Slashdot Log In
Unintended Aural Consequences of MP3 Compression
Posted by
chrisd
on Fri Dec 20, 2002 08:08 PM
from the lossy-compression-obsoletes-itself dept.
from the lossy-compression-obsoletes-itself dept.
zenst writes "A rather interesting read about possible damage to your hearing due to the way most audio compression techneques work. They mainly work by presenting a signal that the brain perceives to be the same as the original and it is this assumption that could effect our hearing and the way we hear."
This discussion has been archived.
No new comments can be posted.
Unintended Aural Consequences of MP3 Compression
|
Log In/Create an Account
| Top
| 670 comments
(Spill at 50!) | Index Only
| Search Discussion
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
woo-whooo (Score:3, Funny)
Its not as crazy as it sounds (Score:5, Interesting)
the nerves in your ear and all the low-level neural processing of sound will fire in response to the gaps, watermarks or subliminal signals in the music stream. It is only the brain that filters these out. But is the brain unaware of the signals?
it's long been known that humans perceive sounds they dont actually hear in the sense that their brain registers it. Ancient church organs have sub sonic and ultra-sonic pipes in them for the purpose of stimulating emotional responses in the audience. It's well known from many pyschological studies that slight , consciously imperceptible, delays introduced into telephone conversation response times causes people to think the person they are talking to is angry.As a kid I could always hear the flyback transformers in TVs and video screen. I could not tell you what the sound sounded like--it was not a high pitch. it was no pitch at all. But I could tell it was present.
The thesis that spectral drop-outs could somehow disrupt neural feedback circuits is an interesting one. Certainly most human made electronic circuits dont handle delta-function responses well: that is the phase lag in any feasible feedback circuit puts an upper limit on the fidelity of the response. Thus the idea that the neural feedback that nulls the unwanted off-pitch sympathetic vibrations in the ear following a loud signal could be disrupted if the waveform was not continous after the loud noise is a valid one. Would this lead to false retraining of the neural net and thus tinitiitus? doubtful. But interesting as an example of an unintended consequence no one thought of before.
Not very much.... (Score:4, Insightful)
Re:Wonder if same true for Ogg as well (Score:4, Funny)
Oh thats what it is... (Score:5, Funny)
Tinnitus (Score:4, Informative)
The article (Score:4, Funny)
As a healthcare provider and someone that works at Mayo clinic, this article does not even merit the cursory speed read.
Re:Tinnitus (Score:5, Informative)
He says "is still unclear whether the consequences of such maladjustments are only temporary (similarly like seeing the world in green/ red discoloured after taking off red/ green 3D glasses) or if the continuous consumption of neuroacoustically datareduced sounds can lead to long lasting or even permanent damage."
and also "I try here in no way to demonize MP3 in the name of the sound carrier industry"
He's not trying to scare people, he's just theorizing, with a educated point of view.
MP3 and other lossy codecs fool our ears, and unlike our eyes, our ears require constant re-calibration to function properly. If we are calibrating to inaccurate/unnatural sounds, he thinks this could be a concern.
Certainly just listening to a few mp3's a day is nothing to worry about, but what about when all of the media we saturate ourselves with is lossy-encoded?
I don't know, and this is not a scientific article. He's just throwing the idea out there.
Re:Tinnitus (Score:5, Interesting)
Interestingly, consider this (lifted from here [rochester.edu] but you can find it mentioned in many places via Google [google.com]):
Fun eh? Makes you think. Possibly calibration of the ears works in a similar way. Presumably if the 'lossy' audio ever became a problem (this is assuming you don't speak to anyone, or make/hear any natural noise for most of the day, of course) then you'd just listen to 'normal' sounds. That's what calibration is, after all.
By the way, out of interest, here's another interesting write-up about the experiment being done in Japan [go.com]. I'm curious to see it was done 'recently', as I remember this experiment being mentioned on TV about 15-20 years ago*.
