Mixing help, (or) throw me a fricken bone

[Dakota]

[It's the author again. Thanks everyone for your advice so far. For real.]

What about a spectrum analyzer? Do you guys think that it's a bad idea to use one in the mixing process? I've been using them on each individual track to see if there is anything weird happening in the sub-lows, to see where to roll it off, but as for the rest of the spectrum, what do you think? When is a good time to use one?

And as for reverb and delay, let me put it this way, do most reverb plug-ins sound bad or am I just terrible at dialing them in! I like how it gives a real sense of 3-dimensionality, but it always comes with a little cheese. Do you think that reverb in mastering should put the band "In a room", and reverb in mixing should be used as, for the lack of a better word, an effect?

Thanks again!

Spectrum analyzer: agreed with MoreSpaceEcho that in the end, it's just a thing of ears and "listen" - but along the way it can be very informative to cultivate a learning-based feedback loop of viewing the spectrum to direct your ears to problematic or nice frequency areas, and listening to better learn what may be helpful in the information the spectrum is telling you. But it's always ears in the end. Sometimes a spectrum looks even (and therefore should be "fine" if spectrum views really were any final arbiter of good sound) but some areas sound crappy and *should* be thinned out. Sometimes a spectrum may look totally wonky, but sounds awesome and should be just how it is.

Most efficient workflow for me (YMMV) - is leaving one spectrum analyzer on the main 2 bus all the time. Only looking at it some of the time, as needed. If individual tracks need scrutinized, just hit solo. It's faster than popping them into tracks, and uses less cpu.

And yes - most reverb plugins sound bad. I'm a reverb and delay freak and very fussy about that - I mix a lot of post-darkwave and art-rock and pop-psych sort of stuff, and ambience has to be really extra cool or it doesn't cut it. Ideally, having a few good outboard for-real reverbs makes a world of difference (springs, and good algo reverbs). Some plugins do make the cut though, when used alongside some hardware: Valhalla Room, Shimmer, and Ubermod - Soundtoys Echoboy (can do verbs too) - Aether, TSAR - and the Lexicon PCM suite if you can justify the outlay.

Agreed w/ MoreSpaceEcho - it's really rare that reverb is added in mastering - and even if so, it's usually the most subliminal little unobtrusive brushstroke. Reverb (and an overall sense of "putting a band in a room") really should be right in the mix.

And thanks, jgimbel! - "where things sound weird but I can't figure out why" - agreed, sometimes I'll get raw tracks to mix, bring them up at reasonable proportional volume, and go "oh gawd, what a pile of incomprehensible muck, maybe this is finally the song I'm going to totally fail to get a nice mix on - my ears are so confused I don't even know where to start" - discouraged and demoralized then, but with due diligence, I'll go down through tracks in order of prominence and do the "cut the uglies" thing and then surface later and listen to them together again - 90% of the time, then it's "oh yeah, no big deal any more, easy to see the way forward, let's just frame all these pieces nicely as per usual".

[dfuruta]

Why would you want the spectrum to look "even"??? Are you recording pink noise?

[ashcat_lt]

Why would you want the spectrum to look "even"??? Are you recording pink noise?

Yes!

That's not really a joke, but seriously...

It kind of depends on how the scaling on the analyzer is set, but most full band music does come out pretty flat looking across most of the spectrum. For years I've been running my music playback application to display an RTA visualization (rather than the "trippier" ones), and have found that a good mix does generally average out to be pretty even.

I don't have any Grammys, but I do agree that an RTA can be a useful tool for the reasons mentioned above as well as a reality check in a less than ideal monitoring situation. Of course it's all about listening, but if somebody's asking a question such as the OP, he probably needs all the help he can get. Some of us, for one reason or another, are unable to just say "that guitar has too much going on at 763Hz". We could try the "boost and sweep" deal, but I have some crap to say about that too.

But first, I wholeheartedly disagree with the assertion that there is a distinction between EQing "for tone" vs "for the mix". There is nothing but the mix! If I could get my bandmates to understand this I would probably drink a lot less. ( actually, I quit a year ago.)

Going along with that is the oft-repeated advice to never solo anything. It can be okay in editing, but unless the instrument is going to be heard all by itself then nobody cares what it sounds like in solo.

