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."
on how long they've been cranking their music up to 11.
Lacking <sarcasm> tags,
Actually... IMO some electronic music sounds better with lossy compression.
As it sounds more crunchy or crisp.
Or maybe it seems just louder. ;)
"I am the king of the Romans, and am superior to rules of grammar!"
-Sigismund, Holy Roman Emperor (1368-1437)
You blame the sound, I blame the people.
I think they should see if there is a correlation to the preferred quality, and how much auto-tuned "music" the people listen to.
Self proclaimed typo king, and inventor of the bear destroying coffee table (patent not pending).
As long as the sound is clean and there is no static, no pops, crackles, or hissing, I could care less what it is encoded at. To my ear there really is no difference.
Here's one such study conducted by the Audio engineering society:
http://www.aes.org/e-lib/browse.cfm?elib=14195
From the article: "We dragged 16 people", I'm no stats engineer but isn't that far too low ?
int main() { while(1) fork(); }
Thats strange, I find it trivial to identify differing qualities of compression when listening to my music files.
You look down at the UI, and it tells you what the bitrate is.
(Joking aside, I have advocated 128 kbps for years, not because of sound quality issues, but rather because most people own cheap computer speakers and/or headphones. You only get quality as good as the weakest link in the system.)
HA! I just wasted some of your bandwidth with a frivolous sig!
Title of article should be: 2/3 of people CAN tell the difference...
AccountKiller
Considering I can buy a 1TB drive for less than $100, I don't particularly care which percentile I might inhabit ... I see absolutely no reason to rip CD's at anything less than 320.
If libertarians are so opposed to effective government, why don't they all move to Somalia?
The tests have to be very carefully set up: double blind, very carefully calibrated audio levels.
Even a 1/3 db difference makes the louder signal sound sharper / higher quality. It's difficult to run a test that won't run into criticism about how it is conducted.
Many technical considerations for this kind of testing but also is the question "Is the difference in quality perceivable?" or is it "Given how people listen, does any difference between the two matter?"
That would be expansion -- lowering the noise floor. Compression would make it more audible. Expansion reduces the volume of all inputs *below* a threshold, while compression reduces the volume of all inputs *above* a threshold.
Also talking of that I wouldn't be too surprised if you're right (in what you meant) and that OGG employs a slight expansion to try and control its pre-echo. I like OGG better than MP3, though my iPod forces me to use AAC or MP3, but I believe it is plagued with pre-echo at lower bitrates.
Are you listening on quality equipment? I don't mean some overpriced audiophile setup with oxygen free cables, but at least a decent set of speakers or headphones and a good DAC?
When I switched from ipod ear buds to Sennheiser cx300's anything under 128kbps sounded terrible, and even 192kbps mp3 has some noticable artifacts. These are $30 ear buds, not anything excessive, but they sound a lot better than the cheap ones that come with iPods.
The key being that the two were encoded with two totally different codecs.
Exactly. And as such they are not comparable.
This certainly does not say a thing about the ability of people to distinguish between a good encoding and a bad one when the only information provided was the bit rate.
At best TFA is a testament to AAC. Says nothing about human ability to distinguish.
You can encode a phone call, typically limited to a frequency response of between, roughly, 350Hz and 3,500Hz at 192kbps. Probably 16kbps would suffice.
Sig Battery depleted. Reverting to safe mode.
I was going to say I have no idea why they would compare entirely different codecs here. Not to mention that lots of people are simply not audiophiles or not folks with extremely discerning ears to quality. Plenty of people show that AAC/Vorbis is situational and sometimes one can work better or vice versa.
As a musician, I've had lots of times where irrespective of my quality that I play people think everything is amazing/fantastic.
but my friend, who has been deaf since childhood, does listen too music but from the point of how it feels. His tastes weren't very different from many others of the time period (album rock which has distinctive beats/etc).
Really didn't crank the base, but it was not loud enough that people around him would stop and point, let alone gesture.
I only asked after asking why he played his music so much and my level ignorance about deafness was high enough to ask.
