EQing a mixdown

[jeffgowins3]

After a song is all mixed down and I go to throw some EQ and compression over it for the final version, I really don't know a good method to EQ it well. I tried cutting muddy frequencies and boosting the ones I've EQed my instruments to be around. My guitar is heavily EQed around 3 khz, bass around 400khz, synth at 5khz, etc. My friend Tom, who works at a big studio in LA tells me to try to isolate each instrument to a set band of EQ so you can just boost it in the mixdown, but I'm listening to the un-EQed version and the post-EQ version and it looses a lot of punch and just sounds flat.
So, how do you go about EQing an entire song?

[roymond]

Here are a bunch of threads to read, each of which have some external links of great value:

How to Make Great Mixes

Drums and Compression

Recording Acoustic Guitar

Band-limited compression

EQ?

[Adam!]

This basically boils down to the eternal question: where does one draw the line between mixing and mastering? I believe that you should trust your ears for the mix, then trust the pro's for the master.

I'd adjust the individual instrument EQs until the mix sounds perfect to you on two sets of speakers. Then mix it down and AB the mixdown with a professionally done song that you want your song to sound like, EQing your track until it has the same tonal balance as the commercial one. I'd use a parametric EQ with low Q values (~1.0) so that when you adjust a band you'll be adjusting two or more octaves at once. If you ever need to boost a band that's less than half an octave wide there's something wrong with your mix.

If you are using a single band compressor I'd compress AFTER you EQ, which will help prevent the bass from pumping the entire mix; if you're using a multiband compressor I'd compress before you EQ so that the compressor doesn't undo all your meticulous EQing. Keep compression light: an undercompressed song is very forgivable, especially in Songfight land. The best thing to do is compress your song until it sounds like your average radio-mix (so peaking the VU meters all the time, but fairly lifeless at low volumes), then mix the uncompressed (or lightly compressed with a high threshold just to tame the peaks) track back in and season to taste. This can produce a final mix that is solid but still dynamic, loud when it needs to be but not loud all the time.

Bottom line: if you have to choose between a mix that sounds punchy and alive and the processed mix that has been hammered flat and lifeless, always, always choose the former. But before you hit submit, check out the third option, which is a mix of the two.

[Southwest_Statistic]

I dont understand mastering. I really dont. Everything I do seems like it sounds better if I *DON'T* have a "final mastering phase". I've got all this nice mastering software and I almost never use it for anything.

I will be in the mastering phaze doing the final eq-ing and I will hear that the mix sounds better if everything around, say, 500hz is suppressed ever so slightly. Well, instead of just doing that in the mastering phase, I end up loading up my whole source file and smacking an EQ on my bass guitar that supresses 500hz.

I think (for me at least, in my uneducated mixing perspective), that doing this results in a better sounding song because you arent manipulating the ranges of every instrument in the mix. Doing that doesnt seem to make much sence to me...

[blue]

well.. mastering a single song is a bit disingeneous, as it is sposed to be the process of making a body of songs fit and flow on media. it was also, as far as EQing goes, designed to prepare a song for working well on vinyl, which has physical restrictions for sound reproduction.

i'm pretty sure EQ'ing was originally meant to be a repair job - if the mix was good, you just left it be. these days, it seems to mean whatever you put on your output bus.

also, for the pros, mastering is a chance to give another set of good ears a final listen to an album.

[Dan-O from Five-O]

I have a friend that worked for Columbia years ago. His forte was servicing reel to reel equipment, but as is the case many times, his job was simply to take "Mastered" tapes and copy them to cassettes. (Hey I said it was years ago) Anyway in most cases, he only made minor adjustments in EQ. He worked the 3rd shift and in the morning the woman who worked 1st shift would invariably zero everything out and re-EQ the whole deal. He says her ear was amazing and that she could pinpoint flaws in frequency that he hadn't heard, UNTIL she did that. I guess that's why she worked first shift.

My point is that Blue raises a good point about getting a second set of ears on your "Mastered" product. As for it being a process for one song or many, it's debatable, but would throw this thread way off course.

To your original question, I've gotten some great advice on the threads here. It's too late for me to seak out links to point you in the right direction but one of the best things I've picked up is the A/B comparison Puce mentioned. Also, listening to the final product at extreme low and high volumes to find any differences and comparing those to a "Pro" mix is a nice tip I've started using.