And that leads to my thing with the "boost and sweep". To my mind and ears, IME and IMNSHO, this is dangerously close to "soloing" individual frequency bands. Again, it doesn't really matter if that little area around 1K sounds terrible and nasal. It pretty much always does. That doesn't necessarily mean that it's a problem in any given situation. I have very rarely had luck with this approach, though I occassionally still try because "everybody else is doing it". I have better luck with "cut and sweep"

[dfuruta]

Maybe I'm talking out of my ass, and I know my personal mixing aesthetic is bizarre. But, it kind of seems to me that if one isn't able to identify a problem by ear & a little trial and error, it's safer to just leave it as is. I mean, if someone really doesn't have an idea of how to identify things that need some EQ, slapping on a spectrum analyzer is asking for trouble. I see people who don't know what they're doing using EQs with dozens of little notches to get rid of what they think are ugly frequencies & high passing everything, and it sounds like an awful, phasey nightmare.

A lot of this stuff is personal preference, isn't it? I like muddy low-mids and dull sounding drums. There's no reason that your music has to sound like everyone else's.

[Gregg Juke]

While I understand the concept of "it's the mix that matters, stupid" (to paraphrase saxophonist Bill Clinton), I disagree re: boost, sweep & cut, and soloing and eq'ing individual tracks are a no-no. Sometimes little bits of crap sneak by during tracking (an open mike with a squeek or a cough here, an instrument in need of eq help there, a particularly nasal vocal spike on an otherwise good take over there), and if the tracks aren't cleaned prior to mixing (or nowadays, at least a volume envelope applied when working ITB), these little crappy bits can have a cumulative effect of big time mix crappiness when the whole thing is put together.

While the overall goal should be for everything to fit in the mix and have its own space, the instruments and vocals should sound good individually, and the tracks should be clean, as that will do a lot of your mix work for you before you even start. If it doesn't sound good on its own, I'm doubtful it will sit right in the mix. But then again, what do I know-- have you heard those soloed tracks from "Gimme Shelter?" But maybe it could have been that much better...

GJ

[Nick Sevilla]

Hi,

Here's a bit of my workflow that I pretty much always do.
This is when I did not record the song I am about to mix.

1.- Listen to every track by itself. I don't "Solo" I mute every single track instead. Semantics...
This is to identify mislabeled tracks, unnamed "audio_n" tracks, to listen for deficiencies in the recording itself, such as saturated mic preamps, unwanted distortion(s), and odd pops and clicks and bad editing. This part has nothing at all to do with the mixing process. It's part of the "discovery" process. What ingredients do I have, and of what quality?

2.- if there are plug-ins in the session I received, I get rid of them. Seriously. If the artist wants the sound of a plug in, either they can print it, or give me the idea so I can use a better plug in, hardware etc to realize the idea in a better way for the song.

3.- After step 1, I can then decide how much artificial ambiance might be needed, so I go and create enough Aux sends in the session that I might need down the road. One important thing I found is to try to include the ambiance that is there. Say for example 1/3 of the tracks were recorded in one studio, which has some ambience to it. I listen to that ambience, and match it on one Aux, for the tracks which do not have that ambience. Then I can "marry" the other tracks together more easily. Of course if there is a shitty ambience, I consider how to reduce it, via gating or EQ.

4.- If I am mixing entirely ITB, I put all faders to zero, and then turn them down to some level below zero so when all tracks play together they still give me about 12 dB of peak headroom on my master buss. This is before any effects or buss mastering is applied. Just the raw tracks hitting at a peak max. of -12dBFS.

5.- If there is a Lead Vocal, I put that fader up to zero. This will be the loudest thing in the mix generally, so I already am giving it a headstart from the rest of the tracks. This gives me a much easier time when I do start mixing, since the Lead Vocal will not need as much gain plug-ins to stay ahead of the pack, and whichever processing I will use on the Lead Vocal will not need to work as hard to be heard.

As to FFT analyzers. If you are new to mixing, you can use one. But ONLY to examine individual tracks, NEVER to make EQ adjustments nor to analyze the entire mix.

I've had more than one artist run my mixes thru an FFT analyzer, and compare that to their rough mixes. And they say they looks similar. Yet they sound totally different.
Like they're trying to un-justify the expense of my mix, or whatever. An FFT analyzer will never ever tell you that a mix is boring or exciting,only your brain can do that.

Cheers

[MoreSpaceEcho]

I have better luck with "cut and sweep"

yeah. boost and sweep has always seemed silly to me. you know you need to cut somewhere, why confuse your ear by sweeping a big boost around first? make a cut and then move that.

I like muddy low-mids and dull sounding drums.

he's not kidding folks.

[dfuruta]

I'm most certainly not!


To the OP: to me, it sounds like the problems are mostly in the recording, not the mixing. A lot of the instruments sound a little dull and distant (perhaps you're recording in a not so nice room?), while the vocals are really forward?that combination feels kind of disjointed to me. Personally, I'd also pan some of the instruments off to the side. Regardless of frequency content, it can be hard to hear a bunch of active stuff in the same place.

