things like this in this communication medium, I don't think we're all that far apart and that it's more of a communication thing.

OK, I think we're on the same page too.

I didn't mean to imply that I never use the trims. I'm with you, if need be, you boost or cut the trim as that sets the overal operating level throught the channel. If the signal is too hot coming into the channel, I trim it down. Too low, trim it up.

To get back to the original post though, I still think it's a bad idea to start with all faders at about Unity. Reason: Not only is it natural to push a fader up when you need a track to be relatively louder, it's easier to puch that fader up than to turn all the others down. If you start with everything at unity, you usually only have about 6dB of gain available, and that might not be enough.

If you start with everything at about -10 dB, or even a little less, that will give you room on the fader to boost where necessary. At the end of the day, after all the balancing is through, you normally wind up with the majority of faders closer to unity. But that is because you started much lower than unity. If you start at unity at the beginning of the mix, 9 times out of 10 you will find that halfway through your mix that you have no where else to go on the faders, you'll already have most of them at the top of their throw and your mix bus is probably pinning like crazy.

[mixsit]

Originally Posted by Raw-Tracks

... I don't see how that could, as a rule, be considered good gain staging practice. Take this example: It's a tambourine track. You want to have it relatively low in the mix and you also need to add a good bit of high end EQ to the track. You propose that I set the fader ad unity and bring down the input trim in order to keep the track low in the mix. However, the input trim is before EQ circuit. By bringing down the input trim, you are no longer sending a nominal signal level into the EQ circuit. It is a less than nominal signal. Now, when you boost with an 8K shelf, you are also really bringing up your noise floor.

That would be my take on it as well. You're setting up for the mixer's internal optimum range.
Either way, as long as the user is clear on which he is compensating (or compromising?) for; in one case individual channel 'internal range', in the other, taming the channel signals to keep the master out of trouble -if you happen to want your faders at nominal. Some people seem to like all the faders up where their range is spread out for example. I guess it would also depend on whether you would be riding a 'low level' fader.
Wayne

[TuoKaerf]

If you start at unity at the beginning of the mix, 9 times out of 10 you will find that halfway through your mix that you have no where else to go on the faders, you'll already have most of them at the top of their throw and your mix bus is probably pinning like crazy.

Nah, just compress the shit out of it. Same difference, right?

[SouthSIDE Glen]

Originally Posted by Raw-Tracks

OK, I think we're on the same page too.
...
To get back to the original post though, I still think it's a bad idea to start with all faders at about Unity. Reason: Not only is it natural to push a fader up when you need a track to be relatively louder, it's easier to puch that fader up than to turn all the others down. If you start with everything at unity, you usually only have about 6dB of gain available, and that might not be enough.

Agreed.

I think where our miscommunication came in was in the definition of "start" . I was talking "start" as the beginning of the channel strip setup, whereas you are talking "start" as the actual start of the mixing process itself. I think we were both right with our own definitions and that we are really in the "end" talking about the same overall technique.

G.

[Raw-Tracks]

Originally Posted by TuoKaerf

Check out the book The Mixing Engineer's Handbook. The info on mixing is super simple, but the back of the book has interviews with 10-15 major mixing engineers. Almost all of them start with the faders at about 0 or -5.

I'm not going to specifically dispute what those engineers have said in their interviews. What I will say, though, is that you should not put too much weight on what you read in those interviews. I've seen plenty of engineers say they "always do this" or "never do that" when interviewed. Then I work with them, and what they said in the interview is quite different that what they actually do on a daily basis. I don't think their intent is to mislead the reader. It's more like they don't want to admit in public that they work in methods other than the publicly perceived norm.

An example, "I rarely use EQ when tracking. I move the mic around and possibly change mics before I EQ." While a good idea, and probably a better aproach than twisting EQ's, it's not always practical in the scope of modern recording budgets and schedules. We're not taking 6-12 months to track a record any more. It's more like a couple weeks if you're lucky. You don't have all day to move mics around. You use your instincts to choose your mic and placement, and if it's close you EQ that last 5-10%.

