drich14 said:

  I tried to put a variety of genres into the test to see if there is any correlation between genre and ability to see a difference in quality.

Click to expand...

.. From where do you get the original FLAC?
 
I think that if the music is very bad recorded, or the equipment used is very bad you cannot tell.
 
I was able to tell using 200$ speakers and under 100$ headphones FLAC from mp3 at 128kbps. 
 
In fact, I am able to tell them apart using my laptop speakers, just tested, and i am totally able. The cutoff is way too early made to not be noticeable, 128 mp3 is too muffled in mids too, i cannot hear parts of female voices and guitars. 
 
Something seems wrong for you to get such results.

drich14 said:

  Actually what I've found so far is that people don't get any better scores than just guessing randomly.

Click to expand...

 
I've never had trouble telling the difference with at least some content, however it will probably depend on the source material, the speakers and headphones being used, and the hearing of the person taking the test. Generally things with lots of high frequency content makes it more obvious that simple vocals, and complex music involving lots of instruments suffers more from the MP3 compression than solo instruments. Different encoders also have different priorities in terms of what they omit, so certain encoders may cause more damage to certain types of music and favor others. (The decoder is standard, so all decoders should play the same encoded file pretty much the same way, but the encoders have significant flexibility in terms of deciding how to analyze the incoming music, and of deciding what to omit.)
 
You will also find that people tend to respond differently depending on how you phrase the question. (For example, some people interpret "can you hear any difference" as meaning "can you hear a significant difference" rather than the question you actually asked. Those people may then fail to mention subtle differences that they can hear but don't consider important to them.) Some people also simply don't pay attention, and/or don't try very hard, and so only report very obvious differences.
 
What I would love to see is a comparison.... Ask one group to do an ABX and see how many they get right, just as a "normal survey", then have a second group perform the exact same test, but offer them$20 if they can get 15 out of 20 correct and see if they do a better job - because they really concentrate and pay attention. I would be very interested to see if a little motivation helps (or not). You should also be very careful not to imply that the MP3 will necessarily be worse. (I read of at least one survey where a significant minority actually heard a difference - and preferred the MP3 files.)

exacoustatowner said:

Prot- EVERYBODY knows that was Lemuria/Mu!
Jeez, have you even aligned your Shakras with your Reference Rose Quartz?

Click to expand...

Guess you are right about Mu, my mistake. In my defense it is pretty hard to differentiate between invisible ghosts. Maybe if I had better cables...

keithemo said:

 
What I would love to see is a comparison.... Ask one group to do an ABX and see how many they get right, just as a "normal survey", then have a second group perform the exact same test, but offer them$20 if they can get 15 out of 20 correct and see if they do a better job - because they really concentrate and pay attention. I would be very interested to see if a little motivation helps (or not). You should also be very careful not to imply that the MP3 will necessarily be worse. (I read of at least one survey where a significant minority actually heard a difference - and preferred the MP3 files.)

Click to expand...

 
That's very interesting and a great idea! I actually made a bet with one of my friends that they wouldn't get above a certain score. He ended up winning the bet so it appears that concentration and motivation is a big part in hearing a difference.

drich14 said:

  Hello audiophiles,
 
I am brand new to this forum so feel free to tell me if I'm doing something wrong.
 
I am a high school senior in AP Statistics, and for my final project I have to perform a hypothesis test. I decided to test if people can tell the difference between 128kbps mp3 and lossless wav music. I made a website to run my experiment and all I need now is some data. Someone who works with my parent suggested this website for getting a lot people to take my test.
 
It would be greatly appreciated if some of you who read this visit my website and take the test. I am very open minded, so if anyone has any constructive criticism I would be happy to hear what you have to say.
 
Here is the url: http://www.audioexperiments.com
 
Thanks for your time,
drich

Click to expand...

This test would be more accurate if rather than simply saying "which sounds better", you had a reference file to compare to. So, for each song, you'd have a lossless file, a sample A, and a sample B, and you had to see if people could tell which one matches the lossless.

cjl said:

  This test would be more accurate if rather than simply saying "which sounds better", you had a reference file to compare to. So, for each song, you'd have a lossless file, a sample A, and a sample B, and you had to see if people could tell which one matches the lossless.

Click to expand...

Yeah I bet that probably would have been a better solution now that I think about it.

That's pretty much the standard way to tell if people can hear the difference, since it doesn't require them to make a subjective judgment of which "sounds better", it merely asks them if they can hear which one matches the reference. In the way your test is laid out, it's entirely possible for someone to clearly hear the difference between the two, prefer the lossy, and then get the answer "wrong" because of it.

I'll have to talk about this alternate design in my paper.

drich14 said:

  I tried to put a variety of genres into the test to see if there is any correlation between genre and ability to see a difference in quality.

Click to expand...

Genre makes no difference... it's types of sound. I've found complex masses of sounds are the hardest to compress... audience applause, certain textures of massed strings. Also pure high tones like flutes. It really doesn't matter if the recording is stereo or mono, digital or analogue, certain types of sounds will cause more problems than others. I have one track from the 1950s that is extremely hard to compress without artifacting. It will clearly reveal artifacts all the way up to 192.

bigshot said:

 
Genre makes no difference... it's types of sound. I've found complex masses of sounds are the hardest to compress... audience applause, certain textures of massed strings. Also pure high tones like flutes. It really doesn't matter if the recording is stereo or mono, digital or analogue, certain types of sounds will cause more problems than others. I have one track from the 1950s that is extremely hard to compress without artifacting. It will clearly reveal artifacts all the way up to 192.

