CAS Quarterly

Spring 2018

Issue link:

Contents of this Issue


Page 68 of 79

C A S Q U A R T E R L Y S P R I N G 2 0 1 8 69 What we found out is that the CALM Act could, and indeed should, very well do what it was designed to do. The problem is that each network and content provider has interpreted A/85 in a different way, resulting in differing criteria for delivery that is directly undermining the effectiveness of this regulation. So, we decided to have a discussion about this, and reach out and see what the other re-recording mixers, layback engineers, and audio professionals in town are experiencing. It's our opinion that if you're going to have a standard, you might want to make sure that the standard is ... well, standard. We spoke with re-recording mixers Andy D'Addario, Joe DeAngelis CAS, Keith Rogers CAS, and Karol Urban CAS MPSE; Jim Starzynski, Chairman, ATSC S6-3 Group on Loudness for Digital Television (and the group that created A/85); Scott Norcross from Dolby Labs; Rick Hart, Layback Engineer at BluWave; and Scott Kramer, Sound Technology Manager for Netflix. It's worth noting that Jim Starzynski and Scott Norcross were both authors involved in writing the A/85 spec that the CALM Act references, so they can be considered authorities on this topic. Let's look at the recommendations for commercials (short-form content) vs. the TV shows we are mixing and watching (long-form content). The A/85 document focuses on commercials in no uncertain terms: "For short-form content, A/85 recommends that the average loudness of the full mix be measured over the entire length of the item." By contrast: "For long-form content, A/85 recommends identifying and measuring the anchor element during audio mixing and ingest. The anchor element is usually dialogue, which the listener would tend to focus on when setting the volume control. The loudness of the anchor element would be reported as the loudness value of the program for a properly mixed program." The intended implementation of this spec for programming is to sample a representative portion of the show and measure it to make sure the dialogue is within a specific range of "loudness," as defined by the BS.1770 LKFS measurement. It's not necessary to measure the entire show. It's not even necessary to measure one representative segment per act, but that would also be a reasonable way to do it. What it categorically does not require is the measuring of the entire program, from start to finish, of all content across all channels of the full program mix. Please read that last sentence again because this is the core of what this discussion is all about. It's also not necessary to limit True Peak to -6 dBFS, which is what a number of cable content providers have also chosen, but that's almost another conversation. Lowering True Peak to -6 dBFS does further restrict our dynamic range and can have the effect of making mixes sound louder, harsher and more squashed. This could also lead to yet another connected issue; that of the inclusion of loudness control devices in the broadcast chain (such as the Evertz IntelliGain ® box)—but that is another discussion entirely. It's also worth noting that people around town colloquially use "dash three" as an interchangeable term for "full program loudness." This isn't strictly accurate, since 1770- 3 can measure dialogue as the anchor, but the user has to specifically isolate the dialogue in order for this measurement to be taken correctly. Whereas in 1770- 1, Dolby Media Meter (the most commonly used tool for measuring LKFS) had separate readings for dialogue loudness and full program or "Infinite All" loudness. The reasons for the exclusion of this are technical and have to do with the inclusion of a gate in 1770-2 and 1770-3, and the phrase "don't gate the gate," but all of this just goes to further show that we don't, as a community, fully understand what it is we're measuring and how. So, herein lies the breakdown of the process: We are, as mixers, asked to deliver shows to specs based on what appears to be a misunderstanding of the A/85 recommendations, and the unintended consequence of this misinterpretation is that a lot of program content is actually made quieter than the commercials, flying in the face of the CALM Act and the intention of A/85 to set loudness but maintain dynamic range and quality of mix. Now, let's hear from some people who know what they're talking about… Jim Starzynski, Chairman, ATSC S6-3 Group on Loudness for Digital Television: "There are two significant misconceptions that plague the delivery of quality DTV audio content: "The first is, 'The measured loudness value should not deviate during the program!' This would yield content with no dynamic range. When content is mixed correctly, the measured loudness value will change with the intended dynamics of the show. However, if the dialogue of the show is mixed correctly as an anchor, isolated as described above and then measured, the desired loudness value should report correctly and consistently. The second misconception is that BS.1770-2 and BS.1770-3 full program mix (Infinite All) measurements

Articles in this issue

Archives of this issue

view archives of CAS Quarterly - Spring 2018