fbpx

Webinar – New Loudness Challenges: Broadcast, Online, Cinema

The media industry is changing quickly, and it can be hard to keep up with changes in standards, regulations, and practices surrounding loudness.
Get up to speed quickly! Watch the webinar replay to learn about:
– must-know specifications and requirements
– loudness in different playback environments,
– lesser-known specifications around online delivery
– how to implement loudness processing to optimize audio quality
– what you need to know and ensure compliance for global distribution

 

Cindy:

Welcome everybody. We’re so glad you’re here. And today we’re looking at new loudness challenges in broadcast, online, and cinema. So today you’re going to hear about loudness challenges and the history of loudness a little bit, and the challenges we’ve come up against in the current day. And we’ll look at the delivery process and also what’s happening with loudness for online and yeah, it’s going to be a great day. So hi MC.

MC:

Oh, hi Cindy. So yeah, thank you for hosting this. And this is our fourth webinar in the summer and we hope to do some more. This particular one, we’re going to focus on loudness. A number of our customers gave us feedback that, whilst it’s great that we do all these different types of audio processing, this is, I wouldn’t say necessarily back to fundamentals, but it is talking much more about loudness.

MC:

So I’m going to start by just talking about, just an introduction to loudness, what is loudness? How it came about and so on. And then we’ll move on to some other things. So basically a bit of history, when television started and we obviously had video and audio, we used to use a technique where we measured the peak audio level to determine what was the highest peak you could get. And there were a number of reasons for it. It was a very crude, but very effective way of measuring how loud the loudest bit would be. And from the analog transmission days, you didn’t want the level to go above a certain amount because we used frequency modulation for the audio, and if you deviated the frequency too much, it interfered with the video. So that’s why it was limited.

MC:

And that practice worked very well. The whole industry evolved and developed around it. Of course there were different standards around the world as to what peaks are. In the UK we had the BBC standard. There were several EU versions, there was a Nordic one, in America we had the volume unit meter, and so on. But in the 80s, people realized that, commercials people especially realized that whilst you weren’t supposed to exceed this peak, you could stay pretty close to it and the result was your commercial was a lot louder than the rest of the program, which I think everyone who watched television knows how irritating it is that people thought it was creative, but it was really irritating.

Cindy:

It’s true. It would be so tough. You’re watching and then just bam.

MC:

Bang. Yeah. And also there was no control if you switch channels, which wasn’t a problem in the old days, but when you have 200 channels, as you switch between channels, if the audio levels are not quite right, you get a sudden jump in audio and not a good experience. And the other thing was that when people made this constantly loud audio, it meant that we lost the dynamics in the audio. Everything was just loud and compressed. So the engineers around the world, the audio engineers in television world had been working on a loudness standard for a number of years. And actually it was about 10 years ago that they came up with the specification. That specs has been evolving since then and it’s 10 years ago, we got into the business.

MC:

So very briefly, program loudness or loudness, there are five measurements that we look at and care about. The first one is true peak, which is actually the highest value that the wave form can achieve. And that’s important because you don’t want the true peak to go into clipping in a system. And if you put compression in audio downstream, then you want to limit the amount of peaking because when you compress it, you will create clipping as a result of the compression process. So -3DBs is a good value for true peak.

MC:

Then the basis of loudness measurement comes out of a block of integration that you do on the audio. There’s a 400 millisecond block, and one 400 millisecond block forms the basis of momentary loud. So that tells you in a short time what the highest audio level is. If you take that average over three seconds, that gives you what we call short term loudness. And then if you take it for the duration of the loudness period, you get program loudness.

MC:

Now the problem loudness is slightly more complicated than just the total average, because in order to get a representative value, they wanted to take out the periods of silence, absolute silence. So if the audio goes to total silence, that measurement block is ignored. And there’s a few more complicated versions of it, which people can read about in our loudness fact book that you can get off our website.

