[MLA] Information about loudness normalization; k-weighting

[florian.camerer at orf.at]

Dear friends,


aa could be "alternative algorithm"... :-)


On a different note (sorry for the long silence, too dense here), let me briefly explain the situation with a potential "change" of the K-weighting curve (has been a topic a short while ago): any major change to the filter curve has to undergo lenghty and profound testing procedures at various test sites, ensuring that the modification is statistically significant. Add to that the potential downward incompatibility with the current K-weighting, and you understand, why the ITU is very reluctant to undergo such a process. The first one took them 3 years.

So the only chance to get a slight modification in (like an LP filter) is if this modification does not change the results of the data base (or at least only marginally), if the CPU load increases only slightly (due to legacy meters; of course current CPUs have no issue, but meters from 2012 might), and if the modification's cause is NOT preventing abuse.

Ian Dash from Australia has actually put forward this point, and I agree. The prevention of abuse is mainly the duty of the broadcaster - so we should (and will) argue that the loudness meter should only measure what we here. There is material out there (plenty, actually) with pilot tones (old tapes) which give totally wrong readings due to this signal way above 20 kHz. It is along this line that we need to argue. PLOUD will do that for at least the autumn meeting of ITU. Once I get my head above the water again, I will draft this input.

Scott Norcross from Dolby has already performed some preliminary tests with filters at 19 kHz and other frequencies, with different orders, and he found very little differences in the original ITU database. But Thomas rightly pointed out that we need to check on current pop/rock material and other stuff which has much more HF content. It would be great to have a test version of a loudness meter where we could dial in different frequencies and orders so that we could perform some testing ourselves.

Eelco, any chance that you get this into Level One V2? If not, I will contact the Nugen guys and Esben from tc electronic.


All the best, Florian




________________________________
Von: MLA <mla-bounces at grimmaudio.nl> im Auftrag von Grimm | Eelco <eelco at grimmaudio.com>
Gesendet: Montag, 05. Juni 2017 22:16
An: Music Loudness Alliance
Betreff: Re: [MLA] Fwd: Information about loudness normalization

Hi Bob,


Apple has been reporting 1770 loudness in afinfo for more than a year. Ever since Eric Allamanche came on board from Fraunhofer.

OK, very cool. I did not know that.

All the best,
Eelco
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://grimmaudio.nl/pipermail/mla_grimmaudio.nl/attachments/20170608/9feb0063/attachment-0001.html>