Are these RMS average ok?

[whitedamp]

This is pretty much a RMS average question, please ignore the tones.

Last week a producer told me that when he mix guitar heavy music he aims for a -6 dbfs peaks, -18 to -20 RMS.
I tried to mix a simple 4 track clip (drums, bass and 2 guitar tracks hardpanned) and got the following results:

Are these levels (and "wave image") ok for a mix? I mean, ready to master? If so...
I remember Andy saying that the last Megadeth album had a -9 to -10 RMS average. So, back to my example, I would have to bang a -12db threshold on L2 for the mix hit a -9.5 RMS average.

Is it normal to boost that much a "good" mix when mastering?

ps: I know that mastering isn´t just about dropkicking with a limiter. Of course not, but as I´ve said, this question is about RMS average.

 

[Mutant]

Last week a producer told me that when he mix guitar heavy music he aims for a -6 dbfs peaks, -18 to -20 RMS.

Bad !

I remember Andy saying that the last Megadeth album had a -9 to -10 RMS average.

Good !

So, back to my example, I would have to bang a -12db threshold on L2 for the mix hit a -9.5 RMS average.

First normalize it to 0.
L2 gives nasty artifacts if used with more than around -4 dB setting.

 

[ReliXKeepeR]

First normalize it to 0.

Bad!

 

[~BURNY~]

Last week a producer told me that when he mix guitar heavy music he aims for a -6 dbfs peaks, -18 to -20 RMS.

Bad !

I think it is rather reasonable. Why do you think it's bad Mutant? He didn't talk about a finished/mastered product, did he?

 

[~BURNY~]

First normalize it to 0.

Normalizing will make a giant mess when trying to get all the songs at the same percieved level, don't you think?

 

[Mutant]

I think it is rather reasonable. Why do you think it's bad Mutant? He didn't talk about a finished/mastered product, did he?

Assuming we are talking about 32 or 64 bit mixing there is no need to worry about occasional peaks over 0dB.

IMHO it is better to have the loudness as close to the finished (mastered) state even during the mixing stage.

 

[Mutant]

Normalizing will make a giant mess when trying to get all the songs at the same percieved level, don't you think?

I always normalize to zero before using L2 to get my wave louder.
Normalizing = no artifacts = you can always reverse the process without losing a single bit of data.

You can always use the volume fader to make it quieter.

 

[~BURNY~]

Assuming we are talking about 32 or 64 bit mixing there is no need to worry about occasional peaks over 0dB.

IMHO it is better to have the loudness as close to the finished (mastered) state even during the mixing stage.

I'd suggest you do a little research about what Bob Katz, Bob Ludwig and Ted Jensen have to say about it. Basically they do not agree with that.

 

[~BURNY~]

I always normalize to zero before using L2 to get my wave louder.
Normalizing = no artifacts = you can always reverse the process without losing a single bit of data.

You can always use the volume fader to make it quieter.

Normalizing is exaclty what L2 before the threshold reach the peaks. So I don't know what you are trying to say. Artifacts occur once your threshold is below the peaks. The value from the threshold is insignificant, It's how much it is below the incoming signal that matters.

 

[Mutant]

I'd suggest you do a little research about what Bob Katz, Bob Ludwig and Ted Jensen have to say about it. Basically they do not agree with that.

I may be mistaken but from what I can remember from reading Katz he is more an analog fan than digital ?
In an analog medium - yes you have to keep it far enough from zero to avoid clips/overs and loss of data.

btw I like to be unconventional ! Always mixing and mastering at the same time.

 

[Mutant]

Normalizing is exaclty what L2 before the threshold reach the peaks. So I don't know what you are trying to say. Artifacts occur once your threshold is below the peaks. The value from the threshold is insignificant, It's how much it is below the incoming signal that matters.

Look at the file he posted.
He has the wave peaking at -5.37 so if he wants to:

I would have to bang a -12db threshold on L2 for the mix hit a -9.5 RMS average.

You know...

 

[~BURNY~]

Let's make it clear: I'm not trying to piss you off.
But...

may be mistaken but from what I can remember from reading Katz he is more an analog fan than digital ?

No, I don't think so.

In an analog medium - yes you have to keep it far enough from zero to avoid clips/overs and loss of data.

You got me lost there.
Oh well...

Whatever works for you dude.