[Raw-Tracks]

Originally Posted by SouthSIDE Glen

I think where our miscommunication came in was in the definition of "start" . I was talking "start" as the beginning of the channel strip setup, whereas you are talking "start" as the actual start of the mixing process itself.

That never even crossed my mind. You are correct, my "start" is referring to the start of the mixing session.

[Cyrokk]

I put the faders at unity on my analog mixer, get the trims to where they register as close to peak as possible (without listening to the mix, just watching the meter at the loudest peak of the wave form), then pull all the faders down before I even start mixing. Then I turn on my speakers and concentrate almost exclusively on the channel and master faders for balancing the mix, only adjusting the trim if the fader clip lights up (which never does). If there were parametric eq adjustments, then before I make last few final passes at the mix, I will make one more static adjustment to the trims.

I have no idea if I'm breaking some sort of mixing law, but my mixes are warmer and have better stereo separation as a result. I always figured I was using my mixer at its optimum operating level.

[SouthSIDE Glen]

Originally Posted by Cyrokk

I put the faders at unity on my analog mixer, get the trims to where they register as close to peak as possible (without listening to the mix, just watching the meter at the loudest peak of the wave form), then pull all the faders down before I even start mixing. Then I turn on my speakers and concentrate almost exclusively on the channel and master faders for balancing the mix, only adjusting the trim if the fader clip lights up (which never does). If there were parametric eq adjustments, then before I make last few final passes at the mix, I will make one more static adjustment to the trims.

I have no idea if I'm breaking some sort of mixing law, but my mixes are warmer and have better stereo separation as a result. I always figured I was using my mixer at its optimum operating level.

No laws broken at all, at least not in my jurisdiction . In fact I think you succinctly summed it up correctly and very nicely .

G.

[Cyrokk]

Originally Posted by SouthSIDE Glen

No laws broken at all, at least not in my jurisdiction . In fact I think you succinctly summed it up correctly and very nicely .

G.

Thanks . We must live in the same jurisdiction, because when I said the same thing here , I was met with a near complete consensus that I am doing things completely wrong as far as gain staging is concerned.

Could've been the same miscommunication issue, though. My mixer's trims are labeled as "gain", which everyone may have misread as "fader".

[SouthSIDE Glen]

Originally Posted by Cyrokk

Then we must live in the same jurisdiction, because when I said the same thing here , I was met with a near complete consensus that I am doing things completely wrong as far as gain staging.

Ok, I remember that thread.

Yeah, I think that thread and this one wound up talking about two different things altogther, or at least different assumptions were made by different parties. I think first there are a couple of lines to be drawn somewhat between the two threads; the first line is the line between the analog and the digital domains. the second line is the line between setting gain staging levels in a mixer and setting final output levels for digital mastering. Or something like that.

The other thread I think was referring mostly to ITB mixing and digital mastering, whereas this one we're talking about gain staging within an analog mixer.

When you used this quote:

Having finished tracking, now you come to the point where you will start to “set-out” your mix. In an ideal scenario, which never ever happens, you would have all your faders, including your masters, at 0dB. Your objective is to turn out a mix with 0dB levels.

it's my guess that they're actually referring to 0dBVU in an analog mixer, not 0dBFS in multitrack editing software. I think that may be the major source of the confusion.

If the gain staging is all set correctly, setting the trim/input/gain on the channel strip so that 0dBVU for that track equates to the maximum signal coming coming out of an analog mixer's channel when that channels' output fader is set to unity gain, and then setting the buss or main out faders so that the maximum signal level coming out of the mixer is around 0dBVU, this should result in buss/submix tracks coming through the converters (assuming they are also gain staged correctly) and into the digital "box" at somewhere around -16 to -18dBFS (give or take.)

That should hopefully make everybody happy. (I hope ).

G.

[Cyrokk]

Thanks for the clarification. I suspected those were the main points of confusion. I only use the computer as a multitracking unit for my analog mixer and I tend to forget that most people are probably mixing in the box.