Mercurial

freddie

“Science Confirms That Freddie Mercury Was Basically The Most Amazing Singer Of All Time” screams one headline. “Freddie Mercury Is The Greatest Singer Of All Time, Because Science Said So” says another. “Queen’s Freddie Mercury Had One Of The Best Singing Voices Of All Time, According To Science” gushes a third. I’m sure you saw them in your social media feed in the last week or so.

So what’s the story then? Have scientists come up with some amazing new method for applying empirical assessment to artistic subjectivity? Has scientific endeavour somehow managed to ‘confirm’ that Freddie Mercury is indisputably the most accomplished singer in human history?

Well, as you might have guessed, not so much.

Let me say from the outset that, by a considerable agreement of subjective reckonings, Mercury is an amazing singer. His vocal prowess is all the more extraordinary in light of his lack of formal training; he was self taught instrumentally and vocally, and claimed not to be able to read music. We really don’t need science to be able to legitimise his talent, per se.

And indeed science is not even attempting to do any such thing. The hyperbole above is generated by the lamest corners of the social media press, after the recent publishing of a paper by Christian T. Herbst et al, from the Faculty of Science at Palacký University, Olomouc, in the Czech Republic. Herbst frames the intentions of the research like this:

“The purpose of this study was to conduct a viable analysis of publicly available data material, in order to arrive at more empirically based insights into Freddie Mercury’s voice production and singing style.”

So the researchers, who self-describe their work as ‘fan science’, take it as a given that Freddie is an awesome singer and are simply gathering some data to help understand – in a technical way – how he did his stuff.

To do so, they examined both his speaking and his singing voice, concentrating on its timbral qualities, vocal range, vibrato and vocal subharmonics. In addition, they had a trained singer emulate some of Mercury’s vocal stylings and observed the active vocal cords with an endoscope. It is all fairly fascinating in its own right, and the breathless exaggeration from some sections of the interwebs really does the science no favours.

So what do the results of their efforts tell us? Well, not a lot that an experienced singing teacher couldn’t have concluded just by listening, in all honesty. They found that Freddie’s vocal style encompasses a remarkable variety of techniques, including a faster than average vibrato, the occasional use of vocal subharmonics (mostly to add a growl or burr to his voice) a proficient control of phonation and an impressive vocal range (although not quite the four octaves often claimed).

The paper provides empirical details of all these things, although, surprisingly in my opinion,  not much investigation into Mercury’s falsetto, which was one of his most dramatic vocal accomplishments (Mercury showed an almost incomparable ability to elide from his upper vocal register into falsetto and back again with virtually no audible transition – singers will tell you that this is very difficult to do).*

So to sum up, we already knew that Freddie Mercury was a pretty astonishing singer. What this investigation does is shed a little bit of scientific light on how he did his thing. What it does not do is ‘prove’ anything about his standing in the dominion of wonderful singers through the ages. And truly, that task is simply not the concern of science.

You can read the full technical paper ‘Freddie Mercury—acoustic analysis of speaking fundamental frequency, vibrato, and subharmonics’ here.

 


*You can hear many examples of this in the early Queen albums Queen, Queen II and Sheer Heart Attack. It is notable on the songs Lily of the Valley, Nevermore and The March of the Black Queen. Mercury’s use of the technique diminished as the band became more conventionally ‘rock’ oriented in later years.

 

Noise Cancelling Shnozphones

Yesterday a friend sent me a link to this crowdfunder on Indiegogo, with the appended question: “This has to be ludicrous BS, right?”

The alleged ludicrous BS is a device called Silent Partner which claims to ‘quiet snoring noise’ (we’ll examine those exact words in a bit) through the mechanism of Active Noise Cancelling (or the popular acronym ANC).

You will be familiar with ANC if you’ve ever worn a pair of noise-cancelling headphones on a plane. Pop those suckers on, and the sound of the plane vanishes almost magically. You can actually hear the in-flight movie soundtrack.

The technology behind those headphones is really quite simple, and has been known for over half a century. In this post we’ll delve a little into how noise cancellation works, what it’s used for, and why Silent Partner is almost certainly BS. Not as ludicrous as it first appears perhaps, but let me just say that I’m not pitching in to the Indiegogo campaign any time soon.

The Acoustic Principle Behind Noise Cancellation

As you probably know, sound propagates through a medium via waves. Most people have seen graphic representations of soundwaves these days. They look something like this:

Screen Shot 2016-03-10 at 11.11.53 AM

That’s a fairly complex soundwave. The simplest manifestation of a sound is a sine wave. This one represents a pure tone, like you might hear if you strike a tuning fork:

sineposgraph

The red bar shows you the frequency (pitch) of the sound and the green bar marks out the amplitude (loudness). I’ve marked the highest and lowest amplitude with an arbitrary value to aid explanation. One thing to understand – and is easy to see in this simple sine wave – is that as a wave propagates, it changes its amplitude value across the zero axis; there is a time when the amplitude has a positive value, and a time when it has a negative value. This is called the phase of the wave, and this is true of all waves, not just sound waves. So in this case, the frequency of the sound stays exactly the same (a constant pitch), but the amplitude is wavering between positive and negative, as it does in all sound waves.