I really don't think you need radical EQing or anything like that. Maybe some compression on the vocals, and maybe a little bit of reverb so it sounds like everything's in the same room. You're using mostly acoustic, natural sounding instruments; work on the miking and embrace the sounds you get!

[Dakota]

Why would you want the spectrum to look "even"??? Are you recording pink noise?

Ha! Indeed. What I meant by "even" as a spectral *reference* is related to pink distribution (-3db per octave slope) but is darker, around -4.5db per octave slope. (But that doesn't apply in the lows, under 200 or so it should be straighter than that, exactly how varies per song). A lot of spectrum analyzers can set for various reference display slopes. Set at -3db slope, pink noise would look flat, white noise would rise at 3db/oct. Setting at -4.5db slope view is most often useful for me.

Again emphasizing that in the end it is only ears that matter (!) - and plenty of great results don't boil down to this - but as a general observation, most good sounding full band or full spectrum orchestral or electronic (or what have you) mixes really do tend to distribute fairly smoothly across the spectrum over time. These kinds of mixes tend to translate better across different systems as well.

Agreed!:

For years I've been running my music playback application to display an RTA visualization (rather than the "trippier" ones), and have found that a good mix does generally average out to be pretty even.

When a mixer has really good ears and lots of problem solving experience, they just tend to do this (smooth total slope) as a matter of course whether or not any RTAs are ever referenced. So we really are talking apples and oranges and personal preferences for personal workflow that's fun and efficient. I have a lot of science education mixed with the liberal arts background, so it's pretty intuitive and fast to reference some spectrum science at the same time as being in a bleeding heart romantic poet beauty right brain listen mode. They are a complimentary team.

and if the tracks aren't cleaned prior to mixing (or nowadays, at least a volume envelope applied when working ITB), these little crappy bits can have a cumulative effect of big time mix crappiness when the whole thing is put together.

While the overall goal should be for everything to fit in the mix and have its own space, the instruments and vocals should sound good individually, and the tracks should be clean, as that will do a lot of your mix work for you before you even start.

Amen!

[Dakota]

I'm loving this thread - lots of food for thought and great and varied perspectives from all concerned!

- A side thought this brings up, how people are really crazy nowadays for saturator plugins, tape sims, decapitator, etc. - back in the tape and transformer days, every track had varying degrees of THD, and every transfer added more. Which of course engineers were generally trying to minimize!

But a side effect of lots of cumulative small amounts of THD is that it tends toward re-distributing simpler spectrums into more complex and spread spectrums, filling in gaps in the mids and highs and tending to saturate and suppress any pure tones that were sticking out - redistributing the energy upward into harmonics (and some intermodulation distortion, sum and difference tones and such). All this tends to automatically push a mix toward a fizzier foamier smoothly filled-in spectral slope.

Digital audio doesn't do this unless you tell it to. I know people often start dropping saturation onto tracks because it's just more Tuff or Organic or some such... but there is that extra bonus of evening and distributing the spectrum.

Thoughts?

[lee]

Has anyone ever noticed, not only the fine wisdom of audio experience in the Tomb, but the excellent grammar? Is there some parallel that I'm unaware of?

[jhharvest]

Has anyone ever noticed, not only the fine wisdom of audio experience in the Tomb, but the excellent grammar? Is there some parallel that I'm unaware of?

The awesome people here are intelligent adults and care about what they are posting.

[ashcat_lt]

Has anyone ever noticed, not only the fine wisdom of audio experience in the Tomb, but the excellent grammar? Is there some parallel that I'm unaware of?

We are some of the nerdiest nerds on the Internet!

Well, of course this is TapeOp. All rules are made to be broken, especially the ones which start with "always" and "never". IISGIIG, do what you need to do, to each their own and all that. I'm not trying to tell anybody that their processes are wrong.

But I think we could all name a number of instances where a track really needs to sound like crap on its own in order to fit into the mix. How bout that extra acoustic guitar track that's really meant more as percussion? Doesn't make much sense to me to spend much time getting a good, full-bodied sound to that track just to go and neuter it via extreme high passing. A DI off the piezo pickup will probably get you there quicker. And how about something like a bottom snare mic? Or that whole thread re: hi-hats where nearly everybody is specifically saying to make the close mic sound like crap on its own? We could go on, but the point is that it seems strange to me that you would put any extra time into EQ moves in solo which you will just undo when you go to fit it in with everything else. I guess it could be argued that it's easier to take out stuff you don't need than to add in things you didn't capture. But, if you know you don't need it...

