Inspiring

Answered

What to do with match loudness + result of loudness radar?

Forum|Forum|9 years ago
January 23, 2017
3 replies
1989 views

I have sound with a wide dynamic range across 3 tracks in the CD track sense. It's sort of music. But, it's generally too low. I need to get the average and increase levels appropriately so match loudness sounds like what I want.

I'm probably far from understanding what the different standards mean. But, I'm hoping to proceed before fully understanding all the details.

If I try to match a standard and loudness radar shows the result is mostly too low (in, umm, loudness) is there a relatively simple explanation of what needs to be done to make the result match the standard?

loudness normalization

This topic has been closed for replies.

Correct answer SteveG_AudioMasters_

If you don't need LUFS, then simply don't use them; use your ears. You'd be better off using the multiband compressor on your mix until it sounds the way you want, and then normalizing whatever you end up with to about -1dB.

[slight rant] Despite all the claims to the contrary, this whole LUFS business, and the reasoning behind it simply doesn't work, and the entire premise of it is seemingly based on messing about with overall dynamic range. It's supposed to even programme material out so that when the adverts come along, they aren't disturbingly loud. This completely ignores the fact that all stations run transmitted audio through a Optimod before sending it out, and the commercial producers know this - so they bang everything up and have no dynamics anyway. You can never match this with programme material over a longer period, so people manipulate bits of material within programmes to skew the results, and it sounds dreadful; you get occasional blasts of sound you weren't expecting, and nobody ever manages to fix the real issues - like music drowning out narrative. The whole thing continues to be a complete mess, now made worse by these ridiculous compliance standards.[/slight rant]

cheyrnAuthor

Inspiring

So, examples of the sort of thing that I imagined are:

1. An option to have amplitude of the source automatically increased to account for the gap between target loudness and loudness recorded by loudness radar.

2. An option to have compression and limiting with some variability applied to the source, automatically, based on the result of the loudness radar results.

Of course, as always, people also have to use their ears eyes brain. One could say there is no point to having any meters at all since ultimately you have to listen to the sound on various devices in various scenarios. But, meters are partly useful.

SteveG_AudioMasters_

Community Expert

cheyrn wrote:
So, examples of the sort of thing that I imagined are:
1. An option to have amplitude of the source automatically increased to account for the gap between target loudness and loudness recorded by loudness radar.
2. An option to have compression and limiting with some variability applied to the source, automatically, based on the result of the loudness radar results.

It's doing processing on this basis that results in audio where the levels jump all over the place, though. Since over any given period, you need an 'average' figure (this is based on a modified Leq measurement), you only need to nudge this up slightly in several places, or a lot in a few places to achieve the same result - same goes with reducing levels. The problem is that if you have a LUFS measurement that measures to spec for a 30 second advert, then that's fine - you won't have much leeway with the levels because there's so little to work with during the duration of it. But if you apply the same figure to a 1 hour programme you can have the levels all over the place during it. To attempt to reduce the effects of this, a 'gating' process is included, so that periods where things are supposed to be quiet don't get included in the measurement. So, to automate any processing, the processor would have to know what your intentions were at any given point in your programme, and if it drops below the gating level it has no idea about it anyway!

Any guesses made on this basis would simply be wrong in the majority of cases. Chances are that any automated processing is going to mess with the relative levels around quiet parts and completely ruin your intentions. I'm afraid that you simply cannot avoid using your ears to achieve a listenable result. If you reduce all of the dynamics in your production, you can achieve -23 LUFS just by dropping the overall level of it simply by normalizing to a lower level - which is pretty much what happens with loud 30-second adverts.

ryclark

Participating Frequently

If you want to hear what automated loudness/level control can do to your audio try the Speech Volume Leveler effect from Effects/Amplitude and Compression. This tool can be useful in evening out levels in unattended recordings.

On your other point about meters. They are required since digital audio recording has limitations and the meters are there to help you optimise the recording levels without going outside of the parameters that the recording/playback system is designed to deal with.

cheyrnAuthor

Inspiring

If anyone has something to add I would appreciate it. For now I will take the answer as meaning there is no commonly known way to use the software to apply the results of loudness radar and instead one would do what they will ultimately end up doing anyway, which is to listen and correct the audio.

ryclark

Participating Frequently

I would agree with Steve. The Loudness Radar concept is really only there to allow users of Audition to check that their programme output complies with loudness requirements for a broadcast or distribution company.

cheyrnAuthor

Inspiring

I suspect it would be hard for someone to know how to answer that question since it is pretty vague. So, maybe this is a better question:

Compared to the level of a dancey discoy pop song from a CD my noises are very very quiet. The pop music is usually in the red of the levels panel (60db range with dynamic peaks).

So, I tried using CD Master in loudness radar and used ITU-R BS.1770-3 Loudness in match loudness. The result of that is that my noises are usually in the dark green area...

If I increase target loudnes from -23 to -15 LUFS then my sounds are too loud. (I picked -15 because the CD master uses -15 LUFS if I switch from LKFS to LUFS).

So, I can keep fiddling with target loudness. But, is there a smarter approach?

SteveG_AudioMasters_

Correct answer

Community Expert

If you don't need LUFS, then simply don't use them; use your ears. You'd be better off using the multiband compressor on your mix until it sounds the way you want, and then normalizing whatever you end up with to about -1dB.

[slight rant] Despite all the claims to the contrary, this whole LUFS business, and the reasoning behind it simply doesn't work, and the entire premise of it is seemingly based on messing about with overall dynamic range. It's supposed to even programme material out so that when the adverts come along, they aren't disturbingly loud. This completely ignores the fact that all stations run transmitted audio through a Optimod before sending it out, and the commercial producers know this - so they bang everything up and have no dynamics anyway. You can never match this with programme material over a longer period, so people manipulate bits of material within programmes to skew the results, and it sounds dreadful; you get occasional blasts of sound you weren't expecting, and nobody ever manages to fix the real issues - like music drowning out narrative. The whole thing continues to be a complete mess, now made worse by these ridiculous compliance standards.[/slight rant]

ryclark

Participating Frequently

And nobody trained to use their ears controlling the actual transmission output anymore.

Sign up

To post, reply, or follow discussions, please sign in with your Adobe ID.

Sign in to Adobe Community

To post, reply, or follow discussions, please sign in with your Adobe ID.