DIY Mastering Clinic #1!!!!!

[SouthSIDE Glen]

Not to get too carried away with this measurement stuff, as I think it get's rather beside the point after a while, but an interesting form of graphic measurement when it comes to relative and average levels is the "cloud graph" used in some of the Elemental Audio plugs. I have in in my copy of Neodynium and I think it's probably in the full-blown version of InspectorXL (though I don't think the demo version of Inspector has it.).

It's tough to explain, it makes a lot more sense when you see it than it does in words, but basically "cloud graphs" dynamically show in a bar graph form the relative amounts of "energy" at any given signal level. They can be instrummental in indicating just where most of your actual volume levels average out to be, but it's more than just a single average value. For examples some songs may have 50% of their energy around -12dBFS, 30% around -8dBFS and 20% near -3dBFS where the average peak level is. A standard RMS rating will only give you one dynamic number somehwere around -9dBFS, but the cloud graph lets you see the real distribution does not cluster around -9 at all.

G.

 

[caryindy]

Here's my final.

A little less high mids than my previous, and less limiting.

 

[masteringhouse]

Just sent an email.

Bob's response was that while my explanation is essentially correct, you must also consider how the audio is monitored. Also that what constitutes "loud" is not just the crest factor of the material, but also frequency composition:

[From Bob Katz:]
"I would suggest avoiding the phrase "Is K-10 "loud"? Because we must
try to teach to avoid implying that we have control over where the
customer sets their volume controls! I prefer to use term "it has a
high average loudness in front of the volume control."

And then there is the fact that currently the K-system meters are all
flat, and distorted material will sound louder than pure-sounding
material even at the same meter setting, and material that has heavy
bass will sound lower."

 

[Dogman]

Here's my final.

A little less high mids than my previous, and less limiting.

I like what you've done here. Sounds pretty good. Right on par with some of the better ones posted. What all did you do?

 

[caryindy]

I like what you've done here. Sounds pretty good. Right on par with some of the better ones posted. What all did you do?

Thank you. I gave some details in my previous posts.

 

[Reggie]

Whut I did

OK, finally here is the stuff I did on my master posted a page or 2 back:

1. PLParEQ:
High Pass @ 25Hz
+1.5db @ 70 Hz ,Q 1.0
-1.6db Low Shelf @ 113 Hz , Q 0.5
-2.2 db @ 140 Hz , Q 0.3
+1db @ 753 Hz ,Q 1.0
+0.8db @ 1700 Hz , Q 0.5

2. Analog Out into GSSL stereo bus compressor
Ratio 2:1 , Attack 10ms , Release - Auto , maybe 3 or 4 db gain reduction

3. Back into computer with more PLParEQ
-1.3db @ 170 Hz ,Q 1.0
-1.2db @ 819 Hz , Q 1.5
-0.8db @ 5775 Hz , Q 2

4. Pultec Pro
Dip @ 200 Hz one notch
Peak @ 1.5K 3/4 notch
Attenuate 100Hz Shelf down one and a half notches
Boost High Shelf: three notches @ 16 kHz , Bandwidth 4 (1=sharp, 10=broad)

5. Spitfish
Some gain reduction when the S hits around 7kHz. The sensitivity is around 1/3 up and the other knob is a little less than half way.

6. UAD Precision Limiter
just catching the stray stuff, maybe 3-4 db max gain reduction (not constant)
Release auto


Kind of a lot of EQing, but I felt it needed it. I probably won't be doing a 2nd attempt, because I already have a lot of other stuff that I need to do right now; 3 bands all tapping their feet waiting . But I would totally be down for a rock/metal DIY mastering clinic.
Thanks for the props, Dogman.

 

[masteringhouse]

I just banged one out between sessions. See sample1 at the bottom of this page:

http://www.masteringhouse.com/masteringclinic/

My intent was to retain as much power as I could from the original, but clean up some of the issues with the bottom end and the synth and make the snare and guitars stand out more. The vocal is a bit mid oriented, but I kinda like it that way. It reminds me of a Floyd vocal sound.

The main technique that I used for this was M/S EQ and compression. Essentially I ran busses from Pro Tools to create an M and S send. Then added a 2.5 db cut at 128Hz Q=1 and a multi-band comp with a low shelf at 100 Hz down to cut back on some of the kick and synth on the M channel. The S channel was untouched. I also reduced the M slightly to try to widen the sound field a bit.

Let me know what you guys think of the above. There were other EQs/comps used as well. I'll give you the details on those if you guys like this, if not then it doesn't matter and they will change anyway.

 

[mshilarious]

Let me know what you guys think of the above. There were other EQs/comps used as well. I'll give you the details on those if you guys like this, if not then it doesn't matter and they will change anyway.

Just tell me how you did that fadein!! Auggh!! That makes me feel like an idiot, because I think I spent 20 minutes on mine and it still sounds horrid.

I like how the MS approach seems to have maintained clarity on the clean guitar.

Making sure I learned something here, you master is K-12, because none of the peak K values goes into the red

 

[masteringhouse]

Just tell me how you did that fadein!! Auggh!! That makes me feel like an idiot, because I think I spent 20 minutes on mine and it still sounds horrid.

I like how the MS approach seems to have maintained clarity on the clean guitar.