Cindy:

That’s true-

MC:

But those are the… Sorry, go ahead.

Cindy:

Oh, I was going to let you clear your throat there or grab a water for a second and just say that, yes, we do have the loudness fact book for you. And there’s a link for you in the chat. But back to you MC on your five parameters. Yeah.

MC:

Yeah. So, the full parameters are true peak, momentary, short term, program loudness, and then we have loudness range. And loudness range is an attempt to describe the dynamic range in a program. Now you don’t want to look at the absolute values because there will always be silence and there will always be a true peak. So you take the measurements that you’ve done as a part of the integration. And then if you do a histogram and you only measure the top 95 percentile and the bottom 10 percentile, then that range of audio that you get is described as loudness range.

MC:

Now, to give you some examples, a really wide high dynamic range would be something like Mission Impossible. It’s a very noisy film, lots of bullets going on, guns, explosions, fast moving, and so on. And then you have others where that may be a lot less. So if you have a mostly dialogue film or a general TV series, like say Friends, there’s not a lot of dynamics in it. Friends, I’d describe as dialogue, a door slam, and a bit of piano music. So not a lot of dynamics.

MC:

But if you look at some of the more recent episodics as he CIS, or, I don’t watch those things, but they have a lot more dynamics and with 5.1 and so on. So loudness range is a measure of that dynamics. And why is that important? It’s important because it helps you understand how that audio will sound in different listening environments. And so in a cinema where you’re seeing them for 90 minutes or something, then you can enjoy that wide dynamic and live with it. If Mission Impossible was in your living room for eight hours a day, that would be kind of painful and not so enjoyable. And then as we’ll talk about later, if you have online delivery where your listening environment is noisier than normal, then the audio has to be a little bit louder and a little bit less dynamic so that you can actually have a good listening experience. So, that in a nutshell is program loudness.

Cindy:

Got it. So the peak measurement system of old has really changed. And so if I got this right, the parameters that now matter are true peak and program loudness, short term loudness, momentary loudness, and then you were talking about LRA, the loudness range. So changed from days of old.

MC:

That’s correct. Yeah. Now why have so many measurements? So if we talk now about the production process, when you’re doing a production process, the audio engineer has what we call a sound budget. So the budget goes along line, if I’m mixing a 90 minute movie, then I want to make sure that I am at a program loudness level, I manage the LRA based on the environment I’m trying to deliver to, and then I also want to know what my short term and my momentary peaks are just, so that you get an idea. Now, this is a purely mental process. You can’t see it across the whole program, but you use these tools to plan certain things. If you want to make an impact, you may look more at the momentary loudness. If you want to have a short burst of something, you will look at the short term loudness and so on.

MC:

So in production, these tools are used, but they’re used in your edit suite. So there are meters inside the edit suite. So as you’re mixing, you can see what’s going on. Now, program loudness itself is the average of the whole program. So, it’s not very easy to monitor it. So typically, the loudness meter, as it’s called, would be giving a running program loudness. So as you do the mix, you know what the average is up to that point. You don’t know it until you get to the end. Now you may say, how the hell do you mix to it? Because if it’s running along? And therein is a problem, but actually most well-trained audio mixers have an inherent sense of balance. And the reason why they came up with -23 in Europe and -24 in you in the US is because that’s what you naturally mixed with, you’re trained for it. Now, you may not hit the number exactly, you’ll be slightly off, but you’ll get pretty close to that.

MC:

So when you finished your production, you will have used these five meters to create the sound that you want and then you want to deliver it. Now, when we go to delivery, there’s only three parameters that really matter. And I will change my story in a little while. So the production side, as I said, you use all five in the delivery generally speaking, if you’re doing a broadcast delivery, program loudness and true peak are the bits that matter. The LRA would have been taken care of if you’re doing an episodic or a full broadcast, because that will be within the guideline given to the production company. So people tend not to… The program loudness and true peak are the ones that are most measured when you’re making delivery for broadcast.

