I've just remastered my first song based on LUFS metering, using the meters in my recently-acquired iZotope Insight. I'm very pleased with the result.
 
To be perfectly frank, I'd been somewhat befuddled by LUFS. All the documentation is about broadcast requirements, which don't apply to me. There is no official standard for music in general. On CDs there is a large window of what's acceptable. Soundclick and SoundCloud do not compress, and therefore impose no expectations on the submitter.
 
That leaves home masterers to find their own target by trial and error. For MP3 music it's come down to YouTube, iTunes and Spotify to establish some de facto standards. Those three can't agree on what it should be, though, ranging from -13 to -16. So I reckon the best target to adopt is one that online streaming servers are least likely to compress further.
 
So this has become my guideline: -16 LUFS. I still set my brickwall limit to -1 dB as I always have, and still enable True Peak (oversampled) detection. For awhile I kept SPAN up alongside Insight for the comfort of familiarity, but it turns out that -16 LUFS yields subjective results similar to K-14. The volume level I set my portable MP3 player to remains the same, and loudness is still perfectly acceptable on my favorite headphones.
 
So far, I've only done the one song, but I plan to remaster a bunch of old projects this way. This is all leading up to a remaster I'll be doing next year for a friend's album, one I'd originally mixed and mastered two years ago. He recently mentioned that he's sold nearly all of the initial run of 1,000 CDs and would be re-ordering another batch (1,000 copies in 2 years - in the homemade record business that's what you call a "hit"!). I saw that as an opportunity to test myself, to see if I was really getting better at this or if it was just wishful thinking. So I volunteered to remaster it free of charge.
 
Whilst thumbing through projects to choose for the next experiment, I listened to stuff I'd mastered 10-12 years ago - holy crap, it was BAD. Which is good. It means I really am getting better at it.

Good post, bit!

I had the same dilemma with SPAN. I probably don't really need it anymore since I can use these other tools. 
I might be wrong about this, but I think Soundcloud allows upload of almost anything, but when they stream they use a 128 mp3.
 
Good luck with that project.

i upload redbook standard wavs to soundcloud,
figure if they are going to dumb it down,
i'd rather they start with a decent file.
 
 
i also am on the path for mastering for levels that stay in the ballpark of 'most' pro releases,
but do not push the envelope but so far...
i'm typically doing mixes that average RMS at about -20, 
and master up to about -12.
 
it all depends on your compression schemes, and how much dynamics you decide to leave.
a LOT of trial and error to dial in your sweet spots,
but i'm there now.
 

that said,
i am without true LUFS metering,
but i have worked around it by doing trial masters,
analyzing them in WAVELAB,
and learning where i'm at with that analysis.....
 
going back, redoing things,
and finally learning how to LISTEN for what it is i need....
setting up goals for max peak levels on a track by track basis is the best way to get there.
 
 
 

You're right, Starise, most streaming services resample. Unfortunately, there's nothing we can do about that except to upload the highest-quality MP3s they'll accept. That way, there'll be a little less damage when they resample. That's good advice even if you're uploading the same bitrate they'll be streaming at, because I've observed that some resample even when it's unnecessary.
 
However, not all streaming sites compress the audio like YouTube does. That's what I was addressing: compressing your files further. They'll still do some processing on it, but if you stay around -16 LUFS they'll smush it less. In fact, if you went lower than -16 LUFS all they're going to do is turn it up.
 
Bat, the EBU Loudness meter from Toneboosters is pretty good, and a cheap alternative to expensive tools like Insight. I only mentioned Insight because I know a lot of folks here got in on the recent sale, so there are a lot more of us now that have it there as an option - and maybe haven't dug into it yet.

I'd love to be able to clip on a clip and get the measurement. Right now I have to play through the clip to get the measurement.
 

I said ages ago that K-14 will put you very close to -16 LUFS so nothing new there. But I bet your song Dave is consistent level all the way through yes. So the K-14 = -16 LUFS rule only works if that is the case.
 
What happens if you have a K-14 master for two thirds of the track but in the middle third it dropped way down and went quiet. That is where the LUFS reading then would not agree to K-14 so much. What would happen then is your LUFS reading would be lower now eg -18 or -19 LUFS or so. Because LUFS takes in the whole length of the tune.
 
So a low LUFS reading might lead you to master the quiet third up more in order to get it back to something more like -16 LUFS or close to it. I have been known to automate the volume of a track during mastering for just this reason. That quiet third might sit well in your hi fi lounge listening situation but it is going to not be heard say driving down a busy road with the window half down. You have to make decisions there. But for music that is more or less constant all the way through it is much easier in general and most of our tracks are.
 
The free Orban LUFS meter is also excellent. Just be advised it is not a plugin that you can just insert. It is a stand alone thing and you have to be able to route audio to it through your underpinning audio routing software such as TotalMix in the case of RME stuff or PatchMix in the case of EMU interfaces.
 
Geoff not sure what you mean but I don't get bogged down measuring LUFS of individual tracks or clips etc. Waste of time. Use a VU meter for that. LUFS really comes into its own on final mixes not so much individual components.

meters rock.
 

...What happens if you have a K-14 master for two thirds of the track but in the middle third it dropped way down and went quiet.

Exactly. This is what I'm having the most difficulty adapting to: dynamic arrangements. My stuff will typically start out at K-20, spend the bulk of the time at K-14 and ultimately wind up at K-12. I like music that does that. As a toddler my favorite piece of music was the 1812 Overture.
 
The problem with LUFS (and its cousin, DR) is that it's an averaged measurement. If I try to hold the crescendo down in order to get -16 integrated loudness, then the intro will be too quiet. Bring the intro up and the big finish ain't big anymore. 
 
Most people who master for broadcast or that make pop music would suggest a simple analysis: the material is too dynamic and needs to be flattened out. Bring the quiet parts up and the loud parts down, and then bring the whole shebang up or down until you hit your magic number. What I'm working to devise is a practical compromise that's informed by the meter but still allows the ears to overrule it.
 
My newest reference is Steven Wilson's Hand.Connot.Erase. That guy nails the technique of drawing you into a delicate passage and then letting loose the canons. I'll have to load some of that up and see what Insight has to say about it.