"ARTE" HD Audio Delivery Specs

[Moofie]

Hi All,


Having some difficulty deciphering the delivery spec for the broadcaster ARTE.

We're mixing an HD prog for them (in 5.1) but only delivering the Stereo version.

Detail we've been supplied via their Jan2011 delivery guidelines contain the following information (annotated to reflect relevant parts of the guide)...

2.5 AUDIO SIGNAL SPECIFICATIONS
2.5.1 PCM FORMAT
The sampling frequency is 48 kHz. Quantisation must be 16 bit.
2.5.1.1 LINE-UP LEVEL
• The line-up level as read by a full-scale digital peak meter is -18dBFS at 1 kHz, in accordance with EBU Technical Recommendation R68-2000.
• The line-up level as read by a 10ms DIN 45406 quasi-peak programme meter is -9 dB.
2.5.1.2 MAXIMUM LEVEL
The maximum level authorized is 9 dB above the reference tone, i.e.:
- -9 dBFS on a digital PPM in dBFS (short peaks up to -6 dBFS can be tolerated) or
- 0dB on a 10 ms DIN 45406 quasi-PPM (short peaks up to +3 dB can be tolerated)

For delivery of Multi-Ch/Dolby-E material they provide the following...
2.5.2 DOLBY E FORMAT
2.5.2.3 METADATA
2.5.2.3.2 DIALNORM SETTINGS
- Measurement should be carried with an instrument using K-weighting (Leq(R2LB)) in accordance with ITU-R BS. 1770-1
- All programme channels except LFE must be metered. - The measured value must be carried in the metadata and should be situated
at an average reference value of -25 LKFS (with a tolerance of ± 3 dB)
For programmes with dialogue and a duration of over 15 seconds:
- In this case the “Dialogue Level” of all dialogue passages has to be measured and entered as a metadata value
- For programmes of this kind, a maximum deviation of ±1dB between the measured value at ARTE and the given metadata value, can be tolerated.

(They offer no guide as to Peak Limits for the Multi-Ch/Dolby-E spec)

Now my questions are... (remembering we're only delivering stereo.)

Do we supply them the stereo mix using the LKFS -25 +/-3db ref?

If so, what peak level is used? -9dbfs as per item 2.5.1.2?

OR do we supply them a stereo mix peak limited to -9dbfs with avg RMS level of -18db (using the PCM spec as described above)

Any other thoughts/experiences would be appreciated as we're having difficulty finding the correct people to discuss this with at ARTE themselves.



Regards

Tom M. NZ

[ErikG]

Arte has the worst written spec known to mankind...

To make it worse there doesn't seem to be any human yôu can communicate about what it actually means either at their co.

We ended up comp/limiting to -9dbfs and pretty much nothing else (started with a feature film mix).
The lkfs reference is also weird as it is Definetly non-standard.

To make it worse they required the messiest m+e delivery ever...

I hope I never need to deliver to them again...

[Jamie Mac]

Stereo mix peak limited at -9dB.


The LKFS part is for multi channel.

[apple-q]

Originally Posted by ErikG

Arte has the worst written spec known to mankind...

To make it worse there doesn't seem to be any human yôu can communicate about what it actually means either at their co.

We ended up comp/limiting to -9dbfs and pretty much nothing else (started with a feature film mix).
The lkfs reference is also weird as it is Definetly non-standard.

To make it worse they required the messiest m+e delivery ever...

I hope I never need to deliver to them again...

Why is it messy? The LKFS specs clearly refer to multi-channel. Since the OP is delivering stereo he needs to provide maximum peak at -9dBFs (on a DIN meter with 10ms integration time). The rest of that list can be ignored.

All pretty simple and clearly written. They even state the tolerance for overs and the type of meter you need. That´s about as specific as you can get. Or am I missing something?

These are the common level standard specs of all german broadcasters (arte is a german/french joint venture public station). At least till Jan 1st 2012 when they switch to R128 instead of peak level based metering.