Tim
* Johnny Ball's 'Think of a Number' for you UK geeks :)
Re:Tinnitus (Score:4, Insightful)
He says
I read it, and I don't buy a word of it. The grammar is faulty. Lay+out is make-believe at most. Pictures are from some first anatomy book. No references, just some weird hypothesis withou any proofs. Sounds like a hoax if I ever saw one.
Besides he even fails to mention most Fourier transform based codecs work. They do not even fullfill the persumptions of his hypothesis, because their main size reductions are based on sliced away frequencies outside the scope of our hearing.
Re:Tinnitus (Score:5, Insightful)
I listen to headphones at resonable volumes, I don't work in industries which have hearing-loss danger, and I don't go to bars or concerts or other loud events very often. My hearing is still great. But this low tinnitus only when its completely and utterly quiet.
So to respond to the original article: Do you drink cola drinks? Coffee? Try quitting caffeine and maybe your tinnitus will go away. And you can still listen to mp3s.
PS: If you read medical sites and other tinnitus support sites, they all say that caffeine aggrivates tinnitus because it constricts blood vessels in the ear. Quit caffine!!
Shhh! Don't Tell the RIAA (Score:3, Funny)
Knowing them... they might try giving money to help the military research sonic based weapons and get something to fight those pesky pirates!
Then again I suppose it will also depend on the quality of the speakers, and what frequency range they can properly output (as well as the soundcard and encoded track).
Music type... (Score:5, Funny)
Re:Music type... (Score:5, Funny)
And I thought that was an oxymoron.
Re:Music type... (Score:4, Informative)
You can buy Bose. Or you can spend the same amount on a quality speaker and get far better sound. And yes, I've heard the difference.
For a starter, you might want to try this site [goodsound.com].
Re:Music type... (Score:4, Informative)
If you've heard the "wonderful sound of Bose" because someone told you they were great, consider this. Budweiser is unlikely a fine beer, but it's the "King of Beers" due to superior marketing and brand awareness. Same thing with Bose. They've gotten their name out there, and made people aware. "Big Bass from a small space" does not equate to better sound. Proper speaker construction, good crossovers, and matched drivers *do* contribute to better sound. The fancy-ass "Bose 901's" that some people drool over? They're made up of (9) cheap 5.25" drivers, of which ONE points forward. The rest fire backward to give you the "reflection" of the Direct/Reflection equation, and make the speaker sound bigger. If you like the 'Reflection' theory, try a pair of Definitive Technology's Bi-Polar Power Towers - they'll impress you.
You want to try some really nice moderately priced speakers? Boston Acoustics, Definitive Technology, and Klipsch are all phenomenal performers without a staggering price tag. Moving up, you've got options that will astound you.
If you want a good comparison, buy a Bose waveradio. Take it to a stereo dealer who sells the little Sony or Yamaha bookshelves - the Bose sucks in comparison.
Remember: "No highs, no lows, must be BOSE!"
Loses all credibility right here. (Score:3, Insightful)
Re:Loses all credibility right here. (Score:5, Informative)
Now have a look at this [fh-hamburg.de] page of his. He appears to think that looking at the colour pink can be dangerous too. How did this stuff manage to get posted? In the pink article he claims to be a "researcher of neuronomy(science about the improvement of the usage of brain and nervous system)". Neuronomy? That's gotta be bogus. Anyone?
Re:Loses all credibility right here. (Score:5, Funny)
As a side-note, my neuromechanomy studies have led me to a potentially disastrous discovery: gravity is the leading cause of death. Preliminary studies involving rodents support my theory -- many died within minutes of being placed in a centrifuge. Therefore, I propose that mankind be fitted with antigravity suits, so that life will not be hindered by the harmful effects of gravity. It's amazing how long the human race has survived with a daily, constant exposure of this magnitude.
Let's see... (Score:3, Funny)
Is there anything left that wont slowly kill or mame you over time? They wont be happy until I'm sitting in a darkened padded room eating a liquid only diet.
Re:Let's see... (Score:4, Funny)
You forgot to mention s*x!
However, I'm not sure yet which takes more damage during the act, my genitals or my ears...
