R128 Loudness and what the networks do with it

[huub]

Well, there is a distinct lack of knowledge on the network side here in Holland.
We (a big OB company) sent a letter to all networks, public and commercial, asking if they were going to implement R128, if they still wanted peak limiting, and if they were going to install R128 processing.
Noone ever answered the letter, most likely because noone at the networks has a clue.
R128 became offical law here in Holland, but the networks never informed us, or any of the other OB companies.
By asking the "eindregie" I found out that they indeed have R128 processing (no more old fashioned peak limiting) and you can hear this if you listen to your mix off-air, there's gentle leveling mostly.
However, it is very possible that some networks, since no one there seems to have a clue technically, take r128 as the new norm, since it's the law now, but do not even realise they need different processing...

huub

[nucelar]

Are you saying you delivered a fine EBU R-128 compliant mix and the network managed to mess up the dynamics, EQ, apply AGC and it's *still* R-128 compliant?? That would be amazing. Not to mention that your reputation as a mixer could be damaged.

Just to double check: are you sure you analyzed the direct mpeg audio stream recorded to the DVR or did you use another method (conversion to analog...) to record the audio?


I still don't understand how so-called Loudness Processors are supposed to work. For example the linear acoustics Aero which seems to be popular among broadcasters. The definition of Loudness is tied to the length of a programme. but these processors work in realtime without any knowledge of the future or the length of the programme. They might be useful for unpredictable live situations, but for preproduced content they should just bypass. In fact they read metadata, so if the metadata is set, they should just read it and scale the audio overall if needed.
Cheers

[Jamie Mac]

Originally Posted by nucelar

Are you saying you delivered a fine EBU R-128 compliant mix and the network managed to mess up the dynamics, EQ, apply AGC and it's *still* R-128 compliant?? That would be amazing. Not to mention that your reputation as a mixer could be damaged.

Unfortunately this is the case.
They apply slight multiband compression (maybe EQ too..) and Limit to -9dB before going to their loudness normalizer (a unit made by JÜNGER-audio) which also does the AGC. so what comes out of that unit will always be R128 compliant

I took the DVR out of the equation by recording different mixes on different networks and they give different results.

All in all the difference isn't night and day, especially when listened to on TV speakers. It doesn't sound bad, it's a vast improvement compared to before, where they processed the living daylights out of your mix with their optimods. But I still have the feeling if they want to play R128, they should do it properly.

I guess the network does it to make up for bad mixes or mixes done in AVID, and also they don't want to let go of "their sound".

[Jamie Mac]

All major Flemish networks installed this machine in their chain (I think).


Jünger Audio

[Jamie Mac]

Originally Posted by huub

Well, there is a distinct lack of knowledge on the network side here in Holland.

Here, they know sort of what they're talking about but some networks give me the impression that they don't approve of it (as do LOTS of 'older' mixers, and video editors who mix smaller shows on AVID or final cut). They all approve of the "lets make everything this loud" but they disapprove of the ability to create more open, dynamic mixes, or mix with less compression and (almost) no limiting.

[philper]

Originally Posted by Jamie Mac

Here, they know sort of what they're talking about but some networks give me the impression that they don't approve of it (as do LOTS of 'older' mixers, and video editors who mix smaller shows on AVID or final cut). They all approve of the "lets make everything this loud" but they disapprove of the ability to create more open, dynamic mixes, or mix with less compression and (almost) no limiting.

That's because the latter takes more work.

From what you are saying, it sounds like R128 requires mixers to even out their shows more than ever, due to the AGC involved in the on air boxes.

philp

[Jamie Mac]

Originally Posted by philper

That's because the latter takes more work.

From what you are saying, it sounds like R128 requires mixers to even out their shows more than ever, due to the AGC involved in the on air boxes.

philp

Problem is it sounds worse with the AGC (on my mixes at least) you can hear the leveling. I think its pretty normal for a talent show like The Voice to have the music stand out from the rest.. but the AGC doesn't let me do that (in the way that I and the director, and music supervisor,... want it) Now it notices a loud part and gradually turns it down and then afterwards gradually turns it back up again when its quiet. Its just a ridiculous process..

