ATSC A/85 - need some advice

[JohnO]

Originally Posted by Meads

Specs are met, I'm just concerned because the overall level of the mix is that low.

-14 sounds about right for a slammed mix. If it is limited hard like you say, there are no real high peaks. When I mix for Europe R128, my limiter is around -13 or -12 (promo so it's also not very dynamic) and I don't slam it too hard. So -14 for a harder limited mix sounds about right.

[Frank Kruse]

Originally Posted by Meads

That's what this mix looks like now. So it reaches peak level a lot, but that peak went down to -14.5dBFS after treating it with NUGEN LM Correct to match A/85.

But if I apply less compression it will be too dynamic for TV I'm afraid. Dialogue level is very low and fighting scenes are as high as it gets.

Did you simply slam it "al gusto" first and left the A/85 thing to LM correct? No wonder it's pulled down that much.
By compressing the hell out of it you increased the loudness so LM correct has no other choice than bringing it all down.

If I where you I'd meter while downmixing and adjust dynamics using the master-stems instead of squashing the full mix.

LM correct will simply measure the loudness and then gain the entire file up or down to match the specs. So if your mix is totally squashed (=loud) it will simply gain it down.

When you say your original has high dynamics it doesn't make sense to squash it all the way and then use a machine like LM correct to simply set the overall gain so that the result matches the LU.

Unless of course that's not what you did.

Often it's quite surprising how a regular theatrical mix is already pretty close to R128 so I was wondering why you had to squash it so much. ATSC is not that much different.

frank.

[Meads]

Originally Posted by Frank Kruse

Often it's quite surprising how a regular theatrical mix is already pretty close to R128 so I was wondering why you had to squash it so much. ATSC is not that much different.

Unfortunately it isn't in this case... Also, I do not have the stems, only the full mix to work with.

The mixer's responsibility is not solely that the mix complies to the standard, but he also has to keep in mind, that the dynamic range isn't too big. I did an R128 mix for ZDF once, and this has been rejected, because they found it to bee too dynamic (and it really wasn't that much). They do want a really slammed mix, just like before R128 - which I do not find to be surprising. Many viewers don't care about the "artistic qualities" of a mix. They sit at home, listening to the mix with their poor flat screen speakers and want to hear everything while eating their chips.

That is why I did exactly as you thought, Frank. I mixed down first and let LM Correct do its job thereafter, knowing full well what it does. And I wasn't surprised that the mix got lower, but I didn't expect it to go down that much.

So, how do you guys do it when you mix for TV? Are your ITU 1770-mixes greatly different from the old -9dB (-8dB) TV-mixes in the past in terms of dynamic range?

[Frank Kruse]

Originally Posted by Meads

Unfortunately it isn't in this case... Also, I do not have the stems, only the full mix to work with.

The mixer's responsibility is not solely that the mix complies to the standard, but he also has to keep in mind, that the dynamic range isn't too big. I did an R128 mix for ZDF once, and this has been rejected, because they found it to bee too dynamic (and it really wasn't that much). They do want a really slammed mix, just like before R128 - which I do not find to be surprising. Many viewers don't care about the "artistic qualities" of a mix. They sit at home, listening to the mix with their poor flat screen speakers and want to hear everything while eating their chips.

That is why I did exactly as you thought, Frank. I mixed down first and let LM Correct do its job thereafter, knowing full well what it does. And I wasn't surprised that the mix got lower, but I didn't expect it to go down that much.

So, how do you guys do it when you mix for TV? Are your ITU 1770-mixes greatly different from the old -9dB (-8dB) TV-mixes in the past in terms of dynamic range?

How can be it be too dynamic and meet their specs at the same time? If the mix is within the loudness-rage specs why would they reject it?

What are the figures you get when you measure the original mix?
IMO, I think slamming it all before even considering the LU metering might not be the best workflow because you can throw just about anything at LMcorrect and it will output a file with the correct LU because it's a machine.

The more you compress your mix the more the plugin will turn it down. That's the whole point of loudness based specs.

[ErikGuldager]

Yes, R128 mixes are a lot more dynamic than the old -10dbfs mixes.

Someone mentioned that there's not a lot to do to get a feature film to get within r128 specs. Hat would IMHO mean a very in dynamic feature film mix and a overly dynamic r128 mix.

I do not agree that just making it within the r128 spec is enough.

If you are within specs and a loudness range of over 22-25, I'd say that there may be intelligibility issues at typical TV playback levels. Personally I try to stay around 17-20 when doing a TV drama.
Iit's easy to stay within the spec while having widely different dynamics. And in a way the old concept of reducing dynamics for TV s still important IMHO. I wish it wasn't, since as a sound geek I'd love to mix for an audience that cared and had the spl when watching TV loud enough so that a nice big dynamic range would work.
Unfortunately that is not the case. I use myself and my everyday TV viewing as a reference. I need to be able to watch a show while the kids are sleeping next door and still hear and understand principal dialog.
This means that the dialog dynamic range needs to be pretty tight IMHO. Wether it's done using fader moves or compression is not the point. Then the rest of the mix is balanced around the dialog to fit within he specs. Does it result in a mix that is less interesting if listened to at a heigher playback level? Yes, but it is still as much of a compromise as it ever was.

