A Deeper Dive into Dolby Volume

Dolby Podcast Episode 49 - October 9, 2008

Find out how Dolby Volume is different from other loudness control tools for home audio in this interview with Dolby’s Alan Seefeldt, one of the inventors of Dolby Volume.

Plus – get answers to 3 listener questions:

1. speaker placement for wall-mounted speakers
2. how to get Dolby Digital to play via a set-top box, and
3. how to play recent content with a Dolby Pro Logic system.

 

[intro music]

Announcer: Streaming to you from our headquarters in San Francisco, this is Dolbycast, the insiders guide to entertainment technologies from the experts at Dolby Laboratories. We're here to give you the straight talk and news on everything you need to know about technologies that excite your eyes and ears.

Craig Eggers: Welcome back to Dolbycast! We have a very interesting show planned for you today. Mr. Alan Seefeldt, one of the inventors of Dolby Volume will actually be joining us later on in the show.

Before we get to the show, I thought we would try to answer some of our listener questions.

You have two places where you can contact us. First of all, you can call us on the phone, 1‑888‑6-DOLBY-C. I'll repeat that, 1‑888‑6-DOLBY-C, or you can contact us via email at dolbycast@dolby.com. We love your questions. We read all your questions, and we try to answer as many of those questions as possible.

So, let's just jump into a couple questions. Received a question from a gentleman by the name of Rajeev, who says, "Went through your chapter on the placement of speakers, and it was very informative, " he went to Dolby.com. We've got a great site on Dolby.com talking about speaker placement.

But Rajeev continues, he says, "One of the points I read was that the left and right speakers are better at ear level. Does the same apply to the surround speakers as well? When the surround speakers are wall‑mounted at a height above you, or to the side of approximately 6.5 feet, can they be tilted?"

Well, basically a couple of answers to your questions, Rajeev. First of all position your rear surround speakers behind you at approximately 110 degrees. That's the best location in a 5.1 system. What we generally advise for height of the rear surrounds is ear level or slightly above.

What's really interesting is you mentioned they might be wall‑mounted. I don't know if you're talking about building them into a wall or a wall‑mounted type speaker, but in both applications a lot of speakers that are either built into the wall or can be mounted on the wall, the speakers can be tilted. I definitely would recommend that you consider tilting the speakers. Listen to it, try it out, see what works best in your home theater environment, what gives you the most satisfying experience, but definitely experiment with that.

So thank you very, very much for those questions, Rajeev. We look forward to hearing about your experiences in the future.

We also had a question from a listener, and I don't have the listener's name in front of me. The question involved a set‑top box, a cable box. The listener was asking us, "I've set up my cable box. I've connected it to my home theater system. It's connected by the digital audio output, but all I'm getting is 2‑channel PCM. My home theater system's telling me, my AV receiver's telling me, I'm receiving a 2‑channel PCM signal. How do I get Dolby Digital out of my set‑top box? I had the same experience with my Comcast box."

What you need to do, is you literally have to go into the setup menu of your cable box, and tell it that you want to receive Dolby Digital 5.1. A lot of cable boxes just default to standard 2‑channel PCM. Go into the menu of your cable box, set the audio output for Dolby Digital 5.1, and if you have any problems doing that, contact your cable provider. I'm sure they will be glad to give you some great insights and information on how to accomplish that.

This is one of those problems that actually was a real life experience for me. I am glad that we were able to solve that for you.

Had another question, from a listener by the name of Ian. Ian writes, "I have an amplifier with the original 4‑channel Pro Logic system, which I love. It still works well and therefore would like to keep going as long as possible." His question, "Am I right in thinking that even recently released material with a Dolby soundtrack, still provides the necessary information to enable my amp to reproduce Dolby Pro Logic output?"

The answer to that question is, "Yes." Ian, you're going to be able to play 5.1 content. You're going to literally though, because your amp is Pro Logic, you're going to connect your player via analog left and right into your AV receiver or your amplifier, if you will. Your amplifier is going to decode that stereo signal and give you a Dolby Pro Logic listening experience. So, your surrounds are not going to be discreet, like you would have if you had a Dolby Digital system.

Because you have Pro Logic, you'll have a left, center, right in your front. The signal going to your 2 rear speakers will be a mono signal, because that's the way Pro Logic was. So if you want to keep your amp going, if you want to enjoy this new content, if you have a Blu‑ray player or anything with a 5.1 channel output, what you're going to do is you're going to select the 2‑channel output. You're going to connect that device to your amplifier via analog left and right, and your amplifier is going to do Pro Logic decoding of the 2‑channel signal, give you a left, center, right and a mono surround listening experience.