Click to expand...

I find that 320 mp3's are pretty safe to use. I'll bet that 256 is probably good enough.

Yes, and AAC is a notch better than MP3 in most cases.

Have you guys seen those reports and studies that show many people actually prefer (pick in blind testing), compressed files over Non compressed? I suspect it's because digital files and digitally recorded music tends to have a strident upper mid and an overall flat presentation and the mp3 softens this and is a possible reason the common man may pick an mp3 out as being more pleasurable a listen. This of course if true is not any justification for compression, only more evidence of the failings of digital of which anyone who spins vinyl (original pressings or selected modern) can attest to. Music if it was analog recorded and kept that way till final listening, would likely reveal that compression actually takes away enjoyment because the analog infomation does not have the strident flatness and doesn't benefit from softening compression. Just a personal theory and one's personal experience.
 
The entire hobby would be better off if it recognized analog vs digital as different beasts and did not criss cross this stuff. Jimmy page putting zeppelin's original analog master tapes into the digital realm so they could be manipulated, then putting them back to analog for a vinyl press is a joke. Go listen to one of those 2014 remasters and compare to a Pallas or an orignal or even 2nd or 3rd pressing from the 70's. It's no comparison. Steve hoffman does good work as well and there are other decent modern presses but they are few and far between. Once it goes to digital, (or if it was digitally recorded), all the good stuff is lost. The ambience, the detail, the ability to hear the tone from 360 degrees around etc. I feel headphones and amps and DAC's and even speakers should be designed for one or the other. This would keep things neat and orderly and end up in better gear for both those who listen to digital and those who listen to vinyl. For example, a DAC or a headphone can be designed to deal with and try and compensate for the digital stridency and flat dynamics (and in fact I feel alot of gear is trying to do this, but it's not done intentionally). By designing for one medium vs the other, it can be made intentional and whenever something is cosncious and intentional, it has more potential for positive end results generally speaking.

thelostmidrange said:

  Have you guys seen those reports and studies that show many people actually prefer (pick in blind testing), compressed files over Non compressed?

Click to expand...

 
I've never looked to deeply into it but I'd suspect it's just because they preferred what they were used to.
 
As to your other point, there's a link in the OP that does a pretty good job of refuting it.
 

prog rock man said:

  9 - Boston Audio Society, an ABX test of Ivor Tiefenbrun, the founder of Linn. August 1984
 

A rather complex testing of Ivor Tiefenbrun himself, who at that time was very pro vinyl and anti digital (the opposite almost of how Linn operate now!). There are various different tests and the overall conclusion was
 
"In summary, then, no evidence was provided by Tiefenbrun during this series of tests that indicates ability to identify reliably:
(a) the presence of an undriven transducer in the room, 
(b) the presence of the Sony PCM-F1 digital processor in the audio chain, or 
(c) the presence of the relay contacts of the A/B/X switchbox in the circuit."
 
http://www.bostonaudiosociety.org/bas_speaker/abx_testing2.htm
 
Even the founder of Linn could not back up claims he had been making when subjected to an ABX test of those claims.

Click to expand...

@drich14 Why are you using 128 kbps mp3 files?

@maverick I've not the time to look into it. But I can say it's possible and quite easy to make a vinyl setup sound like a digital setup. And many modern turntables, carts, phono pres actually tend in the digital sounding direction and so in that case, it may be possible to setup an A/B that makes it hard to discern. There's so many points and variables that could be compromised in testing sessions that it's a waste of time imo. All I know is I had many Cd's of the same album as I have on album and they where totally different. But all my collection is classic recordings that where meant for vinyl the the CD was an afterthought. Still I feel analog tape recording of music is the only way to go. Whether you then play it back on digital or analog is less important. But then again I'm only referring to classic instruments, wind, strings, percussion, etc. Once you add any digital info the whole thing changes and it's a different thing. I'm what's called a purist and believe that music made by people with instruments should be recorded one way and everything else, can be done in another. Have you ever heard coltrane or miles, or hendrix on a vintage vinyl setup? and compared that to a CD? That's the kind of basic thing I'm talking about. Apart from that things can get super messy and I'm sure it's possible to get digital sounding good somehow. All I'm saying apart from the importance of analog recording, is that those records should then be played on gear designed specifically for them specifically. i.e. the designer of such a headphone should only use vinyl as his tuning reference. I'd suggest he use a mix of tables and gear of course, but all analog. He shouldn't use any digital recordings or music at all and should not try and make his gear sound good with such stuff and in fact I think it's impossible because digital music has no reference to the real world and has no limits or boundaries as to what sounds natural. How would an engineer know when he's got a digital sample of a computerized sound, sounding natural? never. But you can have a reference for a saxpophone or acoustic guitar or electric guitar and you'll know when the recording reflects it naturally or not.
 
Think about it, what companies design gear specifically for analog recordings explicitely? I'm sure some do, but it should be much more prominent and intentional, and marketed as such imo. I predict vinyl and analog recording of music will make a strong comeback in the future, and this divide of analog will become more explicit. I've got a buddy who's got an all analog studio and records local bands through it for example and he's dedicated to keeping this particualar sound and technique pure and alive. Id recommend some headphone designers to the same