janimatic

Hello, Thank you for your great work! I am still experiencing the same crash with ofPixels::resizeTo (OF_INTERPOLATE_NEAREST_NEIGHBOR & OF_INTERPOLATE_BICUBIC, OF_INTERPOLATE_BILINEAR is not implemented). In the meantime, i wrote a very...

Hello Michael, I am using the latest master (cmake based project). The crash happened only with `ofPixels_`, `ofPixels_` worked fine if i remember well. I suspect there might eventually be...

Hello @dimitre @johanjohan Here is a minimal code example... thanks! ``` #include "ofMain.h" template void resizeTest(int width, int height, int thumbnailProxy, ofInterpolationMethod interp) { ofPixels_ pixels; pixels.allocate(width, height, OF_IMAGE_COLOR_ALPHA); auto...

Hey @dimitre @johanjohan , I think replacing line https://github.com/openframeworks/openFrameworks/blob/7a09e9eee85b8b98f8447072d6259544a0b9cb37/libs/openFrameworks/graphics/ofPixels.cpp#L1347 `size_t bytesPerPixel = getBytesPerPixel();` by `size_t bytesPerPixel = channelsFromPixelFormat(pixelFormat);` would fix OF_INTERPOLATE_NEAREST_NEIGHBOR (renaming the variable accordingly of course) sizeof(int) was ok...

PS : This seem to work (but should be tested) ``` //---------------------------------------------------------------------- template bool ofPixels_::resizeTo(ofPixels_& dst, ofInterpolationMethod interpMethod) const { if (&dst == this) { return true; } if (!(isAllocated())...

Hello @dimitre , thank you very much for this. For the moment, testing with float pixels, * if i use this minimal test, i get funky logged values ``` #include...

@dimitre great! i just modified it with templated color and it seem to work with floats...perfect... ``` template void resizeTest(int width, int height, int thumbnailProxy, ofInterpolationMethod interp, ofColor_ col) {...

@dimitre for the moment, the resize seem broken with 8 bit pixels too (just tested in the context of Qt thumbnail generation that works in master, I didn't write a...

hello @dimitre Your small fix suggestion on master seems perfect. For the refactor pr testing, I wrote some test code for float images io... It could not be very minimal...

@dimitre hello! * the Qt thumbs are working with int anf float pixels (Qt6 QImage::Format_RGBA32FPx4 thumb looks ultra pixalated.) * ofImage_ / freeimage resize removes the alpha channel * this...