Anyway, as I understand the article, only half of my hearing capabilities will be damaged if my girlfriend screams MP3 encoded, because my ears have been calibrated using analog screaming (the vintage stuff).
I think I can live with it either way...
Re:Let's see... (Score:5, Funny)
They wont be happy until I'm sitting in a darkened padded room eating a liquid only diet
Here in Canada we call that "going to the pub". It's not a bad way to live actually..
speculation (Score:4, Insightful)
Re:Sadly, your assessment is not entirely unwarran (Score:5, Insightful)
Please tell me this is sarcasm.
This guy did little more than quote a college biology book, and scan the pictures to create a web site. On first reading the article I thought to myself, "funny, it doesn't feel like April first."
Also, even if we give this guy the benefit of doubt for a moment, there is still nothing to worry about. When was the last time you listened to MP3's and/or video games in a completely soundless environment for an extended period of time? Last few times I did it, I was at home with the refidgerator humming away, a few computer fans whirring, my chair creaking occasionally, simply put, I had lots of background noise for my ears to filter out, without my speakers adding to it. Sure, I would love to put a sensory depravation tank around my computer when playing Thief, it can really blow yuor concentration when your roomate bursts out in laughter 3 feet away from you while reading his email. But, I don't have one, and so am bombarded with small, often inaudiable sounds.
If I were a betting man, I would confidently bet you were right. But just the same, I hope a few members of the medical community (I think this would take a background in neurolobiology/cog. sci/audiology) see this, and at least consider it. You could probably devise a relatively inexpensive animal study or two that could safely close off this kind of speculation.
There are far better things for that money to be spent researching. Don't waste it on junk like this.
Analogy to vision.. (Score:5, Interesting)
As I read this I couldn't help but thing about RGB displays. The visual world is populated by a wide spectrum of photons of different frequencies, but due to our anatomy, our sensitivity peaks at three wavelengths, approximately red, green and blue. The entire color TV and video industry exploits this fact and achieves huge amounts of compression by transmitting three signals at these peak wavelengths. While I recognize that there are some certain mechanical elements in hearing, it seems to me that if this guy's arguments are sound, then we would have observed similar effects from watching TV-- that the absence of unperceived wavelengths would cause damage. Of course we all recognize that TV's bad for your health, but I don't think it causes the kind of damage he's alluding to.
Hmmm... (Score:3, Funny)
Ohhh Kay then...
Think I'll wait till someone with a smidgen of credibility has something to say.
This is FUD (Score:5, Insightful)
The crux of his argument is that because the psychoacoustic model of hearing allows us to remove some frequencies, the ear will no longer calibrate itself correctly for "real" sounds. Wild conjecture, with no backup. Then he grumbles a little bit about his games using lossy codecs, and finally blames his own tinnitus on these games.
A bit of a loose argument, I have to say.
Re:This is FUD (Score:5, Insightful)
From what he says, the only way that MP3s could damage hearing is if ONLY that type of compressed data could be heard at all times. As long as there are other things to hear (like the world outside of your headphones...) then there's no reason for the ear not to calibrate itself.
It's kind of like saying that apples are dangerous. Your stomach digests apples in a different way than when it digests meat. Eventually, the body will get used to digesting apples and not remember how to eat meat. So, when you finally do eat meat, your body will not know how to digest it. I don't eat apples very often. One day, I ate an apple and then I ate meat, and I had an upset stomach.
Seems to be pseudoscientific drivel? (Score:5, Funny)
He does say that CDs are overpriced though, so it must be worth posting on Slashdot.
what? (Score:5, Funny)
Lovely logic... (Score:5, Insightful)
This guy seemed intelligent all the way up to the point where he wrote that particular line. If it only took that little of exposure to lossy sound caused him to have tinitus, then why aren't people by the millions complaining of hearing problems? I'm quite surprised he'd attribute his hearing problems to his hypothesis. I think it is far more likely there are other causes of his problems.
