You are here

Integrated loudness

Page 1 of 1

Integrated loudness

PostPosted: Thu Jan 10, 2019 7:50 pm
by jaysonzambito
Hello all, here’s a question regarding integrated LUFS. I’m currently working on a trailer track that is very dynamic: a rather low level section in the beginning that builds to a loud and aggressive end. When I Ioop the loudest part, I set the levels and limiter so that the LUFS meter hits -10 LUFS short term and -12 LUFS integrated with a peak of -0.3 dB(I use the LEVELS plugin by master the mix fyi) Every thing seems fine so far, however when I export the track, the track statistics read an integrated LUFS of -16 with a short term LUFS of -12. If I want to have a competitive level(ideally -12 dB integrated LUFS) which level do I focus on for my readings: the -16 dB LUFS integrated after exporting or the -12 dB LUFS integrated loudness that my meter gives me when I loop the loudest section of the track.
I hope the question made sense,
Thanks.

Re: Integrated loudness

PostPosted: Thu Jan 10, 2019 9:03 pm
by Hugh Robjohns
The first thing I have to say is that I suspect you have misunderstood the whole point of loudness metering and/or the concept of loudness normalisation. The notion of 'competitive level' is anathema to loudness normalisation! As is a peak level of -0.3dBTP, by the way -- that sounds like an idea borrowed from peak mastering! The inherent error-margin of true peak metering means the maximum allowable peak level is never higher than -1dBTP, and is recommended to be -3dBTP if delivered via a lossy codec.

The EBU R128 requirements for short-form TV programming -- applying to all material shorter than 2 minutes, which I assume describes your trailer -- states that the maximum Short Term loudness level, measured over the whole item, should be -18LUFS. The Integrated loudness target (over the whole item) is the standard -23LUFS, and peak level the usual -1dBTP (to ensure no inter-sample peaks over 0dBFS). So you're a good 6-7dB too hot to be compliant with the R128 spec at present....

Of course, if you're aiming for a non-broadcast platform then maybe different target levels to those documented in R128 would be acceptable, but you'll need to get the relevant numbers directly from the relevant media outlet(s). However, the true peak level will never be more than -1dBTP though, so you'll still need to address that small issue.

From what youve said, I fear you are trying to mix within a peak-normalised paradigm, while using an entirely inappropriate loudness-normalisation metering system! Here is a link to the EBU's Short-form loudness target document:

https://tech.ebu.ch/docs/r/r128s1.pdf

Re: Integrated loudness

PostPosted: Sat Jan 12, 2019 2:37 am
by jaysonzambito
Hi thanks for the reply. Perhaps I am misunderstanding. But if I may: when I compare my tracks to others in the genre(trailer music) their integrated LUFS levels are always around -12 dB, short term LUFS around -9, and RMS around -11 to -12. I am aware of the -23 dB INT LUFS broadcast standard, and the -16 iTunes and -14 for Spotify. It my question wasn’t necessarily what the LUFS level should be, it was more about proper metering before export. Is the integrated loudness basically an average LUFS reading of the entire track?

Re: Integrated loudness

PostPosted: Sat Jan 12, 2019 11:09 am
by Wonks
Integrated LUFS is an average LUFS reading of the whole track, or at least the section you play which the meter sees. But to stop periods of silence or very low levels from affecting the overall reading (e.g. if there are periods of very high compressed sound levels interspersed with periods of near silence), there is a level threshold below which the loudness integration is stopped until the level goes above the threshold again.

I can't tell you what that threshold is off the top of my head, but I know it's been adjusted at least once, maybe twice, in order to get a meaningful level measurement. So you really need to check that your LUFS meter is using the correct threshold value.

Re: Integrated loudness

PostPosted: Sat Jan 12, 2019 11:33 am
by Hugh Robjohns
jaysonzambito wrote:But if I may: when I compare my tracks to others in the genre(trailer music) their integrated LUFS levels are always around -12 dB, short term LUFS around -9, and RMS around -11 to -12.

Okay. I still don't know what outlet you're using for your comparisons but it sounds to me, from your description, that it is a peak-normalised outlet rather than a loudness normalised one... In which case the LUFS meter isn't really the appropriate tool. A combination of a peak meter and a dynamic range meter might be more useful.

Nevertheless, the LUFS meter will obviously provide a bunch of numbers which you can try to emulate to make your own mixes sound similarly loud. As I said earlier, for a short-form trailer it is the Short-Term meter reading which is the most relevant, and especially so for a track which is very dynamic as you mentioned. It will always be higher than the Integrated value.

Is the integrated loudness basically an average LUFS reading of the entire track?

It is what it says: an integrated value of the rolling signal levels across the entire programme (but with dynamic low-level gating as Wonks mentions). Technically, integration is a specific mathematical function which is not quite the same thing as an mathematical average, but you can think of it that way if it helps.

In the LUFS meter, the Intergrated value starts building at the start of the programme and continues the calculation continuously as long as the programme runs. In contrast, the Short-Term value is based on a sliding 3-second 'window'. So it's delivering a number which indicates the loudness of the last three seconds only -- which corresponds quite nicely with our own perception of loudness 'in the moment'.

H

Re: Integrated loudness

PostPosted: Tue Jan 29, 2019 8:25 pm
by hz37
All good answers! What I might add is this:

- There is no short term loudness reading of an entire track.

- What you see is the mathematical result of having a dynamic track. You can see your loud part, which has its own -12 LUFS integrated loudness, as a measure against which to compare other things. If you also have significant softer parts, the integrated measurement of your entire track naturally drops way below -12 LUFS. So, no worries if that reaches -16 LUFS.

I have mastered a very dynamic album, and this is exactly what I did to get the loudness of wildly different tracks aligned; measure the representative parts for integrated loudness and gain change the entire track so all the representative parts have a similar average loudness. This can easily be done with a dB gain, as a LU (loudness unit) is just a dB by another name.

Hope this helps!

Hens Zimmerman

Re: Integrated loudness

PostPosted: Wed Jan 30, 2019 9:22 am
by Zukan
It's time we had a Hugh Jedi article on this. I get emails almost daily asking me about this subject.

Re: Integrated loudness

PostPosted: Wed Jan 30, 2019 9:34 am
by Wonks
Zukan wrote:It's time we had a Hugh Jedi article on this. I get emails almost daily asking me about this subject.

So why don't you write one, Ed?

Re: Integrated loudness

PostPosted: Wed Jan 30, 2019 10:17 am
by Kwackman
Zukan wrote:It's time we had a Hugh Jedi article on this. I get emails almost daily asking me about this subject.

There was this from a few years ago...

https://www.soundonsound.com/techniques ... udness-war

Re: Integrated loudness

PostPosted: Wed Jan 30, 2019 10:25 am
by Zukan
Wonks wrote:
Zukan wrote:It's time we had a Hugh Jedi article on this. I get emails almost daily asking me about this subject.

So why don't you write one, Ed?

I've created about 4 videos on this subject Si. I don't think I could face going through the same content for an article.

Re: Integrated loudness

PostPosted: Wed Jan 30, 2019 10:26 am
by Zukan
Kwackman wrote:
Zukan wrote:It's time we had a Hugh Jedi article on this. I get emails almost daily asking me about this subject.

There was this from a few years ago...

https://www.soundonsound.com/techniques ... udness-war

Nice 1 Kwack. However, I think a hands-on approach type of article with real world mix examples might be a good idea.