Re: Back to basics: Interpolation teaser
- From: "aruzinsky" <aruzinsky@xxxxxxxxxxxxxxxxxxxxxxxxxxx>
- Date: Sun, 23 Sep 2007 12:24:31 -0400
"> If the
interpolation is intended to be linear, you should look at the impulse
response.
Please explain the above sentence?"
The response to a single white pixel against a black background. For
centered enlargements, the impulse response should be symmetric and not
change with location unless it interacts with the image border, e.g., sinc
interpolation. The luminance of the impulse response should be scaled
(multiplied) and shifted (constant added) for visible inspection.
"Have you experienced any common reasons why that would happen? E.g.
rounding errors perhaps?"
No.
.
- Follow-Ups:
- Re: Back to basics: Interpolation teaser
- From: Pixel.to.life
- Re: Back to basics: Interpolation teaser
- References:
- Back to basics: Interpolation teaser
- From: Pixel.to.life
- Re: Back to basics: Interpolation teaser
- From: Pixel.to.life
- Back to basics: Interpolation teaser
- Prev by Date: Re: fft2 different from FFTW 2d
- Next by Date: Re: Back to basics: Interpolation teaser
- Previous by thread: Re: Back to basics: Interpolation teaser
- Next by thread: Re: Back to basics: Interpolation teaser
- Index(es):