MC:

Now, the thing that matters for broadcast then is, if you’re making content and you want to monetize it all over the world, you have to deliver to all over the world. And sadly, most broadcasts have a slightly different audio spec to the recommended. Now, the recommended is a guideline, the EBU is a guideline, in the US, the ATSCA defines it, is a fairly tight spec, but if you go around the world, they may say, “We want the true peak to be -3 or -1 or -2. We want the program loudness to be -23 in Europe, -24 in America,” but they have different tolerances. So in France, it’s got to be bang on -23, you’re not allowed any deviation.

MC:

So that means on you as a content delivery house, you have to measure these things as they come in to you and as you’re repurposing them, you have to ensure you meet this spec. If you have a theatrical mix, coming from a Hollywood studio for example, like your Mission Impossible, that is an LRA of 30 or there about. And most broadcasters will give you a guideline that the LRA mustn’t be greater than 16 or 18. So you need to reduce it for that.

MC:

And then for social media, there are guidelines, rather than standards, but the big guideline is it’s got to be loud enough. So typically people are talking about -18, -16, as the program loudness. Now, ironically, a content creator may do this, but commercials may be mixed for -23. It may be the one instance where the commercial doesn’t sound as loud as the program, if that spec is met.

MC:

So it’s a lot of fun and games in terms of what you have to do for delivery. The theatrical to broadcast makes it a very, very challenging process because where you have lots of audio dynamics, it’s very hard to predict whether the sounds going to go from quiet to loud to quiet to loud, or consistently loud. So the processing needs to be very, very careful, carefully done. Well, in fact, the loudness processing in general needs to be carefully done because we want to preserve what the audio dubbing mixer wanted to transmit and show to people.

Cindy:

Okay. So in the production process, what it sounds like then, and the delivery process, I guess, really the challenge you’re talking about is you don’t have any idea of what parameters were used when the production took place, right? And then the mixing process might have been designed for a different deliverable because you could be delivering to a German broadcaster, a US broadcaster, Netflix, and you don’t know how that all fits into it. And so that’s definitely the problem. Do you have more to talk about around online? Are you just going to jump right into the solution around that?

MC:

Just a couple of things for the challenges. So we could, as I said, in the production, you use the meters to design the sound and by and large, you’ll be close to the spec for one broadcaster, your primary broadcaster, somebody who employs you for example, but it needs a variation. The other one is you may have content where loudness wasn’t a factor at all, an archive, it was content done 15, 20 years ago, before the loudness pack came out. That’s an issue. And the other one that’s a big issue also is the theatrical content. The theatrical content is built for the cinema because that’s where you make the billion dollars. The broadcast is a long tail.

MC:

So we have spoken to a number of studios and they say, “We only do one mix. We don’t always provide what they call a near field mix.” Or if they provide it, you may not have it. And then the other things that you clearly have, which are not subject of this, but I’ll just mention, is you may have a stereo only mix, but you need to deliver 5.1 and you may have a 5.1, and you need to deliver stereo. Now why this is important is, if you do an upmix or a downmix, you will change the loudness. So the loudness has to be… You have to ensure that the loudness meets the spec if a downmix or an upmix process has occurred.

Cindy:

Got it, got it. Okay. So it means that people… What I’m really hearing is that people who are already using our solutions right now in broadcast and for other types of delivery are now starting to talk to you about online delivery and that’s theatrical mix as well. So a lot of choices there.

MC:

Yeah. So a couple of things I would say is, whilst I’ve described the delivery challenge, if you like, what we do is we have designed a set of algorithms that allow us to meet any deliverable spec and there are plenty of them. So the loudness specs for broadcast is relatively straightforward, true peak and program loudness at slightly different variations. Those we handle by the program loudness is a very simple thing, it’s the average audio level in a program. So if the audio is slightly low, you apply gain. If it’s slightly high, you apply attenuation. Now, if you’re applying gain, you may screw up the true peaks. So we tend to measure all the peaks in the file, put a gate around it, and then we will locally attenuate the peaks and do a little mix in, mix out exactly how you do it in the manual process.

