Loudness-calculating procedures outperform simple SPL for practical noise reduction

Evidence

Using the data plotted in Fig. 8.2 together with those plotted in Fig. 8.1, the equal-loudness contours of the diffuse sound field can easily be constructed. Equal-loudness contours for sinusoidal tones or narrow-band noises indicate the interesting dependence of loudness on frequency. However, loudness depends on many more variables such as bandwidth, frequency content, and duration. Consequently, it is too simple to approximate loudness level by a single weighting such as the A-weight.ed sound pressure level. Because A-weighting has a frequency dependence that corresponds to that of the equal-loudness contours at low levels, A-weighted sound pressure level approximates loudness level only for sinusoidal tones or narrow-band noises at lower levels. Therefore, dB(A)-values of noises or complex tones, or combinations of both, are misleading when used as indications of subjectively perceived loudness. — Professor Dr.-Ing. Hugo Fastl, Professor Dr.-Ing. Eberhard, Psychoacoustics- Facts and Models, 1176 (#corroborates, via linked claim)

Categorization

• XREF: connects to: Noise reduction should target the largest loudness pattern area — Professor Dr.-Ing. Hugo Fastl, Professor Dr.-Ing. Eberhard, Psychoacoustics- Facts and Models, 1330; ISO standardized loudness-based noise metrics because A-weighting proved inadequate; A simple high-pass filter can replace elaborate loudness models (+7 more)

Related

• ISO standardized loudness-based noise metrics because A-weighting proved inadequate (similarity 0.88)
• A-weighted dB values misrepresent perceived loudness for complex sounds. (similarity 0.88)
• A simple high-pass filter can replace elaborate loudness models (similarity 0.87)
• Loudness curves from anechoic pure tones fail on real broadband program material (similarity 0.87)
• Noise reduction claims should use loudness ratios to stay credible (similarity 0.86)
• A-weighting can miss dominant low-frequency energy in broadband sounds (similarity 0.85)
• A-weighted decibels better predict hearing damage than unweighted sound pressure level. (similarity 0.85)
• A-weighted decibels are a rough practical approximation of perceived loudness (similarity 0.85)
• A-weighting is recommended for all sound levels despite fitting low levels best (similarity 0.85)
• A-weighted sound pressure level misrepresents the loudness of broadband noises (similarity 0.85)
• Loudness scales with frequency squared, requiring large amplitude at low frequencies (similarity 0.85)
• Wideband sounds are louder than narrowband sounds at equal dB SPL (similarity 0.84)
• Noise reduction should target the largest loudness pattern area (similarity 0.84)
• C-weighting captures low frequencies better at high sound pressures (similarity 0.84)
• Loudness measures can reveal real improvements A-weighting misses entirely (similarity 0.84)

Broader Connections

• A simple high-pass filter can replace elaborate loudness models #contradicts (0.85)
• A-weighted dB values misrepresent perceived loudness for complex sounds. #corroborates (0.75) — Professor Dr.-Ing. Hugo Fastl, Professor Dr.-Ing. Eberhard, Psychoacoustics- Facts and Models, 1176
• A-weighted sound pressure level misrepresents the loudness of broadband noises #corroborates (0.75)
• A-weighted decibels are a rough practical approximation of perceived loudness #complicates (0.72) — Ville Pulkki, Matti Karjalainen, Communication Acoustics- An…, 988
• Wideband sounds are louder than narrowband sounds at equal dB SPL #corroborates (0.72)
• SIL and SPL always change in proportion even where their absolute values differ #complicates (0.70)