Another problem is that there are no rules to setting up that loudness normalizer and every network can set it up the way they want to and they don't communicate that to us.. So we don't know what the AGC will react to.
I've asked for the settings and they won't give them to me. They say its irrelevant..

[jrhager84]

Yikes. Sorry to hear, man...

Sent from my SAMSUNG-SGH-I777 using Tapatalk

[aTelecine-Lex]

Originally Posted by jrhager84

Yikes. Sorry to hear, man...

Sent from my SAMSUNG-SGH-I777 using Tapatalk

Indeed, this whole situation seems like a nightmare.

How can one expect any sort of reasonable expectation of what shall occur to their work down the line?

A question for those that are impacted by the standard:

Do you feel that your concerns are a product of the standard being somewhat new, and therefore there is a shared failure of understanding of how it should best be intergrated?

I suppose my above question naturally leads me to ask - What would have to happen to reach a balance which would be favorable to everyone involved?

Best,
Alexa

[Jamie Mac]

The problem isn't the standard itself but the way in which broadcasters still have the freedom to do whatever they want to muck up the sound to their liking while still being compliant with the standard.

So communication is key! A lot of this can be resolved (and I hope it will) once everyone openly communicates the way in which they operate, and appreciate the feedback we give them and act accordingly.


I must stress this again, the problems I've encountered aren't as horrible as they might sound. The leveling is only slightly, the compression eq is also only slightly.. Limiting is also only slightly.
But still, I thought one of the benefits of the new standard was that we would hear the mix in the way the mixer intended it. And it's a bit disappointing that this is not the case. In Belgium at least.

[philper]

That AGC thing is ironic. I guess they did it to avoid complaints from viewers having to change their TV volume to accommodate louder and quieter parts of shows, but that's what's going to end up happening as they chase the AGC around! This all is actually worse than the whole "dialnorm" magilla of a few years back.

philp

[audionuma]

Hello,
some experience from France where r128 compliance is mandatory for more than one year now for national broadcasters. Take into account that analog TV broadcasting has ended in France at the same time.
The rules are (summarized) :

• the integrated loudness over 24 hours for a channel must be -23 LUFS.
• all programs delivered since january 2012 must be broadcasted with an integrated loudness of -23 LUFS (+/- 1 LU).
• there's a tolerance for 'stock' (ie delivered before january 2012) programs that can be broadcasted with an integrated loudness between -25 and -20 LUFS.

(You may notice that this raise questions like : how can a channel present a daily integrated loudness of -23 LUFS if half of the programs are r128 compliant and the other half, being stock, is around -20 LUFS).

Several approach are possible :

• You have a strong QC department that is able to check conformance of all new programs, you broadcast a small amount of stock (or you have the tools to do some offline loudness alignment on the stock), then you can just use a true-peak limiter on your channel.
• Your playout system is able to read a metadata indicating wether a given element is r128 compliant or not and appropriately bypass an ALC : you can activate ALC only on the programs that are off tolerance. (Yes, there's a side effect on real-time ALC which is that a r128 compliant program may output not compliant from an ALC !).
• You are lazy, don't have a real QC or just don't want to bother : use an ALC all the time.
• Use the dialog_level dolby e-ac3 metadata to maintain conform output at receiver's side. Unfortunately, SD terrestrial broadcasting in France is MPEG1-LII, no metadata, and besides, most broadcaster prefer to have a constant dialog_level metadata on the e-ac3 stream of HD terrestrial broadcasting because some receiver behave strangely when you play with that (like sound disappears briefly when the metadata changes, or the receiver will hang ...)
• (Other options available)

I think in France a few channels use only true-peak limiting, a few use the 'activate only when needed' option and a majority use an ALC. But I think that the majority of those that use an ALC use it in 'loudness' mode and don't add a 0 PPM limiter. (Maybe some people do that in countries where you still have an analog broadcasting going on. Is that the case in Belgium ?).
Another point to take into account : you are not always really sure of what happens on other distribution paths (IP TV, which is of wide use in France, cable, satellite, ...) and if they don't add a processor ... (besides transcoding from e-ac3 or MPEG1-LII to he-aac for example).
I would ask Jamie Mac from what path he did his DVR record, and also, as mentioned, wether he checked the binary stream or the decoded stream at the output of his DVR.
As time goes by and more main control room go file based and use inteligent automated workflows, you can expect more file based loudness alignment when needed and hopefully less real-time processing.
From the audience perspective, we have much more feedback from people happy that the commercials don't jump in their face anymore than from people unhappy with bad effects of processors.

