Limit first or normalize later ?

I'm leaning toward C.

Well, I always try to avoid recoding any noise or hiss in the first place. But, if I'm mastering a noisy or hissy track for someone, I follow "C".

As for a "standard" I would say whatever works!

Something sounds amiss to me personally. There shouldn't be any hiss introduced under normal circumstances but who can tell based on the info provided so far. It might help to provide some more in-depth information about your mixing and mastering philosophy, signal chain, plug-ins, etc. I've never had hiss introduced as the result of mixing or mastering unless there was something horribly out of whack. Usually it was always easy to diagnose and correct.

If you can provide more information, that might help.

The best way to remove hiss is to use some sort of noise reducer. EQ can reduce it but it also reduces the sparkle. Use Sound Soap or an old copy of Cool Edit.

To alltheseworlds:

Ditto here...Not enough info provided...If there is "hiss" or other "noise" on the individual tracks then you have problems...

I have done almost all my audio work for the web, but now I'm finishing a CD. I'm finding it a totally different beast with much higher - and different - mastering standards...

Not only is it a totally different beast, you must understand that there is a difference between mixing & mastering...They are two entirely different (although inherently related) stages of the recording process...Just because you have a DAW program that is capable of "mastering" doesn't mean that you can do it correctly...Mastering is an art (as is mixing) and you're going to have to spend months (not a few hours) learning it...In addition, there are room dimensions, acoustic treatments, and monitoring chain (just to mention a few) factors that affect how your final mix/master translates to a quality product...

b) Apply a limiter to Sonar's master output so the wav for mastering is louder to begin with

This will make the noise floor louder as well.
Better to relook at the mix and deal with the hiss there.

Do all the tracks hiss?
Maybe some soft gating on the one track that does hiss would do it.
Or as mentioned above, is the hiss in the physical signal chain?

Or Sound Soap as mentioned above.

Mastering is not the place to deal with problems that could and should be handled in the tracking and mixing stages.

Good luck. Persist, and you will accomplish it.

Apply a limiter to Sonar's master output so the wav for mastering is louder to begin with

I wouldn't. But then again I'm not a huge fan of using limiters in any stage of the mix process, either on the master buss or worse, on individual tracks.

There are many pro mastering engineers who stipulate that when receiving a final mix for mastering, that the RMS db level shouldn't exceed -22 to - 18db, and, that the peak values should not exceed -6db.

Essentially, when asking for this, they are asking for "room to work" within the dynamic range.

You could approach this the same way by having a final mix that is roughly at the above specs, and then master it to the way you would like, after the fact, much like if you had sent it out to a pro mastering house, using whatever limiting, EQ, or imaging manipulation you desire...but you still have your original final mix, un-effected, to turn to and experiment with.

Do understand though, that mastering is an art in itself, much like any area of recording.
I was a pro engineer for many years and rarely did I handle the mastering stage.
There are many pro recording and mixing engineers who can and do, and I applaud them for that, but I wasn't one of them. And, I don't think I was all that alone either..otherwise you wouldn't have seen thousands of albums over the years that said:
"Engineered by Tom Lord Alge, Mastered by Bob Ludwig".

FWIW

D.

the question i have is this: what is the sonic difference between the process i've described and turning down the busses so they don't peak and then mastering later?

Not much, IMO. And I certainly for one can't say to you that you are doing something definitively "wrong" per se'.

If, by adding the limiting on the master buss to the same degree that you would in that extra/added step of mastering, then you are saving yourself a bit of time, except...

If you then decided, after the fact, to still master the song using a 2 track mastering prog, you would have to be very careful of what you did at that stage, and not just in regard to "volume". If you were that close to digital zero on the final mix/pre master, even adding gain to particular frequencies could push you over, zap your dynamics into mush, or at the very least result in possible brittle or muddy sonics. But if you were really attentive to these possibilities, you'd be fine.

And you are right that mastering is more than making it "loud", LOL, although sometimes in this day and age that's largely debate-able I'm afraid... Especially when working with a project or album... there's continuity between tracks as to not only levels but to tone, dynamics and ambient feel.

I guess, if this is working for you, that I would certainly be the last one to say what you are doing is wrong. If it's working for you, the go for it. But I always liked having a version of the mix unencumbered by any master dynamics processing.
You can always add, subtract, whatever in the mastering stage and stuill have an unmastered final mix to fall back on should you want to, but by adding this processing directly into the final mix (pre master), well, it's kind of like trying to put the toothpaste back in the tube, IMHO.

FWIW

D.