[Moofie]

Cheers for the input guys and much appreciated as you've all been involved with 1770/1&2 a lot longer than any of us in this part of the world and you're also more familiar with the local nomenclature used to describe various metering/reference specifications.

You're all confirming what we're thinking and good to get some outside guidance with the LKFS spec as it's not a spec we've seen in NZ as yet and this is the first project I've seen request it.

After 12 years of mixing long-form film/tv drama/doco a particular way, with no mix being rejected at QC (to multiple delivery spec's/broadcasters) you want to make sure you understand the next gen of tech spec pretty thoroughly.

As an aside, whilst this programme is being delivered Stereo (on HD-CAM with no Dolby-E component), we are sourcing that stereo output from a 5.1 mix which will be done referencing the LKFS portion of that delivery guide. Obviously we meter the stereo fold down separately to ensure that's still adhering to ARTE guidelines.


Cheers All once again.


Tom
--------------------------------------------------------------------------------------------------
Main System: 2 x 2.26 Quad-Core, 10.6.5, 12gb RAM, PT 9.0.2, HD-Accel4, Magma PE6R4, 2 x 192, 1 x 192 Digital, SyncHD, Decklink HD Extreme 3 + Satellite Link + D-Control Main Unit. Satellite System: 2 x 2.66 Dual-Core, 10.6.5, 5gb RAM, PT9.0.2, HD Accel2, 1 x 192 Digital. SyncHD

[apple-q]

Originally Posted by Moofie

As an aside, whilst this programme is being delivered Stereo (on HD-CAM with no Dolby-E component), we are sourcing that stereo output from a 5.1 mix which will be done referencing the LKFS portion of that delivery guide. Obviously we meter the stereo fold down separately to ensure that's still adhering to ARTE guidelines.

I can asure you that the stereo-mix will positively not be QCed for loudness specs. ONLY the peak level specs have to be met.

Just make sure you don´t waste headroom by using a sample-accurate peak meter. 10ms is rather slow so if you use a digital brickwall limiter set to -9dBFs the result on a 10ms DIN meter will be well below possible maximum.

Loundness specs haven´t hit public broadcasters in europe fully yet. None of german based stations want R128 specs to be met for stereo yet, AFAIK.
This will change next year.

[ErikG]

If the spec is -9dbfs that IS referenced to digital full scale and not DIN scale. Thats why it's messy as the din scale reference says one thing and the -9dbfs is another. And finally a non standard lkfs level ie not EBU r128.

[apple-q]

Originally Posted by ErikG

If the spec is -9dbfs that IS referenced to digital full scale and not DIN scale. Thats why it's messy as the din scale reference says one thing and the -9dbfs is another. And finally a non standard lkfs level ie not EBU r128.

Quote from the specs:

2.5.1.2 MAXIMUM LEVEL
The maximum level authorized is 9 dB above the reference tone, i.e.:
- -9 dBFS on a digital PPM in dBFS (short peaks up to -6 dBFS can be tolerated) or
- 0dB on a 10 ms DIN 45406 quasi-PPM (short peaks up to +3 dB can be tolerated)

I can´t follow what you´re saying..The DIN scale is just a scale. In the old days it referred to +6dBU electric level at 0 dB on the meter but that refers to an analog signal.

Before that they state that reference tone is at -18dBFs.

as a result the 0dB mark on a DIN meter shall equal -9dBFs.

Makes sense to me.

I have a digital RTW meter in front of me right now. It shows 0dB on the stereo DIN scale when I feed it with -9dBFs digital signal.

[ErikG]

I see what you are writing.
You are correct that this version of the spec is actually not incomplete as the one I received a year or so ago.

I think my memory clouded my reading abilities a bit...

I still don't think it's a very good way to set up a spec but Germans will be Germans (I was delivering to the French part of arte) perhaps the language barrier did it even funkier to deal with then.

