Thursday, November 08, 2007

Experiment on Your Friends (or Simple Psychology Experiments that Destroy People's Whole World View)

Here's a little experiment you can try on your friends.

Take a true lossless file and use iTunes or whatever you like to create an inferior version (160kps or whatever you like). If you are unsure of the source of your file, take a CD and rip a 320kps and a 160kps version.

This next part is critical -- make sure they are normalized to the same volume.

I'd also suggest that you use regular computer speakers or iPod style headphones and turn off any and all eq, and effects in your software, but that's really up to you... whatever you do, be consistent for each song (volume, eq, listener position).

Make a playlist that plays :30 seconds (arbitrary number, but why listen to the whole damn song?) of one file twice, and the same :30 of the second file once. Make sure that your subjects can't tell which is which by reading the file names or whatever, perhaps rename the three files with random letters (Q, F and J or something like that).

Have your subject listen to all three files and ask them to tell you which one is different. It's also a better design if you have yet a third person run the experiment who doesn't know which file is which and if you are out of the room or at least out of the subject's view. Repeat the experiment, ideally with at least 10 friends. If you don't have 10 friends or don't like letting people in your home, round up some tech friendly hobos on the street.

Random chance says 33% will get the right answer just by luck -- those people will take this not as sign of random probability but of their innate genetic superiority -- don't ruin it for them. If four or more of your friends can tell the difference -- congrats they did better than random chance. I'd still guess that more than 50% will get it wrong.

You could then ask them to tell you what files were the higher quality and lower quality files -- this is a more difficult question then just asking which was different as "better" is somewhat subjective. I'd expect the number to be even lower than before.

I stole this design from market researchers who have been running this experiment using Cola for almost 30 years -- subjects as a whole have done no better than chance and the experiment has been run literally thousands of times.

Incidentally, be prepared to have people hate you for this... people do not like being wrong.

Labels: ,

3 Comments:

Anonymous Jasper said...

There should be more people arguing here. Arguing is awesome!

The car I'm leasing came with a free Super Awesome Audiophile Mega Woofer High Fidelity 20.1 Surround Sound George Lucas DVD Audio upgrade. Or something. That is the only system I own where I can hear the difference between low bit rates and higher bit rates. In fact, I'm back to burning CDs because the FM transmitter from my mp3player sounds awful on it. It is my new car that has made me an mp3 nazi.

Everything that comes out of my brand new laptop that I paid a lot of money for sounds like 64kbps.

And, although I don't have an iPod, I do know they are regarded by the scientific listening community as "sounding like absolute shite". Something about a dB falloff. So I doubt you could tell on an iPod, either.

Bottom line: All our electronics are cheaply made, yet widely accepted as "OK!" so it doesn't matter how most people rip their shit because you're not going to get something quality out of crap equipment. It's like making eggs in the microwave. Or something.

We should go to the drive-in together sometime so you can check out my "system".

6:52 PM  
Blogger Gorilla said...

Rooms make a huge difference in how you hear this stuff too. A regular untreated room in your house is full of acoustical problems... of course this stuff does get out of control, I just saw a pair of $2500 connectors, like to connect your turntable to your amp or whatever... so stupid.

1:23 PM  
Blogger Yale Bloor said...

as an aside to your arguments I feel your hypothesis maybe flawed...reason various controlled substances may enhance and/or detract from this experience, that being said I was just curious about your control group and if any and what controlled substances were involved, the type of music used ect...as a 40 and a blunt could make Jay Z sound good on an 8 track and some x can make most Lounge music sound like Beethovens 5th.... I really think you need to deleve much deeper into your research......love you gorilla

11:36 AM  

Post a Comment

<< Home