RGB Color Channel Decoding

A project log for itanimulli code crackers

cracking the code of the itanimulli riddle

stephenstephen 12/18/2015 at 05:460 Comments

Using the color channel FEZ images @Moritz Walter extracted from the waterfall PNG I decoded the code for each channel based on the FEZ alphabet. I assumed the orientation marker in each image to be in the upper-left and worked in a coordinate system with an origin in the top-left moving down-right.

The color channel PNGs were converted to BMP format. Fortunately the PNG was a 1000 x 1500 px image which proved extremely convenient for calculations.

Because of the sheer number of characters and the likelihood of transcription errors, a script was used to decode each channel map. The procedure starts with a color channel BMP and converts it to binary pixel data (1 for black, 0 for white). The data is arranged into a matrix and reduced using the dimensions of each FEZ glyph. Fortunately the glyphs are all the same size and evenly spaced!

Once reduced, each glyph is compared against the FEZ alphabet and converted to a letter. The end result is a letter matrix representing the original FEZ-encoded color-channel PNG.

Most of the output doesn't make sense, but there are some hidden messages in the output. An interesting note about the Red image is that the structure of the letters that spell out the message actually looks like a waterfall! Also interesting is the lack of information in the green and blue channels (perhaps I missed something?).
Red Channel:

a novice was enlightened

a novice should so find the river

a novice should search through water

a novice should so find the sky

a novice should search through air

a novice should so find a boulder

a reply from a master was

a solution to riddles hard

a novice asked how to find

a novice came to a master

Green Channel:

this is the wrong place

reading this text just right now

Blue Channel: