Can you hear the difference between 128kbps mp3 and 320?

[Patrick Burns]

As a rule, I always choose incorrectly on these tests. I have subsequently begun to care less about bitrate and searching for uncompressed audio (unless I want to do some audio editing/canceling).

My high frequency hearing loss is average for my age (19), I spend a reasonable amount of time working on my mixing/mastering skills, and yet these tests usually fool me.

[dannthr]

You've got it backwards. Just because a spectral analysis shows loss does not mean there's an audible difference. Do an analysis of a 256kbit MP3. There will be loss compared to the original, but you will not hear the difference. MPEG Layer III represents audio perfectly (to the human ear, of course) at that bitrate.

Stop while you're ahead. If you're only using software analysis tools then you're wasting time. Whether your hypothesis is right or wrong, it's useless information. More/Less frequency loss does not equate to less/more accurate perception. Perception is what matters.

Subwoofers are intended to represent the last octave of human hearing, i.e. 20-40hz. Most subwoofers will roll off anything below 20hz because our ears don't interpet it as a tone. So what's the problem with sub-20hz inaccuracy in MP3s? Answer: There is no problem. It didn't matter in the first place.

I hope that was enlightening.

Not really.

Perception is subjective. What I hear, maybe someone like Snappleman can not.

What I can do, however, is quantify audio degridation through a scientific analysis.

Whether you're interested in that or not is irrelevant to my pursuit of the study.

Whether you can hear above 15khz or not is irrelevant to my study completely.

[analoq]

Your problem is that the encoding process works like magic to you. Voodoo. You think you can quantify audio degradation in a perceptual codec using spectral analysis because you don't know any better.

It's not just my opinion that your tests are worthless, you simply don't know how to test your hypothesis. It's as if you're using spectral analysis to see if the guitars are in tune or not.

There is a scientific way to test perceptual codecs: Double-blind listening tests. Until you're doing that, you're just wasting your time and everybody else's.

cheers.

[Lunahorum]

There is a scientific way to test perceptual codecs: Double-blind listening tests. Until you're doing that, you're just wasting your time and everybody else's.

cheers.

That's a little harsh. He hasn't wasted any of my time. I enjoyed looking at his graphs whether or not it mattered.

[analoq]

As a visual cue for educational purposes, the graphs are innocuous. What you quoted was in reference to dannthr's so-called "study". Peer review is a harsh mistress.

[Lunahorum]

Ah I listened to it on my monitors and noticed the difference. Yay. I guess my $5 radioshack headphones aren't good for this stuff. You would think they would be good enough to hear a difference, but I guess not. Also my ears aren't very attuned to this sort of thing so that could be the issue.

[Nekofrog]

Yes.

There's no other answer for me; yes.

Got it right on the first try. It's really simple with decent headphones. As soon as the voice kicked in on the second clip I could hear the lousy compression; as well as with the tambourine hits.

[starla]

Got it, the tambourine gave it away. Like skrypnyk mentioned, rides, hi hats or similar percussion will make compression stick out like a sore thumb.