In the end I think it comes down to ears.. Some people seem to be able to hear things others can't. Bouncing it off other sets of ears and taking or ignoring their advice as you see fit is the best piece of advice I can offer.

[EDIT] Oh and something that's sure to throw this off track is, what you're listening to your mix through. An issue I'm currently researching as I seek to upgrade my monitors. Sober, I've seen your Tapco's in pictures, your thoughts?

[the idiot king]

the only advice i can offer on mastering is the following:

-eq as little as possible! if massive eq is needed, then there is probably a blatant (read: easily fixable) problem with the mix itself.

-when compressing the mix, set the threshold to a nice, low number like -25 or so. this will prevent that first major drum hit or vocal passage from sounding like it is being forced down.

-keep the compression ratio low, as well. in fact, i typically start at 1:1 and slowly creep up until the mix is a bit more "in your face", but still has some dynamics to it.

-compression attack and release times DO matter! these can actually be set to the tempo of the song (which tends to provide the best results), and there are charts online that compare bpms to time in milliseconds. always remember that you can cut these times in half or double them or anything similar for different results.

-finally, do you really need to master the song? if you have a song that is completely reliant upon dynamics, it may just make more sense to send it through a LIGHT compression (if at all) and then normalize the final product. a great example of an unmastered recording that sounds great is "my father my king" by mogwai (out of print, but i'm sure you can download it somewhere). my old band had begun work on our final cd before we broke up, and our plan until the bitter end was to have it normalized and have levels adjusted, but to not touch any song with any amount of compression. but then, we were dynamic. we weren't a hip hop group. it's all about serving the song.

[boltoph]

Here are a couple of my mastering schemes. I'll preface this by saying I don't know what I'm doing and just play and experiment with all the toys. I like to compare my final mixes to commercially produced stuff of similar sound / genre. The method of mastering depends on the song, for me. I also always tweak the knobs from the original stock settings. Twiddling knobs...

First, an AU setup. The chain starts at the top left, reverb, eq, dynamics, multiband, and limiter. I like this setup because it seems like if I do it carefully, everything sounds more natural than the TRackS's analogy squishyness. And if I want that squishy tight sort of bass drum, I use the "Hard" setting on the dynamics plugin there. Sometimes I even just use the limiter if I have a good mix or if it's not a full rock band sort of recording, might just have a bit of eq and limiting. Maybe a little multi band, too.

Second, a TRackS setup. That soft clipper is a dangerous thing, I just got this TRackS and am still playing with it. It tends to get more "slappy" I think, so I feel the challenge is to keep it punchy and loud but avoid the super slap. I have yet to understand the full workings of this plugin, but I like it so far.

I used to use Bias' Freq and Squeez, but that Squeez seems to alter the sound in a choppy sort of way I'm not crazy about.

I also try to bounce as little as possible and only use a little of each plugin, I usually run 5 channels on the AU setup, but the TRackS has it all in one plugin, like a big analog mastering setup. With the TRackS, I'd run the hair of reverb first, separately, or put it on the output bus of my mix.

I try not to vary eq more than a db or two but sometimes go up or down around 3. I add that touch of reverb usually, not really enough to even notice but when you're recording an home studio with alot of close mics and stuff, I think the final mastering reverb can help bring it together...or screw it all up and muddy it....

and from here on out I normalize at 99% as the last step, and check it out in Inspector. It's different when you're putting together the whole album because I need to measure the average RMS's of all the songs and bring them all together so in that case each song will have different loudness settings, etc.

And Roymond rocks!

[Mostess]

I dont understand mastering. I really dont. Everything I do seems like it sounds better if I *DON'T* have a "final mastering phase". I've got all this nice mastering software and I almost never use it for anything.

I always thought that mastering is for albums or other song collections. The goal isn't to improve the sound of an individual song (which can be done by remixing), but to make several well-mixed songs sound more alike, or more complimentary. I could be wrong, but my impression was that a studio sent mixed tracks to a mastering facility where special masterers used special mastering speakers and equipment to string the songs together into an album. That distinction is nonsensical for SongFight!, and probably for most home recording. But I imagine it's built into the professional system for a reason, most likely financial.

But that's my uninformed, novice opinion. Anyone have real knowledge?

