more precision for interpolation
This commit is contained in:
parent
c936389bfc
commit
2b1adda96c
@ -20,6 +20,8 @@
|
|||||||
- interpolators use type introspection
|
- interpolators use type introspection
|
||||||
- added vips --list classes, does formats too
|
- added vips --list classes, does formats too
|
||||||
- include sys/param.h to get PATH_MAX in more places
|
- include sys/param.h to get PATH_MAX in more places
|
||||||
|
- VIPS_INTERPOLATE_SHIFT bumped to 12, we need the extra precision for u16 gel
|
||||||
|
data
|
||||||
|
|
||||||
11/9/08 started 7.16.3
|
11/9/08 started 7.16.3
|
||||||
- oop typo in manpage for im_project()
|
- oop typo in manpage for im_project()
|
||||||
|
@ -94,10 +94,10 @@ int vips_interpolate_get_window_size( VipsInterpolate *interpolate );
|
|||||||
|
|
||||||
/* How many bits of precision we keep for interpolation, ie. where the decimal
|
/* How many bits of precision we keep for interpolation, ie. where the decimal
|
||||||
* is in the fixed-point tables. For 16-bit pixels, we need 16 bits for the
|
* is in the fixed-point tables. For 16-bit pixels, we need 16 bits for the
|
||||||
* data, 4 bits to add 16 values together, another bit for the sign and some
|
* data and 4 bits to add 16 values together. That leaves 12 bits for the
|
||||||
* other stuff, so say 24 total. That leaves 8 bits for the fractional part.
|
* fractional part.
|
||||||
*/
|
*/
|
||||||
#define VIPS_INTERPOLATE_SHIFT (8)
|
#define VIPS_INTERPOLATE_SHIFT (12)
|
||||||
#define VIPS_INTERPOLATE_SCALE (1 << VIPS_INTERPOLATE_SHIFT)
|
#define VIPS_INTERPOLATE_SCALE (1 << VIPS_INTERPOLATE_SHIFT)
|
||||||
|
|
||||||
/* Convenience: return static interpolators, no need to unref.
|
/* Convenience: return static interpolators, no need to unref.
|
||||||
|
Loading…
Reference in New Issue
Block a user