Oh nice! Sorry for the slow reply. It looks like Mateon1 might have already solved it, but I’m going to take an independent crack at this before looking at the solution they came up with.
So far I’ve established that it appears to be
a pattern of bits that is 686 bits of header, followed by 500x600x56 bits of message, followed by 2 0′s. The 500 and 600 length sides appearing to have neighboring cells have similar value / behavior, and the side of length 56 representing maybe channels. Every 7th channel seems to be different than the other 6. There is some easily visible structure in channels 1-4, 10-13, 29-32, and 38-41.
Alright, I managed to figure out what the message is and where it came from. I have not managed to write a better compressor than gzip yet, and expect that doing so will require picking up some computer skills that I don’t currently have. But I think it’s something modern software and hardware is capable of.
Edit: Answer: extremely far behind. I am very impressed. Also I take back my statement that I could probably beat the best general compression algos here given a reasonable amount of time, because
it’s a blurry picture of something that had lots of pretty diffraction patterns, which was then converted to jpeg. Deconvolution is a thing, and can sometimes recover information from blurry pictures, but the conversion to jpeg destroys some data and I don’t think deconvolution algos are robust to lost data. And also the data that looked random actually was random.
I probably could tell you some details about the camera used, focal distance, etc though after some playing around with Blender. And someone who knows a lot more physics than I do could maybe tell you interesting things about the light source by looking at the interference patterns in the layer of bismuth oxide in the parts of the photo that aren’t blurry.
Interesting choice of message to encode—it’s one that (aside from the lossy compression aspect) would actually be quite a bit more informative about the laws of physics in the universe it came from than a picture of a falling apple would be about the laws of physics in our universe.
Oh nice! Sorry for the slow reply. It looks like Mateon1 might have already solved it, but I’m going to take an independent crack at this before looking at the solution they came up with.
So far I’ve established that it appears to be
a pattern of bits that is 686 bits of header, followed by 500x600x56 bits of message, followed by 2 0′s. The 500 and 600 length sides appearing to have neighboring cells have similar value / behavior, and the side of length 56 representing maybe channels. Every 7th channel seems to be different than the other 6. There is some easily visible structure in channels 1-4, 10-13, 29-32, and 38-41.
My progress so far is at https://imgur.com/a/SM0gY1o
Alright, I managed to figure out what the message is and where it came from. I have not managed to write a better compressor than gzip yet, and expect that doing so will require picking up some computer skills that I don’t currently have. But I think it’s something modern software and hardware is capable of.
Update on progress: https://imgur.com/a/6r7VcDV
Time to see how far behind Mateon1 I was.
Edit: Answer: extremely far behind. I am very impressed. Also I take back my statement that I could probably beat the best general compression algos here given a reasonable amount of time, because
it’s a blurry picture of something that had lots of pretty diffraction patterns, which was then converted to jpeg. Deconvolution is a thing, and can sometimes recover information from blurry pictures, but the conversion to jpeg destroys some data and I don’t think deconvolution algos are robust to lost data. And also the data that looked random actually was random.
I probably could tell you some details about the camera used, focal distance, etc though after some playing around with Blender. And someone who knows a lot more physics than I do could maybe tell you interesting things about the light source by looking at the interference patterns in the layer of bismuth oxide in the parts of the photo that aren’t blurry.
Interesting choice of message to encode—it’s one that (aside from the lossy compression aspect) would actually be quite a bit more informative about the laws of physics in the universe it came from than a picture of a falling apple would be about the laws of physics in our universe.
Thanks for running this challenge.