You can’t hear this constant amplitude change. You are only aware of the overall amplitude between the peak and the trough of the sound – this is the volume, or loudness, of the sound.

Somewhere, some time ago – no-one really knows when, but we can guess it was likely to around the beginning of electronic sound recording – someone discovered that it is possible to artificially ‘flip’ the phase of a recorded sound wave.*

sineneg2

Here’s the thing: this flipped sound wave will sound to your ear exactly the same as the one we depicted above. EXACTLY. If I played them to you one after the other, you would not be able to tell which one was which. However, if we artificially combine both these waves so that they are coherent – that is, they start at precisely the same time, and stay in step with one another – then the positive value of the first peak of the sine wave is completely cancelled out by the negative value of the first trough of its phase-reversed counterpart. It’s just simple mathematics: you add +1 to -1 and you get zero. The result is total silence. The positive ‘loud’ bit of one wave totally cancels the negative ‘loud’ bit of the other.

The first time this was demonstrated to me as a sound student, I was astonished. It seems like a weird kind of audio magic. Nevertheless, it’s just a property of wave mechanics, and we use this phase cancellation trick in numerous different ways in the pro audio business.

How Does Phase Cancellation Work in ANC Headphones?

Even though we’ve been looking at an example of a pure sine wave, phase cancellation works on any complex audio wave you care to submit to it. You take a recording of Kanye West, flip it out of phase and recombine it with the original and zap, perfect silence. An exemplary case of nature balancing itself for the better good.

The only condition is that the two waveforms must be completely in step with one another – coherent.

One day it occurred to someone that it might just be possible to accomplish this trick in real time, with live sound. This is how it would work in theory: you’d aim a 180º out-of-phase microphone at a sound and pump it through a speaker in the vicinity of the actual sound. The amplified phase-inverted sound would interact with the real sound and cancel it out. And indeed, it does work. Sort of. It’s very dependant upon where the listener is positioned in respect of both the phase-inverted source and the real sound. Too close to one or the other and the necessary coherence between the two sounds gets out of step. The effect completely vanishes. It works, then, as long as you tell your listener to sit exactly there, and don’t move your head even the tiniest bit. Not particularly satisfactory for most practical purposes.**

If only you could stick the inverted phase sound right at the listener’s ear drum, where it would arrive at precisely the same time as the real life sound… and of course, you can do almost exactly that, providing that the listener is wearing headphones.

So ANC headphones have a little microphone on each earphone, which captures almost the identical sound that is heading toward your eardrums. That sound is then amplified to the exact same amplitude as the original, and inverted in phase. It arrives at your eardrum at the same time as the real sound, effectively cancelling it out. It’s a very cool trick, but it’s highly reliant on the microphone and the speaker diaphragm being very close together, and both those things being as close to your eardrum as is technically – and biologically – feasible.

That All Sounds Plausible! What’s the Problem with Silent Partner?

This is where things get a little more complex than they have been so far, but you already know most of what you need to know to follow me.

I’m speculating that this is how Silent Partner is intended to work: the sound of the sleeper’s snoring is picked up by some microphones – I assume this is what is designated as the sensor in the rather lite ‘technical specifications’. The captured audio is then phase-inverted, amplified and played back through the speakers and ‘resonance chamber’, where it presumably combines with the sound of the actual snoring with the intention of cancelling it out.

You can see one big problem already: as I’ve mentioned above, the most effective sound cancelling happens if the coherent waveforms arrive as close to the listener’s ear as is achievable. In this case, we have something of a distance between the snoring source and the beleaguered sleeping partner – plenty of distance for the waveforms to dis-cohere, and also plenty of distance for another wave-related phenomenon to come into play. I’m talking here about polarisation. Without getting too technical about it, waves are not just two-dimensional like the representations in the graphs above. They are fully 3D, and so for exact noise cancellation to happen effectively, the waves and their anti-phase cohorts must be fully coherent in those three dimensions. This is really hard to achieve in a real-world environment, even with precise, point-source sound. It’s especially hard to achieve when you consider the nature of snoring, which doesn’t actually come from just the nose, as anyone who’s slept with a snorer knows. So, even in the best possible technical case, with the Silent Partner reproducing the snoring sound in every tiny detail of frequency and amplitude, it’s not going to be aligned in proper polarity and coherence with the sounds coming from the nose, throat and chest cavity of the snorer. It just can’t be.