Making sure I learned something here, you master is K-12, because none of the peak K values goes into the red

Sorry but I didn't really spend much time on the fade in 'cause that would be something that I would need to work on with the client. I just though that the intro should climax where it repeats at the end and threw on a power fade from Pro Tools up to that point. The end fade is probably not exactly where it should be either.

To be honest I didn't really watch the meters on this, just went with gut feel on where it felt right while still kinda pushing it a bit. I'll need to get to back to you on where this one is K-wise.

 

[giraffe]

maybe i'm a retard, but i can't find the link to pingu's attempt, which i'd like to hear.

 

[Reggie]

Haha, never seen someone so impressed with a fade-in!

Cool beans.
The one thing I regret is the way I did my intro. I thought I put enough space at the beginning to avoid losing anything with slow players, but I didn't put in enough. For some reason, I didn't really want to do a fade-in in the beginning, I liked to hear the build of all the instruments from a cold start. Just my preference, not knowing the artist's wishes. I did, however, put some thought into the ending fade and tried it a few different ways and settled on a big slow fade a little earlier than a lot of people did; fading out while the last solo is still going.

Nice sound mang

 

[masteringhouse]

sweet, seems to me that you took most of the harsh out of the top end, (by which i mean the 2-9K region) and really improved the sound of the lead (to me). like the vox a little better too.

you basically did everything i wanted to do to the top end, but didn't know how.

Actually I didn't cut any of the top end at all. What you're hearing is the de-esser (a Weiss DS-1). I know that most on the forum don't have something like this, so I'm at an unfair advantage. The main thing is to learn the technique and apply it with what gear you have available.

For de-essing setting was essentially set to 1 octave at 8.37K, threshold -26 db, ratio 20:1, soft knee .2. The Weiss has 3 different settings for release time, fast was set to 20 ms, ave was 50 ms, slow 200 ms. Attack set at 3.15ms.

 

[giraffe]

i'm the ultimate retard, i just listened to the unmastered version...
and liked it much soooooo much better than my mastered version

seriously
back to the drawing board, it seems that any alterations in the hi's bring out the plinkyness of the gits, and make the sibilance in the vox stand out and harsh the lead a bit.

maybe the answer is to get the bottom worked out and leave the top the heck alone?

i feel like such a heel

 

[giraffe]

come on, gather round laugh and point.

 

[giraffe]

soooooooo having actually listened to the proper one.....
i really like what you've done with the snare, the arpegiated gits still kinda have that plinky thing, but less so than other submissions.
can't really judge the low end, i'm on phones.
the vox have a more "finished" sound and are definitely a little more listenable than other versions, but still sit outside of the song a little.....
that's just the way it is i'm afraid.
i still say the lead is a little better, owing probably to the vox and lead problems being in about the same place.
not sure i like the compression, but that's just me. (and you heard my submission)

 

[Reggie]

dang man, you are like, really harshing on yourself.

dumbass-ahedron

i'm the ultimate retard

Why am I still so bad at this stuff?

It's all good in tha hood.

*********************

I know that most on the forum don't have something like this, so I'm at an unfair advantage.

Yeah, I love Spitfish and all, but I sure wouldn't kick a DS-1 out of bed. The Finalizer has a decent "dynamic EQ" de-esser too when I have access....

 

[masteringhouse]

maybe the answer is to get the bottom worked out and leave the top the heck alone?

Everything is relative with EQ I call it "the politics of equalization", I think in BKs book he calls it Ying and Yang. Anyway, a dip or boost at one location is not only going to affect that one area, but the relationship of all the frequencies. A cut in the bottom is going to make things sound brighter and less powerful, but it also helps to clean up mud. To help soften the sound in the upper mids, it's sometimes useful to try boosting the lower mids rather than cutting the upper mids. It just depends on how boosting the lower mids is going to interact with the low end, etc ...

Try both and see which gives you the sound that you want.

 

[SouthSIDE Glen]

The main technique that I used for this was M/S EQ and compression. Essentially I ran busses from Pro Tools to create an M and S send. Then added a 2.5 db cut at 128Hz Q=1 and a multi-band comp with a low shelf at 100 Hz down to cut back on some of the kick and synth on the M channel. The S channel was untouched. I also reduced the M slightly to try to widen the sound field a bit.

Tom, could you expand (pun intended) on the mechanics of this this technique a little more?

How did you create the M/S sends from the stereo mix to begin with? Did you manually sum and differ the two channels to determine what was common and call that the "mid" (and v.v. for the "side"?) or did you use a plug or device?

What is the theory behind knowing how shaping an MS signal is going to affect the final stereo mix sound; that is, why would shaping the mids and not the sides and vice versa? Is it just a matter of listening to each channel and shaping it to taste, trusting that it will still sound "right" when matrixed back together?

What tool did you use to matrix the MS back into stereo? Did you use PT or an external MS decoder?

G.

 

[masteringhouse]

What tool did you use to matrix the MS back into stereo? Did you use PT or an external MS decoder?

G.

G. everything was done in PT in with regards to M/S. It would take me a while to explain the technique and I would like to create a reference on my site rather than a post buried here for future use.

I PROMISE I'll write all of this stuff up and have a link for everyone. Just looking for feedback at the moment.

 

[giraffe]

G. everything was done in PT in with regards to M/S. It would take me a while to explain the technique and I would like to create a reference on my site rather than a post buried here for future use.

that would be awsome man.