So yep, keep that old amp working, but there's a lot of great new products out there that give you an even better, more compelling experience. So think about that in the future. Thank you for the question, Ian.

We're going to take a quick break. We will be back in just a couple of moments here, with our very, very special guest, Mr. Alan Seefeldt of Dolby Laboratories. Standby.

[music]

Announcer: Hey listeners! Got a question for Dolbycast? Contact us at dolbycast@dolby.com or our new toll free number 888‑6-DOLBY-C, "C" as in Dolbycast.

Craig Eggers: We are back at Dolbycast, and hey, we're glad you are with us! During our last Dolbycast presentation we were speaking with Mr. Ara Derderian of the HT Guys. The topic of that Dolbycast was [LINK: "New Innovations, New Technologies We Saw at CEDIA."] Well as you know at CEDIA, we introduced some of the first Dolby Volume equipped AV receivers, products that included AV receivers from Harman, Arcam, AudioControl.

As a result as some of our discussions about Dolby Volume, we got a lot of listener mail. So we thought we'd do a dedicated show about Dolby Volume, and talk to one of the inventors, if you will, of Dolby Volume. Alan Seefeldt is a Senior Staff Engineer here at Dolby Laboratories; he works in our research department.

Alan, you are responsible for the development of Dolby Volume. Welcome to Dolbycast.

Guest.

Alan Seefeldt: Thanks Craig, happy to be here.

Craig: Hey, we're glad to have you here. So, Alan, tell our listeners what problems does Dolby Volume solve? How do we make the entertainment experience better?

Alan: Right. So Dolby Volume really solves 2 different problems for audio listening in the home, and the first is one that I think a lot of people can connect with. You've got audio coming from all sorts of different sources, or you're watching television and you're changing between different channels, and the audio levels are all over the map. You're constantly riding the volume control, and it's very annoying.

So, one of the main aspects of Dolby Volume is to level out all those differences in the audio. The idea is you turn on Dolby Volume, you set the volume control once to a comfortable level, and then you just forget about it. Dolby Volume takes care of the rest for you behind the scenes in an extremely high quality manner.

Craig: Dolby Volume obviously is a great solution for display devices- TV sets - but when you consider that the AV receiver today is becoming the digital switching system for the entire home, that you've got content coming in from CD, music files, Internet, radio, broadcast, Blu‑ray, all of a sudden a technology like Dolby Volume that gives you consistent level listening experience becomes very compelling.

Alan: Yeah, it's great. I mean, the quality of the content coming in from all those different sources just varies so widely and there are no standards for setting a level, so you get levels that are all over the map when you've got that much content coming through your AVR.

Craig: So with Dolby Volume, you basically choose a preferred listening level, and what happens?

Alan: You choose a preferred listening level, and then what Dolby Volume does, the algorithm behind the scenes, it's in real‑time monitoring the audio that's streaming through, listening to the audio in a way that simulates how the ear actually perceives the sound, and then in real‑time is adjusting the loudness to keep everything consistent, so that you don't have to reach for the volume control.

Craig: I understand that having a good understanding of how humans hear - how we perceive audio - is a very critical element about Dolby Volume. I want to get into that a little bit later, but there's another element of Dolby Volume also besides just the leveling, right?

Alan: Right. There's another feature that we call the "modeler" and the idea here is that we're taking into account the way that we hear things differently as the volume changes. In particular, when you start to turn the volume down on a mix, your ear perceives the sound differently. One of the things that happens is that we are less sensitive to bass and treble frequencies as the volume turns down.

In addition to that, soft parts of a mix will just start to disappear more quickly than the louder parts. They basically fall below the threshold of hearing. So what that means is that as you turn the volume down, all the details and subtleties of the mix start to disappear.

That is particularly important for surround sound mixes where you might have a lot of details and ambience in the surround channel.

Craig Eggers: Sure.

Alan: So what happens is that Dolby Volume again monitors the audio in real time and it figures out how the ear is going to hear those things differently as you turn the volume down, and it compensates for all those effects in real time and maintains what we call this "reference level experience" as you turn the volume down. So you get the full impact of the mix even though you are listening to it at a much lower level.

Craig: I think that is one of the things that has always confounded me is when you are listening late at night, you start to turn the volume down, and what you end up basically with is the dialogue coming out the center channel.

Alan: Right. Often times in a surround sound mix it is the dialogue in the center that is the loudest part of the mix and all those little details that are in the other channels, you just lose those as you turn the volume down.