I also don't think, from what I've read here, that we're in any real danger of suffering noticable hearing damage from MP3s. The the main reason is that we don't listen to just MP3s 24 hours a day. Not even close! We'll be surrounded by compressed sound for years to come, but it'll never replace the natural every day sounds we hear all the time. Right now, as I write this, I can hear things happening all around me that definitely are not digital. As long as that noise is there, I can't imagine that our brain would focus in on the compressed sound itself.
It's an interesting hypothesis, but it doesn't hold up against real world data.
Arrgh! My Eyes! (Score:5, Funny)
How MPEG Audio Compression Works (Score:3, Funny)
Persistence of Vision (Score:5, Insightful)
Furthermore, television, movies, and computer monitors are based on persistence of vision-- the idea that the eye and brain can be fooled into perceiving motion if the pictures are switched fast enough (in the case of NTSC TV, 30 frames per second). This is a significant "compression" of the data, far larger than the amount of data being thrown out by psychoacoustic compression. NASA uses cameras that record 10,000 fps to examine explosions and things of that nature that occur far too fast for us to perceive.
Reality occurs at a rate that technology currently finds impossible to record in full. That doesn't mean it's damaging us.
This is pure idiocy... (Score:5, Interesting)
This looks like one of many crackpot "religions" based on a few scientific terms and some mystical psychobabble. These are people that believe microwave radiation or EMF from power lines slowly poisons your soul, the world is coming to an end becuase of evil american weather control machines, the aliens have visited us from dimension Z, the ancient Mayan calendar is the key to all knowledge, astrology is a real and important force in our lives, and so forth.
Mix varying amounts of scientific-sounding nonsense, mysticism with references to eastern religions, profound realizations about the nature of space and time, and maybe a few terms like "asymptotically" to really fill the minds of morons with awe and fear, and you have yourself a religion [bluehoney.org], or more appropriately, a cult [scientology.com].
Wow.... (Score:5, Insightful)
This is just ridiculous bull crap. So is the brief mention of "subliminal messages". Normally, I would elaborate further and explain, but on this I think not. Anyone stupid enough to not immediately realize that this is bullshit is beyond reason anyways.
Other research from the author -- Pink is Evil (Score:5, Interesting)
From the author's web page:
Warning: Pink can be dangerous for health! [fh-hamburg.de] about the stress generating, sick making and learn- hindering effect of long exposure to pink in the viewfield
I sure am glad someone is finally focusing on these severe health risks! Where are the Surgeon General's warnings about the risks inherent in MP3s and the color Pink? Why isn't CNN covering this?
I mean, it's obvious that pink must be bad for you -- just look at the grammar in the abstract. The author is obviously a severe sufferer of pinkitis, poor man.
Re:Other research from the author -- Pink is Evil (Score:4, Informative)
Why are most hallways tope? It's soothing.
Bright colors give us an "open" feel in rooms while dark ones close us in.
Yellow is an alerting color, which is why it's used in stop lights, as is red. Colors do effect us in certain ways.
Example, drunk driver sees cop car on the side of the road with it's red blinking lights. Drunk driver can't remove focus from them and crashes into cop car. It happens almost every day in the U.S. Is it because of the color of the lights? Blue lights have proven to have a different effect. Maybe it's just that their flashing? Flashing blue lights had a different effect.
Anyway, the guy seems like a crack pot, but colors can effect us in minimal ways. Very minimal however.
If technology directs evolution... (Score:5, Funny)
...then everybody only heard mono before stereo was invented.
...then there was no math before the Babbage machine. Thus, Pythagoras, Archimedes and Newton are frauds.
...then video game players couldn't hear human voices before the mid 1990s because games didn't have much speech before CD-ROM.
...there is no such thing as depth perception because TV is still 2D. Thus no one is qualified to drive a car, or at least the people who watch TV aren't. Nor are Slashdot readers, I'm afraid.
Calvin and Hobbes has evidence that the same thing happened to color vision:
Calvin: Dad, how come old photographs are always black and white? Didn't they have color film back then?
Dad: They sure did. In fact, those old photographs are in color. It's just the world was black and white then.
Calvin: Really?
Dad: Yap. The world didn't turn color until sometime in the 1930s, and it was pretty grainy color for a while there, too.