* Winners compare their achievements to their goals, losers compare theirs to that of others.
We did the same, ohh, 7 or 8 years ago. Took four tracks (a solo piano, a new Rolling Stones piece, a classical piece, something else), encoded them to 128/192/256 kbps CBR using the Fraunhofer codec of the day, converted them back to WAV files and burned them to an AUDIO CD. Each piece was put on the CD 5 times: The first was the raw track. The following four tracks were the raw track (again) and the 128/192/256 bit versions, in random order.
/frank
Everyone at work was invited to take the disk home, play it on their home stereo, and tell me what each track had been encoded as. This took "computer" items (sound cards, speakers, etc) out of the loop, and let them evaluate on the best system that they had. Being as this was an engineering company with a lot of high-ego types, there was some pretty impressive equipment out there.
50% of the people who took the challenge were unable to tell the difference between the encoding methods - they simply said "I listened to all five versions of each song, and they sounded exactly the same to me". Most of the others tried to assign bit rates to the various versions, but their results were essentially random - none of them reliably detected even the 128 kbps version. One guy was fairly confident in his results, and reliably detected the 128 kbps version of each song, but didn't make a guess on the higher bit rates as he couldn't tell the difference between them. One guy spent the evening with his spectrum analyzer trying to cheat on the test, but gave up.
That's when I stopped worrying about bit rates, especially when I spend most of my time these days listening to music in my car over the factory sound system.
And the worms ate into his brain.
I was actually a participant in this study, and despite their best efforts, their test was seriously flawed. Too many heads of other participants blocking the treble; the speakers were too directional; the amp was not even grounded properly, and as a result, there was a 60-cycle hum throughout. The only way I would trust a study like this is if they did it on people who actually have the listening experience, and in a controlled environment (headphones, and in an environment as close to anechoic as possible).
I can fairly easily tell the difference between V0 MP3 and the PCM original release, but I spend hours doing close listening on good equipment.
At the same time, I think people spend way too much time worrying about this shit. If you enjoy your music, and you've heard the difference between the CD and the MP3 rip and still don't care enough to re-rip in a higher bitrate or a lossless format, then good for you. Storage space is getting so cheap now that the argument that lossless formats aren't worth the space they take up no longer holds water. I'm a FLAC convert for many reasons, but most of all, the peace of mind that I (a.) am not missing anything, and (b.) won't be screwed over if a newer format gains popularity.
I want an electronic compression detector that will severely reduce the volume of all compressed audio. I would hook that to my tv speakers so that I don't get blasted out of the room every time commercials come on.
...
People deteriorate over time, too.
As a young man, I could tell a difference in the quality of recordings on vinyl or on tape. Today? Meh. I lost a good portion of my ability to hear just by being around 5 inch guns. Lost some more in industrial work environments. Lost some more to big trucks and heavy equipment.
When it comes to music, I'm a rundown old wreck of my former self. *sigh*
All that said, I suspect a lot of other people have similar histories, and yet more people have health problems that contribute to crappy hearing.
"Windows is like the faint smell of piss in a subway: it's there, and there's nothing you can do about it." - Charlie Br
Not to mention how poorly the results were considered. TFA states that over 1/3 of participants thought 48Kbps sounded better than 160 Kbps, but automatically assumed that meant 1/3 of participants couldn't tell the difference.
I'd be more willing to believe that 2/3 of participants can't tell the difference, and half of the 2/3rds just guessed wrong...
That is a sad state of affairs....
All your 09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0 are belong to us
Pop music is engineered to be played on cheap equipment. After all, that's what most people have. Practically nobody has ever heard Michael Jackson without a ton of electronics between them. You want a real comparison, use classical or jazz, where folks know what a *real* live performance sounds like.
It's also notable that the people who liked the lower bit rate recording said "more bass == better". "More bass" has been the "gold standard" in pop music for a good number of years -- the harder it punches you in the stomach, the "better" it is.
Welcome to the Turing Tarpit, where everything is possible but nothing interesting is easy.