And again, as I said, Dolby Volume adapts its processing to maintain the audibility of all those other components, but it does it in a smart way. If all those channels are loud for example, it turns them down by the same amount so you get the right balance. But if there are other channels that happen to be softer later on in the mix, it will boost those up above the threshold of hearing so you can hear everything.

Craig: So tell me. I would like to understand why Dolby developed Dolby Volume. What did you see in the past that were problems that weren't being solved with regards to addressing those inconsistencies in volume leveling? What past attempts have been made at doing this, and what did they get right and what did they get wrong?

Alan: These problems that we are addressing have been recognized for quite some time, but the problem is that dealing with them in a really high quality way requires some pretty sophisticated technology.

So there have been attempts in the past to address these issues with simpler technologies. So for example, with level variations, there are compressor limiter solutions out there on the marketplace. Many of them are quite simplistic in the way that they work.

You can kind of think of it as an automatic volume control, just kind of turning up and down as a function of an incoming audio signal. But what often times happens with those systems is that you can hear them working and they start introducing annoying artifacts into the audio. We often refer to these as kind of "pumping and breathing artifacts" where you can hear the system kind of riding the gain, and you can hear parts of the mix changing in real time as it is trying to adjust the audio.

So to do that kind of processing really well, it requires a more sophisticated understanding of the audio. That is why we have this model of how humans perceive sound built into Dolby Volume, so that we can really understand how a human is hearing the audio and then at the same time understand when it is appropriate to adjust the level of the audio to maintain a consistent loudness.

Craig: So these pumping and breathing artifacts, they are really related to the fact that you have a compressor that is pushing down on the peaks, if you will, in a sound mix, then trying to elevate the elements of the mix that might be much lower and below the reference level.

Alan: Exactly. If you can think of it in the simplest sense, it is looking at the level of the audio signal, and if the system says "Oh, it is too loud," it is going to automatically turn it down, and if it is too quiet it is going to start turning it up.

But there are some basic cases in audio where you can really get into problems just applying that simple rule. For example, let's say if you have got a piano chord in your music mix that is just decaying naturally...

Craig: And it is supposed to decay, right?

Alan: It is supposed to decay naturally. A simple system, like a simple compressor, will say "Oh, the audio is getting softer. I need to boost it up" and you are going to get this really unnatural rise in the decay of that piano note.

Craig: We often talk about that. If a piano chord is decaying and then somebody hits a very percussive slap on a snare drum or a bass, what would happen in a compander-type technology?

Alan: Well in that case it will have boosted up the decay, so it has got a big gain riding on the system. Then all of a sudden this loud sound comes in and the gain gets sucked back down and you will hear all the ambience of the signal get sucked down with that hit of the snare drum.

Craig: So literally your average audio level, if you will, is going up and down and up and down.

Alan: The problem is that you can hear that as unnatural modulation of the audio in one of these really simple systems.

Craig: Alright. So the big question; the $99 question if you will: How does Dolby Volume solve this?

Alan: Right. So there are a number of pieces of technology inside of Dolby Volume that address these issues. Again, one of the main things is that we are using this model of how humans perceive loudness to first of all, monitor the loudness and then adjust it.

So we have got a better understanding of how truly loud something is or if it is not loud at all. So we are able to track that loudness better.

That model is fairly sophisticated in that we are tracking the non‑linear growth of loudness in a multitude of different frequency bands. Because we have all these frequency bands, we can actually adjust the loudness somewhat independently in these different bands.

That allows us to make some adjustments where you don't get these kind of...for example, the example you had where a snare hit might be unnaturally modifying other parts of the spectrum. We can kind of isolate our changes to certain parts of the spectrum. So that is one really important aspect of the algorithm.

Another really important aspect is something we call "auditory scene analysis", where we actually break up the audio signal across time into elements that a human would actually pick out as being important.

So in the case of that piano chord, we kind of pick out that one chord as a single audio element. The system recognizes that it is supposed to be decaying naturally, and in that case, it doesn't allow Dolby Volume to unnaturally boost the decay.

So it keeps the system very well behaved, because we are picking out these perceptually important cues.

Craig: So it sounds like the system is very intelligent then.

Alan: It is. We have got years of research into how humans perceive sound built into this system. Again, the model that we are using is quite sophisticated.

Craig: You were talking about multi‑band signal processing. I know in a lot of our marketing literature, we talk about loudness domain signal processing. Is that what you were referring to?

Alan: Right. I mentioned that we have this model that models how we perceive loudness. Basically Dolby Volume analyzes the signal and transforms it into a space that is kind of a representation of how humans hear loudness. This is something we call loudness domain signal processing.

Craig: And this is something that is new.

