Nice Post, Ive been using a combination of bob katz and T racks metering and common sense, the less is more approach I also use, but have a taste for tape+Tube, and EQ+Dynamic tools that add some mojo to the groove or sound. Always leave the master free apart from maybe a soft clip to spare my speakers. At mixdown I take all FX off master and render out a file. That is when I can work on the same file 2-3 time at once trying different techniques out My favorite at the minute is mid/side processing and I am definitely looking at TBs Loudness meter but all the sample packs have gradually come down in maximum volume levels to a level taht is just right. I tried a loopmaster sample from an artist set and it was good and I was glad. I listened to the sampleism demos and they were mastered very hot and so the house trax I had based upon the Loopmasters volume were so much quieter than the sampleism samples but a lot better quality. I did them again to compare with the best of sampleism to see if I have a chance selling some loops but the audio was massively clipped and the EQ gain must have been massive, the sampleism loops sound better first off loud and proud, but they are clipped with no dynamic range available so to use them in a mix means you can add no further compression or clipping because the samples are maxed out altogether and would simply become mush. With the loopmasters samples they are good all round volume level with between 5-10 db of gain and or EQ boost attenuation left to play with. This is where the amatuers and the Pro's are split, The loudest might always seem best first off, until you try mixing them into your projects. I loaded one drumloop up and the headroom was nearly all gone, so how I was going to add any more to the composition without messing around with subtractive EQ to the sample which then is completely changed in tone or texture, This is my one gripe concerning the sale of music and samples_
"My Tracks/samples are louder than yours"

Really, I would just rather have a universally accepted range that everyone could work to like back in the good old days of the 1960's. I compared a hendrix mix from the 60's to a Dance CD, give me the warmth and smoother hendrix mix anyday of the week.

RMS, DR (Dynamic Range, like the tools from Brainworx), use your ears

Loud enough means in most cases "loud as the competitors"
With DR most pop/trance/dance ambient stuff is about 6 to 7. With RMS it's very different, there are tracks with -4dB RMS others with -6dB.

I use often a reference track and would also recommend to have some reference track not only for loudness comparison, also for frequency balance and stuff. A good analyser with RMS, Katz Metering or DR is of course also very helpful.

This seems loud as f*** to me. Could you name an example of a track where this is the case please ?

Infected Mushroom. Cannot remember about the track. I ask also Erez that he earned a reputation for most loudness in the IM forum. But he had not really a idea about this high RMS values.
But they don't use a limiter, they using clipping (against a Prism converter).

But most of the tracks I compared are about -6dB RMS.

That's interesting.



Hell, that's still annoyingly loud for my personal taste. I feel everything above ~ -10db RMS or so is fatigueing to listen to, but than again maybe I am what you would call audiophile in this regard.

Wow, now that following post was hard to read. Maybe consider some paragraphs?




The only tools in my master channel are usually a VU, maybe a console emulation (if the production asks for it) and ARC post fader.

Since I use heavy gain staging (like in the old days: leveling in the signal at -18dB RMS/-9dBFS peak maximum, then do a quick mix with pan and volume, add EQ/Compression, compensate for the gain loss, mix to taste), I usually never tough -3dBFS digital (or -16dB RMS) on the master bus.






I stop it right here and say, that all of these samplesets are too loud. Granted, it's the "sound" the producers asked for. This usually a quick solution if you're after a specific sound and don't want to mess with it yourself. But I prefer less over-processed sounds to begin with.

Else, I stick to the gain staging rule.






Trust me, certain engineer groups (alliances) are already working on it. And this is why I mentioned -16dB RMS - because this will be the ideal hotspot for Loudness Normalisation schemes which will be introduced eventually - to finally stop the loudness war.

In terms of broadcast, it's planned to be over by end of this month. The ARD, ZDF and ORF (working pool of the broadcasting corporations of the Federal Republic of Germany) already implemented the EBU R-128 (or ITU-R BS.1770-2) standard and gradually go down in loudness while broadcast. Though the effect is still hard to hear on longer programs.

Again, I'd advise you to stick to K-14 currently as absolute maximum. As soon as hardware/software players implement a Loudness Normalisation routine, your productions will benefit from it compared to those that constantly ride the loudness train.






The problems with DR Metering (before Brainworx added weighting filters) are these:
1) it's measuring too strong on the bass side
2) the "offline" metering uses 10-20 random points to declare the DR value
3) depending on the source material, a techno production can have DR-6 while another one has DR-8

This is what Bob Katz critisized with his own meter as well (it's also unweighted). The solution: a weighted loudness meter - and here, EBU R-128 bridges the gap perfectly.

I wrote a white paper about it - it's on my technical blog (see my sig).






Actually a combination of a good FFT and a well setup loudness meter (example: K-System) works indeed well. Though I wouldn't focus on the same frequency courves as other so called "modern" productions. Especially if they are squashed to bits.




BTW:
The loudnest productions I've encountered:
Metallica (-5dB RMS)
Rammstein (from "Rosenrot" times, -5dB RMS)
certain broadcast streams (-4dB RMS)
and I think one track from Noisia

Paul McCartney's album from end 2000s wasn't also great. Overemphasized high frequencies and pushed up to -6dB to -5dB RMS as well.

What do you think about the free Steinberg SLM 128 Loudness Meter?

Since I'm not on Cubase 6.5 or Nuendo, I can't use it. It's not working in Wavelab either.

Then again, it's a basic R-128 meter, fixed setup in a Steinberg shell that Steinberg doesn't even support themselves. There are other alternatives out there. Both in terms of freeware and (affordable) commercial.

I do the same thing I use my K-meter and my T-racks meter to tell me perceived loudness and phase correlation. and My klanghelm VUMT for gain staging. some subtle saturation and good gain staging in the mix with nice eq cuts and good compression control while mixing seem to be working the best for me. but people need to really start thinking about gain staging. a lot of analog emulation stuff relys on it now. and its just easier to use. the gain staging part is agreeing with the other post Im just to lazy to find out how to multi quote

but one of the most important things is proper gain staging with proper listening and doing things in the mix I don't usually have anything on my master maybe a mast bus comp doing like nothing most of the time but that is because of gain staging.
@Compyfox your awesome for mentioning gain staging. I forgot to mention that.

How do I know.... ?

--> If your your neighbours come out, it's loud !!!

that's when I turn it up