The first clue we have that the makers of Silent Partner are either unaware of this issue (odd, since they supposedly trialled a proof of concept in May 2015), or worse, are being disingenuous about it, is in the video clip. It pertains to their demonstration of the ‘Silent Zone’ around the snorer. In the clip you see a ‘scientist’ holding a decibel meter close to the nose of the snoring person, and showing how the sound level drops as you get further away. Well, yes, this is exactly what you’d expect with normal snoring – sound falls off with distance according to the well known properties of the Inverse Square Law. But this is not what you’d expect to see if the Silent Partner was working as claimed. What you’d actually observe is no change in the sound level. The noise-cancellation effects should be in play – and in fact optimal – right at the source of the snoring. The sound level test in the video is showing us nothing at all but what you’d expect to see with natural snoring.

In fact, the whole meter thing is quite amusing if you know anything about sound. If the meter is measuring decibels (and I’ve no reason to think that it’s measuring anything else – that’s what sound level meters usually use) it’s showing that the snoring volume is dropping from about 84dB to 60dB. To give you a sense of that in terms that might be more familiar, that’s about the level of noise in a moderate factory or a busy road, down to the level of a normal conversation. I’d expect average snoring to have less effect on a dB meter! It’s certainly not an endorsement of a Silent Partner working as claimed.

And all this assumes, as I said, a best-possible case: that the Silent Partner is exactly reproducing the snoring sound. I’m skeptical that it could do even that. Snoring is made up of highly complex human sounds, and is not just one rhythm or frequency. And, as I’ve also pointed out, it doesn’t just come from the nose and mouth, so the entire upper body can contribute to the frequency characteristics. I would be very surprised if tiny components such as those in the Silent Partner – even with resonance chambers – could deliver an acceptable frequency response to enable the kind of phase and frequency matching required for this technique to work.

Aside from all the sound stuff, something just feels really fishy about this whole thing. First of all, the snoring-relief market is chock full of pseudoscience, which indicates an area of high subjectivity and therefore exploitable credulity. You can get anti-snoring pills, anti-snoring rings, anti-snoring pillows, anti-snoring homeopathy, anti-snoring acupuncture and so on and so on, none of them based on even the most delicate whiff of science. So it’s a field rich for the pickings, as they say.

Furthermore, Silent Partner deploys a technique beloved of purveyors of said pseudoscience: it piggybacks on something that to the layperson sounds credible and possible, but brushes a lot of actual dirt under the Carpet of Mystery: “It’s like magic!” boasts the hyperbole – a claim which is actually probably quite true, because it would be magic if it worked, since it defies the laws of physics. Like many pseudoscientific gadgets and gew-gaws, it fails to provide any real science to substantiate its claims, just a lot of hints at how it might work based on other things that do work. ***

There are many other red flags:

• The Silent Partner website is full of exhortations to Order Now! I don’t know about you, but for me this may as well say Give Us Your Money Quick, Before The Truthful Reviews Start Rolling In! Although the Indiegogo timeline includes a proof-of-concept as of May 2015, there is no indication that this showed any kind of convincing result. If it was my product, I’d really want to demonstrate that the thing worked. Not that it seems to matter, evidently, since the campaign is already way over-funded.

•There is a vagueness about the whole affair that is offputting. As we saw at the outset, the gadget claims to ‘quiet’ snoring – but that’s a very subjective thing. Nowhere do we find any definitive terms about how much it will quiet the snoring. The one piece of solid empirical data comes from the decibel meter in the videop clip, and I can tell you, that’s not showing any ‘quieting’.****

•The ‘team members’ for the campaign include far too many marketing and advertising people for my liking. Because that’s what the whole thing smacks mostly of – marketing.

•There are some uncomfortable legal problems. The Indiegogo campaign for Silent Partner holds a disclaimer that says

“Participation in this campaign is in accordance with the Terms and Conditions listed on the Silent Partner website.”

As of this writing, there are no Terms and Conditions listed on the Silent Partner website. There is a link there, alright, or at least words saying ‘Terms and Conditions’, but clicking on them does nothing. I could not find any such terms elsewhere on the very spare site. I think you can see that this is a legal disaster. People who have contributed to the Indiegogo campaign have agreed to do so under non-existent terms. Or terms that don’t exist yet. The makers of Silent Partner could, at any time, write a contract that simply says “I agree that the product might not work as stated and I understand that I will not get my money back under any circumstances”. And then make the link active.*****

In this legal respect, I also feel that Silent Partner might be wandering into the dubious area of being a medical device. This is a highly-controlled and legally rigid field. I’m sure they will vehemently deny it is such a device, but snoring can be a sign of underlying medical problems, and it’s an issue that you should take up with your doctor if it’s a significant detriment to your sleep [to be fair, the FAQ on the Indiegogo campaign site does indicate that you should consult your doctor if you have sleep apnea, but sleep apnea is only one of quite a few serious snoring-related problems].

To Wrap Up

My professional audio experience and technical knowledge tells me that Silent Partner is quite unlikely to show results in the manner claimed. Indiegogo provides little redress for failed campaigns, and especially for investors getting their money back. This, and the fact that there are no Terms and Conditions listed anywhere on the Indiegogo page or on the Silent Partner official site should make someone think twice before spending their money on this product.