Alan: It is new. Yes. We are using what we call psychoacoustic model of loudness perception. Psychoacoustics just means the science of how humans perceive sound. What we do is we break the audio signal up into a multitude of different frequency bands, we can do either 20 or 40 bands, and then we are looking at perceived loudness in each of those different frequency bands, and we can adjust those independently if we need to.

Craig: Channel independently, too, correct?

Alan: And across channels as well. So for example, if you have got some loud sound that comes crashing into the center channel, we can pull that down without necessarily pulling down the surround channels.

Craig: So it sounds like Dolby Volume, in terms of providing a level consistent listening experience, is A: highly intelligent, and B: revolutionary - in the fact that we are really processing the signal and looking at the signal in ways that just has never been done before.

Alan: Yes. I think we have a really unique product.

Craig: So Alan, after the break I want to talk to you about this thing called Dolby Volume modeler and how we compensate for the fact that as we turn the volume down on our home theater we tend to lose everything.

We will be right back with you at Dolbycast right after this break.

[music]

Braden Russell: Hi, I am Braden Russell.

Ara Derderian: And I am Ara Derderian. Together we are the HT Guys.

Braden: You are listening to the Dolbycast. If you have got a question about audio, where do you send it?

Ara: To dolbycast@dolby.com.

Craig: And we are back at Dolbycast with our very special guest Mr. Alan Seefeldt, Senior Staff Engineer in the Research Department, Dolby Laboratories.

Alan, before we talk about the whole modeler aspect of Dolby Volume, I want to talk about how it would be implemented or integrated into a consumer product - a product like an AVR, a display device, or even a set-top box.

Do you see Dolby Volume as being a very sophisticated, complicated process for the customer to figure out and to have a user interface experience with or as a simple on/off switch? How do you see Dolby Volume being implemented in products in the future?

Alan: Yes, so the implementation can be extremely simple, I know we were talking about a lot of details of the technology. Behind the scenes it's doing a lot of sophisticated things, but for the user, it's going to be dead simple. In the most basic implementation, there's just an on/off switch for Dolby Volume, which is just "Do I want leveling or not?". So if you want leveling, you turn it on and at that point you just use your volume control as normal, you turn it up and down to achieve the desired listening level.

Craig: So, you would basically set the gain level or the volume level on your TV, your home theatre to your desired setting, turn Dolby Volume on. Then all the content that's above that threshold would obviously be processed by Dolby Volume to give you consistent level listening experience and the content that would be under your desired setting would be elevated, correct?

Alan: That's correct, but again for the consumer, they just set the volume control to their desired level and Dolby Volume will keep it there.

Craig: Apologies to Ron Popeil, "set it, and forget it," right?

Alan: Yes.

Craig: [laughs] So, let's talk about what happens in home theatre late at night. I turn down my volume in my home theatre and basically, if I'm watching late at night, those incredible soundtracks, that you know Hollywood spends millions of dollars on creating, is lost.

Alan: Yes, that's right. Again, that has to do with the way that your ear starts to hear things differently at these lower levels. Again, there has been lots of studies to look at those aspects of our hearing and we know that we're just less sensitive to softer sounds as the volume gets turned down.

Craig: Alan, we often say that if we're going to a theatre or if you're on a mixing stage, and you're listening to a movie being created or played back in a theatre cinema, it's loud!

Alan: That's right.

Craig: And it's loud for a reason isn't it?

Alan: Well, there's a defined kind of reference level for cinema mixing and the idea for these mixers is to get across, you know to use all the dynamics that they have. They want really loud things, really soft things and to really give this great experience inside of the cinema. And there, you can afford to have these really wide dynamics and really loud sound, because you have this dedicated theatre.

Craig: So in a louder environment, obviously we're gonna experience the full essence of the soundtrack. But once we start to get the content at home, once we start bringing down the volume down to measurable levels that we would want in the home theatre, things start to happen.

Alan: Yes, that's right, so, your impression of the mix starts to change as you turn the volume down and if your perception of the relative balance of the different frequencies changes, so again for example, you start to be less sensitive to the bass and the treble frequencies and things start to sound very midrange heavy and again, in addition to that, all the softer elements, the little details of the mix start to disappear and that happens across all the different frequencies.

Craig: So I guess in previous attempts to solve this, I mean, I've got stereo receivers at home that have a loudness button.

Alan: Right, so there have been attempts, some pretty simple ones, to address this problem in the past and the most notable one is the so called "loudness" button that you've seen...

Craig: What does that do?

