What LUFS are you guys mixing to?

What happens to commercial releases that are louder? Or do the mastering engineers master for Spotify at - 14LUFS as well as CD?

-16 and -14 seem to be the competing standards, anything hotter will be turned down, AES recommendation is minimum of -20 for streaming, which acknowledges the weak wattage and gain of most consumer systems when combined with the -14 to -16 standard. -23 is broadcast. I don't have any LUFS tools and don't think about it, that's the mastering guys job.

-16 for stuff I am uploading to youtube myself

cheers

Wiz

What's a LUF and why should I care about it?

Loudness Units relative to Full Scale

It's hard to go wrong with a VU meter set for 20 dB. of headroom.

I'm not entirely sure the value of LUFS as questioned, in regards to mixing. I shoot for peak levels at or under -6dBFS, and any question of LUFS is downstream...mastering guy, etc. Where am I off? There's nothing I deliver directly to streaming, some things I occasionally deliver to broadcast, and even broadcast will have another engineer after me who adjusts levels.

I also do pretty dynamic mixes by modern pop/rock standards, many would say under compressed.  So I'll be at a lower relative LUFS anyway.  There's never been a time I've tweeze compressed a set of tracks and a mix that I didn't go to mastering and hear all the problems I'd created for any listener with a hi-res system.  

So pretty up until now, I've been using the TT Dynamic Range Meter to check my "mastered" mixes (which go to SoundCloud). For continuity's sake, I've been trying to get in them in the 10 DR range, give or take 1 db, depending on the style, etc.  Recently however, I've been looking at them using Logic's LUFS meter, and for the most part, with the more folk/rock oriented stuff, I've noticed that the "Integrated" LUFS reading comes out in the -11 to -12 range with LU (dynamic) ranges around 7 or 8. (compare that to 10 on the TTDR meter) But things don't always line up nice and pretty.  For instance, I just remastered a more orchestrally oriented track to LUFS -12.5 that the TTDR meter pegged at 9db DR.  I did notice that the Medium range part of the LUFS meter was hitting -9 during the loudest sections, but I'm not sure that isn't coincidence.

On the one hand, and for my purposes the numbers are somewhat meaningless. Ultimately I let my ear decide the balance of loudness to dynamic range, but on the other hand I do like to check the numbers - and now I'm wondering which numbers I should be looking at.  My first impressions are that the TTDR's meters align better with my perceptions, but maybe that's because I've trained/calibrated my ears to that meter.

Thanks for this info. It's helpful to know how the DR meter works.

I have to say, the DR calculations as you have described them seem somewhat arbitrarily constructed (e.g. the 3 second window), but I've been reading up a bit on it, and yes, as you say audiophiles use it as reference, but traditionally, so have engineers.  If I may play devil's advocate, it still seems to me that DR was designed for and may still be a better reference for "perceived" loudness, while LUFS was designed to ensure that a mix passes through the broadcast chain as unfettered as possible.  Now I will agree that for engineers the value of both in practice are not mutually exclusive, but I have to admit that LUFS doesn't correlate to what my ears hear nearly as much as DR does.  Case in point. I just passed two mixes through Logic's adaptive limiter, pulling down the input gain 0.5 db on one of them.  Now to my ears, that's a lot and the DR meter shows it to the tune of ~0.25, but the integrated LUFS number didn't budge.  Honestly, I'm a bit baffled. I feel like I should go back and run it again to make sure.

There's no question that LUFS is the more valuable of the two from a practicality standpoint, but I plan on using both until my ears figure out what's going on.  Right now my sense is that DR is more easy to discern aurally precisely because it's weighted/focused on the top 20% of "windows," which is where most of the compression has been applied, while LUFS looks at the entire range and weights things more evenly, which makes much more sense when it comes to bringing different program materials in line for broadcast. My reasoning may very well be flawed, but that's where it is right now.

It's not really worth researching. Far more of an "online audiophile" thing, than actually used by any jobbing audio engineers I know.

LUFS ain't perfect, for sure, (and one should of course always rely on ones brain/ear/room/monitor interface), but it's the best we have so far if you want to play the matching numbers game.