alan little's weblog

currently listening to ...

16th July 2004 permanent link

Friday night, and Alan’s Friday night excitement consists of babysitting while Maria goes on a Girls’ Night Out. I get the baby to bed early and celebrate by listening to Yehuda Hanani’s marvellous recording of the Bach Cello Suites while I’m doing the dishes. And I find myself thinking “this sounds bloody good”: meaning, not the playing which already know is good, but the sound quality of the CD-burned-from-mp3. I had only listened on cheap headphones on the computer before having just downloaded the thing from emusic [1] – now I’m listening to it on my cheap'n'cheerful but decent kitchen stereo [2] and thinking I really wouldn’t notice it wasn’t an original CD if I didn’t already know.

Wouldn’t I? Time for a test.

Methodology: very scientific:

  1. Pick something I would normally want to listen anyway, and that I already have on mp3 from emusic and also on CD. I liked the mp3 a lot but thought a CD must sound even better, and it was cheap, so I bought it. The Hungarian Quartet performing Haydn’s String Quartet opus 64 no. 5, “The Lark”.
  2. Take an uncompressed version copied from the CD, make two compressed copies of it using AAC at 128 kbps and 320 kbps.
  3. Take the mp3 copy from emusic. Emusic say they use Lame 3.92 to produce their variable bit rate files. From what little I know of these things, this is an up to date version of reputedly the best encoder. Average bit rate for the movements of the quartet is around the high 180s.
  4. Put the four versions of the first movement into a playlist, hide the bitrate column and shuffle.
  5. Burn to a CD and listen using the best available equipment [3].
  6. Come to the following conclusions:

This is not a masterpiece of the recording engineer’s art. The CD original doesn’t sound that great when listened to closely. Nevertheless it’s clearly better than the compressed versions.

The 128 kbps AAC is a clear loser. Generally flat-sounding. Treble is thin & shrill, everything else muted. AAC “as good as mp3 at the next bitrate up”, as it is generally reputed to be, seems not to be the case here. [4]

The 320 kbps AAC and the mp3 (variable bit rate, average 185 kbps) are hard to rank relative to one another. Both have a rounder, fuller sound than the 128 AAC, but both sound somewhat flat and dull compared to the CD original. If pushed I think I would pick the mp3.

What I should do in the interests of more scientificness is repeat with the CD player on random and see if I get the same results again. Which I might one day if I ever have the time & inclination. Also repeat with something that’s actually a good recording in the first place.

[1] One of quite a number of very good pieces of classical music I’ve picked up on emusic. Coming Soon: Alan’s emusic classical gems.

[2] A portable CD player from iRiver, feeding a NAD 310 amp and Mission 731 Pro speakers. The CD player about €100 from amazon, the speakers the same from ebay, the amp years old and I can’t remember what I paid for it, but now also available for about €100 or less on ebay. And voilà – a very decent sounding stereo system (as long as you don’t want loud) for about 300 bucks plus a few bits of cable.

[3] Bottom of the range Cambridge Audio CD player, Marantz PM-80 amp & Sennheiser 580 headphones.

[4] I’m well aware that 128 kbps from iTunes might not be the best AAC can do, even at that bitrate. I read somewhere that iTunes doesn’t use the best possible encoder settings, and that the 18 kbps files from Apple’s music store are sourced from higher quality originals than CDs. If at some point I actually manage to find something interesting in Apple’s music store that I also have on CD, in the interests of science I shall buy a track or two and do another comparison.

related entries: Music

all text and images © 2003–2008