Alan: Alright, so most loudness buttons on receivers, basically, they just apply a fixed bass and treble boost that starts to change as you turn the volume down, so basically as you turn the volume down, it starts to apply more of this bass and treble boost. And that's trying to compensate for our decreased sensitivity to those frequencies, but it's doing so in a fairly simplistic and inaccurate manner. Really to do that kind of processing correctly, you need to take into account what's happening in the audio signal.

Craig: OK.

Alan: You actually need to monitor the audio signal in real‑time and look at what's happening in those different frequencies. Because what happens is that, if you have less bass in the audio signal, you actually need to turn it up more, turn the bass up more, than you do for a signal that has more bass. So, really the processing that you're doing needs to be a function of the content that's in the audio signal. And that...

Craig: That's what Dolby Volume does.

Alan: That's what Dolby Volume is doing. It's monitoring the perceived loudness of the audio in a multitude of frequency bands, like I said - either 20 or 40 frequency bands - and it's adjusting the gain in each one of those bands as you turn the volume down, based on the content in those different frequency bands. And this allows us to maintain the relative perceived levels between all those different bands as you adjust the volume. So it's not just applying a simple bass and treble boost, it's adjusting all these different frequency bands, depending on the content in those bands.

Craig: Now I know the answer to this question, but I'm going to ask you. OK, so I've turned the volume down on my home theatre, I lose my surrounds, why don't I just turn my surrounds up?

Alan: Right. That's a good question. So this notion of just turning the surrounds up, that's really applicable when you have soft surrounds in the mix and many times the surround signals are quite a bit softer than the front sound stage, so there's this kind of conception that "Oh, I lose my surrounds when I turn the volume down". So in that sense, yes, you can boost your surrounds, but really the correct thing to do is to adjust the level of the surrounds based on the content that's in the surrounds.

Craig: Exactly.

Alan: So, if you have a mix where there's really loud sounds in the surrounds and there are plenty of portions of action movies where you have all the channels lit up by the same amount. In those cases, you want to turn down all the different channels by the same amount to maintain the proper spatial balance between all the channels. It's only when the surrounds are really soft, that they need to be turned up relative to something that's loud in the front and Dolby Volume does that, it dynamically adapts to the content of the audio signal. So, if the surrounds are soft and you turn the volume down, Dolby Volume will automatically turn up the surrounds by the proper amount relative to something that is louder in the front. But if the surrounds are really loud, it'll turn them down by the same amount as the fronts.

Craig: That's interesting, because I believe there have been solutions that have been tried in the past in which you basically do elevate the surround sounds to compensate for what you perceptually lose. But when you think about a movie track these days, musical swells in the past used to be dedicated to the left and right channel. But a lot of times musical swells in a mix now today, are in the left surround, the right surround, or left front and the right front. So if you're artificially elevating the gain levels - the volume levels - of the surround, you're literally changing the integrity of the entire mix.

Alan: That's right. We've seen plenty of examples where you got loud signals in the all the different channels and if you take the simpler approach of automatically boosting the surrounds, you can really kind of destroy the mix, where you get this very surround heavy sounding mix as you turn the volume down.

Craig: So not only do we have a technology that gives you consistent level listening experiences across all of your content, across all of your programming, but we also have a technology that enables you to have a reference listening experience at any volume level.

Alan: That's correct, so basically it gives you consistent level and it'll give you a nice sound, enveloping surround sound regardless of the level you want to listen to.

Craig: You know we began this conversation talking about, you know, this is not something that hasn't been attempted in the past to be solved, there have been a lot of solutions that have been applied and some of them worked half way OK and [laughing] some of them didn't work at all.

My question to you is, why Dolby? Why was Dolby able to accomplish this revolutionary technology and bring this to market? You know, what tools did you have available to you to really develop Dolby Volume?

Alan: Right, so the key thing that makes this work is really modeling the way that humans perceive audio and Dolby has a long history in using these kinds of psychoacoustic principles to make our technology better. Going all the way back to the noise reduction days, Ray Dolby used principles of psychoacoustics to make that system work. And then Dolby Digital, again, uses psychoacoustics to figure out what data can be thrown away to make the coded audio transparent to a listener.

And so, we took all this knowledge in psychoacoustics and basically applied it to this new problem for us of maintaining consistent levels and maintaining the intent of the audio as you turn it down. So it's really that understanding of psychoacoustics, that's allowed us to create this unique solution.

Craig: [music playing softly] Great technology, an incredible solution for the market place. Alan Seefeldt, Senior Staff Engineer, Research Department at Dolby Laboratories - one of the inventors of Dolby Volume. Alan, thank you for joining us. Thank you for being a part of Dolbycast.

Listeners, thanks you for being a part of Dolbycast. Join us next time.

[music plays]