Issue link: https://digital.copcomm.com/i/1524736
C I N E M A A U D I O S O C I E T Y. O R G 17 I've written in previous articles about how LRA as a brick wall delivery requirement is a terrible idea whose existence mostly demonstrates a lack of understanding about how LRA works, and recently, I've seen this get amended to a "guideline LRA," which is to the credit of the people in charge of those specs. For clarity, the tricky thing about LRA is what is excluded from the measurement—it doesn't count the loudest 5% or the softest 10% of the dynamic range toward the measured value—for the good reason that a few loud sounds mixed for emphasis shouldn't skew the whole mix. In practice, however, this creates the reality that mix moves designed to reduce dynamic range can result in a higher LRA value, and we can end up chasing this number ad nauseam when it's a required delivery parameter. Along with LRA, the inclusion of a maximum range between dialogue loudness and full-program loudness is gaining popularity. In the AES paper, we refer to this relationship as "Program to Dialogue Loudness Ratio" (PDLR), so I will do so here. (In the EU, they are currently using the term "Loudness to Dialogue Ratio [LDR]," but myself and some of my colleagues on the AES Committee thought this term was worth another shot.) Despite the questionable merits of including this as a delivery requirement, it turns out there's solid data to suggest that paying attention to this value can provide insight into how the mix might translate to viewers in the home. One of the most common complaints about home theater mixes is the "reaching for the remote control" factor, where the viewer sets the level of their system based on the loudness of the dialogue, but then has to spend the duration of the program chasing the level up and down based on the dynamic range of the overall mix. This is, however, highly subjective and influenced mainly by their living environment and personal preferences. Someone living in an apartment building with nearby neighbors or someone with kids asleep in the next room will have a far lower tolerance for a wide PDLR range than someone living in a house and watching on a beefy home theater system. The latter is more likely to be perfectly happy with a more dynamic, cinematic home listening experience. I, myself, am in this latter category, but I understand I'm likely in the minority. Both Netflix and Disney+ have an LRA guideline, and Disney stipulates a maximum PDLR of 7 dB. The EBU's PLOUD group has recently released an update to their R128 - s4 (the European equivalent of ATSC A/85) that recommends the value not exceed 5 dB. Netflix doesn't specify a delta, but in conversations I've had with Mr. Scott Kramer, he has expressed concern over a value as high as 8 dB. Interestingly, internal data shows that only around 3% of titles on the Netflix platform exceed the EBU recommended value of 5 dB, but I'm helpful, and fostering the mindset that on-set sound is important and requires time and resources to correctly capture would help get us all to put our best foot forward. The same time and budget restrictions apply to post, and we often find ourselves shortchanged in that regard. In sound editorial, the balance between noise and noise reduction is a delicate one that can have ramifications—too much noise and the dialogue may be obscured, but too much noise reduction can create an unnatural effect that distracts the ear and brain, which can steal focus from the spoken words. The aforementioned Mrs. Urban pointed out that clients can sometimes get used to audio that is very noisy, and once cleaned up, they feel like it is somehow missing something emotionally. Or, they may even get used to something that has been cleaned up badly, resulting in excessive artifacts, and then request the other version be put back into the mix. Resist! I say. Best of luck. LEVEL AND READING CONSIDERATIONS And so, the mix. There's a lot of information to be gathered from the different loudness measurements we can take and the relationship between them. You may have noticed some streaming specs that have started to stipulate both a dialogue- anchored loudness and a full-program loudness, or a max LRA range, or even both.