But hey, you did learn a thing or two about noise cancellation and acoustic science, right, so your time here has not been entirely without benefit…


Footnotes:

*It is possible to hear naturally-occurring phase related phenomena. For this to happen, a sound needs to be combined with a copy of itself in such a way that it has some semblance of coherence and phase alignment. The most common manifestation of this is with reverberation or echo. You may have experienced this as a kind of ‘dead’ spot with your tv sound, perhaps; when you sit in a particular spot in the room, the sound seems to be muted and quiet, compared to most other places in the room. This is because the phase-changed reflections off the wall are combining with the actual tv sound to effect a phase-cancellation situation. It will never be as perfect as an electronically contrived version, but it can be quite profound given the right circumstances.

**Car manufacturers in concert with electronics firms like Harman and Siemens – have been playing with ANC for a fair few years now. Because the concern here is mostly with inhibiting road rumble, it’s actually not out of the question that it can work; longer, lower waveforms are less directional for the human ear, and so the necessity for the listener to be in a ‘sweet spot’ is not nearly so critical. Nevertheless, the technical requirements are still complex, and it hasn’t so far caught on in a big way.

***It also uses technical waffle: “subwavelength active noise cancellation” is an entirely meaningless term. Sub what wavelength? The clip with the decibel meter is technical waffle too, which as we’ve seen means precisely nothing.

****An expert audio colleague of mine points out that the sound meter display in the video clip is actually an added visual effect. It’s hard to know what to make of this – the inference is of course that the creators of Silent Partner consciously chose the numbers that are being displayed – numbers that are, as we discussed previously, really only indicative of rather loud normal snoring. We could speculate that this a a kind of ‘cover your ass’ move, inasmuch as they can always point to it and say “This is all we promised it would do…”

*****There are a few non-active links on the site. It opens up the possibility of jiggery-pokery. In the past I’ve seen this problem and commented on it and then, mysteriously, the links work – and the content under them is as outrageous as you expect. I’m not saying the Silent Partner people are doing this, just that it does happen with rather alarming frequency with these kinds of products.

 

 

 

 

 

Typical Audio Misrepresentation

A couple of friends have pointed me to this piece of egregious nonsense over the last week. If you haven’t already seen it, you can watch the [largely ridiculous] clip here (I can’t embed it, since it’s some kind of proprietary format).

In any case, the image above tells you everything you need to know; basically, a piece of Vivaldi plays while you watch a spiky and glitchy circle wobble (on the left), which is meant to represent a ‘typical’ CD recording. Meanwhile, on the other side of the screen, a circle supposedly representing a ‘typical’ analog recording also does its wobbly thing, only with curvier lines and colour.

This piece of claptrap offers up a much perpetuated myth about digital recording: that somehow the discrete sampling bits make digitally recorded music sound spiky and colourless, and inferring that analog must certainly be better because it’s more ‘natural’ (or something). Worst of all, the video is useless even as an analogy because it literally doesn’t convey ANY information. It is the equivalent of the old Colgate ‘Mrs Marsh’ advertisement of the 1970s:

It’s bogus ‘science’, appealing only to those who think they understand something, but really don’t.

I’ve said it here before, but I’ll say it again: the quality of a well mastered CD recording is not, in any way, inferior to a vinyl LP. Vinyl is different, and that is all. If you think you can hear a difference in high frequency content, you can’t. If you think vinyl sounds ‘punchier’ and ‘warmer’, you’re probably hearing the effects of the substantially different mastering techniques that are used for vinyl (and that’s mostly equalization and compression).

And before you start to take exception to what I’ve just said, go here and witness a whole lot of people who think like you do failing to be able to tell the difference between cables, amps, DACs and all kinds of formats in controlled blind and ABX listening tests. Mostly to their own disbelief.

Brainless

Over the last few months I’ve had numerous articles pop up in my Facebook news feed of the kind that promise to reveal some new ‘scientific’ discovery or other that will just amaze the socks off me. One of the favourites – going by the number of times I’ve seen it shared – is this one:

“According to Scientists, This is The Most Relaxing Tune Ever Recorded”

The one page sound-bite rich story is widely dispersed over numerous ‘health and wellbeing’ type websites (like this one), where it is for the most part uncritically accepted as a proper piece of news. It outlines the results of ‘a study’ that has honed in on the single most relaxing piece of recorded music ever to have graced human ears. I know – quite a bold claim – but it’s science, my friends, so you couldn’t just pull that kind of thing out of your ass, right?

The ‘tune’ in question turns out to be a piece called Weightless by ambient music trio Marconi Union. You should probably listen to a little bit of it before we go on – I hope it won’t relax you so much that you can’t come back after the break.

Well. Cute idea for a video, but in my considerable experience of the genre, that’s fairly generic electronic ambience.