As for the "boost and sweep" thing: For me it just doesn't work in most cases. Among other things, it has a pretty big impact on the overall level of the track in question. Maybe not such a big deal if you've got the solo button engaged, but we've already discussed what I think about that. "Cut and sweep", at least with a narrow Q, will have far less impact on the overall level (there was an article somewhere which showed the math...) and give you a better idea of the impact of your moves in context. This way, too, you don't get hung up on one small band which sounds nasty on its own, but is actually essential to the overall sound. Instead you listen to see if things get better, or if you end up missing those frequencies.

I personally have my default DAW template come up with an instance of Voxengo SPAN across the mix bus. I freely admit that I use it as a crutch. Well, maybe more like a walking stick. I don't depend on it to get me where I want to go, but I do lean on it now and then when the going gets tough. My situation has me sort of "pre-mixing" - narrowing things down to pretty close - at home. I'm either on headphones where I just can't hear the low end or on a 2.1 system in an untreated room where the bass is just generally loose, and there's some weirdness in the crossover region. I'm not going to sign off on a mix till I get it up on the big monitors in the studio and check it on a couple different systems, but if I combine what I know about the deficiencies of my system with some visual cues from SPAN I can usually get pretty close and save some of my precious studio time. It can also help save some time zeroing in a problem area, or to get a quick idea of where a couple of tracks might be overlapping or fighting for space. Yes, the ears make the final decision, but I am not ashamed to use all of the tools at my disposal. And then there's the thing where it can help newer folks to train their ears to recognize various frequencies by both seeing and hearing.

Does that come across as overly defensive? I'm just saying here's some ideas, somewhat contrary to what other folks have said. I dont think that any of it is completely whacko. If you've never tried these techniques then go ahead and try both and see what works for you. If you are one who's set on one methodology maybe you could just try something different? Maybe one way works in one situation, but the other will work better somewhere else. Fill your toolbox and don't rule out anything by default.

So onward...

The thing about the reverb leads to an idea that I think applies to a lot of the techniques discussed in the thread: often subtlety is key. For the most part a little dab'll do ya. My rule of thumb with just about anything which is not meant as a special effect is to turn it up (or down, in the case of "corrective" EQ) until you can just hear it, and then back it off a little bit. It's pretty standard advice when we're talking about compression, but I think it works just about everywhere. You may not be able to consciously hear the reverb, but if you mute it you miss it. You don't notice that little dip at 1201Hz, but if you bypass the EQ it suddenly sounds a little funky. We often advise metalhead guitarists to turn the amp gain down a little bit.

And I think this can apply to entire mix elements as well. Maybe that shaker track doesn't really need to stand out and say "I'm a shaker!" Maybe we don't need to hear the rosin on the bow of the cello track or every breath that the flutist takes. Sometimes we can just tuck those way down under to where you can't even name the sound, but it adds a little something to the overall vibe. One might ask "why bother if you can't hear it anyway?" But not everything can be big, and sometimes a lot of little things can add up to sound even bigger than a couple of really big things. Not that we're always shooting for big...

I think that this then sort of leads back around to arrangement vs eq. Sometimes you just can't get one single track to do all of the things which you might like it to do in the mix at the same time. You can have glassy highs or nice tight punchy low-mids, or you can shoot for the middle ground and end up compromising both. Or you can look again at the arrangement, maybe double the track with each one taking care of a separate area of the spectrum. I end up very often with three or four guitar tracks - each playing one or two strings and with sound optimized to suit the part of the mix that it needs to handle. And yes, sometimes these individual tracks sound like almost nothing on their own, but added all together it gives an impact which just can't be had with a single track.

[Gregg Juke]

Great post Ash, and it doesn't sound defensive at all. One thing though, and you basically already answered it for me, but I have tried the "If it's supposed to be a crappy sound, why bother recording and EQing it well?" approach, and in every case, it has always come back to bite me! It really is easier/better to make a good sound crappy, than vice-versa, and this is strongly tied into arrangement too.

I have pretty awesome instincts and darn good "audiation" (hearing the completed song in my head), if I do say so myself. But sometimes, the part (more specifically, the sonic aesthetic of the part) I heard in my head just doesn't work in the physical world. In these cases, it's always better to have a good, strong signal, solidly recorded track to go back to and try to salvage something from. The half-steps because "it's only a percussive acoustic part" get me nearly every time. I'd much rather record a good take with the right mike technique, and get it sounding good, and then try to "mess it up," rather than the other way around.

GJ

[Gregg Juke]

Unfortunate double-post deleted by the author. We shall not speak of it again.

GJ