Search:

i'll see if i can get this in the next build.

here's a workaround until we fix the JoinX problem:

go to More Options. under General Operation, go to the Macro Expressions page.
click on "_LinkFiles".
replace the definition with the...

looks like a bug. i'll see what i can do about it...

yep.



you shouldn't need to do that. the DLLs are built with static CRT linking.



no. the DLL and the static libs behave identically

libs are never compatible between versions of MS compilers. so, no, don't bother trying any existing IS libs with VS2010.

and, no, we don't have any VS2010 binaries available for any version of...

we don't currently have any VS2010 static libs, for any version of IS.

(though, if you have the ImgSource source, you can build your own)

the algorithm is:

1. get a grayscale version of the image
2. if you're using mode 0, do a histogram stretch on that grayscale image, using loPercent = lowThresh / 255 and hiPercent = (255 -...

nope.

nobody wanted the job, and time i have available to do development has been reduced to barely a trickle.

i suspect you should be able to fix the problem by moving your call to _is5_Exit.

the delay, i suspect, is in the thread pool cleanup. in is5, worker threads are deleted as a result of _is5_Exit....

ok, then it looks like KFactor is 1/k, in IS*_ApplyConvolutionFilterToImage. and you can just add KOffset to each pixel with a LUT.

there will be slight differences on the edges of the image,...

since KOffset is just added to the output, you can duplicate it with a LUT:


BYTE lut[256], *LUTs[1];
for (int i =0; i < 256; i++)
lut[i] = i + KOffset;

is5_ApplyLUTsToImage(...LUTs)

interesting!

and just FYI, there's a function in is5 (and ISEffects 4) which will calculate stdDev for you.

what do KFactor and KOffset do ?

what are the HSV values of your target?
what are the HSV values of a typical background pixel?
what are you using for tolerances ?

that's really cool :)

would it be OK if i linked to that from the ImgSource front page ?

i'll take a look at the code and get back to you ASAP. (don't have the source with me at the moment)

what are the values of the parameters you're passing ?

i'd say the Windows text smoothing gives slightly better results. but that's only because the character shapes are generated to display exactly at the size you specify, while the SmoothText functions...

no. this doesn't apply to the is5_DrawSmoothText functions. it only applies to:

is5_DrawTextOnRGB
is5_DrawTextOnRGB2
is5_DrawTextOnRGBA
is5_DrawTextOnRGBA2

what does is5_GetLastError say ?

there's a uFlags option in is5_DrawTextOnRGB* (and the v4 versions) which tells the function to use Windows font smoothing. we made some improvements to the performance when using that path.

w * h * 3 * sizeof(double)
sizeof(double) = 8 bytes

"double" in C/C++ lingo is "double-precision floating point", not "double-width integer".



i've been trying to change those to...