fix(signal): refuse filter designs whose coefficients overflow
Assisted-by: GLM 5.3 Flash
This commit is contained in:
+3
-1
@@ -1054,7 +1054,9 @@ Each design returns the direct-form coefficients `b` (numerator) and `a`
|
||||
band edges prewarped to the bilinear axis and the answer exact at the mapped
|
||||
frequencies. All of them refuse an order below 1, a non-positive or infinite `fs`,
|
||||
and an edge outside `(0, fs/2)`; the band forms additionally require
|
||||
`0 < edge1 < edge2 < fs/2`.
|
||||
`0 < edge1 < edge2 < fs/2`. An order whose coefficient arithmetic
|
||||
overflows the float64 range is refused as well, never returned as a
|
||||
numerator of zeros or NaN.
|
||||
|
||||
| Call | What it does |
|
||||
|---|---|
|
||||
|
||||
Reference in New Issue
Block a user