Steep-filtered noise produces pitch matching the filter’s cut-off frequency

Evidence

Noise with steep spectral slopes can elicit pitch sensations. The pitch corresponds closely to the cut-off frequency of the filter for both low-pass and high-pass noise. … Very steep filters with a spectral slope of at least 120 dB / octave were used in both cases. … The dashed line in Fig. 5.14 corresponds to equality of matching frequency and cut-off frequency, indicating an almost perfect approximation. … Only at cut-off frequencies above about 800 Hz, high-pass noise elicits a pitch that is relatively faint, whereas low-pass noise produces a more distinct pitch at all cut-off frequencies. — Professor Dr.-Ing. Hugo Fastl, Professor Dr.-Ing. Eberhard, Psychoacoustics- Facts and Models, 734

Noise with steep spectral slopes can elicit pitch sensations. The pitch corresponds closely to the cut-off frequency of the filter for both low-pass and high-pass noise. This result is illustrated in Fig. 5.14 where the frequency of a pure tone matched to the pitch of filtered noise is plotted as a function of cut-off frequency. Circles illustrate pitch matches for low-pass noise, triangles indicate pitch matches for high-pass noise. Very steep filters with a spectral slope of at least 120 dB / octave were used in both cases. The dashed line in Fig. 5.14 corresponds to equality of matching frequency and cut-off frequency, indicating an almost perfect approximation. Data for low-pass noises (circles) show small interquartile ranges, but larger interquartile ranges are found for high-pass noises (triangles) especially at lower frequencies. Only at cut-off frequencies above about 800 Hz, high-pass noise elicits a pitch that is relatively faint, whereas low-pass noise produces a more distinct pitch at all cut-off frequencies. This effect on pitch strength will be discussed in more detail in Sect. 5.7. — Professor Dr.-Ing. Hugo Fastl, Professor Dr.-Ing. Eberhard, Psychoacoustics- Facts and Models, 734

Categorization

• XREF: Connects to the broader virtual pitch models (Ch. 5) being built up; also relates to why high-pass noise fails below ~400-800 Hz cut-offs, which recurs in later paragraphs.
• QUESTION: Why does high-pass noise only elicit pitch above ~400-800 Hz cut-offs while low-pass works at all frequencies? Section 5.7 promises elaboration on pitch strength.
• corroborated: 2 independent passages support this claim — Professor Dr.-Ing. Hugo Fastl, Professor Dr.-Ing. Eberhard, Psychoacoustics- Facts and Models, 734