Slashdot Mirror
1/3 of People Can't Tell 48Kbps Audio From 160Kbps
An anonymous reader writes "Results of a blind listening test show that a third of people can't tell the difference between music encoded at 48Kbps and the same music encoded at 160Kbps. The test was conducted by CNet to find out whether streaming music service Spotify sounded better than new rival Sky Songs. Spotify uses 160Kbps OGG compression for its free service, whereas Sky Songs uses 48Kbps AAC+ compression. Over a third of participants thought the lower bit rate sounded better."
Are these the same people who prefer MP3 Sizzle?
Remember RFC 873!
(although not as low as 46kbps) and reached the same conclusion. Most people vastly overestimate their ability to distinguish tracks encoded at different bitrates. And I've seen study after study that backs this up. This includes self-professed audiophiles, the original authors of particular tracks of music, and so forth.
Mr. Wizard... why is this place called the Cave of Hopelessness?
So, 1/3 of people eh? Hardly a damning assessment when your sampling size is 16 people. Besides, most people I know including myself have some sort of hearing damage from the past or don't really know what to listen for when presented with different types of sound.
I say the only valid comparison is listening to the live music, vs the digital format. This way you compare to the original and your not just saying which sounds better (which is subjective). I once worked with a audio system designer and everything was tested using analogue formats with various types of music preferably classical because of it's range in sound.
"Of the 16 people tested"
Good-bye.
I do not believe in karma. "Funny"=-6. Do good and forbid evil. Yours, Oft-Offtopic Flamebaiting Troll.
And I've been telling people for years that the "weakest link" concept in audio reproduction is an oversimplification and therefore wrong.
There are orthagonal distortion components introduced by various devices. An MP3's digital distortion (sizzle sounds, to borrow from another article somebody linked to) would be IN ADDITION TO poor frequency response and mechanical distortion. It isn't "masked" by it. And it doesn't take significantly more bitrate to go from "crappy" to "great." 128kbps CBR MP3 is pretty crappy, but 160kbps VBR MP3 is indistinguishable from the source "even on great systems." I don't intend to argue what bitrate you consider "sufficient," just that "Listen to a low bitrate because you have crappy speaker" implies that crappy speakers mask MP3 compression artifacts.
If I were to go out on a limb, I'd say its possible for crappy speakers to distort even more with overcompressed MP3s than good speakers do.
CNET - Owned by Rupert Murdoch.
Sky - Owned by Rupert Murdoch.
CBS, actually:
http://www.techcrunch.com/2008/05/15/why-cbs-bought-cnet-and-not-the-other-way-around/
If it'd been a Myspace survey or something from the Times, the Courier-Mail or the WSJ, you'd have had a point.
Except that most of the compression gained from mp3 is gained by removing frequencies we can't hear anyway, speakers with poor frequency response absolutely 100% do mask this.
Jesus I'd never seen those. Just for those of you at home, I'm a professional sound designer for films, and I use ethernet cables that I bought at Fry's for a couple bucks a piece.
But seriously, can you make a sweeping statement like "People can't tell 48k audio from 160k"
The issue isn't "can I tell 48kbps from 160kbps" -- the real question should be: can I tell the difference between 48kbps AAC and the original uncompressed recording? AAC can sound "better" or "good" under a lot of situations where it's significantly distorting the original program material. AAC was designed specifically to choose "good sounding" over "accurate" as the bit rates get lower and lower. Also, keep in mind that a side-effect of compressing an audio stream like this is that you'll strip away noise and unusual harmonics from the original, which might cause a lower-rate recording to "sound better," when in fact stuff that the producer actually has in his mix is being removed.
Don't blame me, I voted for Baltar.
Over a third of participants thought the lower bit rate sounded better.
Another thing is that majority of people actually have quite crappy speakers, atleast on computers. Lower bitrate sounds "better" on cheap speakers because it dumbs down highest frequency changes in the song.
Blaming people is a pretty good start.
To be more specific, the physiology of humans. We can hear up to about 20kHz. Basic theory tells us that we require a sampling rate of twice the highest frequency, so that's 40kHz. You throw in 20% extra and you're sitting at 48kHz.
Anything more is filler.
Now, let me qualify that for a moment -- some codecs are terrible, and I can often hear the phasing on higher pitches, notably on cymbals. I have excellent hearing and have more than 20 year of musical training in both brass and choir. (I have to listen to mp3 players on minimal volume.)
However, when music is background music it doesn't really matter how perfect the sound is. If there's a hearing loss epidemic caused by the prevalence of, say, cheap mp3 players, then most people are probably halfway to tone deaf anyway.
---
ECHELON is a government program to find words like bomb, jihad, plutonium, assassinate, and anarchy.
it said 48Kbps, not kHZ.
Most lossy music formats totally submarine a lot of detail at 48Kbps and I would wager that almost everyone has the auditory acuity to recognize it. They simply don't have the mental acuity to care.
I agree, so much auto-tone (big air quotes) "music" and they hardly notice gross clipping and drastic tone flattening. :-)
That would be true, except that even crappy computer speakers these days can produce high frequencies just fine. Consider the following speakers that are among the least expensive on Newegg. They have an advertised frequency response of 100hz to 20,000khz, plenty of range to reveal encoding flaws. Yes, the actual frequency response might not be as good as advertised, but if they're anywhere close, they will not have any trouble revealing encoding flaws.
In my experience, medium-high frequency reproduction is probably the chief problem with poorly encoded music. From the article, "Some also noted that cymbals, hi hats and vocals in particular sounded better" (referring to the better encoded stream). Cymbals and hi-hats are dead on - they end up sounding like 60s sci-fi if encoded badly. Even the most modest of computer speakers and earbuds will reproduce a cymbal frequency range without breaking a sweat.
The grandparent is dead on here - sound reproduction is not a chain, it's a relay race. Any particular member of that race can single handedly improve or worsen the reproduction.
No. This would only be true if the speaker masked the exact same frequencies as the MP3. In this case, you are losing frequency content at the source (lossy MP3 file) and are AGAIN losing frequencies at the speakers. I have found, at least in my experience, that low bitrate stuff is even more unbearable on low end gear than on better systems.
To the haters: You can't win. If you mod me down, I shall become more powerful than you could possibly imagine
This test isn't a complete experimental fiasco (like some of the Microsoft-sponsored listening tests that deem WMA to sound as good at 64k as MP3 at 128k).
But there are a couple of significant flaws with it, that make the results pretty useless:
If you want to know about some methodologically-better comparisons of audio codec quality, please see the Codec listening test page at Wikipedia. Full disclosure: I wrote most of this article, and have attempted to compile the results of all the carefully-conducted independent tests that I could find.
Finally, none of this is to say that we should all demand 160kbps streaming audio if 48kbps can be made to sound just as good. It's just that this study doesn't establish that, not by a long shot. The headline is also wrong in claiming that 1/3 of the participants couldn't distinguish 48k from 160k audio: in fact, they preferred the 48k audio. And preferring one format is very different from claiming that it is of a high-fidelity: for example, audio with a compressed dynamic range is by definition degraded, and yet it persists in commercial rock recordings because uniformly loud music grabs listeners' attention more easily.
My bicyles