MC:

So the big question is, so why not do it manually? And the reason is for one piece of content, you may need to make 30, 40 deliverables if you’re a studio delivering to multiple clients. And so that becomes almost impossible to do manually in a cost effective manner. So by creating the algorithms that we have, we then have created an automation layer on top of it and that automation layer allows you to do this thing unsupervised and we will deliver to any spec and it’s guaranteed.

MC:

And so the tool, we have two tools, there’s a desktop tool for low volume stuff, where you basically say, every deliverable I have, I create a profile. So I say that I’m delivering to a German broadcaster so it’s -23 program loudness, the true peak is -3, the LRA has to be 16. And we don’t worry about the short term and the momentary for that, that spec defines that deliverable and we will measure the audio to say, “Does it meet that spec?” If it doesn’t, we’ll make the necessary adjustment. Now, if you do it in the desktop, it literally is pick a file, pick a profile, press correct. And it does it. Generates a nice report to tell you what it’s done so that you’re accountable, “It was okay leaving me.” And then if you need to make changes to it, you can edit and rename that spec. Now you can have as many of these profiles as you like. Now, what this means is that the operator doesn’t have to be skilled in audio to do this deliverable.

MC:

The other form of course is, if you’re doing more than a few files a day, you don’t want to employ lots of people to do this, you want to automate the process. So now, you could have our product engine, which has a number of automation strategies though. You could have watch folders, very simple. If you have 16 clients or 16 deliverables, you have 16 watch folders, throw the files in the relevant watch folder and out it pops corrected. If you have a Telestream Vantage or an Aspera Orchestrator, you can have a plugin to that. So whilst you’re preparing your video deliverable in your Vantage, you could have the audio taken care of by engine, and it will do the loudness correction for that.

MC:

Obviously that gives you more than 16 profiles if you need it, because it’s as many as you like. And you could also have this driven by a MAM system. So the MAM system can do this. We publish a rest API, so you can have an API that your in-house automation system can drive. So there are many possibilities. So what we’re trying to do is say that we can do the deliverable for any standard and we can also do a deliverable in an automated fashion through any automation system that you have. But I haven’t spoken about online.

Cindy:

That’s okay. Before you speak about online, I just want to see if I got your key points there. I feel like what you’re saying is with all the different processes that need to be done, they can all be done in an automated way and that means that people can go off and be creative and do what they do best. And then automation makes it repeatable. And with the watch folders, that really makes it super easy and straightforward. And of course, as you mentioned, MAMs and Aspera and Telestream, if you want that integration and the rest API as well. So does that sound like what you were talking about around the automation part?

MC:

Mm-hmm (affirmative). Yeah. So, easy to use, flexible, scalable. That’s the idea. Now, the online, as I said, it is basically… Let me talk a little bit about online. So as I said, you’re in an environment, if you’re on a plane and you’ve downloaded a movie on your iPad and you want to watch it, the ambient environment is noisy, it could be on a plane, on a train. Why you’d want to watch Mission Impossible in that environment, I don’t know, but hey, people do. I watch it on planes and I’m quite often annoyed about the audio. So people are now coming to us and saying, “Hey, we’ve got a… We do the broadcast because we have to meet a specification. We may get fined. We may get our content rejected. Now we’ve got these online platforms and we need to do that because we want our audio to be heard.”

