Mastering

It depends a bit what order to put things in EQ or compression. I actually do it the other way around and EQ first. Especially if the track I am mastering has an excess of energy in the spectrum eg suppose there is a lot of low end present or too much 300 Hz present. With the compressor first it is going to act on those frequencies and that will effect how the compressor reacts to the rest of the spectrum. By EQ ing first you are actually correcting what you are feeding into the compressor and hence the compressor will react more evenly to the overall sound. So there is an argument against what has been said so far.
 
When the mix is very well balanced energy wise though then it is slightly less critical and either way can work. But even in a very good mix there is still some very extreme parts of the spectrum that needs a little work eg very low and high end and I prefer to get those right first then the compressor is not having to deal with an excess in those areas. It is usually happier. Be good to ask Danny what order he puts things in. I think he runs the mix through the LP64 first to clean the extreme parts of the spectrum before anything. I tend to agree with him.
 
The idea is to present the compressor with the most well balanced EQ energy across the spectrum you can. You are only boosting areas that are lacking. Once right you don't boost above that. You also cut areas that are excessive.
 
Lately I have been mastering in Harrison Mixbus and using thier mastering multiband comp and it is very very good. It has a very interesting display as to how much gain reduction is taking place in each band. I find when I EQ first it is easy to get all the bands to gain reduce the same amount. Very important in order to maintain the EQ balance across the spectrum when using a multiband comp. But is does feature a low end replace boost type control for putting back some bottom end which can be sucked away if you are not careful with a multiband comp and this must obviously be after the compressor.
 
Watch out listening to before and after grabs of mastered tracks. Listen to the after versions at exactly the same volume as the before version and you will really hear how well the mastering is done. (ie remove the obvious louder is better syndrome) When you do this you reveal things like the snare sound losing all it's attack and snap. Poor mastering IMO. Good mastering keeps that all intact and delivers the volume at the same time.
 

I second Jeff's advice that separating the mixing and mastering process is a good approach. In days of yore, the person doing the mastering was almost always different from the person doing the mixing. This created a kind of "insurance" that there would be another set of objective ears before hitting duplication. Although taking a break between mixing and mastering isn't the same, it's more of a step in that direction.
 
Many clients want at least a semi-squashed sound, but that can change the mix and after hearing the master, some people want to go back and make changes. I've found that I can cut down on that by saying to people go ahead, put a compressor or limiter or EQ or whatever in the bus when you mix, but then 1) send me a mix with ALL the bus processors disabled, and 2) send me a copy of the mix with the processors so I can hear what you like.
 
Finally, I master a really wide range of material, running the gamut from country to techno to Christian to narration to restoring a cassette recording of a play from 1986. People often ask what my setup is for mastering but it seems like it's never the same twice. The various masters have different problems and unique characteristics. To relate this to your original question, I usually do EQ after compression (if compression is needed) but before limiting. However, for me most of the time the final limiter is more about catching transients than affecting the sound. I try to do all the "loudifying" before it hits that stage by techniques I've talked about before, e.g., redrawing selected waveforms in Sound Forge or WaveLab.
 
[Edit: and of course, you can't do that kind of surgical editing just by putting processors in the bus.]
 

And now I'm going to disagree with myself...sort of...
 
For some people, mastering is not about detailed surgery but simply enhancing a tune as much as possible using available tools. In that case, mixing through processors can pretty much accomplish that more limited goal. However, even in this case, it's a good idea to re-listen several days later with fresh ears. You can then re-open the project, do a few tweaks, and try again.
 
Another thought i'd to throw in is that mastring in today's era of "singles" is very different from mastering back in the days of albums. With albums, there was a whole element to mastering of sonic consistency as well as song order and flow. With vinyl, it was even more demanding because you couldn't have loud songs toward the end of a side due to inner groove distortion, and also, the mastering engineer had make decisions about tradeoffs. For example, if an artist handed the engineer a 50-minute album, did you lower the overall level in order to keep the bass intact, or cut some bass in order to be able to cut at a higher level? Or compress it more so levels were less of a concern? At least we don't have to worry about those issues any more when mastering for a digital delivery medium.

Something obvious—but maybe not to new engineers—is that master bus compression will change the mix a lot more in those songs with a wider dynamic range and instruments that come and go.  By this I mean, if you start with an acoustic guitar, then add voice, later add bass and drums, and finally build with sweetening, you may find that compression really alters the way your mix sounds.  After you add that lift and limit you may go, "Whoa!  What happened?"
 
If you have recorded a rock band that goes balls-to-the-wall from start to finish, the master bus compression is less likely to make your mix sound radically different.

I.....nevermind, it would take me 4 pages to cover all the stuff quite a few of you are missing. No one wants to see me do that anyway. LOL! I will say this...as a silent mastering guy who has never bragged about any clients or tried to sell mastering services....it blows me away how wrong people are about this subject. What's even more disturbing are the guys that think they know what they are talking about posting about it with confidence. There are so many things that go on in a real mastering situation before you even get to the plugin/hardware stage, it would spin your head around if I started rattling things off to you. 
 
Quickly....and this is my opinion only. It is not meant to discredit, belittle or bring anyone down.
 
1. Anything that *usually* sounds drastically different than the original mix once mastered....was not mixed right. When someone sends a mix off to be mastered, they pressed the "Export" button for a reason. That reason is...they liked what they heard. What YOU as the mastering guy think sounds better isn't always the answer. The client usually has an agenda. So I'm never really blown away when I hear a before and after. To me, it makes the engineer that tracked/mixed it look like a fool. It tells you "don't go to that studio!" However, it DOES make the mastering engineer look good. This is why I do not provide before and after on my site. It makes me look good, but it makes the studio look terrible....and well, to me that's not fair. I've done those types of surgeries but only when the client can't give me a remix and all we have is what we have.
 
