Reverse C.A.L.M. - Would love your thoughts

[admix]

So here's a new one...

Shipped a commercial which consisted of one on camera talent, with minimal dialog consisting of a few hand claps, some sighs, and an "alright, ready?" at the end.

To my ears, it was mixed properly.

It measured -34LKfs.

Until it was trafficked by Extreme Reach.

Apparently, regardless of what's sent to them, they run through a normalizer of some type, that aims to spit out only -24LKfs product.

Since quiet commercials aren't generally a problem, this has never come up. Their process however, destroyed this commercial. It aired 10db louder than it was meant to.

According to a tech at Extreme Reach, there is no way around this, and no opportunity for a human to intervene if they think something is wrong. I further learned that ER measures the full spot, INCLUDING THE SLATE, which is a measure of stupidity I'm having difficulty grasping.

My question to you guys is, does CALM, or A/85 have information regarding this type of situation? In my cursory read of it, I didn't find anything relating to audio that was quiet. My understanding was that CALM was created to combat overly loud spots. I fully understand and appreciate -24 +/-2 as a concept, but shouldn't there be leeway for soft?



Thanks in advance,

[nucelar]

¿How exactly did you measure it? AFAIK up to date meters use a relative gating algorithm that takes soft passages out of the measurement, so regardless of the quantity of dialog, the reading should be consistent.
If your meter measured -34, why do you wonder it airs 10 dB hotter? what is your reference? Do you have a calibrated monitoring?

A hack would be to mix in a high pitched (say 18kHz) tone at enough amplitude, but of course this could lead to a whole lot of other problems :)

[admix]

Measured with the TC LM5D plug in a calibrated, dolby certified mix room.

[Chief Technician]

The EBU has encountered this issue since their adoption of R-128. I think their upcoming solution to this problem is to have humans verify that such commercials and promos are indeed supposed to be this quiet for artistic reasons, in which case they are effectively granted immunity from being too quiet.

The ATSC will adopt ITU-R BS.1770-3 (gated) measurements in June. Those gated measurements will not solve the problem you described here.

If what you said about ExtremeReach is true, then not only are they extremely processing audio, they are overreaching with their measurements. A/85 does not call for measuring the slate.

[philip_purcell]

As far as I am aware, loudness standards such as A-85/R-128 are about consistent loudness at broadcast, not just loud commercials. If your spot is 10dB low, a lot of viewers will turn up the volume (because that's what people do) so the next spot will be too loud, and they will complain. So you've just defeated the whole purpose of the loudness standard. It has to be delivered to -24 LKFS (or -23 LUFS if using R-128). That's the whole point. Dunno what you do if there is not sufficient content. Make the claps and minimal dialogue unnaturally loud?

[soundworx]

Would it help to insert a 20 Khz tone loud enough to get it into spec? Poor dogs...