MC:

And so we have two approaches to this, one, we say, “Tell us the spec and we can deliver it.” The other one is where they say, “Hey, we are not sure what we really want.” And so we do this very frequently, we’ll just interact with the supplier, give them some files, give them some ideas. And a classic one here is not just increasing the volume, it’s really increasing the volume and managing the compression or the dynamics, the LRA. And so the LRA, which was originally we used for taking a theatrical mix and making it broadcast ready. We can now use that to take it even further down. And now what we’re doing is we’re restricting the dynamic range and we’re making it louder, so that if the environment isn’t friendly because it’s noisy, you can still hear the dialogue. And so that’s really where we are with online.

MC:

Now, nothing changes as far as the product itself concerned, you don’t need to do anything different, you just change the settings. And probably the one I have to mention is Netflix have their own ideas about how they want their audio to be presented. So they’ve taken a variation of loudness, which actually is an older version, the loudness spec is evolving. And it is something that threw a lot of people, because it was a technical change.

MC:

Now, when we had the peak program meter in the UK, we had BBC PPM6. So you just didn’t exceed PPM6 and you were in good shape. Here we have four or five parameters and I haven’t mentioned some of the other subtleties in it. And so when Netflix came up with its chain, it was something that non-audio people found challenging. But now remember, at delivery, you don’t have a lot of audio people. You have machine operators, I guess and so on. So those people are sitting there saying, “How do we do this?” So what I was saying to you is we have a service as part of what we do is come and talk to us about your audio needs and we’ll either set it up for you or help you set it up or interact with you to say, “This didn’t work out so well.” So we’ll say, “Okay, let’s try this variation and that variation.”

Cindy:

And does that work for any size company? Do I need to be a big company to do that? Or do you work with smaller companies as well?

MC:

No. We are a small company. We deal a lot with the larger companies and we deal a lot with the smaller companies. So the desktop tool will do everything that the main engine does, it just doesn’t scale. So if you need, we’ve had customers… We had a customer who basically was their first entry into movie mixing. They were big noise in commercials mixing and they said, “We want to mix movies.” So they know their audio, they know things, but they have a deliverable problem. They said, “We have a 7.1 mix, we want to deliver 5.1 in 25 frames. Then for US delivery,” this was a UK company, “we want to time compress the audio and pitch shift it so that it sounds right. And we want 5.1, we want stereo.” And so the guy said, “So, what would you do?” He said, “Well, we’d sit in the Protools suite and knock these out one at a time.” So I said, “Well, what we can do is create a workflow, which will allow you to do this in a single pass.”

MC:

So for them, what they’re really saying is, what will take me a day to do, I can do in an hour. So they look at a saving in a very different manner to somebody who says, “I need 10,000 hours a month for processing.”

Cindy:

Nice. Nice.

MC:

And it’s really that scalability. The algorithm came first, then came the nuts and bolts to sit there and say, “How do we make scalable?” And then came the flexibility of what else can you do with the audio? Because I want to do more than loudness.

Cindy:

Got it, got it. We are going to go to questions in minute because we do have a couple questions, but before we do that, I just wanted to recap. We looked at loudness and the history of loudness, and then really what’s changed and how we’ve moved into those parameters that you talked about. And then talked about delivery challenges and online challenges, which you just hit on. And then of course, as you’re saying, the solution and the automation around it, really no matter what size your facility is. So yeah, that’s where we’re at. So we’d like to take your questions. What questions do you guys have? And the first one, oh yeah, so I’ve had a couple people ask about examples. And so MC, if you could give some examples, that would be great. What customers are doing this now?

MC:

So I already mentioned a little bit about that little films audio house that was going from commercials mixing to film mixing. A lot of post companies do it when they deliver commercials. So we have Smoke and Mirrors in the UK, the Mail. Smoke and Mirrors have offices all over the world as do the Mail. So they use them. We have a little post house in the US, Leo Ticheli in Midwest… I’ve forgotten now.

Cindy:

I think they’re in the Midwest, I’m just trying to remember where Leo Ticheli is as well. Yeah.

