33

pan up and to the left

Agent: OK, let's test this out... pan over to that art on the wall... zoom in... Enhance!
Agent: [under breath] I've always wanted to say that...
Agent: Excellent, it's crystal clea-- Wait... What's that reflection?
Me: It looks like a face.
Agent: Ok, pan up and zoom in... Now, enhance!
Agent: Wow! Incredible! There's a slight fringing, but the detail is unbelievable! How does it work?
Me: Well it's based off a recent discovery I made about the fundamental recursive nature of the quantum membrane--
Agent: Shut up. I don't care. Zoom in on that reflection in the eye... ENHANCE!

[...8 enhancements later...]

Agent: Ok, pivot by ninety degrees... There! What's that hanging on the wall? Zoom in...
Me: *sigh* It's still just the same cross-stitch art as before, only from yet another slightly different angle. Nothing interesting.
Agent: What would you know about detective work? It's probably some sort of clue... Enhance!
Agent: OK, crop that and print it out... Let me have a look.

corrupted cross-stitch

Agent: What the-- That makes no sense. Clearly your software is rubbish. Get out! You're fired!
Me: Hmmm... Looks like the enhancement algorithm just has an off by one error in the fractal decompression subroutine--
Agent: What? Are you still here? Security!!

What does the cross-stitch say?

Hints:

Although you need to focus almost entirely on the cross-stitch, there's still a couple of clues to be found in the other parts:

  1. Take another look at the first picture (though, as I mentioned in the comments, you can ignore the text at the bottom right)

  2. "the enhancement algorithm just has an off by one error"