ORIGINAL: mixman22000
i have been looking at so many comercial cd mixes lately and my god the wav fills the whole track, in fact ive extracted a lot of songs from cd into soundforge and the wave was so huge all you see is solid color and the song clips imediately and stays that way all the way through the song.

this is typical for "professional" cds. squashed to hell. and when it hits the radio, believe it or not, they'll squash it some more. this ensures that the guy driving his stock chevy truck from 1978 with one speaker blown and the other praying for deliverance has the same experience as the guy driving a shiny new acura tl. but then most people aren't paying enough attention to notice.

there are a lot of things you can do to get your mix hotter, but compression/limiting should be the last step (in my humble opinion).

the first thing to do is hunt down bass frequencies and kill them. bass frequencies rob you of headroom and you probably can't hear them anyway. everything below 40hz goes, no matter what. on everything but bass and kick, you can start at 150hz and go up until things start to sound thin.

next you should hunt down the fundamental and resonant frequencies of each instrument. go find the "mixing tips by chaz" thread in this forum for a list of starter frequencies. or you can download voxengo's free plugin called span, which can show you what's popping out on each track. dip these frequencies until things start sounding bad.

do not solo tracks when you are doing this. listen to the whole mix. if it sounds good solo'ed you have a lot of work to do.

use complimentary eq to fix problems with instruments competing with each other (turning up one track drowns out another). this means you find out which track is "masking" the one with problems and at what frequency. a narrow band eq or the previously mentioned voxengo span vst is good for this. dip the offending frequency on the offending track instead of boosting it on the masked track.

use high pass filters on tracks that don't need it. this would be rhythm guitars, kick drums, basses, toms. start high and work your way down until it starts affecting the mix.

don't do this all in one session. you'll get mix blindness. mix at low volume to make sure everything you need is getting the attention it is supposed to get.

as you do this you'll notice you can turn up your tracks more and more. your mix gets "louder" but the peaks are the same. why? because you've gotten rid of the frequencies which you can't hear or are supplied by other instruments.

the next step is gain envelopes (right frank? ). got an instrument or vocal that peaks out one place? slap a gain envelope (this is a clip envelope, not to be confused with the volume track envelope) on that clip and dip it. the idea here is to get all your instruments (except drums) even in volume. the section where the vocalist is breathy should be approximately the same volume as where they are screaming their lungs out. your brain thinks the breathy section is quieter but it really isn't. just different frequenices are brought out.

if there's something in particular you want to bring out, like a solo, turn down the other non-bass/drums tracks a little bit and turn up the solo track a little bit. for example, once the vocalist starts singing you can dip the instrument bed (but not the bass or drums!) 1 or 2db. no one will notice. you've already played that killer guitar riff 8 times, and it's people's heads. they don't really need to hear it again until the next vocal break. concentrate on keeping your instrument level (except drums) fairly level.

now put a single band compressor on your master bus. turn up your other busses until things get too squashed or the kick drum/bass is "ducking" the rest of the mix too much. this is basically the same thing you'd do in sound forge. listen for rude instruments (bass, kick and chugging electrics are likely to be culprits here). if you get your levels high enough and still like the sound, then good for you. if not, start at the beginning and take away more frequencies.

it's not a mortal sin to boost frequencies, just be sure you've tried dipping first.

use compression on a track as a last resort for evenness. don't compress too much (~5/6db gain reduction) because you'll be doing that enough later.

if you still want more gain, or not all of your song is heavy guitars and yelling, you can use a technique called multing (props to yep for posting this before) which i'll briefly summarize here:

each track has something in it that you want people to hear (if not, get rid of it). the trick here is to find that frequency and isolate it with a fairly narrow band. put each of these frequencies on it's own track. you'll end up with several tracks of one instrument (for instance, the thump, body, click and "air" of the kick drum). i would also keep the original track, eq'd so it sounds good solo'ed (in case you actually have to solo it). throw some gain envelopes on those clips and when the mix gets heavy turn down/off the original track and experiment to taste with the others. now you can just have the thump and click, if that's all you want, and save yourself a bit of space in the low-mids and highs.

now here's where things get tricky. if your mix sounds good now, compressing may screw it up. this is why i put a multiband on my master buss, because i'm not intending to send my mixes off to be mastered for cd, i'm going to "master" them myself for mp3. if i ever get to the point where i'm going to burn a cd, my mixes are going to a real mastering engineer.

if you are going to master the mix yourself, i would recommend doing it in sonar to save yourself endless loops between the mixing stage and the mastering stage. this way when you nudge the volume on one track you know exactly what will happen to the end product.