Calvin: That's really weird.
Dad: Well, truth is stranger than fiction.
Calvin: But then why are old paintings in color?! If the world was black and white, wouldn't artists have painted it that way?
Dad: Not necessarily, a lot of great artists were insane.
Calvin: But... but how could they have painted in color anyway? Wouldn't their paints have been shades of gray back then?
Dad: Of course, but they turned colors like everything else did in the '30s.
Calvin: So why didn't old black and white photos turn color too?
Dad: Because they were color pictures of black and white, remember?
[Calvin leaves, meets Hobbes]
Calvin: The world is a complicated place, Hobbes.
Hobbes: Whenever it seems that way, I like to nap in a tree and wait for dinner.
Partial English Damages the Brain (Score:3, Funny)
This effect seems magnified if subjects have been sitting in front of CRT all day reading headline websites and not generally excercising their physical body in any way.
(BTW-Tongue firmly in cheek, no offense meant to these researchers in any way.)
downhill (Score:3, Insightful)
I stopped reading after he started going on about dictators slipping propaganda into the inaudible cracks in your media.
And it started off with such promising analysis! I bet the slashdot moderators didn't read to the bottom of the article before approving it.
This article is obviously part of a RIAA plot (Score:3, Funny)
In fact, the newer formats, such as DSM (SACD), have so many more frequenceis, that listening to these formats will, in fact, improve your hearing. So, everyone, listen to SACDs instead of normal CDs. beecause even CDs may cause brain damage.
Never mind the fact that SACDs are copy-protected 15 different ways, and that our methods for copy-protecting normal CDs have been shown to be ineffective. We want people to listen to SACDs for, well, their hearing.
Again: Do not listen to MP3s! they only damage your hearing (and promote bands which are not approved by us).
- the RIAA
Jpegs damage your eyes (Score:4, Funny)
Maybe only because what your doing while veiwing those jpegs..
Drag the RIAA Into It... (Score:3, Insightful)
While reading this story, ironically enough, I played an MP3 I had downloaded from Gnucleus (a Gnutella client) using a multi-host download. One of the hosts seems to have been one of the RIAA servers that sends out static; for a few seconds in the middle of the song, there's this horrible (loud) clicking and popping. I have no desire to be the one to try it, but how cool would it be if I sued them for damage to my ears (when listening to the MP3 I downloaded from them) and won. It's actually not as ridiculous as it might sound -- if I steal a candy bar, and it turns out to have cyanide and razor blades in it, I'm almost positive that I could still sue / file criminal charges -- you can't 'booby trap' things if they cause injury.
As I said, it's a stretch, but I'd love to see the RIAA ordered to pay a tremendous fine for causing hearing loss / damage to speakers.
ferrous meweler (Score:3, Funny)
So it must be the author's contention that glaring irony doesn't compress well and so intensive consumption of his infotrash is juuuuust fine.
Did you guy read the rest of his pages ? (Score:3, Informative)
Just go to the above... The guy is a proponent of "logologie" "religion from the cybertime" (sic). The other article are nice too, like the one about natrium glutamate in cantine food ("nervengift"...). There is nothing to see there, just an Informatik Student in hamburg having fun (heck he isn't even a neurologue !) : Quote : Ich bin ein Kind des Cyberzeitalter, geboren im Jahre des Pong, und ich studiere Softwaretechnik an der Fachhochschule Hamburg (was leider den Großteil meiner Zeit kostet).
I am a child of the Cybertimes, born the same year than Pong, and I study Programming (software technic?) in the Highschool (not university something else) Hamburg (which cost me the biggest aprt of my time).
Move along, ntohing to see here.
Don't RTFA (Score:4, Interesting)
MP3's and other lossy compression (loss of quality through compression) methods change the distribution of frequencies along the sound spectrum, and maybe, just maybe, because nobody has proven otherwise, it might be the case that this can possibly have permanent effects on one's hearing. Maybe. Possibly. We dont really know. Neither do you. Or so we might think. Maybe. Oh yeah - here are a bunch of pictures from a biology textbook that look really cool, but are only connected to our speculation in a weak tangential unscientific way. Maybe.