Alconja
  • 37,250
  • 16
  • 116
  • 164
  • 6
    I like how the enhancement algorithm got the border perfect and only screwed up the letters – question_asker Mar 17 '16 at 10:32
  • @question_asker how do you know the border is showing perfectly? :-) – Carl Löndahl Mar 17 '16 at 10:46
  • @carllöndahl it's too regular not to be! – question_asker Mar 17 '16 at 10:53
  • 7
    @question_asker - Yeah, the bug must... uh... only affect certain colours... Or maybe because the border is so regular it was easier to recalculate... Or something... ;) – Alconja Mar 17 '16 at 11:12
  • @Alconja Very cool puzzle though, as usual. Wish I had the tools to work it out! – question_asker Mar 17 '16 at 11:13
  • 1
    @question_asker - you don't strictly need tools to solve it (though tools would allow short cuts). Print it out and grab a pen. :) – Alconja Mar 17 '16 at 11:20
  • 2
    @question_asker - come on, this is a puzzle site, think outside the box! Copy the grid into excel? Online image editor? Sharpie on your monitor? – Alconja Mar 17 '16 at 11:37
  • @Alconja OK, I'm gonna throw this into Excel but I'm pretty sure someone will have it answered before I get halfway done :) – question_asker Mar 17 '16 at 11:45
  • @question_asker - Awesome. Good luck. :) – Alconja Mar 17 '16 at 12:10
  • most cross-stiches are samplers, so 'ABCD..XYZ'? – JMP Mar 17 '16 at 12:54
  • @JonMarkPerry - nope. Makes a fairly well known phrase. – Alconja Mar 17 '16 at 13:05
  • 3
    If I scroll the browser window up and down, quickly, I can almost read something. But I can also tell you that some of the marks don't belong to the letters (or they've been tossed away from the letters, as fragments)... Oh, and if you try this, don't be surprised when the letters jump around. Your brain will keep switching between noticing the blue as primary, and the red as primary. It's annoying. – Khale_Kitha Mar 17 '16 at 14:20
  • 1
    is it possible to have a look at the 'enhancement algorithm just has an off by one error in the fractal decompression subroutine' - this might give us a clue – JMP Mar 17 '16 at 14:42
  • Yah. I have a feeling that just lining up the grids isn't going to give us an answer. @Alconja makes hard puzzles. My bet is the letters and numbers in the bottom right mean something rather important. Not just the "zoom level" and "location within the picture." – Z. Dailey Mar 17 '16 at 20:57
  • @JonMarkPerry, I'm afraid the algorithm itself is proprietary and can't be disclosed. But you're right, that line does give some information away.... – Alconja Mar 17 '16 at 21:38
  • @Z.Dailey - not all my puzzles are super complex, and this one is on the simpler end of the spectrum... Though I suppose simple doesn't necessarily equate to easy. – Alconja Mar 17 '16 at 21:43
  • @Alconja, I'm picking up on the word recursive, just not quite sure how recursive.... Maybe I'm not being recursive enough to come up with anything other than jumbled up stuff haha. – Z. Dailey Mar 17 '16 at 22:08
  • Does it say: https://www.youtube.com/watch?v=jofNR_WkoCE – Z. Dailey Mar 17 '16 at 22:10
  • @Z.Dailey - Don't want to drop hints just yet, so won't comment on recursion, but I will say (re: your previous comment) that whilst I have scattered some clues, you can safely ignore the lettering at the bottom of the first picture. That's just a nod to blade runner. :) – Alconja Mar 18 '16 at 03:59
  • @Alconja ok. I gave it shots on doing 2 different things I know one of which is how you solve it. Just couldn't figure out exactly what I needed to do from the clues. I'll be watching for someone craftier than me to give a whack at it. Nice added touch btw. As usual. – Z. Dailey Mar 18 '16 at 04:05
  • 1
    So I did manage to get all the reds plotted yesterday (do you have any idea how long it takes to do that in Excel?) and started toying with those, but I'm starting to feel like I can't do it without also (separately?) plotting the blues. – question_asker Mar 18 '16 at 14:13
  • @question_asker - you can probably work with just one colour initially... I think you should be able to tell when you're on the right track. – Alconja Mar 18 '16 at 15:37
  • I wasted an hour trying to move around the different red and blue layer x's in Paint.NET, and I'm still as confused now as I was then. The only slightly legible thing I can make out is a "t" and maybe a "0" in red, and a backwards "g" in blue..... but I could be completely on the wrong track. – Martin Walker Mar 18 '16 at 22:13
  • are the colours used significant?, AND is the phrase 'Conway's Game of Life'? – JMP Mar 21 '16 at 01:39
  • @JonMarkPerry - Only in that they're different. I could have used any two colours. And, no, nothing to do with the Game of Life. – Alconja Mar 21 '16 at 02:09
  • Hint #1. I think @Alconja is trying to get us to notice the faint grid in the background being the same as the one in the second image. --- I could be way off base, though. – Khale_Kitha Mar 21 '16 at 17:44
  • @Khale_Kitha - I'm afraid you're way off base. What other similarities are there? ;) – Alconja Mar 22 '16 at 01:42
  • Pity - the other thing that's been poking me in the eye is that the zoomed out version of the image appears to all be blue lettering. – Khale_Kitha Mar 22 '16 at 01:45
  • @Khale_Kitha... I would've said more purple than blue... – Alconja Mar 22 '16 at 02:11
  • has the photo been transposed left 2 right? and 93 221 is the drug risperidone, used to treat schizophrenia - is this a clue? – JMP Mar 22 '16 at 10:45
  • @JonMarkPerry - here's the original image if you want to compare. And as I said in a comment above (and in my hint), you can ignore all the text in the image. – Alconja Mar 22 '16 at 11:30
  • Going by "off-by-1" implies that we have performed the same error 11 times to the colours in the picture. Might it be safe to assume that the 90 degree rotation is also to be accounted for? – John Mar 23 '16 at 14:30
  • @JJChivers - might want to count again. :) Also, if by "90 degree rotation", you mean the fact that the final picture is portrait, that's just because the agent asked to crop the final image. If you look at the original picture, you'll see there's been no rotation. – Alconja Mar 23 '16 at 18:04
  • @JJChivers - oh, wait... Just realised you're probably talking about the 90 degree pivot in the text. That's just me adding to the ridiculousness of the situation (i.e. implying that you could pivot a 2D image through 3D space). You can ignore that, it's just flavour. – Alconja Mar 23 '16 at 18:41
  • I really worked on this one for a day or 2, but I've been busy... going to have to take another stab at it... – Z. Dailey Mar 23 '16 at 19:17
  • Is it an off by one or off by 2 problem? :) – Z. Dailey Mar 23 '16 at 23:02
  • @Z.Dailey - I guess it depends on how you count... Have you found something? – Alconja Mar 23 '16 at 23:21
  • Yah. Gotta finish it tonight tho. @Alconja – Z. Dailey Mar 23 '16 at 23:23
  • 1
    I took the route of plotting in excel too. Takes a long time. – Z. Dailey Mar 23 '16 at 23:24
  • @Z.Dailey - Well if you do get it with excel, you'll have definitely earned the bounty. :D – Alconja Mar 23 '16 at 23:28
  • Yep. Got a bit done, not 100% sure on the pattern so I may end up having to anagram it. But I'm coming up with letters. XD – Z. Dailey Mar 23 '16 at 23:30
  • @Z.Dailey - letters are good. Re: the pattern, see if anything you've done matches up with my hints (and comments above about counting). – Alconja Mar 23 '16 at 23:44
  • 1
    Never finished plotting the blue... im definitely doing the right thing now but so tired.... I'll work on finishing up the blue and I may have it solved tomorrow... we'll see. – Z. Dailey Mar 24 '16 at 03:26
  • can we have more hints please? – JMP Mar 24 '16 at 08:35
  • Would I be right to guess that blue pixels are offset horizontally and red ones, vertically? – GettnDer Mar 24 '16 at 14:39
  • @JonMarkPerry - I think a few people may be getting close, so I'll hold off on hints for another day or so... – Alconja Mar 25 '16 at 04:15
  • @GettnDer - if you have evidence of something by way of hints, then it's probably worth pursuing... – Alconja Mar 25 '16 at 04:26
  • @Alconja I do have evidence of how to move the dots, but any method I try just ends up moving the dots of the canvas, so I'm assuming it must be wrong :( Great puzzle by the way! – DaveBensonPhillips Mar 25 '16 at 21:08
  • @HumphreyTriscuit - perhaps not the best assumption... How could you deal with stuff that goes off the edge? – Alconja Mar 25 '16 at 21:14
  • @Alconja I feel I know what you mean, and I'm trying to account for it but am going very crosseyed :) I thought I knew everything about any offsets, and excitedly plotted it all in Excel -- as of yet to no avail, so I think I'm missing something else – DaveBensonPhillips Mar 25 '16 at 22:21
  • Ha! I totally missed out on this puzzle. Great stuff, as all of your work! – BmyGuest Jul 08 '16 at 22:58

