Louder playback can bias listeners into rating a device better

Evidence

The non-linear processing in the ear causes sound with different sound pressure levels to be perceived differently. The loudness differences between the samples may cause undesired effects if the loudness itself is not studied. Thus, when conducting a listening test, the effect of the level of presentation of the signals should be taken into account. For example, if the audio quality provided by different loudspeakers is tested, and if one loudspeaker delivers slightly but noticeably higher SPL to the listener, it quite probably will be rated to provide the best quality of the tested items. — Ville Pulkki, Matti Karjalainen, Communication Acoustics- An…, 866

Categorization

Category: Causal Relationships

• SEED: Could extend to an essay on confound control in subjective evaluation — the loudness halo effect in any perceptual ranking study.

Related

• Louder playback is unconsciously rated as better quality in audio listening tests (similarity 0.93)
• Uncontrolled loudness differences confound audio quality listening-test results (similarity 0.91)
• Louder sounds are perceived as having more bass and bias subjective comparisons. (similarity 0.86)
• Seeing the loudspeakers during evaluation actively stresses trained listeners (similarity 0.83)
• Monophonic listening is the most demanding test of loudspeaker sound quality (similarity 0.83)
• Most listeners can recognize audio excellence when comparing loudspeakers side by side. (similarity 0.83)
• Brand, cost, and prior exposure bias subjective audio judgments (similarity 0.83)
• Perceived loudspeaker sound quality depends only on amplitude responses (similarity 0.82)
• Visual appearance of loudspeakers shifts listener preference ratings (similarity 0.82)
• Flattening sound power via bass-shy or treble-hot designs harms perceived quality (similarity 0.82)
• Sighted loudspeaker evaluations are contaminated by non-auditory visual and social cues (similarity 0.82)
• Sound power alone cannot predict listener preference for loudspeakers (similarity 0.82)
• Loudspeaker evaluation requires mono, while enjoyment happens in multichannel (similarity 0.82)
• Loudspeaker evaluation must be done in mono because multichannel adds forgiveness (similarity 0.82)
• Randomizing identifiers, positions, and music prevents experimenter bias in listening tests (similarity 0.82)