gain reduction when mastering

[mixaholic]

what is a good gain reduction level when compressing during mastering. i pretty much know about 6 db gain reduction or a littl less is good for vocals. thanks

[bblackwood]

Whatever sounds best...

That being said, I rarely do more than 1dB GR.

[VTgreen81]

Originally Posted by mixaholic

what is a good gain reduction level when compressing during mastering. i pretty much know about 6 db gain reduction or a littl less is good for vocals. thanks

I don't think you can use a rule of thumb like that. -6dB may be good for one voice, another may be able to handle -15-20dB. as far as mastering I'm clueless, but I'm sure it'll depend on the track and the producer's ear.

[masteringhouse]

Originally Posted by bblackwood

Whatever sounds best...

That being said, I rarely do more than 1dB GR.

Second that. There are no rules for compression/limitng and it varies greatly among engineers.

Compare your results with a reference that you respect and let your ears be the guide.

[SouthSIDE Glen]

Start with your attack and release both set to "auto". If either of them does not have an "auto" function, set them both to medium speed. Start with threshold set to 0dBFS (or the analog equivalent) and a compression ratio of 1:1.

That is often a very good setting for mastering the two mix. You can always tweak from there.

G.

[Massive Master]

I'm in the "whatever works for the mix" category - Rarely more than a dB or 2 on compression. If limiting is involved and the clients is looking for "stun" volume, perhaps a bit there also.

[Robert D]

Originally Posted by SouthSIDE Glen

and a compression ratio of 1:1.
G.

That's a little extreme, isn't it Glen?

[SouthSIDE Glen]

Originally Posted by Robert D

That's a little extreme, isn't it Glen?

Actually it's starting right down the midde; no compression, no expansion. Tweak from there as needed.

It's just a smart-assed way of saying the same thing Brad, Tom and John have said; do what the mix needs.

At the very worst, it's no less accurate of a place to tell someone to start than any other number that can be put out here.

G.

[Robert D]

Yeah, I got that. It's always pissed me off that most efx plugins and program resident efx launch with other than a neutral initial or default state.

[SouthSIDE Glen]

It does make me wonder...

It's probably more of a rarity these days, but I'd be curious to hear Tom, John and Brad respond to this: When was the last time, if ever, that you applied expansion instead of compression to a two mix to breathe life back into it instead of squashing it out?

G.

[bblackwood]

Originally Posted by SouthSIDE Glen

When was the last time, if ever, that you applied expansion instead of compression to a two mix to breathe life back into it instead of squashing it out?

Never - expansion doesn't do a very good job of that, imo...

[Massive Master]

Same here. I've tried a little "transient modeling" (SPL, etc.) with mixed results, but mostly just mixed.

[flatfinger]

Why are you compressing the 2buss? is it to get the RMS up??? Or do you really want to change the inner structure and make the softer stuff more prominent? Be sure to use a soft knee ( and program dependancy if it's available) if you want to keep it less obvious.

[masteringhouse]

I've tried using the upward expander from the Waves Linear-Phase MultiBand.
Eh, not great.

Once transients get thrown out the window you can pretty much wave goodbye IMHO. I usually try to get a remix/unsquashed version before trying to add prosthetics.

[SouthSIDE Glen]

Brad, John and Tom, thanks for the rapid replies. That's pretty much what I figured, but it was a thought that crossed my mind and got me wondering. Thanks .

G.

[masteringhouse]

Originally Posted by Robert D

That's a little extreme, isn't it Glen?

Actually G's technique is close to the one that I use.

There are generally 2 ways I approach compression. First I decide if it's transients that I want to reduce, or if I feel the music needs more average level/add density.

If I'm going for transients I will use a high ratio (10:1 or more) and start with a threshold at 0dBFS lowering it until I find where it get at a level where I want the comp/limiter to react. Attack/release settings are set to something that reacts very quickly.

For averaging I'll use a technique similar to G's and start at a threshold of around -10 dBFS (or wherever the average level seems to be at) with a ratio of 1:1 gradually increasing it until I get the desired effect. Attack and release settings are slower for this.

Please keep in mind these are only starting points, from there I'll adjust attack/release listen to the results and make adjustments from there including gain makeup.