. I haven't heard so many maybe's and 'might be the case' equivalents since the last 'In Search Of' marathon. And the article didn't even have Spock. .
like Mark Levinson (Score:3, Interesting)
It's like this: the ear is able to pull a lot of information out of natural, acoustic sound. There's regular features to such sounds that are distinctly different from plain random noise. The ear can dig into the random noise very deeply to get information out.
What these guys are saying is this: with certain types of distortion, the noise becomes opaque, the information just ain't there when the ear tries to dig for it. Soon it stops trying- or just gets out of practice. It atrophies.
There are a few points that are established (some recently) to support this, though the whole chain of evidence isn't there, and in fact it's a bit alarmist.
(1) Ears do adjust. If your hearing isn't symmetrical, your brain WILL construct a coherent picture from the sound field, despite the ear inputs not matching.
(2) Digital noise floors are NOT the same as natural white noise.
That last helps to support these wilder theories, but nobody that I know of had tested it until recently. I got in an argument on Usenet where I had to establish this. The argument was that dithering and truncation produced a noise floor different from the same signal with exactly equivalent white noise overlaid onto it. Basically, that quantization can be heard as a distinct character to the noise floor.
I had people very huffy about me even arguing this, because their digital audio theory demanded that dithered digital was perfect in every respect, and specifically that it behaved the same as analog noise w.r.t. detail retrieval beneath the noise floor.
I was given matching files- one being signal plus random-amplitude noise, and one being the same thing but quantized to the level of the noise, resulting in a normal TPDF noise floor, entirely uncorrelated. There was a 2 bit and a 4 bit example for me to try, because I was arguing that this difference was obvious at coarse levels, not that I could consistently hear it at 16 bits or something.
I did a computer ABX double-blind test, using both the examples, and got 40 out of 40 trials correct, establishing beyond reasonable doubt that these types of noise DO sound different. It's not even subject to debate anymore- that's what ABX is for- not asserting a negative but proving a positive beyond serious doubt. Dithered noise floors measure a lot like broad-band noise, and they may be uncorrelated, but they are absolutely not the same as simple random-amplitude noise (like you use for the dither signal prior to quantization).
I'm not aware of anyone doing this test before, but now it's been done and the point proved.
I am inclined to agree with the lunatic fringe here that it's the results of these very 'unnatural' processes which cause problems- they damage musical enjoyment, and they're part of why modern music is so commodified and worthless. The only serious mass media formats are prone to these problems. As a result, mass media itself seems less important- a self-destroying process. The sound alone contributes to a lessening of interest.
That said- anyone who had their hearing actually damaged by this effect would have to either live in an anechoic chamber or wear Walkman headphones every waking moment. The world is FULL of acoustic sounds- hell, traffic alone is an acoustic sound quite capable of 'recalibrating' the ear, and any face-to-face human contact often involves sound, which also 'recalibrates' the ear. So the alarmism is entirely foolish. Maybe Mark Levinson lives in an environment entirely free of any outside sound, I don't know :)
Pseudoscience (Score:3, Informative)
Perhaps our author really is state-of-the-art, but I see nothing in his article to indicate that. Everything cited can be found in beginner's texts on the subject. Nor is anything cited particularly relevant to his conclusions.
Let's not forget that the CD itself is a 'data reduced' sampling of a real world signal, at best an approximation of the original. And so was vinyl. I don't see many claims that the harsh approximations of the 33 1/3 LP are damaging ears by the very nature of their artifical reproduction... Unless, of course, you play them too loud
Living in a modern city, it's nearly impossible to not end up with some level of permanent tinnitus, and it worsens with age. However, there's an interesting paradox here: Background noise is required for the auditory system to function properly. Perfect dead silence, for prolonged periods, will also damage the auditory system-- through atrophy due to lack of stimulus (an unexpected discovery from a few fascinating experiments)
Monty
"Extraordinary claims require extraordinary evidence"
"Put down that uninformed pontificating before you poke out an eye"