2. Pre-mastering: No one EVER makes a mention of this because they simply don't have a clue....unless you're in the industry. No one throws a song in an editor and starts running plugins until it sounds right. So let's stop right there. The material needs to be analyzed. Ends are trimmed to get rid of unwanted space because most engineers could give a rats @ss and leave you with 15 seconds of intro and outro dead space. 5 seconds is enough, I'll handle it from there. :) The material gets listened to closely to see if there are any artifacts such as hums, hiss, noise, rumbles, oscillation, rise or fall in audio, pops, clicks, guitar hum due to the mix engineer not doing his job during slip editing, punch in artifacts, drum sticks whacking things other than drums, engineer editing laziness, and anything else that should not be heard during play back. Everything is logged and written down in one listen. This forms your pre-mastering work order.
 
3. Along with what I just told you, when that stuff is all done, you can move on to the next phase of the pre-master. What you did above took about an hour. Now you get a fresh start and remove any DC offsets so you are starting fresh. From here, we analyze and remove any rogue peaks we find in the audio. We do this because the material will only be as loud as its loudest peak no matter how much you use a limiter. Remember, the limiter will threshold itself to that peak. So leaving loads of peaks will always stop your material from being as loud as it can be. You don't want to use a limiter to control the peaks for this because it will destroy your attack on your snare drum everytime. Create a mix with a snappy snare and try to remove the peaks by limiting. Then, try it by hand manually....huge difference for the better every time, but unfortunately, incredibly time consuming to where if you do it once, you'll probably never do it again or you'll hand it off to a dope like me that doesn't mind doing it.
 
After we take care of all those peaks (usually snare drum or kick drum transients because the mix engineer didn't know how to control them) we level the audio by ear using automation. This preserves dynamics and makes things gradual. This also allows your compression and limiting to work as it should without smearing the stereo field or creating artifacts. Keep in mind, all the stuff I've mentioned so far above, is NEVER and I repeat NEVER seen if the project comes in to you from a major label or an engineer that has a clue.
 
Have you noticed we haven't even started the mastering procedure yet? Does everyone do things this way? No, but by rights they should and I know big name guys that do. Why do I know this? By talking to them as well as seeing the numbers they achieve when you look up the stats on the audio they master. You don't get those numbers that consistent by feeding stuff into Ozone. You work the material the right way....or....you're so rich you have your own programs that do it based on your old work habits that are not for sale. You'd be surprised at the stuff they have that they DON'T share with you. A trip to Katz studio will show you that, although he has started to sell his personal contraptions.
 
FX chains: For what it's worth, I'm eq first as well and prefer all ITB as it has proven itself worthy enough for me to get rid of my outboard gear. You eq to get the song right...then compress to control the little peaks and valleys you may have just created. As far as the rest of my chain, it varies from job to job. The one thing I rarely do is add reverb to a mix during mastering unless the material is crying for it. I've done it with success, but it's not something I agree with or feel the need to do unless absolutely necessary. But rest assured, at any given time, there are anywhere from 9-11 plugs in my chain during a mastering procedure. Sometimes I use them all, other times I do not. The reason for the amount is several eq options, so they add up quickly.
 
After all this is done....there's a post mastering procedure....oh joy! This finalizes everything just right, converts sample rates, dithers, removes any final residing DC offsets and earns my DZL stamp of approval. :) There's more, but you get the idea.
 
Stem mixes: Quite a few mastering guys master using stem mixes. This gives the ME way more control and allows he/she to paint a much better picture. You'd be surprised at how often an ME is NOT mastering using a stereo audio track. It depends on the situation as well as who the producer is in a big time situation. Most of the hobby guys will just provide a stereo track...which is fine. But when you find out you have problem areas, then you either have to fix the mix or send stems. This is why all my clients are screened beforehand. Not screened to the point of declining (well sort of but not due to "you suck") to the point of being honest. 
 
It stops me from polishing turds and stops me from taking money from them when mastering will not make a major difference in their audio. The better you mix the song on your end, the better the master will turn out. I even help people for no additional cost as long as they let me master it. You learn something during the process and you get back the best material possible that is not doctored up, fake or synthetic....it's you being you. I wouldn't be able to sleep at night if I took on a job that was so badly mixed I'd be taking money "for the sake of" just to make it 2% better if that. Yet, I see this happen every day because other mastering guys need to eat. Uggh. What a world we live in. :( I'd get another job before I'd take someone's money "for the sake of being in business".
 
Mastering today vs. mastering of yesterday is something only a pro that has been on both sides could answer in my opinion. I believe it would be different procedure wise if we were still going to vinyl, but I have no clue and am not embarrassed to say that. I do know this...I'd be willing to bet way more editing is done TODAY than when they mastered years ago. I don't know how they dealt with pops, clicks, noises etc. But this takes up quite a bit of my time for some clients because of how they have misused or misunderstood the tools that were given to us/them. Not every project is a project done on a major label, so I'm sure they had to deal with this too. There are so many things today that contribute to noise and people totally not understanding things that they are literally destroying the audio before it gets to the ME. This is something that was unheard of in the past.
 
Whatever the case, the principal *should* be the same for most situations. The major difference is volume and the amount of compression, distortion and limiting entering into the scheme of things. They would have had a mess on their hands back in the day dealing with this volume and excessive compression/limiting/coloration stuff the way things are today.
 
Ok, not quite 4 pages...but long enough for some of you curious about this to get an idea. :) Back into my hole I go. :)
 
-Danny