It's empirically true that some people prefer MP3 artifacts in blind listening tests, some of the time. (Most people don't.) "Why" is an interesting question, though I don't know the answer.

Heh. Maybe it's because the people under discussion in this thread have only recently learned to program (i.e. in the last 10 years). Certainly anyone self-taught who learned in the 80s or 90s couldn't avoid learning how to implement data structures or write algorithms.

Many can't do pseudocode or understand what big-O notation means because you never encounter it unless you've taken an algorithms class.

Not true at all. I'd expect anyone who was competently self-taught to understand both.

i don't think that's true. People should be expected to raise themselves to minimum standards, not meet them ahead of time. After all, it's basically effortless to look it up and learn what it means, and lazy evaluation in reading slashdot doesn't have any negative consequences I can think of.

Stallmane and GNU ***DO*** have the last word on what the GPL says. What do you think the "G" stands for?

I don't know. I tried a depth-first evaluation, but it's not done yet!

Were you able to change her back!?

*My* grandmother can use them just fine. My mom and grandmother are both technically inclined. (My mom is 71 now, and is a grandmother herself.) Being elderly does not by itself cause computer illiteracy, just as being young doesn't magically cause someone to understand how computers work.

Great! Now you just need 49 more fiancees to prove GP wrong!

In all seriousness, a list of requirements by state are online here. Although many states do require ID, I didn't find any that required a driver's license.

I can attempt to explain two things. First, you can beat the time-frequency uncertainty principle if you're willing to be wrong sometimes. The ear does this, functioning foremost as a wavefront detector.(*) Second, most sounds including the human voice follow an approximation of the harmonic series. (Always an approximation; sometimes, it's not a very good model at all.) So you can detect the upper partials and reconstruct the fundamental if the audio in question fits the model well enough and the harmonics are present and measurable. Again, this works by being wrong some of the time.

I found an article detailing how the Guinness record was measured here. It was only measured for nine seconds; this gives us a (minimum) bandwidth of .1Hz, which at .0189Hz would be within error around 10 semitones up or 30 semitones down (though I had to clobber the numbers pretty hard with the error bar), keeping in mind semitones are separated by a factor of 2^(1/12). The transform to frequency domain was further inaccurate due to the window size, and the 2270 is only specced down to 3Hz in any case, so the measured numbers probably contained a generous helping of error.

So while I'm no expert, it looks like the the bandwidth of the measured sound definitely exceeds half a semitone in either direction, probably by at least one order of magnitude.

(*) Hartmann, W. H. (1995). "The physical description of signals," in "Hearing," Edited by B. C. J. Moore, San Diego, Academic Press, 1-40.

And no, you can't tell the difference between a vinyl and a digital recording, and if you can, I have some gold-plated audio cables to sell to you.

I know what you mean, but I want to clarify your point:

• You can't tell the difference between vinyl and a digital copy of that vinyl;
• You can tell the difference between a digital recording and a vinyl copy of that digital recording, because it degrades in a characteristic manner.

Quite a lot. I come from a family of readers, and my mother in particular collected science fiction and fantasy for most of her life. She's still alive (and now 71 years old!), but I took a lot of the books with me when I moved out.

On the other hand, my mother listens to nothing but church hymns, and my father nothing but marches. I like a lot of music, but answer to your question is zero in that particular column.

True, though the selection of "difficult" material doesn't fit neatly into "classical stuff or electronica" like you suggest. And of course, 320kbit MP3 usually works pretty well on both of those.

(It's also true that some extremely trained listeners can beat chance in distinguishing 320kbit mp3s from originals. But this is only barely true.)

More usefully, it's nice having uncompressed audio so you can do things to it without noticeable degradation, as many activities involve an encoding step, and MP3 isn't designed for tandem encoding.

Just a nit: is nit not? Or is it not?

I suppose that depends which frequencies you mean by "high".

Noise-induced hearing loss doesn't usually hit the very high frequencies first. Usually it fits Fletcher-Munson, so you get a notch around 4k first. (It varies.) That's still fairly high, of course, but it's distinct from the very high end loss (15k+) that tends to come with age.

MP3 does suck compared to more modern codecs. You haven't done the research yourself, though.

It's true that extremely trained listeners can beat chance comparing a 320kbps MP3 to a CD, but that doesn't mean even they can consistently tell the difference. It's very tough, even for these people. And you're almost certainly not one of them, since you think you can "definitely" hear the difference, which is not supported by any research I'm aware of.

You also meaninglessly specify a bitrate (300+) without mentioning which codec you mean--is it AAC, OGG, or "others"? In the case of AAC, for example, you're unlikely to be able to tell the difference above 160kbit/s.