According to the ‘news’ article:

This eight minute song is a beautiful combination of arranged harmonies, rhythms and bass lines and thus helps to slow the heart rate, reduce blood pressure and lower levels of the stress…

…and

The results showed that the song Weightless was 11 per cent more relaxing than any other song and even caused drowsiness among women in the lab.

The more you read, of course, the more you begin to see the divergence between the hyperbolic claims of the headline and the actual content of the piece. “11 percent more relaxing than any other song”? What does that even mean? 11 percent more relaxing? Calibrated how? Any other song ever written? Any other song of the same genre? Or just any other song on the iPod of one of the ‘women in the lab’ (is it just me, or is the implication here that the men in the lab were far too macho to be influenced by the wafty charms of the New Age ditty?).

Oh, here we go:

this song induced the greatest relaxation, higher than any other music tested till date

Right. So that would be how many songs, exactly? Not all the tunes ‘ever recorded’ is my guess.

Having worked for many years writing music for advertising, I can easily spot a marketing campaign when I smell one, and this one veritably reeks of fish. But a marketing campaign for what, exactly? There’s no product mentioned in the article, and it doesn’t seem like a promotion for the band. But I guess that’s possible. Let’s see what the internet has to say about Marconi Union.

OK, not much to be gleaned out of the obviously self-penned Wikipedia entry, other than that Marconi Union uses the flimsiest of pretexts to contrive to have their name linked with that of ambient music guru Brian Eno (in several different ways).

What about the ‘scientists’ involved in this astounding breakthrough? The experiment (we’ll call it that for expediency’s sake) was run by an entity by the name of Mindlab International who describe themselves as  a ‘neuro’ marketing company. Their website is slick and snappy.

But here’s one thing I’ll tell you about marketing – it’s not science. The sole goal of any marketing company on the planet is to gather enough data to convincingly support exactly what their client already believes to be true, extracting a generous fee in the process. Not convinced? Consider this: the marketing company is hired by the product maker. The product maker already thinks their product is the best thing since sliced bread. The marketing company who tells their client otherwise is therefore wrong, and gets fired. Marketing companies are the definition of the messenger who really doesn’t want to be shot. So, being a successful marketing company is all about finding convincing ways to make the message appealing to the client. If you can also convince them that you had a hand in increasing sales, that’s a bonus, but don’t think for even a second that that’s their raison d’être (notwithstanding the fact that many of them apparently believe that it is).

In case it needs to be said, none of this is the job of science. Science is about determining truth, no matter what the consequences to human sensibilities or dog food sales. Good science is out of a job when it comes to the advertising business.

The fact that this story involves a marketing company and a bunch of exaggerated claims  immediately raises red flags. It’s quite obviously not concerned with giving the reader factual news. No,  it’s solely about selling something. The more observant among you will have noticed that the video of Marconi Union’s Weightless – linked in the article (and above) – is interestingly named. It has the peculiar bracketed codicil ‘Radox’ in the title.*

Ah, yes. It’s all starting to make sense. So factoring ‘Radox’ into a search with ‘Marconi Union’ and ‘Mindlab International’ gets us straight to the nitty gritty. The first sentence on this page at a site called musicactivation.com reads:

Radox Spa brand has made an ambient track that is scientifically proven to be more relaxing than a massage.

Oh, now what a coincidence! The ambient track Weightless, which was made by (I think we can assume ‘commissioned by’) Radox Spa bath salts just turns out to be ‘the most relaxing tune ever recorded’.

Are you smelling any science here, sports fans? Or are you getting an overpowering aroma of horseshit? For the full experience, you might like to watch this:

Well, that was the biggest load of bollocks I’ve seen in a long while. I wonder how many completely unscientific instances of prestidigitation you caught there? If you needed proof that Mindlab International is entirely unconcerned with actual science, there’s plenty of ducking & weaving going in that clip.

What’s happened with this whole thing is that hundreds of websites have mindlessly snapped up a piece of marketing propaganda (something that was, no doubt, fed out to the media machine as a press release from Mindlab International) and completely credulously echoed it far and wide across the social networks as if it actually had meaning. I probably don’t need to rephrase it to reflect actuality, but for amusement’s sake, this is how that headline should really have read:

“According to a Marketing Company, a Piece of Music that their Bath Salts Client Commissioned is the Most Relaxing Tune Ever Recorded”  

Not something that’s going to rapidly circulate on the social media, is my guess.


*Update: since I penned this piece, Marconi Union seem to have disassociated themselves from Radox, so the original video I linked here has disappeared. References to Radox don’t appear on the re-uploaded version of Weightless

The Sound of Distant Drumming…

bilbo

With Peter Jackson’s movie version of The Hobbit killing box offices around the world, most people can hardly have failed to notice that, like Spielberg and Lucas before him, Jackson is using his considerable clout to spearhead another leap in technical quality for the cinema experience. I am talking, of course, about The Hobbit‘s release at the projection speed of 48 frames per second, which has already become popularly known by its acronym, HFR (high frame rate).

