Beyond Loudness: Does Narrow Dynamic Range Bias Listener Preference When Loudness Is Held Constant?
Abstract
The “loudness war” — the decades-long escalation of average levels in commercial masters via compression, limiting and clipping — has historically been justified by the belief that louder recordings are preferred and sell better. Universal adoption of loudness normalisation by streaming platforms (now targeting roughly ‒14 LUFS) removes loudness as a competitive variable at the point of playback, yet heavily limited masters remain the norm. This study asks whether the dense-master aesthetic retains any perceptual pull once loudness is neutralised: specifically, whether reduced dynamic range and audible clipping bias listener preference independently of loudness, and whether that bias is absolute or relative to genre expectation. Using a double-blind design with stimuli matched on perceived loudness (ITU-R BS.1770 integrated loudness as the starting point, then a listener-adjusted and model-based perceptual trim, with true peak deliberately left free so dynamics survive), the study compares dynamic, transparently limited and hot-clipped masters all derived from a single owned source per song. In density-native genres a relaxed remix serves as the common parent from which every processing level is produced, so the only difference between conditions is the final-stage dynamics treatment. The controlled provenance permits public release of all stimuli, addressing a reproducibility gap in the existing literature.
Research Questions
With perceived loudness held constant, does dynamic-range reduction (transparent brick-wall limiting) shift listener preference relative to a dynamic master of the same source?
Does audible clipping distortion exert an effect on preference that is separable from dynamic-range reduction per se?
Is any such effect absolute, or does it interact with genre schema — i.e. is reduced dynamic range penalised in dynamic-tolerant genres but tolerated or preferred in density-native ones?
Does the effect differ between trained (audio-professional) and naïve listeners, and does it vary with each listener’s familiarity with and affinity for the genre?
Over longer continuous exposure, does heavy limiting/clipping produce measurable listening fatigue independent of momentary preference?