[Archimago]

Hey there everyone... Those who hang out at Audio Asylum (Computer Audio Forum) may know I've put up a survey / test on my blog this week.

My hope is to capture data from music lovers, especially those with good sounding gear, to see how well a "well encoded" (AKA "properly molested") MP3 can be discriminated from the unmodified original CD audio rip.

I know some of you may have engaged in similar tests over the years so this is not for everyone of course. But as for any poll / survey, it's always good to get more data points and especially for those who have never tried this, it may be useful to have the experience of going through this as a challenge for oneself.

Feel free to spread this test around to other audio sites. In time, I'll spread this test around the other audiophile 'watering holes' as well.

I figure I'll collect the data till the end of January or so then publish the data. I certainly welcome discussion here and on the blog site (I'm not a frequent visitor here at the forum but will try to pop by once awhile). For those with audio editors, remember, use your ears, not your eyes . Also please do not 'let the cat out of the bag' if you know the answer until after the survey's conclusion! Thanks!

Get the test here:
archimago.blogspot.com

Note that the test file (ZIP'ed together) is quite large ~75MB. Also, the survey will ask things like approx. price of the audio system, type of system (speakers, streamer, headphones...) so I get a sense of what kind of equipment was used.

Have a wonderful holiday season everyone!

[Archimago]

So far the survey is up to >70 responses. Thanks to everyone who participated.

2 weeks left to go for anyone who still wants to try this before I close off and do some data analysis!

Please pass along the website to other audio forums, friends, family, and enemies

http://archimago.blogspot.com/

[Archimago]

FINAL CALL!

Survey closes next week. So far, I've received 124 responses. Thanks everyone for the participation

Data analysis should be interesting.

[Archimago]

Study complete!  Total respondents - 151.

Follow on my blog as I post the procedure and analysis in the days ahead.  Thanks for all the input.

[stereocilia]

Surprising results, right?  Thanks for doing this, I wish I had seen this sooner.  I wonder what ACers think about this.

[toxteth ogrady]

Great work Archimago

If every audiophile took this kind of test, I suspect we'd have a lot more 'phools than 'philes. Professional reviewers should be the first in line.

[wushuliu]

Well first thing to note is that a majority found audible differences between formats. The whole 'there's no difference' argument is always the first mantra that gets thrown around for encoding wars so I consider this the most important element of this experiment. As for the preference for the lossy, this has come up from time to time. Way back when about 10yrs ago one of the creators of the mp3 format gave an interview explaining why felt some people preferred the high bitrate lossy to uncompressed. I have long forgotten where the interview is though.

In the past I also preferred some lossy 320 to uncompressed when compared. I did my own comparisons way before I knew anything about audiophiles - I just knew I could hear a difference and wanted to know why I liked one and not the other. The conclusion I came to was the lossy 320 emphasized a certain midrange clarity and snap vs uncompressed. However, this was at the expense of treble and HF information, and information in general - and there was the 'glassy' quality at the top end. Not unlike the boomsizzle frequency response of cheap commercial speakers at a Best Buy that sound good at first demo, but are fatiguing over the long haul.

[stereocilia]

Well, I'm certainly not going to convert my whole library from flac to mp3 based on this, but I do wonder what the heck is going on here.  When my iPod classic ran out of space for lossless I converted it to high-rate mp3, and I didn't notice any decrease in fidelity.   But then again I use my iPod in my factory car stereo.