Jackon’s decision to use HFR for the film was almost certainly prompted by the desire to increase the quality of 3D – something which it indisputably does. Unfortunately, HFR got off to a bad start when news got out that audiences at the CinemaCon preview of The Hobbit in April of 2012, were finding the high resolution experience rather disconcerting. Mostly, the criticisms centred on the HFR looking ‘stagey’ or ‘fake’ or ‘cheap’. Peter Jackson brushed off the criticisms, saying we’d all get used to it, but there is no question that HFR has some ‘issues’. Since the release of the full movie, there have been many dissections of the curious nature of the 48fps effect in the film, some smart and some fairly far-fetched. I have my own hypothesis as to why we have trouble with the HFR experience, but I mostly want to talk about one aspect of that here. And it’s an aspect that no-one seems to have picked up on yet: the sound.

It’s not really that surprising to me, I have to say, that no-one has thought to scrutinize the part that sound plays in the perception problems for a high frame rate experience. People hardly pay attention to sound at the best of times, and in this particular case the whole issue has been entirely one of ‘the look’. A movie, however, is not all about what you see.

For you to follow my argument in this post, I need to just give a brief overview of the usual experience of sound in the cinema. Currently, most movies you see are delivered to an audience in a format that was first formulated in the 1970s by Dolby Laboratories and then widely adopted in the 1990s for the cinema. It is called 5.1 surround sound (some high end modern movies use a variation of 5.1 called 7.1, but for purposes of this discussion it’s essentially the same thing). 5.1 is arranged in such a way that three speakers across the screen deliver most of the critical sound. These are designated as Left, Center and Right. Usually these speakers are placed just behind the screen, along a horizontal center line halfway down. Two further speaker groups (they are usually groups, but just think of them as single channels, because that’s how they are treated) are designated Left Surround, and Right Surround. These are placed to the side and behind the audience. This array, the L,C,R,Ls,Rs is the ‘5’ of the 5.1, referred to in the profession as ‘5.0’. The ‘.1’ is the Low Frequency Extension, or LFE, which is situated behind the screen with the LCRs, usually on the floor. The important thing about the 5.0 array is that this is what determines the spatial placement of sounds for the audience.

soundlayout

Over the last couple of decades, the surround sound array has proved to be a versatile and durable system, providing considerable enhancement of the 2D experience. Here’s the thing to keep in mind, though: the key to the effectiveness of the whole notion of surround is that the sound can be made to appear to be – fairly convincingly – anywhere inside the theatre in front of the screen. The whole surround sound concept is based on the idea of enveloping the audience.*

The increasingly fashionable use of 3D over the last five or six years has, however, introduced something new into the movie experience: depth. 3D imaging is, in a way, the complete obverse of surround sound. 3D doesn’t bring the image out into the theatre as much as it expands the depth behind the screen into which the viewer looks. This creates an obvious problem for sound: how do you make the sound feel like it’s back there with the image?

If an object – let’s say a racing car – is on a flat screen (in 2D), then no matter where it is on that screen, it’s always on the same plane as the the speakers reproducing its sound. If  the car appears to be receding into the distance, we have learned by cinema sound ‘convention’ that its sound will get quieter (fade away) to make it seem like it’s attached to the car, but in realistic terms it simply can’t appear to be coming from a point source a  hundred meters beyond the screen (it’s crucial to understand here that our ears don’t judge the location of a sound solely by its loudness. In addition to loudness, we detect small differences in delay times to place an object aurally in space, as well as as different kinds of reverberation textures and times, and comparisons to other sounds we might be hearing. Our brain’s processing of audio information is complex and detailed, and artificial sound reproduction has pretty much always been a kind of ‘cheat’). As it happens, it’s perfectly acceptable in 2D because the image of the racing car and its sound never actually go off that 2D projection plane.

In 3D in the movies, though, our brains are forced to deal with a kind of cognitive dissonance. The racing car appears to be much more realistically moving at some distance away from us, but we hear the sound that it’s making on the same screen plane as every other sound we’re hearing! In fact, the 3D ‘world’ space and the domain of surround sound hardly overlap at all.

3dcover

This odd paradox is a legacy of the stumbling technical evolution of the cinema. The whole point of introducing surround sound into the movie environment was to attempt to create another level of involvement for the audience. Because the image on the screen was flat, no-one really thought of having the sound go beyond the screen; like the wall of the cinema to which it was attached, the screen plane has always been treated as a hard and absolute boundary (aside from anything else, there would be the technical and economic consideration of having speakers – and a whole room – on the other side of the screen). The difference between 5.1 and everything since old fashioned mono, was that surround was largely about pulling the sound forward and off the screen, by creating the ambiences of things that you did not see.

With 3D, a whole new space opened up – a space full of things we can see and that can be anywhere from slightly in front of the screen plane to an infinite distance away. And if those things make sounds, then our brains quite understandably expect those sounds to be glued to their matching apparent position in space.

