A little bit ago the Audio Engineering Society issued an update to their 2015 document TD1004, now titled TD1008 “Recommendations for Loudness of Internet Audio Streaming and On-Demand Distribution.” So, let’s talk about it (and LUFS) a little bit.
The AES is, “the largest community of audio experts” and is comprised mostly of audio engineers and mastering engineers (if not entirely). I’ll let them say a little more:
The Audio Engineering Society is the only professional society devoted exclusively to audio technology. Founded in the USA in 1948, the AES is now an international organization that unites audio engineers, creative artists, scientists and students worldwide by promoting advances in audio and disseminating new knowledge and research.
12,000+ members are affiliated with 90+ AES professional sections and 120+ AES student sections around the world that provide members valuable opportunities for professional networking and skill and career growth.
The AES is a 501c3 non-profit corporation, registered in New York City, New York, USA. The bylaws of the Audio Engineering Society can be found here.
It should be clear that, if anyone has the ability to have an informed opinion about anything related to audio, it’s the AES.
Phone and other devices manufacturers
When you create a phone, let’s say the iPhone, one of the conversation you have to have is how loud things on the device should be or seem to the humans using those devices. LUFS is a measurement of loudness as perceived by human and when you say that a sound is “mastered to” X-LUFS what you’re saying is that the sound, whatever it is and however long its duration, seems to be this loud to humans. This doesn’t have anything to do with the volume during playback, it has to do with the loudness of the experience relative to human beings perception of loudness.
Apple has decided that its sounds, like text message alerts or rings or whatever, should be mastered at -16LUFS. Other devices agree or disagree, but they’re free to do that because they own the environment soup to nuts. It’s their box, they make the rules.
Podcasters have been told, for a long time now, to master their audio to -16 LUFS (for Apple Podcasts)
But there’s a slight problem with doing that because Spotify wants audio submitted to its platform at -14LUFS and, sometimes, -11LUFS (though that’s music related value and I should skip over it).
Well how does a podcaster master their podcast audio to -16 LUFS and -14 LUFS? They’ve only got one mp3 file in their feed, right? So which do they choose? I’ve taught my students to aim for -15 LUFS because it splits the difference and no normal person can discern a loudness difference of 1LU either way. But I don’t like saying this, I’ve never liked saying this, because it’s a compromise and I, for better or worse, am not a fan of compromising on technical things.
What does this have to do with the AES and their -18 LUFS recommendation?
Because neither Apple, Spotify, Google, or any of the rest of them seem the least bit interested in conforming to this standard. Though, to be clear, this recommendation is not yet a proposed standard (we don’t have an established standard for loudness in podcasting yet), but it may well wind up being something various bodies work to make the standard for “Internet Audio Streaming and On-Demand Distribution.”
Why is this important? Is it important?
The -18 LUFS value isn’t the important thing here, the important thing here is that it seems unlikely that any of the platform or device developers would be likely to comply with a standard were it formalized. I take issue with that because, as is usually the case with me, I feel it makes things more difficult for creatives because it means we can never get our loudness dialed into a value which befits all platforms… which means it creates a different listening experience depending on where you’re listening to what those creatives create.
I don’t feel that the device manufacturers or the platform developers should not be beholden to a loudness standard. It’s not as though a television manufacturer has the final say in the standards of the medium it conveys. Nor does the manufacturer or the radio in your car have the final say in the medium it conveys. Why then, is podcasting different?
Well, for one, because we deliver our audio through RSS and no one has control over what we do (something most podcasters think is of great benefit re: freedom of speech — and I don’t, at least on that point, disagree). The only entity between our audio and the platform is the hosting provider we upload the audio to; like Mark Asquith’s Captivate or Alban Brook’s Buzzsprout. They could apply modify our audio to meet an arbitrary loudness measurement, but even they could only do it once. In order to satisfy each platform they’d have to build some sort of mechanism that created multiple versions of our file, then, before delivering them, check which platform was requesting them, and, in real-time, suggest the appropriate version.
They could probably figure that out, maybe.Or Adam Curry could do something with podcasting 2.0 that made this possible — but is it fair to put the onus for meeting a would-be standard like this on the hosting providers?
You might ask, in return, is it fair to ask the developers or manufactures to modify their current proprietary standards to meet an agreed upon standard? I would say yes, I think it’s more than fair since developers and manufacturers are in the business of standards compliance anyway — at least far more than intermediaries like podcast hosting providers are.
AES isn’t proposing a standard yet, -18 LUFS is just a suggestion, so should you abide by it?
Ultimately, probably not. I’m a grumpy antagonist who rarely sees the practicality on issues like this. I want a standard, and I feel quite resolute in everything I’ve said up til now, but, at the same time, I know full well that the monumental effort required to get an entity like Apple to change how it builds its ecosystem is, at the very least, immense.
I think the best I can hope for is that platforms and devices will begin to build in programmatic solutions that perform normalization on audio which is out of range of its proprietary preference… that’s what Spotify does right now for music, and Tidal does it as well (and as it happens, so does Apple Music) but I’m unsure if this can be accomplished if we continue to deliver our audio through RSS because it would require any platform performing the normalization to download and modify our audio files (and store them locally, as is the case with music)… which most podcasters don’t like the idea of.
It’s another reason that I worry about the destruction of the podcasting medium over time. If these platforms (Spotify, Apple Podcasts, Google Podcasts) become walled gardens — which they seem to be more so every day — they may stop allowing delivery via RSS entirely and begin to demand the same thing they demand from musicians: that files be uploaded directly to them (or through an intermediary like Distrokid) so that they can “standardize them” as the see fit. Imagine having to do that as a podcaster? For multiple platforms? That would suck, and it’s not something every podcaster is equipped to understand how to do. That means fewer podcasters, and, if the bigger platforms go walled-garden, the bulk of podcasters will conform and if they don’t … they’ll be relegated to the increasingly proliferated platforms that have a smaller and smaller share of the listening market.
It’s like the Hollywood-ification of film…. but with podcasts.
Anyway, I’m trailing off into the brambles now. Hope this helped. Take care.