Below is an English translation of a recent article "Gefahr aus dem Kopfhörer" (The Danger From Headphones) written by Matthias Hohensee over at Valley Talk. His article refers to my recent investigations into whether younger generations prefer lossy MP over higher quality music file formats. The preliminary results of that study were reported in the article I recently posted called, “Some New Evidence that Generation Y May Prefer Accurate Sound Reproduction”.
Matthias makes a good point about listener preference for MP3 becoming a moot issue with higher quality file formats becoming the standard, as bandwidth and music storage costs drop. I only briefly mentioned this in my slide presentation (see slide 7), but it deserves repeating. The days of low quality music downloads are numbered, I hope. Then, the main sound quality issue will become the recordings themselves, and the quality of the headphones and loudspeakers through which the recordings are heard. What are your thoughts on this matter?
The Danger from Headphones
by Matthias Hohensee
from Valley Talk 6.30.2010
Can the Germans be really proud of MP3 or has the digital stroke of genius desensitized the hearing of a complete generation?
When Angela Merkel recently visited the prestigious Stanford University in Silicon Valley and enumerated German technology services, she also mentioned the data compression method MP3. The technology that was largely developed by scientists at the Fraunhofer Institute has changed the music industry, even though it’s mainly U.S. companies that profit from the MP3 player market.
But can the Germans be really proud of MP3? Or has the digital stroke of genius desensitized the hearing of a complete generation? At least the observations of Jonathan Berger suggest this. Over the years the Stanford professor of music has been asking his students if they are satisfied with compressed music files, or if they prefer the full Hi-Fi sound.
He came to a surprising result: For years, the number of those who preferred the sound of ‘packed’ music to the uncompressed audio spectrum seems to grow steadily. Berger concludes that the taste of sound has changed.
Good sound is measurable
Sean Olive, on the other hand, considers Berger's insight as nonsense: "Good, accurately reproduced sound is not a question of taste, but scientifically measurable." And this is the way he should see it. After all, Olive is the head of acoustic research at Harman International. The U.S. manufacturer is considered to be THE address for sophisticated sound systems.
Alarmed by Berger's observations Olive recently invited Los Angeles high school students to the Harman studios for extensive tests. "Everyone could hear the difference between different compacted sound files - and preferred less compressed songs," says the scientist relieved.
Danger from headphones
Now, Olive is not really unbiased, after all Harman sells nearly three billion dollars worth of high-end audio technology per year. But in fact, technical progress makes Olive's worries already obsolete. In times of high-speed Internet, data compression does not play the same role as it did in the nineties when the music piracy supplier Napster made MP3 popular.
The songs that were exchanged back then were extremely compressed in order to distribute them via the still slow Internet connections – but also to spare the limited memory of computers and MP3 players. Today the vendors such as Apple and Amazon are selling songs which are formatted in such a way that only real audiophiles can hear the difference to music CDs.
And so the real dangers for the hearing of ‘generation iPod’ aren’t the highly compressed music files, but simply the volume adjustment of their headphones.
Acknowledgements: Thank you to the author Matthias Hohensee for permission to repost his article here, and to Alena Winterhoff for the English translation.
Hi Sean,
ReplyDeleteI've heard anecdotal evidence that lousy headphones put listener's hearing at risk because the listener cranks the volume up in some attempt to hear the sound.
I suspect this would almost certainly be true of open earbuds vs. in-canal headphones. But, does the difference hold for nice, flat in-canal headphones vs. lousy frequency response in-canal headphones?
Stated another way, what are the average SPL levels that people adjust to across various headphones and environments? Are there any good studies on that?
Actually the temptation to listen at significantly higher volume levels compared to speakers is commonly attributed to the low levels of distortion produced by high-quality headphones even at high volume. Thus, one would expect the endangerment of hearing to be positively correlated with the technical quality of the headphones.
ReplyDeleteHi Sean
ReplyDeleteThere was an interesting line of thought I came across somewhere that argued that selling music online would become commoditized over the next 5-10 years. Then like any fiercely competitive capitalist marketplace, companies would be forced to compete on customer service, product excellence or brand. This would naturally drive up audio quality over time. However, I do feel like there is a quality barrier which companies providing music online may not need to cross. When it is difficult to tell the difference between a 320KB AAC and a uncompressed WAV then there may not be any incentive to offer higher quality...
I always tell people to make an MP3 version of a CD tune they love and to phase reverse one of them in ProTools, GarageBand or whatever they use, for them to hear the difference of what information they are missing from the CD version. I personally don't like MP3's, and I'm a young person. I'm on my way to converting people away from MP3's. It might be a good idea for you to do a quick blog post with some short audio bytes of CD quality AIFF, MP3, then the difference of information missing. Most people are shocked how much is missing!
ReplyDeleteHi Caleb,
ReplyDeleteI have not heard that people crank up their headphones to compensate for the fact that they are lousy. Perhaps they might do this if they lack bass (a common problem), but who knows for sure. One would think the added distortion at higher levels, would want you to turn them down - not up.
I would tend to agree with you that people will listen a lower SPLs if the phones provide good isolation to background noise, and I think there is data to support that. The well-sealed in-canal types like Etymotic ER4 provide > 25-30 dB noise isolation, which is what I wear on planes to avoid having to crank up the level. If I need more bass I tend to use the bass EQ in the Ipod rather than turning up the overall level.
There is some research indicating that some adolescents listen to Ipods at high levels for extended periods because they are almost addicted to listening that way and cannot stop because it makes them sad or depressed. Interviews with these people indicate they do not clearly understand the risks of permanent damage. It's an educational problem.
