Filling the gap- community mixing/mastering fx project(s)?

[acg2010]

This is a topic I have done a lot of research on as part of attempts at mastering my amateur music I work on from time to time. However, the more I read articles on mastering, it became apparent that there were lots of schools of thought on how to approach it and many different suggestions to the "perfect" chain of vst plugins for mastering. Here are some links to mastering articles on the net (picked at random):

https://www.izotope.com/en/community/bl ... tutorials/
http://music.tutsplus.com/articles/14-t ... audio-7753
http://productionadvice.co.uk/how-to-ma ... usic-loud/
http://www.remix64.com/tutorials/faq-tu ... ering.html
http://getthatprosound.com/mastering-like-pro/
http://productionadvice.co.uk/mastering-basics/
http://prosoundformula.com/how-to-master-a-song/

As far as FS vsts, as previously mentioned, need a good EQ, a multi-band compressor, other types of compressors, a transient shaper, a limiter, and stereo widener. Not everyone of these is always need when mastering though. Some of these have already been created at either the old SynthMaker website or at FS.

As for unmastered music tracks, I pulled some from this link and practiced mastering and comparing to the already mastered versions and came pretty close to sounding the same.

http://www.analogdimension.com/audiosamples.html

Would be interested to see what people come up with.

[tulamide]

However, the more I read articles on mastering, it became apparent that there were lots of schools of thought on how to approach it and many different suggestions to the "perfect" chain of vst plugins for mastering.

The truth is, there is no one correct way and certainly not one perfect chain of plugins. And each tutorial that wants to convince you from the opposite should be avoided.

There are techniques that are useful, but they are not related to certain tools. Instead, the most important tools are your ears. The second most important tool is a real world tool, that you should be very well used to. Your headphones of which you need to know exactly how a well balanced song should sound like. This is important, because you need to compare your song to others. Some Hifi-headphones tend to emphasize the bass, which could mislead you to over-strengthen the "airspace" (Spogg told me it's called so )

If people wouldn't make it such a mystery, it would be pretty easy: Mastering tries to give your song the best sound on a certain media. Quite different requirements have to be met to master for a CD, for vinyl, for the radio, for Youtube, for a club, etc.

Try to spread frequencies linear
cut low frequencies below 40 Hz (or 30 Hz when going for a club master, or 50 Hz and even higher when going for a vinyl master, etc.) to increase your dynamic range
give the kick drum room (via sidechaining or by cutting all other instruments from the kick's base frequency, etc.) and make the hi percussions "sizzle" (for example with slight distortion or like noisenerd with an exciter, or by emphasizing with a narrow band, etc.)

In the end you should have the impression that nothing stands out and everything is working together. Then you have a good generic master.

As for unmastered music tracks, I pulled some from this link and practiced mastering and comparing to the already mastered versions and came pretty close to sounding the same.

http://www.analogdimension.com/audiosamples.html

Would be interested to see what people come up with.

I listened to a few examples and they are heavily "maximized". Little DR so that the overall loudness is high. The problem is, that human beings perceive loudness as better sounding (you know the effect if you ever went to a club, or, for people of my age, a disco). So always compare at the same level of loudness. You can achieve this by decreasing the volume of the mastered versions to the level of your own master and vice versa. Else you will always prefer the louder one.

[noisenerd]

The truth is, there is no one correct way and certainly not one perfect chain of plugins. And each tutorial that wants to convince you from the opposite should be avoided.

+1, clapping smiley, etc.

[Spogg]

I've read some of the stuff you guys have suggested (thanks to all) and tried my best to distill out what's important and useful.
It seems there are 2 areas:
(1) Getting individual sounds optimised on their own tracks before mastering.
(2) Mastering to provide the desired overall result.

Am I right in thinking that if you get (1) "right" there will be far less mastering work at the end - just getting relative levels correct?

Am I right in thinking the most generally useful arrangement is as follows for mastering:
A multiband compressor with various options for knee, ratio etc per band, with a variable look-ahead to offset the envelope followers' response times (i.e. a delay of say 0-20mS). This followed by an equaliser (graphic/parametric etc.) and finally an overall limiter/compressor to get the "loudness" right?

I don't see any real problem in assembling such a module or modules from what we already have, but I want to know if I'm on target or completely deluded before I start any experiments...

Cheers

Spogg

[nix]

in a generalised way, maybe I would prefer eq, compress, limit
more than compress, eq, limit

[tulamide]

Am I right in thinking that if you get (1) "right" there will be far less mastering work at the end - just getting relative levels correct?

Well, it's still more than just the levels (to be honest, if the levels don't fit, mastering won't help much, it's more about getting the frequencies right), but yes, the more accurate you work while mixing, the less work you have at the mastering stage.
One example: I use reverbs in the mix already regarding the mastering. So, instead of having a reverb on an instrument and then panning that track, I pan the track and then apply reverb. This way it fits better in the final mix, because the reverb does not pan unnaturally.

Am I right in thinking the most generally useful arrangement is as follows for mastering:
A multiband compressor with various options for knee, ratio etc per band, with a variable look-ahead to offset the envelope followers' response times (i.e. a delay of say 0-20mS). This followed by an equaliser (graphic/parametric etc.) and finally an overall limiter/compressor to get the "loudness" right?

Just like nix said, you want to prefer equalizing before compressing. Here's one reason why: The stereo sum will contain a lot of unneccessary frequencies. Especially the low ones (below 40-45 Hz) carry a lot of energy that leaves little headroom to use subtle compression ("pumping" will most likely appear). So filtering them out before compressing makes sense. But that's just my personal view.

I don't see any real problem in assembling such a module or modules from what we already have, but I want to know if I'm on target or completely deluded before I start any experiments...

I do. The compressor that I found in my toolbox is a nightmare, not working nearly as good as needed for mastering. A limiter should not just cut everything, like the one I found in my toolbox, it should have a variable attack/release. An equalizer could be built with the flexible-band peaking filter that Martin created (I think he also provided an example once for an x-band eq based on those filters). And all of that should not use 100% CPU power

[Spogg]

I'm not going to argue about the signal chain routing because I have no experience BUT I did read somewhere (Sound on Sound?) that the eq should follow the multiband compressor. Reason given was that the compressor will react to and offset (fight) the eq if it comes before the multiband compressor.
Any comments?
Maybe it would be better to have a routing system so you could choose the chain order?

Cheers

Spogg

[tulamide]

My example is about something else than the eq'ing that you mention. For example, say you attenuate 60 Hz for your kick to come alive and then compress, you will get a pumping effect (in EDM that's actually wanted!). But my example wasn't about attenuating but about cutting off frequencies. This should be done pre-compressing.

So, how about a low/hi cut with variable cutoffs before the compressor, followed by an eq, and finally limiter?

[RJHollins]

Many of the compressor and limiters we use for mastering have SIDECHAIN FILTERS built-in, and or can be accessed via External input.

[nix]

here is my q8 eq->

freQ_1_308.fsm

(491.25 KiB) Downloaded 1340 times

and functional limiter(author unknown)->

limiter_rc_2015.fsm

(61.61 KiB) Downloaded 1352 times

Maybe the mastering device could be a suite of plugins?
I use an audio editor(SoundForge) and apply the plugins one at a time.

Something I have thought of that would aid me is a visual of the waveform as the plugins are adjusted.

You can have a look at the UIs of the freQ(it's delicious) and limiter here in 'effects'->
www.phonicsaudio.com

I would eq before compression to basically attenuate any unnatural characteristics imparted by the eq,
if you have heard of gluing a track, a compressor will do this, a bit like blurring a .png