It feels to me that the Tucson images with gamma correction applied are overall a little bit darker then the real image. I feel like the non gamma corrected version is more matching. Could you analyze that? Maybe it's just me / an optical illusion.
Well, there's a lot of subjectivity to it, and using a geometric function (square root) is an over-simplification. If we wanted to REALLY get it right, we'd have to look up the DICOM-14 spec or something like that and get an accurate perception curve.
Perhaps that top left corner artifact of Floyd-Steinberg dithering could be reduced by copying the left edge and top edge outwards. Like to make the picture 10-20 pixels wider and taller by adding 10-20 copies of the left edge and the top edge.
Thank you for this! I was looking for a good video that explained dithering and this really helped.
If you take a step backwards and squint your eyes.
It looks normal, no matter how close I get.
It feels to me that the Tucson images with gamma correction applied are overall a little bit darker then the real image. I feel like the non gamma corrected version is more matching. Could you analyze that? Maybe it's just me / an optical illusion.
Well, there's a lot of subjectivity to it, and using a geometric function (square root) is an over-simplification. If we wanted to REALLY get it right, we'd have to look up the DICOM-14 spec or something like that and get an accurate perception curve.
@@theosib guess there is the next video xD
but that makes sense thx
Perhaps that top left corner artifact of Floyd-Steinberg dithering could be reduced by copying the left edge and top edge outwards.
Like to make the picture 10-20 pixels wider and taller by adding 10-20 copies of the left edge and the top edge.
Oddly, I prefer most of the toucan outputs without the gamma correction. I feel like I can see more meaningful detail the lighter images
I probably should use an image with more lighter stuff in it. I think that toucan is overall too dark.
very nice video
I wonder who watches the video early enough that they get to hear a flub I edited out using youtube studio. 😁
the one where there is a 5 second silence?
@@paulwesley3862 Yeah. UA-cam said it could take ours to take effect.
Is there a reason behind the gamma corrected dithered images being consistently darker than the ground truth?
Neat.
Why is each pixel's value multiplied by 4 before rounding?
can u also make one on RGB and YUV