 

[Mutant]

I just did a test with L2 on his mix.
Normalizing (+5.4) + small threshold (-6.6) vs no normalizing + big threshold (-12) then inverted one wave and mixed with the other one.

The result was not a digital silence.

Then to confirm what you said about L2:
L2 at -4dB threshold then volume -4db then inverted the wave then mixed with the original.

The result was not a digital silence.

I did the test in Adobe Audition 1.0 if it matters.

 

[Mutant]

No, I don't think so.

OK as i said "I may be mistaken"

You got me lost there.

I never worked with analog tape, mixing desk and recorder, so I can only imagine that I would be worried that if I had the mix too loud I could actually go over 0dB somewhere later in the song and that would go to the recorder and I would have to start over - yes I know "limiters and compressors", but afaik there is no analog limiter with infinity to 1 ratio, so it still could record a clipped peak.

Whatever works for you dude.

Yeah and for you too !

 

[~BURNY~]

Well, I did a similar test and you are right. No absolute null with L2 (disabled dither) vs volume.
I also tried to put a dummy L2, then I got a null, but it doesn't prove my point. It's important to notice that the difference wasn't altered transients or unstable levels of any sort but just a volume difference afaict. (could be internal values calculation vs daw values calculation...)
And cancellation was considerable.
So I for one don't think it's a real issue and I still think that what people will point as audible artifacts is about the way L2 reacts once the the threshold is below the incoming signal (harmonic distortion) .

 

[Mutant]

It's important to notice that the difference wasn't altered transients or unstable levels of any sort but just a volume difference afaict. (could be internal values calculation vs daw values calculation...)

Let me see...
Yes you are right !
Strange.
[edit]
Let me see the first experiment again:
+5.3 + threshold -6.7 vs only threshold -12
The difference is very quiet: and it is dither [edit2]quantize not dither because dither was turned off[/edit]noise + very distorted original.
So no you are not right ! Hehehehe
[/edit]

So I for one don't think it's a real issue and I still think that what people will point as audible artifacts is about the way L2 reacts once the the threshold is below the incoming signal (harmonic distortion) .

OK you got me convinced !

But I'll still use a volume fader before any loudness maximizing plugins.

 

[Mutant]

[edit]
Let me see the first experiment again:
+5.3 + threshold -6.7 vs only threshold -12
The difference is very quiet: and it is dither [edit2]quantize not dither because dither was turned off[/edit]noise + very distorted original.
So no you are not right ! Hehehehe
[/edit]

I'll check it in Reaper (64bit) too.

[edit]As I expected I got a distorted (bass and transients) track without the (quantize ?) whitish noise:

+4.7 + threshold -7.3 vs inverted only threshold -12

You may want to normalize it to better hear the difference.

 

[ReliXKeepeR]

Mutant, you always miss things.

My advice is to write a book now. Mutant's guide to mastering. It'll blow Bab Kotz.

 

[Mutant]

Mutant, you always miss things.

Which things ?
Maybe that's because of the language barrier ?

My advice is to write a book now. Mutant's guide to mastering. It'll blow Bab Kotz.

Hehehe I did not say that mixing and mastering at the same time is the only one proper way - only that I like to do it this way.

 

[whitedamp]

Wow, calm down, no need to heat the thread. Thanks for the replies!

I think that there´s no difference between boost with normalizing (or fader) or L2 as long it doesn´t clip. As I set the out ceiling to -0.2 on L2, I think that if I normalize to 0db before L2 the sound will already go clipping 0.2dB.

Anyway, this is not a big problem as clipping will always occur when trying to hit a -9 RMS average. My main doubt is if that´s how a "good" mix would look like before mastering: -6 peaks and -19 RMS average (or 0 peaks and -13 RMS average, if you normalize it). I mean, a mastering engineer would get this mix and say "oh, that´s exactly what I would expect from a pro mix. I LOVE to work with levels like these".

It´s pretty hard to me to have a visual referrence of how a pro mix looks like before mastering simply because I don´t have access to such material. I don´t know how much dynamic the "pro" songs used to have and how much of that was smashed by digital clipping to hit a -9 RMS average, I just see the finished product on CD (that brickwall with little headroom).

As people always tell to send your mixes for others to master, it´s pretty important to me to know what a mastering engineer is expecting in terms of RMS.