[UnderTow]

Originally Posted by ErikG

I think my memory clouded my reading abilities a bit...

Well the "Maximum level is -9 dB FS but short peaks up to -6 dB FS can be tolerated" is rather vague... What is a short peak?

Ah well, hopefully everyone will switch to R128 soonish.

Alistair

[apple-q]

Originally Posted by UnderTow

Well the "Maximum level is -9 dB FS but short peaks up to -6 dB FS can be tolerated" is rather vague... What is a short peak?

Ah well, hopefully everyone will switch to R128 soonish.

Alistair

It´s actually quite easy. Get a real DIN meter with 10ms. Set your limiter so you´re peaking at 0dB on the meter (not the digital brick wall limiter) and all is fine. Common sense really.

[UnderTow]

Originally Posted by apple-q

Common sense really.

Not when you write dB FS. ;-)

Alistair

[apple-q]

Originally Posted by UnderTow

Not when you write dB FS. ;-)

Alistair

Did I?

[UnderTow]

Originally Posted by apple-q

Did I?

No the spec does but never mind, it isn't important. The OP has the answer to their question.

Alistair

[DaVogi]

Originally Posted by UnderTow

No the spec does but never mind, it isn't important. The OP has the answer to their question.

Alistair

no, the spec doesn't confuse dbFS with DIN, you did ... the spec defines both clearly (dBFS AND DIN-scaled limits) and a engineer should be able to tell the them appart... BUT I totally understand that this can be really confusing if you're not used to the 'habits' and measurement-tools of the broadcasters home country...

so I really look forward to an open and more globally used measuring system... and of course loudness metering, not peak like now in german broadcasting

[UnderTow]

Originally Posted by DaVogi

no, the spec doesn't confuse dbFS with DIN, you did ... the spec defines both clearly (dBFS AND DIN-scaled limits)

"-9 dBFS on a digital PPM in dBFS (short peaks up to -6 dBFS can be tolerated)"

PS: I deliver stuff for NL, BE, IE, UK, DE, AT, CZ, PL, HUN, US (off the top of my head) and have never had anything rejected. I can read specs. :-)

Alistair

[Steven1145]

I recently delivered an ARTE version of a film I mixed. The woman who handles the technical deliveries was there to oversee the mix.
Basically they dont care about loudness, LKFS or LU for now. What they care about is that their audio doesn't go above 0 on their PPM meter (however she did accept a few overs by up to +2 for gunshots, or very loud scenes, as long as they were short). Also ARTE likes dialog loud. As loud as the other sounds actually, meaning that your dialog should peak as much as possible between -9 and 0 on the meter.
For metering I used the PPMulator in DIN 45406 mode which is a perfect match for their requirements, and is exactly the same as the RTW meters they usually check their program content on. And the meter calibration for a constant sine wave at 1kHz is -9PPM (DIN45406) = -18dBFS.

[apple-q]

Originally Posted by Steven1145

I recently delivered an ARTE version of a film I mixed. The woman who handles the technical deliveries was there to oversee the mix.
Basically they dont care about loudness, LKFS or LU for now. What they care about is that their audio doesn't go above +9 on their PPM meter (however she did accept a few overs by up to +2 for gunshots, or very loud scenes, as long as they were short). Also ARTE likes dialog loud. As loud as the other sounds actually, meaning that your dialog should peak as much as possible between 0 and +9 on the meter.
For metering I used the PPMulator in DIN 45406 mode which is a perfect match for their requirements, and is exactly the same as the RTW meters they usually check their program content on.

+9?!
I guess you mean -9dBFs or 0 on the DIN scale. A regular RTW DIN meter only goes to +5 so +9 would be 0dBFs

[Steven1145]

Originally Posted by apple-q

+9?!
I guess you mean -9dBFs or 0 on the DIN scale. A regular RTW DIN meter only goes to +5 so +9 would be 0dBFs

Yes sorry, the morning coffee had not yet permeated my brain. I have edited the previous post.