FA-91695 / Digital signal filters / Member archive

Resampling plan normalizes the cutoff to the input Nyquist · case 05

The normalized cutoff handed to the filter designer is L times too large.

Member previewVariant 5 · 3 implementations · 7 checks per implementation

Case contract

Input [fin, fout, rolloff] (positive integer rates, rolloff in (0, 1]). Reduce fout/fin to up/down = L/M by gcd, intermediate rate = fin L, anti-alias cutoff = min(fin, fout)/2 * rolloff in Hz and normalized to the intermediate Nyquist. Return {"up", "down", "intermediate_rate", "cutoff_hz", "cutoff_norm"} or "bad-spec".

Why this case matters

Sample-rate converters are planned from these numbers; a wrong cutoff reference aliases or dulls the output.

One recorded failure

Sample boundary fixture

This sample comes from the broken implementation of a controlled reproducer.

Boundary fixture	Actual	Expected	Outcome
regression: upsample by 2	`{"cutoff_hz": "3200", "cutoff_norm": "4/5", "down": 1, "intermediate_rate": 16000, "up": 2}`	`{"cutoff_hz": "3200", "cutoff_norm": "2/5", "down": 1, "intermediate_rate": 16000, "up": 2}`	Failed

MEMBER ARCHIVE

The complete case is available to members.

This record includes three runnable implementations, regression fixtures, execution results, and source hashes.

Member access is invitation-based. Sign in with your invited account to inspect the sources.