1 Answers1

23

Solution:

The more things change, the more they stay the same!

Looking at the first picture,

the red lines are duplicated and offset vertically, while blue lines are duplicated and offset horizontally.

"the enhancement algorithm just has an off by one error"

The image is enhanced 12 times in total, and assuming the off-by-one error is an off-by-one-pixel error, you would expect any offset to be 12 pixels.

"There's a slight fringing"

This suggests that the offsets are in both directions (up and down for the red pixels, left and right for the blue pixels).

So the trick is

to duplicate the red pixels, shift them vertically 12 pixels both up and down, and then do the same for the blue pixels, but left and right, and wrapping around pixels that go off the edge.

Here's the result (made in GIMP):

pixel-based solution

And here's the "original" cross-stitch (from OP):

"original" cross-stitch

Alconja
  • 37,250
  • 16
  • 116
  • 164
Volatility
  • 5,099
  • 1
  • 21
  • 42
  • 1
    Whoa, nice find! – Deusovi Mar 27 '16 at 23:39
  • 1
    This is fantastic, well done. I was caught up on the fact that the image sugguests the blue and red offsets are down and right (or up and left, depending on how the algorithm works) by the orthogonal red/blue fringing lines. I wouldn't thought of duplicating anything! I'm not sure how "There's a slight fringing" suggests the offset is in both directions, either, but clearly you have the right answer. Great puzzle. – DaveBensonPhillips Mar 28 '16 at 00:24
  • 1
    @HumphreyTriscuit thanks. Re: the fringing, I didn't pick up on the hint initially and only recognised it after I figured out the solution, but in photography, fringing occurs on both sides of an image. – Volatility Mar 28 '16 at 00:49
  • 2
    I had the red about done and the blue half done. (Notice my off by 2 comment referring to 2 chunks of the grid) You beat me to it. Nice solve! – Z. Dailey Mar 28 '16 at 00:50
  • @Volatility yeah I realised this after I typed my comment. Again, just my misinterpretation of the image. Did you just try manually drawing stuff? – DaveBensonPhillips Mar 28 '16 at 00:51
  • @Z.Dailey I think you severely handicapped yourself by using Excel :-P – Volatility Mar 28 '16 at 01:17
  • @HumphreyTriscuit I plotted the pixels onto two layers in GIMP (one for red and one for blue), then played with them a bit (duplicated them, moved them around). – Volatility Mar 28 '16 at 01:25
  • Very well done. Glad to see you picked up on the hints. I think the only other clue was simply the references to fractals, which contain lots of self-similarity, hinting at the copying, rather than moving of the "pixels". – Alconja Mar 28 '16 at 09:52
  • 3
    Just edited an "original" version of the cross-stitch into your answer, and while I was in there, I scaled up your pixel art a little to make it easier to see. Hope that's ok. – Alconja Mar 29 '16 at 05:37