Just-noticeable changes are tenfold worse at masked threshold then normalize 20 dB above

Evidence

Few data are available for just-noticeable differences determined with additional background noise. In most cases, only white noise has been used as background. The data available indicate that just-noticeable difference in frequency shows large values in the neighbourhood of the masked threshold. However, 10 dB above threshold, the just-noticeable difference in frequency is only a factor of two larger than without masking. About 20 dB above masked threshold, just-noticeable differences in frequency reach almost the same value as measured without masking. The data available for just-noticeable difference in level lead to similar results. In this case, not only broad-band noise but also low-pass and high-pass noises have been used as additional maskers. The data available for the latter conditions show relatively large scatter. There is general agreement, however, that JNDL is higher by about a factor of 5 close to masked threshold, and seems to reach values corresponding to those measured under unmasked conditions for levels of the test tone of more than 20 dB above masked threshold. It can therefore be summarized that just-noticeable changes (variations as well as just-noticeable differences) are larger by about a factor of 10 close to masked threshold and reach almost normal conditions for levels of test tones about 20 dB above masked threshold. — Professor Dr.-Ing. Hugo Fastl, Professor Dr.-Ing. Eberhard, Psychoacoustics- Facts and Models, 1117

Categorization

• DEFINE: Defines the quantitative relationship between masking level and just-noticeable difference/frequency/level perception.
• QUESTION: Why exactly 20 dB above threshold as the normalization point — is this tied to critical-band excitation patterns?