A few weeks ago I was complaining about the audio quality of MP3s that I was listening to, when approached by Erik J. Heels (@ErikJHeels on Twitter) with a challenge that I couldn’t actually tell the difference between a well encoded MP3 and a CD. Since everyone in the audio world “knows” that the lossy encoding of any MP3 is far inferior to CD I took him up on the challenge that I would easily be able to tell the differences between MP3 and CD audio. It seems that he had a pretty good test put together (details in link) and a few days later I recieved a CD in the mail.
The CD contained a wide mixture of commercial, high quality, well recorded material with 12 songs, each with both CD and MP3 format on the the disk but in random order. The MP3s were originally encoded at 256kbs VBR, and at 48khz even. I felt that the shift from 44.1 to 48khz would give it away, as well as the encoding. I felt confident.
I originally intended to go into Taylor Barefoot’s recording studio and listen to everything in a well treated room, with stereo subwoofers, multiple monitoring options, and a Benchmark DAC-1 converter. However times scheduling that didn’t work out, so one day I at work I pulled out some pretty high quality Sony in-ear monitors/headphones and gave the test a shot on my Macbook Pro. I listened to each track, quickly writing on a sheet of paper which I felt was MP3 and which was originally CD. Threw it all into a quick email and sent it back to Erik. Only two of them did I mark as “hard/unsure”. I felt that the other 10 tracks were pretty clear as to which was MP3 and which was CD. The Macbook Pro with good headphones isn’t the perfect listening environment, but I felt that it should be ok for telling CD vs MP3 if there is such a huge difference.
I got my results back later that day. Out of 12 songs/24 tracks, I had gotten half right. Yes, I had lost. Half right is the same as rolling dice or a monkey choosing them. I felt pretty confident about my answers, and consider myself to really have an ear for audio. I was shocked, but Eric admitted that few people, himself included, pass the test with 100% right. This all goes to show, don’t believe everything you read in audio forums. Go with your own listening ability. I might have been able to tell the difference between 192khz/24-bit audio and the MP3, but between MP3 and CD I can say that there were differences, but I couldn’t tell which was “better” and certainly not which was which.
You can find all of the details on his test on his blog at: http://www.erikjheels.com/?p=1236