In my case I also have to hit multiple specs as Sweden isn't yet using R128 for TV so I have a few more things to consider as well (-8/-10dbfs OR max peaks to +11 on the Nordic ppm scale, depending on what network it's going to).

[Frank Kruse]

Originally Posted by ErikGuldager

Yes, R128 mixes are a lot more dynamic than the old -10dbfs mixes.

Someone mentioned that there's not a lot to do to get a feature film to get within r128 specs. Hat would IMHO mean a very in dynamic feature film mix and a overly dynamic r128 mix.

I do not agree that just making it within the r128 spec is enough.

Just for the record: I never said that there's not a lot to do to get a feature to meet R128 specs. I said that an average film (as is "dialog driven film without a ton of explosions" etc.) is often surprisingly close to -23 in R128 specs. More ore less wanting to say that normally you don't need to squash the hell out of a mix to meet the specs. Unlike in the old days of peak metering where this was much more the case.
No doubt there are exceptions.
Of course this also depends on the range-specs of a given station.
I only used R128 as an example knowing the OP needs to meet ATSC because the two standards are not extremely far away.

It does get a bit weird when you have a documentary with no voice at all and they want you to reach -23. That's where the new specs fail to make sense I think.

[Meads]

Originally Posted by ErikGuldager

Yes, R128 mixes are a lot more dynamic than the old -10dbfs mixes.

Someone mentioned that there's not a lot to do to get a feature film to get within r128 specs. Hat would IMHO mean a very in dynamic feature film mix and a overly dynamic r128 mix.

I do not agree that just making it within the r128 spec is enough.

If you are within specs and a loudness range of over 22-25, I'd say that there may be intelligibility issues at typical TV playback levels. Personally I try to stay around 17-20 when doing a TV drama.
I have done that and it's easy to stay within the spec while having widely different dynamics. And in a way the old concept of reducing dynamics for TV s still important IMHO. I wish it wasn't, since as a sound geek I'd love to mix or an audience hat cared and had the spl when watching TV loud enough so that a nice big dynamic range would work.
Unfortunately that is not the case. I use myself and my everyday TV viewing as a reference. I need to be able to watch a show while the kids are sleeping next door and still hear and understand principal dialog.
This means that the dialog dynamic range needs to be pretty tight IMHO. Wether it's done using fader moves or compression is not the point. Then the rest of the mix is balanced around the dialog to fit within he specs.

In my case I also have to hit multiple specs as Sweden isn't yet using R128 for TV so I have a few more things to consider as well (-8/-10dbfs OR max peaks to +11 on he Nordic ppm scale, depending on what network it's going to).

This is exactly what I was saying, Erik. And that's the reason why I reduced the dynamic range of the original mix by a lot - to a degree that I think is suitable for "being able to watch the show while the kids are sleeping next door and still hear and understand principal dialog.", as you put it.
But that leaves me with this low overall-level mix, and my question is just that - is that okay and a common thing, or am I way off and need to allow a greater dynamic range therefor gaining more over all loudness?

Frank, as I said, I do understand what LM Correct does, that you can feed it anything and it will either raise or lower it's volume to achieve the specs. That is exactly how I used it, and it did what I intended it to do.
I didn't measure the original 5.1 mix, because that doesn't really help. I have to deliver a stereo downmix of the show, so the loudness of the 5.1 is not a reference.

[Frank Kruse]

Originally Posted by Meads

I didn't measure the original 5.1 mix, because that doesn't really help. I have to deliver a stereo downmix of the show, so the loudness of the 5.1 is not a reference.

I meant the stereo downmix without any compression added.

[ErikGuldager]

Meads.

If you don't have stems or original project I would not squash it that hard anyway. It's better to first reduce levels during the loud parts (where any dialog will likely be loud anyway) and raising the lower sections before adding any compression, this will sound a lot better than just squashing the loud stuff. And it will require less over all compression and going you a nicer mix.

If its a VERY dynamic feature I would consider treating the centre differently to the other channels, but this takes time and is of course another compromise, but still better than squashing it all IMHO.

[Meads]

Originally Posted by ErikGuldager

Meads.

If you don't have stems or original project I would not squash it that hard anyway. It's better to first reduce levels during the loud parts (where any dialog will likely be loud anyway) and raising the lower sections before adding any compression, this will sound a lot better than just squashing the loud stuff. And it will require less over all compression and going you a nicer mix.

If its a VERY dynamic feature I would consider treating the centre differently to the other channels, but this takes time and is of course another compromise, but still better than squashing it all IMHO.

That's pretty much what I did. Were compression sounded badly I used volume automation. As I already said, I'm happy with how the stereo-mix sounds, that is not the issue here.

Frank, I just measured the Lt/Rt and it reads -12.8LKFS.