Hi Jeremy:
ReplyDeleteThat's a pretty effective trick to demo how well perceptual masking works. George Massenburg has been doing those type of demos, as well lately.
It's a good idea to put some demos on the website of poor/medium/high quality lossy versus lossless. Maybe we can incorporate this into our listener training software as well.
hi Anonymous,
ReplyDeleteThat makes sense to me. When I hear speakers with low distortion (e.g. JBL Everest) like find myself listening at higher SPL because there are not the usual distortion cues to tell me its too loud or too unpleasant.
We've also found that removing low frequency whole body vibration from the reproduction can result in the listener turning the bass up in the headphones 2-3 dB - depending on how much vibration is missing.
Hi Nyal,
ReplyDeleteThere are record companies now offering albums downloads in 4-5 different formats, and you pay according to the bit depth/sampling. I recently purchased this album in 24/96 FLAC which was 1 of 5 choices:
http://www.linnrecords.com/recording-highwire.aspx
Of course, there are lots of controversy whether there are audible benefits beyond CD-quality but that is a topic of a future blog.
There is not much sense paying for 24bits if the effective dynamic range of the recording is 2-3 dB from loudness war processing.
A bad recording doesn't sound any better whether its 256 AAC or 24/192. Right now, there's just a lot of bad recordings out there. Quality must begin well before the recording ends up at the mastering phase.
Nulling a degraded signal with the original is not a valid way of determining or proving audibility of a degradation. Specifically: provided the masking model is accurate and the bit-rate is adequate (two provisos often forgotten both by pro and contra camps!) the difference is fully masked when the actual signal is present. That is what the whole lossy coding thing is about! Of course masked content becomes audible when you remove the masking content. Listening to the difference alone (or exaggerating it over the actual content) is good to sensitize test subjects to what they should be listening for, but in the end determining actual audibility of a coding method should be done with the whole signal.
ReplyDeleteBruno,
ReplyDeleteAgreed. I assume people who do these nulling demos know this, but thanks for pointing this out.
it also depends on what codecs are used, some lossy formats can be nearly transparent at certain bitrates, for instance, itunes offers AAC in 256kbps, if you rip a CD to AAC at 256kbps and then do one of those out of phase listening tests you will hear mostly non musical sounding data, similar to white noise with a rhythm. i am unable to tell a difference between CD's and AAC above 160kbps, AAC is entirely different from MP3 as well, with MP3 i find anything below 192 to be noticeably different from the original. as far as people cranking the volume i would suspect this is due to the weak bass capabilities that cheap headphones or speakers have, the threshold for low frequencies is much higher then mid and highs so therefore in order to even hear those notes the volume must reach a level that the headphones are able to play those frequencies at the threshold so they are heard. I myself know that before i had a subwoofer when i used speakers only good down to like 80hz i would crank the volume much higher, with a sub, i don't feel the need to do that.
ReplyDeleteSean
ReplyDeleteI eagerly await the time when one can download their choice of music online in any desired bit depth/sample rate up to that at which the studio master runs at. Unfortunately beyond pure audiophile music e.g. Linn there is little hi-rez content available. For example I bought the Bing Crosby box set by Mosaic. I saw on the packaging that it was transferred over from these pristine analog masters at a high bit rate yet only available to us on CD :(
The controversy I am refering to is whether we can judge the difference between 320Kb AAC or WAV on something like a iPod. I have a good pair of in ear headphones and it is REALLY difficult. Obviously plug it into my system at home and I can tell the difference.
Nyal
Hi Sean,
ReplyDeleteFantastic blog. You wrote that you use the sealed Etymotic ER4 headphones on the go. I know Harman owns AKG but I'm curious what headphones you use at home.
Thanks,
Scott
"In times of high-speed Internet, data compression does not play the same role as it did in the nineties when the music piracy supplier Napster made MP3 popular."
ReplyDeleteWow... what an unjustified and inaccurate statement, with no evidence behind it. This writer should see - and hear - as many iPods as we do. We use a shop iPod using Lossless tracks for demo purposes, and then we test the vehicle owner's iPod. This is to forestall complaints about why the iPod doesn't sound as good as the CD player.
The main agitator for higher compression rates now is lower cost nano players and higher-cost, but lower-storage, iPhones used as iPods. It's not the download speed, it's the stoage capacity.
My recent AES presentation, "The Loudness War: Background, Speculation and Recommendations," included a number of slides about headphones and possible hearing damage. A video based on the presentation is at http://www.sfxmachine.com/docs/loudnesswar/ .
ReplyDeleteHi Earl,
ReplyDeleteThanks for the reference and link. I missed your presentation at AES so this is useful.
Hi Ken,
ReplyDeleteYou make good point regarding small flash-memory storage capacities in Ipods resulting in higher compression. I still use uncompressed Apple Lossless on my Nano and Classic Ipods -- but I am probably the exception.
Don't you think though that even on these limited storage devices the average bit rate of music has increased over the past 3 years given that iTunes is now standardized at 256 kb/s AAC ??
Eventually internet streaming (assuming the bit rates are decent) and/or larger storage capacities in Iphones will result in people listening to higher quality digital music files -- at least we can hope.
Sean,
ReplyDeleteYou are not the exception. I converted my over 2000 classical CDs to Apple Lossless, listening to some of them at a time on my iPod Nano and iPod Touch. Good article and information. Thank you so much!