Consider with a smooth nonnegative phase positive in the interior and zero only at the endpoints. Suppose, in inward endpoint coordinates ,
where . The rescalings and and the Gamma integral give
The first two terms therefore come from different endpoints. The absent quartic phase term and linear amplitude term at the cubic endpoint are essential for the stated remainder: either can produce a correction. Smoothness, the compact interval and positivity away from the endpoints justify localization and the termwise error estimates. Without those cancellations the two leading scales remain the same when their amplitudes are nonzero, but the next error is generally .

Articles by others on the same topic (0)

There are currently no matching articles.