Audio News for December 14, 2008

by | Dec 14, 2007 | Audio News | 0 comments

Flawed Listening Test of MP3 Samples – A seriously flawed listening test – which supposedly found that few people can tell the differences between “average” and “best” MP3 sampling rates – was carried out and published by Cognitive Daily.  About 700 people completed the study online, using either their computer speakers or headphones. Only two very short excerpts were provided: one by rock guitarist Carlos Santana and the other from an orchestral work by Aaron Copland.  Each was encoded at three different MP3 data rates: 64, 128 and 256 kbps. One of the main conclusions of the testers was than since almost no one could tell the difference between 128 and 256 kbps, there is no need to use the higher data rate since it takes up more space on hard drives.

The list of obvious errors in the listening test which make it invalid is a long one. Knowledgeable listeners would say that any data reduction is going to have a negative effect on the music.  Why was the “best” sample still a data-reduced 256 kbps – not even the highest MP3 data rate possible?  The choice should have been between a couple of different MP3 files, standard 44.1K/16 bit CDs, SACD, 96K PCM and 192K PCM – if restricting it to only two-channel reproduction. The musical samples were too short, didn’t have some of the sounds readily causing data reduction artifacts – such as cymbals – and the Santana was a heavily compressed track which destroys most of the musical sense to begin with. The data set was simply too small to draw the proper conclusions.  And there was no way to quickly do A/B/C switching between the samples if desired.  Just because most of the people couldn’t tell the difference between the data rates doesn’t mean that with different selections of music and/or a different setup they wouldn’t be able to.

However, due to other problems, they still probably wouldn’t be able to tell differences.  The main reason is that there is no control over the equipment with which people are listening to online music samples.  There is a very wide range of audio quality with personal computers.  Few users feed their PC audio into a top-quality audio system; most are using the tiny speakers built into their monitors or PCs or at best a tiny system with a so-called subwoofer and two small satellite speakers without tweeters. The lack of tweeters means the highest frequencies – which contribute to the transparency and realism of the music – are not being reproduced, thus causing the music samples to sound more alike. Also, the interior of every computer is filled with noise-making and EMF-exuding components which cannot achieve top-quality audio playback.

Would the use of headphones solve the computer speaker problem?  Perhaps, if they were highest quality hi-end phones powered by dedicated headphone amps, but most computer users use throwaway headphones plugged directly into the audio jack of their PCs. To conclude, the best summation probably was left by one of the comment posters on the Cognitive Daily article: “The only way this test can be done legitimately is with a reference grade system placed in a reference quality room with people that have trained ears.”


Related Reviews
Logo Pure Pleasure