even on my budget monitors, i can hear pro cds (audioslave as an example) which to be honest, don't sound that good. they are loud to be sure, but i can hear what sounds like high-end distortion and thin-sounding guitars and vocals. the point of this is to say that your mix before compression may sound quite thin and nasty, and even afterward, may not sound that great. but put it on an ipod, and no one will know. so that's why i recommend doing all the mixing and mastering of individual tracks at once. doing the whole album is a slightly different process which has also been covered in this forum. search for posts on "mastering" by yep to get a detailed version.

that's all the time i have for today. gotta go.

jack-t-e-c, that is a particularly well articulated description of some good general mixing/mastering techniques and principles. It should be expanded and published as a brief guide for amateur (and quite a few pro) engineers. I'd love to hear what other tidbits you may want to share - do you have a website or blog?

On a related matter (this may actually be the start of another thread), should any standard EQ (e.g., Baxandall) be applied at the mastering stage for CDs? for MP3s? Or is this dependent on the nature of the material and intended audience?

-bmcmanus

Thanks for the words-of-wisdom and recommendations, papa2004. I've read and re-read S-O-S, Digital Domain, and every other article and book I could find on mastering.

I understand the limitations of self-mastering, which are exacerbated by my pitiful audio environment (sidebar: perhaps we should start a thread with pictures/descriptions of the worst home/project studios - I've got a head start!). I still find it fascinating to see the results of even minor tweaks at the final stages. I listen to my material on as many different types of systems as possible, and use a varied support group of musicians and casual listeners who readily give me their biased opinions - mastering by chaos theory. This is strictly amateur stuff, so I don't worry too much about "radio-ready". I just want it to sound reasonably balanced (in all dimensions) with no uncontrolled transients to blow anybody's speaker.

IMO, unless we're at the level of Chessky and similar audiophilia, certain mastering standards will generally apply, e.g., at least a minimal touch of limiting to mitigate and control aforementioned transients. I am still curious as to the use of Baxandall curves and related standard EQs. Where are they used? Should they be applied to all material to "standardize" the final product? The documents I've read suggest that they are intended for low-level listening, but should they be applied if the material may be heard at higher levels? Or do I misunderstand the usage?

I welcome any and all enlightenment and opinion.

ORIGINAL: bmcmanus

jack-t-e-c, that is a particularly well articulated description of some good general mixing/mastering techniques and principles. It should be expanded and published as a brief guide for amateur (and quite a few pro) engineers. I'd love to hear what other tidbits you may want to share - do you have a website or blog?

thanks both to you and n.y.s.r., but it would be quite pretentious of me not to mention that yep and chaz have basically provided me with all the ideas i currently have on mastering (which have been mostly summarized above). also as papa2004 mentioned, it's not a good idea to master (or even mix) your own stuff, and being cheap i've ignored that advice to the peril of my own projects. definite proof that you can grasp the fundamentals and still be woefully unable to apply them. i find that i'm able to spot everything wrong with my "final mixdowns" but unable to get them to a happy place.

sadly, no website or blog, i'm too busy with my full-time job and wife to really keep one up. eventually i might have one and fill it up with rants against overcompressed radio pop... =)

On a related matter (this may actually be the start of another thread), should any standard EQ (e.g., Baxandall) be applied at the mastering stage for CDs? for MP3s? Or is this dependent on the nature of the material and intended audience?

here's what i've learned from others and observed listening to albums:

one of the goals for your cd should be getting all the songs to sound natural togather when listened to from start to finish. it's also nice if they sound good on random (which is probably how many people will listen to the cd after they've run it straight through a few times).

how this relates to e.q. is that you want all the songs on the same cd to have the same "feel", the same warmth or air or presence or grit or whatever. this makes all the songs cohesive, so one isn't super bass-heavy and the next thin, and the next present and the next hollow and distant. unless of course you are doing it on purpose. (when you are experimenting, do whatever comes to mind. when you are finalizing, you should do everything on purpose - you should know why you're adding an effect or dipping e.q.)

you'll have to do this in the context of the entire album, which is why some people master by throwing all their tracks back to back into [mastering program] so the "global eq" affects everything and so you can quickly find out whether what you did was a good idea.

basically, there aren't really any rules to post-mix e.q., but in general if you are fixing a song rather than shaping the album it's time to go back to the mixing stage.

as for mp3s, if it's a "single", then i'd say that similar rules apply (shaping the sound and not fixing it) but you don't have the rest of the album to worry about.

also mp3's take out frequencies we supposedly can't hear (or rather, can't pick out at 85mph/on an ipod) so i'm sure that affects mp3 mastering as well, though i don't know any specifics.