This was never a problem in 2D, because 2D is a stylized way of looking at the world which we have learned to accept as ‘reality’ through massive exposure to 2D images over more than a century, and to which we have become habituated. Up until The Hobbit came along, there wasn’t even much of a problem with 3D, because the quality of 24fps 3D is not really that great and you just didn’t notice the ‘gluing’ error that much. But with the extraordinary detail that is available in HFR, the 3D begins to push a level of resolution that approaches reality. And, as that happens, our brains start to tell us that something is out of whack: “That horse galloping off into the distance! It’s very quiet, but its sound is coming from right in front of us!”

Not that we process it consciously, of course, but it adds to the sum feeling that there is something kind of wrong with what we’re experiencing.

While I was watching The Hobbit this strange dissociation of sound and image kept catching me at every turn: Bilbo runs behind a rock but his footsteps are unattached from him; an arrow hits a distant ledge in a cave and it sounds like a small stick being thwacked next to my ear; a warg runs off into the distance and its growling doesn’t manage to go with it. In general, the whole soundtrack is somehow ‘smeared’ and its details diffuse. Now, I know the guys who did the sound on this film, and I can tell you – they really know their stuff. And I’m willing to bet that when I get to view The Hobbit in 2D here in my studio in 5.1 on BluRay, it will sound spectacular. I believe that the problem is not with the way the sound is done, but with the way the sound is done for an HFR 3D film. The sound for The Hobbit has been created in the manner in which we normally create sound for the surround cinema environment and that process is now approaching a point where it’s simply not adequate to create an illusion of reality.

This difficulty is not confined to the sound alone. In The Hobbit, a similar predicament exists with the lighting and other aspects of its cinematic ‘world’. We’re at a stage with this technology where the whole technical way we make films needs to be rethought.

For sound, I don’t really know how we can easily remedy this situation. One of the other technical innovations that was rolled out for The Hobbit was a new sound format by Dolby Labs called ‘Atmos’ (which unfortunately I didn’t have  a chance to experience, since there are no cinemas in the southern hemisphere – other than the Embassy in Jackson’s home town of Wellington, NZ – equipped with it). Dolby understands that as cinema image resolution increases, the 5.1 (or 7.1) array will start to struggle to hold its ground. The Atmos system is an effort to expand the level of detail of, and control over, the aural environment. Atmos employs a speaker array with an astounding 64 programmable point sources throughout the theatre – that’s 58 more speakers than 5.1 – but I’m sure you’re ahead of me here: this hasn’t changed the overall concept of cinema sound. The sound is still inside the auditorium, not back where the 3D is happening. It is likely that Atmos will give a new level of detail to what you hear in the theatre, but it still needs to somehow address the sound on the other side of the artifical boundary that is the screen plane.

Of course, there is a way to tangle with this puzzle, but as I said, it’s not likely to be easy. Or cheap. It involves computers, and sound modelling and other kinds of new tech. It’s rather too complicated to go into at the end of this long post, but maybe I’ll attack it at another time.

_____________________________________________________________________________________________________________

*Director John Boorman, an outspoken critic of surround sound, bemoaned the fact that he’d spent his whole career attempting to get his audience involved in what was happening on that flat screen and now he was expected to embrace a technology that was trying to pull them out of it.

By the Numbers

Today on Hummadruz we’re going to look at the mysterious phenomenon known as Numbers Stations: radio beacons that broadcast strange shortwave radio transmissions that consist of strings of numbers read in sequence and usually identified at the start with some kind of audio ‘logo’ such as a sound or a piece of music. No-one really knows the purpose of Numbers Stations (well, I say ‘no-one’ – obviously someone does), but it is likely that they are used to send coded messages from government security agencies to their undercover operatives out in the field. In other words, this is spy stuff. It is thought that the first Numbers Stations started up shortly after World War 1, and it is certain that there were many operating after World War 2 and throughout the Cold War.

The most well-known of the Numbers Stations is probably The Lincolnshire Poacher, so called because of its use of the folk song of that name as its identifier.

Audio link: The Lincolnshire Poacher

The Lincolnshire Poacher was a very powerful shortwave broadcaster, purportedly operated by Britain’s MI6 out of Cyprus. It had a similarly powerful sister station called Cherry Ripe which was believed to be in Australia.

Audio link: Cherry Ripe

Both these stations ceased broadcast around 2008/2009, but there are still dozens of Numbers Stations in operation, especially in Eastern Europe and South America. The mysterious repeated sets of numbers are almost certainly a type of code, most likely a system called a ‘one-time pad‘. A one-time pad code is completely uncrackable if the people employing it stick to the strict protocol. Short wave radio is an ideal method for transmitting the information, due to its long reach, and the relatively low tech and easy availability of portable shortwave receivers.

