[devel] Revised commentary about png_rgb_to_gray coefficents

and fixed spelling (truely -> truly) in comments.
This commit is contained in:
Glenn Randers-Pehrson 2011-08-11 09:01:57 -05:00
parent a45ffbefb9
commit 0bc79772c8
3 changed files with 7 additions and 2 deletions

View File

@ -3093,6 +3093,11 @@ png_do_gray_to_rgb(png_row_infop row_info, png_bytep row)
*
* Y = (6968 * R + 23434 * G + 2366 * B)/32768
*
* Starting with libpng-1.5.4, if the image being converted has the
* sRGB chunk, then the sRGB numbers are used by default:
*
* Y = 0.33000*R + 0.60000*G + 0.06000*B
*
* The calculation is to be done in a linear colorspace.
*
* Other integer coefficents can be used via png_set_rgb_to_gray().

View File

@ -99,7 +99,7 @@ png_set_gAMA_fixed(png_structp png_ptr, png_infop info_ptr, png_fixed_point
* possible for 1/gamma to overflow the limit of 21474 and this means the
* gamma value must be at least 5/100000 and hence at most 20000.0. For
* safety the limits here are a little narrower. The values are 0.00016 to
* 6250.0, which are truely ridiculous gammma values (and will produce
* 6250.0, which are truly ridiculous gammma values (and will produce
* displays that are all black or all white.)
*/
if (file_gamma < 16 || file_gamma > 625000000)

View File

@ -2104,7 +2104,7 @@ modifier_progressive_read(png_modifier *pm, png_structp pp, png_infop pi)
png_error(pp, "store state damaged (progressive)");
/* This is another Horowitz and Hill random noise generator. In this case
* the aim is to stress the progressive reader with truely horrible variable
* the aim is to stress the progressive reader with truly horrible variable
* buffer sizes in the range 1..500, so a sequence of 9 bit random numbers
* is generated. We could probably just count from 1 to 32767 and get as
* good a result.