MC:

Yeah. Yeah. And they have a couple of edit suites that they use it for. And then on the other scale, we have Viacom in the UK and in the US were very large installations, because all their content needs to be loudness processed. And in the case of Viacom, they are archiving all their content in 24P, but some of their localizations are done at 25. So they’re bouncing between 24 and 25, a multiple number of times, and still doing that.

MC:

We have a number of companies, play out centers in Australia. They’re a very large one Channel Nine and Channel Seven play out operations. They have two engines and they’re doing all their loudness with us, but they’re also doing all their other audio processing, Adobe encoding, upmixing, downmixing, track mapping. As I said, once we get the audio onto the video file, there’s many things we can do with it. And more and more people are coming to us. Comedy Central actually were the ones who came to us with the online, they said they wanted it, primary playout was Alexa for them.

Cindy:

Yeah. I love Comedy Channel. I’m all in.

MC:

Yeah. So that was an example of interactive stuff, because the guideline wasn’t clear from the platform, they came to us and said, “Hey, what would work?” And we interacted with them. And the way the interaction typically works is we have a chat about it, we make some suggestions, they try it out on their own content. If there are issues with it, then they come back to us and they may send us the file over, we’ll do some analysis, we’ll readjust the settings and give them back.

MC:

Now, the reason for this is it is actually a very exact science, but as with audio, it’s extremely subjective. So people have ideas about how they want to present their audio and what’s important and what isn’t important. So, we also have in India, ZTV and all those listening, and I worked together for a couple of years for Z to be able to deliver globally to the world. And we had some very interesting interactions on how we wanted to do this. So, yeah, it really is something that we find pretty much every broadcaster has a need for and use for. And the more with the present look down and things, there’s a huge demand for file based content because live based content is nonexistent or very low in volume. So the people are digging out archives for presentation. So it’s a big, big requirement.

Cindy:

That’s true. You were telling me the other day about the new resurgence around archives. And I liked some of those examples you gave from Comedy Central to Leo Ticheli to Z. You’ve got a nice range, which ties into the next question, which somebody posted and you kind of touched on this. The examples you gave seem like big companies, but I have a small facility. How can this help me in a small facility?

MC:

So yeah, as a small facility, as I said, Leo Ticheli do commercials, so they don’t do many and they got a rejection from one of the people that they were delivering to. The rejection was by way of a report from us, because these guys had our measurement tool. So they talk to us and they bought a couple of desktop products called F. Now we have customers who say, “Well, all I want to do is make sure I comply and not worry about whether I’m going to make adjustments or not.” And so we have a product that starts at $1,000. And it does the measurement and it gives you a report telling you what’s wrong or not wrong, if it’s good. And you can start at that point. So we have a solution really for the very small customers as well.

Cindy:

Nice. And whoever you are, whatever size facility you have, we have a trial for you that you can download it and try it yourself and see what it does for you. So, there’s the link in-

MC:

Yes. If you go on our website and click the try button, which is on every page of our website really, it will take you to a place where you’ll give us your details and we’ll give you a 10 day evaluation license. And you will get some helpful emails about loudness. And at that point, this is our standard way of working, call us, email us with questions. We are happy to help people get going with this.

Cindy:

Nice. Well MC, it’s been a wonderful discussion on this topic and I think we are wrapping up here. Any closing words?

MC:

Well, no. All I can say is, thank you to everyone who attended. We will do some more of these. And actually, if you want to do a download, if you want the fact book, please visit our website. If you have specific questions that you want us to answer, you can email me, MC@emotion-systems.com, or support and you’ll get our experts helping you with questions on loudness.

Cindy:

Perfect. All right. Well, thank you everybody and have a beautiful day. Thank you MC. See you later.

 

By continuing to use the site, you agree to the use of cookies. To change your mind please go to our Cookie Policy Page. Cookie Policy

The cookie settings on this website are set to "allow cookies" to give you the best browsing experience possible. If you continue to use this website without changing your cookie settings or you click "Accept" below then you are consenting to this.

Close