Not all bits have equal significance. Example: Suppose I have a perfect-fidelity 30-second recording of a pure 440 Hz sine wave. On a CD this would require 16 bytes/channel * 2 channels/sample * 44,100 samples/second * 30 seconds = 42,336,000 bits.
This same signal can be compressed to the formula "L(t)=R(t)=sin(440*(2*pi)*t)*u(30-t)" (where u(t) is the unit step function { 0, t=0 }). Using string representation, this compressed version requires 25 bytes, or 280 bits. In other words, we eliminated 151,199 of every 151,200 bits—and the compressed version actually has better fidelity than the CD version, since it can be losslessly decoded to an arbitrarily high sampling rate.
This is obviously a contrived example, but it suffices to demonstrate the reducing the bitrate, even drastically, does not have to negatively impact the quality of the audio.
"The state is that great fiction by which everyone tries to live at the expense of everyone else." - Bastiat
It's not even "FM" quality, since even on FM, a song that is mastered well will sound very decent and authentic.
It's that the advent of compression in mastering, *in general*, has been so gradual that unless you knew about it beforehand, you wouldn't have noticed it.
Anecdote: I have two friends who heard me complaining about the shitty mastering and compression that has been happening in hard rock and metal for a while now. They finally made me sit down and explain what was happening, and then I found some examples on YouTube for them to listen to (of songs they knew, best of all).
Now they hate me, because now they *notice* it, but can't do anything about it :P
You can have all the greatest audio equipment in the world (leaving aside the argument about audiophiles and their dubiously useful accessories and beliefs about what results in "better audio quality), but the real problem is at the mastering stage, not the consumer output stage.
It's not just bitrate, at this point. They're comaparing 160kbit Vorbis to 48kbit AAC+.
AAC+ utilizes two tricks to make low-bitrate audio sound better: parametric stereo, and spectral band replication.
The first, parametric stereo, stores the audio as monaural with an extremely low-bitrate sideband (2-3kbit/s) to store stereo information.
The second, spectral band replication, stores half the frequency explicitly (low and midrange). The upper frequencies are then recreated from shaped noise, which works quite well.
These two techniques are psychoacoustics taken to the extreme. They're incredible at low bitrates, but useless at higher bitrates.
While I agree that part of it can be blamed on the Loudness War, and that is one of the reasons I prefer buying CDs from local artists as they don't compress the hell out of their music, I have always thought that the bitrate should depend on what you were going to do with the music.
For example, while I'm sure the audiophiles would have a coronary all the music on my MP3 player I've got encoded at 64k. Why? Because I use it when I'm out and about and frankly with all the outside noise I really can't tell enough of a difference to make the larger files worth it. With 64k I have 75 hours worth of tunes on my 4Gb Sandisk, with 1.5Gb left over for adding new stuff. The same goes for my Truck's Sony MP3 CD player, which sounds great but with me needing to hear traffic noise it is more "background" music than actually listening, and the smaller size means i can have just about every song I like to listen to on the road fit onto two CDs.
Now when I'm at home listening on my nice phones I'm listening at 256k, but that is because ALL I am doing at that time is enjoying music and can actually listen to it closely and tell the difference. But if most folks are using these services like radio stations it is probably more "background" music than anything and frankly why waste the bandwidth on something you can't really tell the difference on? Most folks aren't gonna be using audiophile equipment at the office anyway.
I think for most of us, especially those that grew up in the age of scratchy records and fuzzy radio stations most of this stuff is "good enough" for the things we are doing with them. But after helping a friend convert his old LPs to CD and comparing them to the CD "remasters" I have to say what they've done to older music is obscene and has nothing to do with bitrate. They have compressed the older rock albums so damned much thanks to the loudness war that when compared to the analog original they sound like they are being played through a bad compressor foot pedal. All the dynamics are gone and it just sounds...well it just sounds nasty. Compression artifacts from using low bitrates is one thing, but when my 64k rip of his analog still sounds better than the CD thanks to all the compression they used on the master, well that is just sad. Listening to classic Rush and Styx VS the CDs were just like night and day.
ACs don't waste your time replying, your posts are never seen by me.