XnView Software

stolarz

so there will be at least a public method in gflCBitmap.h that returns m_gflBitmap (a private field) so I can invoke other functions from libgflE that require a GFL_BITMAP*

stolarz

there is a minor bug in file gflCBitmap.cpp,
method GFLC_BITMAP::create(const GFLC_BITMAP & bitmap),
there is return create(bitmap.getType(), bitmap.getBitsPerComponent(), bitmap.getWidth(), bitmap.getHeight(), bitmap.getBitsPerComponent()) ,
should be return create(bitmap.getType(), bitmap.getWidth ...

stolarz

or:

Code: Select all

function gflAllockBitmap(BType: GFL_BITMAP_TYPE;
  width, height, line_padding: GFL_INT32;
  var color: TGFL_COLOR): PGFL_BITMAP; stdcall;

stolarz

is:

Code: Select all

function gflAllockBitmap(BType: GFL_BITMAP_TYPE;
  width, height, line_padding: GFL_INT32;
  var color: PGFL_COLOR): PGFL_BITMAP; stdcall;

should be:

Code: Select all

function gflAllockBitmap(BType: GFL_BITMAP_TYPE;
  width, height, line_padding: GFL_INT32;
  color: PGFL_COLOR): PGFL_BITMAP; stdcall;

stolarz

post scriptum: is there any way to use color index when putting pixels on bitmap? i mean sth like gflSetColorAt(bmp, 0, 0, 5) where 5 is the 6th color in the palette (0 is origin)

stolarz

hi :)
i'm trying to create a 256-color bitmap. when i do it with libgfl functions everything is fine, the code:

GFL_BITMAP *b2 = gflAllockBitmap(GFL_COLORS, wh+2*margin, wh+2*margin, 4, NULL);
GFL_COLORMAP cm2;
cm2.Red[0]=0; cm2.Green[0]=0; cm2.Blue[0]=0;
cm2.Red[1]=255; cm2.Green[1]=255; cm2 ...

stolarz

oops, sorry, there is one in gflsdk, i was referring to nconvert...

stolarz

if i'm not wrong there is no possibility to apply Floyd-Steinberg error dispersion algorithm when changing color depth to 256 colors. and i think it works somehow analogous to the 256 greys version...

stolarz

I've tried but have no crash...
Yes, but the sizes in mem and disk differ and it's just a matter of luck if there is a crash or not because of different memory allocating algorithms: your first malloc in a program actually allocates smallest allocable block, like f.e. 0x1000 bytes, so char* a ...

stolarz

Could you send me your test program?
emm.... d'you mean the source? or the executable?
here's the source:

#include <stdio>
#include "libgfl.h"

void main() {
printf("hello\n");
GFL_ERROR init = gflLibraryInit();
if (GFL_NO_ERROR != init) {
printf("error: %d\n", init);
return;
}
printf ...

stolarz

yeah, and what if length is less than the size of allocated memory? gflRealloc will crop my data

stolarz

unsigned char* data;
unsigned long length;
gflSaveBitmapIntoMemory(&data, &length, ...

I don't know if it's only for tiff but with tiff and lzw 'length' is bigger than the actual allocated block of memory. therefore
unsigned char * new_block = malloc(length);
for (int i=0; i<length; i++) new ...

stolarz

btw, is there any lossless compression in gfl? like lossless jpeg? i see there is GFL_LZW for tiff. any other?

stolarz

:( you are right... i tested it once again, the results differ. it is 1288438 bytes for quality 95 and 99025 bytes for quality 5. i haven't noticed that fact before cause i was using gflSaveBitmapIntoMemory. But what worries me most is why i thought Quality parameter doesn't change anything - i was ...

stolarz

hi people, i'm using the jpeg 2000 plugin with the gfl library and i don't know how to change the quality when compressing to this format. i'm using gflSaveBitmapIntoMemory() and filling GFL_SAVE_PARAMS structure. been trying to modify the Quality member but it seems that it works for regular jpegs ...

XnView Software

Search found 15 matches

GFLSDK 2.90 gflCBitmap.h wish

GFLSDK 2.90 gflCBitmap.cpp bug

Minor Bug, GFL SDK 2.82 for Delphi

problems using colormap with gflC

...

no Floyd-Steinberg for 256 colours?

Re: bug in gflSaveBitmapIntoMemory() (tiff only?)

Re: bug in gflSaveBitmapIntoMemory() (tiff only?)

what if...

bug in gflSaveBitmapIntoMemory() (tiff only?)

lossless

:(

gfl + jpeg2000 + changing quality