How bad is MP3? (Part 2)

lego mp3 player
Image courtest of

Last week I wrote up a post about my opinions on the MP3 format which included a few listening reviews of various MP3s in my collection. I admitted the weakness of this experiment, in that I wasn’t comparing these recordings to any other format, so I concluded that I needed to do a review the same recording in both MP3, CD and vinyl forms.

I chose to review Van Cliburn and the Chicago Symphony Orchestra’s 1961 performance of Beethoven’s Piano Concerto No. 5 in E-Flat, Op. 73 “Emperor”. Specifically, the third and final movement. I own this recording both on vinyl as part of the Great Men of Music boxset I reviewed a while back and as an RCA Living Stereo SACD . I ripped an MP3 copy at 160 kbit/s using iTunes, which was sourced from the CD layer of the SACD disc. Before each test, I calibrated the sound level for each source to -25 dB using the trial version of Electroacoustics Toolbox 2.0 , a versatile sound analysis program. To do this, I placed my laptop in my listening position and then played a 1 KHZ test tone from the Stereophile Test CD , test MP3 file in iTunes, and the Cardas Test record . The laptop microphone picks up the test tone and displays the current SPL (sound pressure level) ,and I would adjust the volume knob on my amp until it reached -25 dB. Performing this adjustment for each source reviewed ensured that the music was being pumped out the amplifier at a consistent level so that differences in the sound pressure wouldn’t color my judgment. So, here’s what I came up with:

CD- I made sure to play the CD layer of the disc for this test, as it wouldn’t be fair to compare the MP3 it was sourced from to a hi-rez version of the same recording. Bass frequencies are very present (if not fairly boomy) and Cliburn’s piano is fresh and lively but demonstrates some “smearing”. The strings are dense without sounding excessively stodgy and the soundstaging is good. Everything is in its proper place, and since I have actually attended a concert in that hall before, I could really reimagine it.

MP3 – A marked flatness is apparent right off the bat, the dynamics are severely reduced. Just as I had remarked about some tracks in the previous week’s part of this review, the music just can’t seem to escape or transcend the speakers. The orchestra is trapped in a wood and kevlar prison. In contrast to the CD recording, the piano seems a bit more defined and runs together less. But despite the improved clarity, the piano still suffers from a loss of energy. The soundstage has shifted to the left somewhat…somehow. Overall, the MP3 was not unlistenable but still not ideal.

LP- Well, to start off, hissing, rumble and clicks are apparent in quiet portions (no shit) but it’s not bad enough to be distracting. This recording shows excellent soundstaging and the bass is subdued when compared to the CD and MP3 edition. The piano isn’t as shiny as the other two tests, it lacks in high treble which can be a limitation of vinyl records. This was craploads better than the MP3 and rivaled the CD for clarity, warmth and overall enjoyability.

So to conclude, MP3s still kinda suck. But as I wrote last time, they have their definite advantages if you’re concerned about space and price. I understand that this wasn’t a double-blind ABX test , the gold standard for hi-fi experimentation, but the point of these reviews wasn’t to determine if one could tell a difference between the formats (which ABX does), but instead to assess the subjective qualities of each format. Sound is still always a matter of taste and preference, so you might not mind or even care about the sound of an MP3 if it seems okay to you. But for me, i’ll stick to MP3s cranked out of a portable device for when i’m on my way to work, and i’ll use CD and vinyl for when I really want to sit down and enjoy the music.

4 thoughts on “How bad is MP3? (Part 2)”

  1. You should really try using LAME encoded V0 OR 320 mp3s. They are still lossy but its not really fair to compare a low quality rip to the source media. You could give the mp3s every advantage that they are afforded. This is a poor test.

  2. I’m glad you’re trying to test these things for us.

    However…”Concerned Reader” is right: you need to extract higher quality MP3s using better encoding applications for this test to be useful. MP3 is just a format, and some implementations of it do a better job than others. iTunes is a notoriously shabby MP3 ripper—that “flatness” is characteristic of iTunes’ MP3 ripper. Try XLD, a proper mp3 ripping frontend in Mac OS X. I definitely second Reader’s recommendation for bitrates: either -V0 or 320kbps would be more representative of MP3’s potential quality.

    I’ve done my own A/B tests of MP3s vs FLAC (which is Lossless, and so equivalent to the original CD). I couldn’t tell the difference on many (if not all) tracks. I have an EMU 0404 DAC going to a pair of AudioEngine A2s. I ripped everything using XLD. I’ve done the same test on a pair of Grado RS-1s…still can’t tell the difference. I can, however, tell the difference between iTunes and LAME mp3s at 192 or lower.

  3. I performed a smilar test using 320 mp3s and experienced the same results. It has always surprised me that more isnt said of mp3s significant shortcomings.

Leave a Reply

Your email address will not be published. Required fields are marked *