obrhubr

Initially I didn't realise historical data was available for free... I was also interested in learning more about the system itself, writing a diy decoder, etc... which is why I bought one. But yeah, kind of lost track of explaining that in the post :)

Thanks for pointing this out :), I fixed it!

I really enjoyed your post, it’s a very nice write up!

Your monitor and your browser 100% affect the appearance. After calibrating your monitor, try opening the image in full resolution and take a few steps back.

For me, viewing the images on my phone makes them look off.

Thank you for pointing that out. The Atkinson dithering I was using was indeed messing with the results. I'll be updating the post shortly :)

I’ll try them as soon as I get the chance, I have perceptual luminance implemented already. I’ll compare :)

I am weighting each of the channels according to the formula in my post.

I’ll try OKLab and compare, thanks for the comment :)

You’re right it kind of isn’t finished… I had it done, then had an exchange with the author of didder and I’m still in the process of rewriting :)

Thanks for your comment! I'm glad you're seeing the same thing :) I re-implemented the linearised dithering in python and got similar results. I checked and rechecked the colour profiles in GIMP, nothing... At this point I can only hope for an expert to appear and tell me what exactly I am doing wrong.

Great catch! I would be interested in how scammers keep coming up with these new methods? How would they even validate their new attack vector? I would like to think that there’s scam A/B testing or something similar…

Thanks for the kind words!

This is highly specific to C# which was the language imposed on all participants. But I agree that some languages might be especially adept to these kinds of tasks and it would be interesting to see which.

I didn’t think of that, but I guess the conclusion as to performance of the model would have remained the same.