[kosmokrator]

What I do is mix to R128, as is required, and call it a day. There's nothing I can do about what happens on the network side.

That being said, having the chance to deliver dynamic mixes where appropriate is VERY cool, although heavy compression/limiting is still an option, again, where appropriate (trailers f.e.)

What concerns me more is that here some mixers, mostly in the ad-world, are trying to 'circumvent' R128 by applying ridiculous processing chains to their spots, so that it seems it's louder than the rest.

[MusicTrax]

Man, I've been listen to the NY FOX station (WNYW) on DirecTV, and the new ATSC spec (essentially identical to R128) doesn't seem to be helping. The local & national promos still blow me out of the room -- I'm guessing the levels are at least 3-4dB too hot. Spots are at least a dB or so hotter than program levels. I don't think it's working.

Other networks, like CBS, are damn near perfect from what I hear -- but they generally weren't too bad even before the new specs came into effect.

[Fredo]

I think we should add that the concerned Broadcaster in Belgium also has "customized" the R128 specs. As for the commercials, they also expect their program material to be delivered with a momentary loudness of -15LUFS. (!?)

They have a long history of "doing it their way" (before R128, they required that dialog should stay within a 6dB range). They are most concerned about their older audience, which should be able to hear everything loud and clear, even with their TV set pretty low in volume.

Of course, there is something to be said about R128 or any other (or lack of) loudness control, and it is understandable that they want to use the specs/rules in a way they are happy with it. However, if this means that all of the good work of the Sound Engineers is ruined and becomes pointless, then it is not OK. That being said .. not much has changed. The only difference is that we don't have to screw up our mixes ourselves anymore.


Fredo Gevaert
Temple Of Tune
Belgium.

[Steven1145]

We recently had to go to France Television because one of our feature film TV mixes did not pass QC there. I came out realising that the whole R128 thing had been sold to sound people as a way to get a better sounding mix on peoples TVs, when in all truth it's the absolute contrary.
Mastering highly dynamic film mixes to the TV medium, within French specs, can not be done while preserving the artistic intents of the director for many scenes. Asking for an overall Loudness of -23LUFS is one thing, but asking ALL the dialog to remain within-7/+7 LU also is a recipe for crap sound and disappearing dynamics. The opposite of what was promoted with the introduction of loudness metering.

And most TV channels resort to final on air processing anyway to avoid being fined by the CSA (Audiovisual Authority) for straying from the -23 LUFS measurement over 24 hours.

Disappointing to say the least.

Sent from my XT910 using Tapatalk

[philper]

I'm finding it kind of tough to get a PBS mix (-24, peak to -3) to pass all the requirements for R128, at least as far as the metering I have goes (Tone Boosters, Loudness Difference, DMM). The mix is pretty compressed compared to what I had before....

philp

[Oliver.Lucas]

Originally Posted by Fredo

The only difference is that we don't have to screw up our mixes ourselves anymore.

Hmm, that's sad. I have better experience from German TV. (now)
Last autumn I heard one of my shows mixed in R128 being broadcast within a -9dBfs context! Well, actually it was even worse, because the commercial after my show was -2dB true peak with a loudness of -12! Wild west!

Now, however I am under the impression that everything is working as expected. I'll double check and grab a stream off the cable network next weekend to be sure.

Ollie

[ErikG]

And if they run a AGC to conform to -23 during the course of a day, shows that was mixed and conforming to r128, they will no longer conform after the r128 compliant AGC.
But the channels output over a day will conform...

Between a rock and a hard place...

The only way this will ever work properly is when ALL delivered material complies (and the networks remove AGC) or at least that they handle non compliment material separately from the stuff that is compliant.

[huub]

I think everyone agrees that there should be some kind of flag in R128 compliant material, causing the processing to be bypassed.