Maximum-likelihood estimation can predict categorical loudness functions from only 15-30 trials

Evidence

The catalog can be used to compute a ML estimate of a listener’s MCPF. This may be particularly useful when the number of experimental observations is relatively low, but an accurate estimate of a listener’s loudness perception is needed, as is the case in most adaptive-level CLS methods that would be used in the clinic. We denote a listener’s raw CLS data as where N is the total number of experimental observations. The likelihood, of these observations is computed for each catalog MCPF, maximizing over all potential parameter sets (i.e., all ). The maximization is computed over the functionally-equivalent log-likelihood. (3) Once the ML parameter set has been determined, it can be used to construct the listener’s MCPF, via Eqs. 1 and 2 . — Dijk, Pim van, Physiology, Psychoacoustics and…, 2530

Categorization

• QUESTION: How well do ML estimates from catalogs perform against direct measurement at various observation counts? The claim that it works with ‘relatively low’ observations invites a validity check.
• XREF: Connects to adaptive-level clinical testing (e.g., Bekesy tracking, categorical loudness scaling used in hearing-aid fitting), a known clinical psychoacoustic instrument.
• corroborated: 3 independent passages support this claim — Dijk, Pim van, Physiology, Psychoacoustics and…, 2530