[Me$$iah]

A number of years ago I worked at very pro studio...worked with a lot of big acts, recorded a number of tracks etc...
We used to record everything on a given to a 2" reel, we had at the time the early digital editors and hard disk recorders etc but they were used not as the main recording system, but as a way to move vocals, or set up massive choruses etc....not the stuff I was into....I liked the hard wiring chains of effects back through side linked synths etc....studio geek stuff...... but I ramble... we used to put then all the tracks of any given song (chorus harms often reduced to a stereo mix) on to a 2" reel
Then we would take the reel and mix it and record the final mix to a DAT
if we were recording and album, then all the songs were seperatly mixed and generally put to a DAT each, along with all the various other mixes, edits etc.... then all the tracks to be put on the album were transfered to a single DAT. This DAT would then go off to the mastering company, to be tweaked, sorted so that all the tracks ran at the same level...ie track 1 is quiet track 2 is louad as shit track three medium volume...they used to sort that out so that all the tracks were at the same level.....
it would usually be then sent directly from the mastering suite straight to the pressing company, who would print the cds or records
If we were recording a single then all the individual mixes and edits of the song that was required (b-side inst mix.... or different song)were compiled on to a DAT and that was sent to the mastering company, and again from there to the pressing company, where the single was made.
then theres distribution etcetcetc

The mesure of a good mastering engineer is those that get the most from the mix without taking away any of the original sound or dynamics

Me$$iah

[boltoph]

This DAT would then go off to the mastering company, to be tweaked, sorted so that all the tracks ran at the same level...ie track 1 is quiet track 2 is louad as shit track three medium volume...they used to sort that out so that all the tracks were at the same level.....

This is measuring the average RMS levels, and equalizing them. Another part of it is to shoot for a certain average RMS for all the tracks, to get the tracks loud enough, and up to modern standards.

Here at songfight, the "dynamics arguments" will ensue. Everyone has their own opinion. The point isn't just to mercilessly squash everything to be "loud". The point is get a good sounding frequency balance and punch from the track, as well as loudness.

I've seen way too many tracks where one or two seconds of some loud 1 khz vocal will eat up all the levels because that part hits 0, while the rest of the mix is under -10 db. In this case, one would consider using compression or limiting to reduce that vocal peak or to limit that frequency at a certain level so it's attacking the rogue vocal and not the rest of the mix. This could be done in a mixdown or as a mastering process.

Same goes for tracks where there is such loud bass and bass drum that those frequencies should be compressed to allow the other frequencies to be all they can be...This is what I'm talking about. But take it or leave it. Everyone's got their own opinion and mine is that I love dumping the mixdown into a 2 channel audio editor to finalize stuff.

- clip beginnings and endings to make a clean track
- get a good average RMS using compression and limiting
- possibly adjust eq, depending on if this is needed

Take it or leave it. When an amateur, or older, track comes on in my playlist and I can't even hear the damn thing because it's unmastered and all it's peaks never touch -10db except for that one high B on the vocals, I might turn up the volume for that one track, or I might just let it play and not really hear it. Or I might skip it. If I like the track enough, I'll master it myself...I actually did this with some Allman Bros, Luke Henley, and live Bob Marley tracks in my playlist. It just pissed me off too much when those tracks would come on and I couldn't even hear 'em in the context of my modern playlist...

[the idiot king]

i just saw a bob ludwig-moderated panel at the audio engineering society convention in nyc this weekend, and the general consensus among the four grammy award-winning mastering engineers aboard was that if any more than a half a db or 1 db of eq is needed for anything, then they would actually send the song(s) back to be remixed.

i guess the bottom line is mix it right, and there shouldn't be a huge need to eq the overall master. but then, this is just opinion, too.

[HeuristicsInc]

I've been getting good results on vocal understandability by cutting the eq of other instruments a small amount near the 'dominant frequency' of the vocals. It makes almost no difference in the sound, but the vocals jump right out. If you do it right.
Starfinger taught me these tricks - to find the right freq at which to do this, put an unnatural narrow eq spike on the vox, slide up and down until it sounds the loudest. That's the spot.
Thanks Craig!
-bill

[blue]

The point isn't just to mercilessly squash everything to be "loud".

Yes it is.

Heretic.