Some Numbers Stations use a type of phonetic alphabet system that can result in a very surreal effect. Here’s a station called Nancy Adam Susan:

Audio link: Nancy Adam Susan

The robotic hypnotic quality of the echoic female voice reciting names over and over seems to me much more disconcerting than the EVP phenomenon that we talked about recently on Hummadruz. Here’s perhaps the creepiest of all the currently broadcasting Numbers Stations, dubbed Swedish Rhapsody (the name of the melody):

Audio link: Swedish Rhapsody

The tinkly music box tune and the sampled voice of a young girl are surely the stuff of radio nightmares. One of the most fascinating things about these Numbers Stations is that on a purely aural level they are evocative and intriguing, and speak of a Cold War era that is surely fading as we move into the world of high speed internet and digital encryption. If you’d like to hear some more of these strange radio relics, there’s a substantial collection of recordings here on archive.org – these are part of The Conet Project which is available through Irdial Discs. Irdial also makes the entire collection available for free as mp3s.

One Trick Pono

The ‘big’ news in audio this week is that legendary musician Neil Young has introduced a new music player to the world – the Pono. The word ‘pono’, apparently, is Hawaiian for ‘righteous’. The principal selling point of Pono is, according to Mr Young, that it presents the listener with ‘the best quality audio available’. Here he is pitching the concept (badly) to a fairly underwhelmed David Letterman:

Flea, bassist from the Red Hot Chili Peppers, effusively spruiked the Pono experience to Rolling Stone:

“It’s not like some vague thing that you need dogs’ ears to hear. It’s a drastic difference.”

Rolling Stone reports that Flea discerned this ‘drastic difference’ after hearing Aretha Franklin’s ‘Respect’ played in Mr Young’s car. Now, I suppose Neil Young has a pretty good sound system in his Cadillac, but trust me, a car is not an optimal listening environment for a phone conversation, let alone for judging music quality, so I for one am taking the audio assessment of a bass player from a rock band criticized for the egregious loudness of its recordings with a grain of scoff.

See, the problem with this kind of thing is one of perspective. Even if we accept that Pono will deliver an appreciable difference in fidelity to what is already available – and for the record, I don’t – Neil Young thinks enough people care about that to make his idea a commercial viability. He obviously doesn’t go to cinemas, have teenage daughters, listen to the radio, or pay attention in any way to how the great majority of people consume music. He’s failed comprehensively to understand the reason that compression codecs like mp3 caught on in the first place, and, worst of all, he’s possibly the only person in the world not to have learned a business lesson from the VHS/Betamax format war of the late 1970s (which, in case the point needs to be made, showed that people don’t give a flying fuck about quality when it comes down to it). As much as I admire Neil Young as a musician, I think his business acumen sucks.

Let me put it to you from my personal perspective as a prospective Pono punter: I’m a trained sound professional with a love of music – new and old – and an appreciation for the amount of work that goes into the craft of getting it to my ears. I love good quality sound. Occasionally I buy music for the fidelity of its recording. But mostly, I don’t. Mostly I buy it for its content. I buy it for the songs, or to play while I’m making dinner, or to listen to in my car when I want to be able to ignore the hum of the city. I rarely have the time to sit and just listen to a recording in the relatively superior listening environment of my sound studio. I like to take my music with me, so I have some on my phone, and some on an iPod in my car. I have re-purchased music I already owned so I can do this, and have also digitized my not-insubstantial CD collection. Now – WHY ON EARTH WOULD I BUY A PONO AND ALL MY MUSIC AGAIN? I know that, theoretically, mp3 and AAC are inferior to uncompressed digital (whether that extends to the stratospheric192kHz/24-bit sound that Pono offers is arguable…) but I don’t care. I bet the Pono music won’t be as cheap as the iTunes store, and I bet the Pono won’t interface with my car. And I already carry around music on my phone – why would I want another gadget cluttering my pocket? It’s one of the cool things about the iPhone: I have music, a phone, a diary and a camera with me at all times in one unit. What I’m trying to show you here is the vast hurdle that Neil Young is proposing to leap, on the basis that people care about superior sound quality.

An interesting aspect of the reporting of this story is that the press seems to have picked it up under variations of this leader: “Neil Young Expands Pono Digital-to-Analog Music Service”, which, aside from being an entirely inaccurate appraisal of the way the gadget works, rides on the coattails of the hoary old myth that analog recording is somehow magically ‘superior’ to digital. Analog is different to digital. That is all. You may even prefer the sound of old analog recordings over modern digital ones, but that has, these days, nothing at all to do with tehnical quality. It’s merely fashion. And it’s a fashion that can, in fact, be reproduced adequately – for the great majority of listeners – under existing digital audio codecs. Why, if you really want very high quality audio, it’s already available in iTunes (not quite the 192kHz offered by Pono, but Jesus, people – Aretha Franklin IN A CAR???)

My prediction? One year on from the official launch of Pono and you’ll be buying the things on eBay for 50c. Come back and tell me I was wrong.