What the Research Says About Hearing Lossless: Two Studies and Their Limits

Debates about audio quality produce more heat than evidence. Two studies are worth knowing, because one is famous for finding almost nothing and the other is careful about why.

• In NPR’s 2015 online listening quiz, listeners picked the lossless file 36.0 percent of the time against 33.3 percent expected by chance ^[1].
• Meyer and Moran’s 2007 Journal of the Audio Engineering Society study found no significant correlations on music at normal levels when a CD-standard loop was inserted into high-resolution playback ^[2].
• Detection happened only with an exceptionally quiet recording played about 20 dB above reference level, with peaks above 111 dB SPL ^[2].
• Neither study says lossless is worthless; both say the benefit is subtle and conditional.

Study one: NPR’s open quiz

In 2015, NPR ran an online listening quiz comparing a lossless WAV with 320 kbps and 128 kbps MP3 versions. Listeners chose the lossless file 36.0 percent of the time, against 33.3 percent expected by pure chance ^[1]. This was a self-selected audience listening on whatever equipment they owned, so the study is better as a reality check on casual listening than as laboratory evidence.

Study two: Meyer and Moran, 2007

E. Brad Meyer and David Moran tested whether a CD-quality (16-bit, 44.1 kHz) conversion loop inserted into high-resolution playback could be heard. They used four different systems, from professional mastering rooms to a high-end audiophile setup, with multiple source players and music chosen for dynamic range and detail. The study was published in the Journal of the Audio Engineering Society in September 2007 ^[2].

Their finding: no significant correlations on music at normal levels. The only positive detection occurred with one extremely quiet recording played 20 dB above the standard reference level, which produced peaks above 111 dB SPL ^[2]. As the authors note, the vast majority of productions have a noise floor that swamps whatever the CD link adds ^[2].

What the studies do and do not show

• They do show that for typical listening, the difference between CD quality and higher resolution is hard to detect.
• They do not show that MP3 is as good as lossless: the NPR quiz included a 128 kbps file too, and picks were spread fairly evenly across all three files, which is a reminder of how hard blind identification is for casual listeners.
• They do not address practical reasons to prefer lossless: archiving, editing and the lack of generation loss when converting.

A sense of scale for those loud peaks

The one test where listeners did hear a difference used playback about 20 dB above the standard reference level, with peaks above 111 dB SPL ^[2]. For scale, the safe-listening rule in NIOSH guidance halves allowed exposure time for each 3 dBA above 85 dBA [{n1}]. A sustained level 26 dB above 85 would allow only about a minute on that rule (my arithmetic: eight hours divided by two raised to the 26/3 power), although peak sound pressure and an averaged A-weighted level are different measurements and the study’s peaks were brief. The point is that the one condition where the CD-quality loop became audible was a condition few people choose to listen in.

A sensible conclusion

Buy lossless because you want a master you can convert from, not because you expect a revelation. Buy hi-res when it is offered for a similar price, since the research says it is harmless and occasionally audible in extreme cases. And be wary of anyone who claims the difference is obvious on every system.

Questions readers ask

Can people hear the difference between lossless and MP3?

In NPR’s 2015 quiz, listeners picked lossless only slightly more often than chance (36.0 versus 33.3 percent) ^[1].

Is hi-res audio audibly better than CD quality?

Meyer and Moran’s 2007 double-blind tests found CD-quality conversion inaudible at normal levels, with detection only under extreme conditions ^[2].