# "Add one" to every color in an image

You would simply take this image and make every color added one to every hexadecimal digit. For example, #49de5f would become #5aef60 (with the 9 looping to an a, and the f looping to a 0.)

This would also mean that all white (#ffffff) would become black (#000000) because all f loops back to 0, but all black will become a lighter shade of black (#111111).

Scoring is based on the least number of bytes used, as this is a question. If you want to, you can also use this other rainbow image: • What assumptions can we make on input/output format, if we would like to use a programming language designed for hardware/FPGA design? Jun 19, 2016 at 15:52
• @hexafraction I think by default you can take as a file or as hex code iirc. Jun 19, 2016 at 16:03
• @hexafraction meta.codegolf.stackexchange.com/q/9093/8478 Jun 19, 2016 at 16:03
• @Peanut, it would be good to have a test case which includes ef bytes (which should become f0 as opposed to 00 which happens when you just add 17 and take mod 256). Jun 19, 2016 at 17:44
• You should post an image for sample output, rather than just providing sample input. Other than that, this is a very nice first post! Welcome to the site! Jun 19, 2016 at 18:59

## Mathematica, 78 bytes

Image@Apply[16#+#2&,Mod[IntegerDigits[#~ImageData~"Byte",16,2]+1,16]/255,{3}]&


Takes and returns an image object (to create an image object, just paste the image into Mathematica).

Result for the test case: Taking input and returning output as a 3D array of integer channel values, this reduces to 51 bytes:

Apply[16#+#2&,Mod[IntegerDigits[#,16,2]+1,16],{3}]&


But those meta posts don't have an overwhelming amount of support yet, so I'm going with the 78-byte version for now.

## Pyke, 17 13 bytes

.Fh16j%ijcjs 8{


Try it here!

.F            - for i in deep_for(input):
h16%        -    (i+1)%16
+   -   ^+V
i16+    -    (i+16)
8{ -  unset_bit(8, ^)


Takes input as a 3d integer array of pixels and outputs in the same format • Can you provide the output image? Jun 26, 2016 at 16:03

# Verilog, 220 bytes:

• Programs may take input as an array of RGB pixel values with dimensions
• Programs may output via an array of RGB pixel values with dimensions

It's currently not clear as to how the dimensions are to be provided and if the array is to be streamed or provided all at once. I'm going to stream it 8 bits at a time using a clock signal (with a valid-data flag that goes low after the entire image has been processed) and input/output the dimensions as 32-bit integers:

module a(input[31:0]w,input[31:0]h,input[7:0]d,input c,output[31:0]W,output[31:0]H,output reg[7:0]D,output reg v=0);assign W=w;assign H=h;reg[65:0]p=1;always@(posedge c) begin v<=(p<3*w*h); p<=p+v; D<=d+17; end endmodule


## Python, 226 bytes

Now, it's valid !

Use the Pillow library.

from PIL import Image
m=Image.open(input()).convert("RGB")
for y in range(m.size):
for x in range(m.size):
t=m.getpixel((x,y))
h=t+(t<<8)+(t<<16)+1118481
m.putpixel((x,y),(h&255,h>>8&255,h>>16&255))
m.show()


Thanks to @TuukkaX for saving 9 bytes !
Thanks to @mbomb007 for saving 18 bytes !

• Is it necessary to use 0xFF instead of 255? Jun 20, 2016 at 23:23
• @TuukkaX Woops i haven't noticed that thanks you Jun 21, 2016 at 7:10
• There is one more 0xFF in there :D Jun 21, 2016 at 7:12
• I think you can save more bytes by saying from PIL import*. I also think that Image.open can be changed to just open after that. Jun 21, 2016 at 17:00
• @TuukkaX Yes, it can be changed to from PIL import*, but i can't change the Image.open Jun 21, 2016 at 17:23

# Dyalog APL, 21 15 bytes

Programs may output as a matrix of RGB pixel values

I assume output may be in the same format.

### New solution takes matrix of values [[r,g,b,r,g,b],[r,g,b,…

16⊥16|1+16 16⊤⎕


Explanation

⎕ get numeric input
16 16⊤ convert to 2-digit base 16
1+ add 1, i.e. 0 → 1, 1 → 2, 15 → 16
16| modulus 16, i.e. 16 → 0
16⊥ convert from base 16

Example

      ⊢m←2 6⍴90 239 96 255 255 255 0 0 0 239 239 239
90 239 96 255 255 255
0   0  0 239 239 239
16⊥16|1+⎕⊤⍨2/16
⎕:
m
107 240 113   0   0   0
17  17  17 240 240 240


### Old 21 byte solution takes matrix of [["RRGGBB","RRGGBB"],["RRGGBB",…

{n[16|1+⍵⍳⍨n←⎕D,⎕A]}¨


Needs ⎕IO←0, which is default on many systems.

Explanation

{}¨ for each RGB 6-char string, represent it as ⍵ and do:
n←⎕D,⎕A assign "0…9A…Z" to n
⍵⍳⍨ find indices of the individual characters in n
1+ add one to the index, i.e. 0 → 1, 1 → 2, 15 → 16
16| modulus 16, i.e. 16 → 0
n[] use that to index into n

Example

      f←{n[16|1+⍵⍳⍨n←⎕D,⎕A]}¨
⊢p←2 2⍴'5AEF60' 'FFFFFF' '000000' 'EFEFEF'
┌──────┬──────┐
│5AEF60│FFFFFF│
├──────┼──────┤
│000000│EFEFEF│
└──────┴──────┘
f p
┌──────┬──────┐
│6BF071│000000│
├──────┼──────┤
│111111│F0F0F0│
└──────┴──────┘


# C - 11411370666172 67 bytes

Here's the code (with support for Martin Ender's test case (without it's 60b)):

main(c,b){for(;~(b=getchar());putchar(c++<54?b:b+16&240|b+1&15));}


And here is less obfuscated version:

main( c, b ) //Defaults to int
{
//Get characters until EOF occurs
for ( ; ~( b = getchar( ) ); putchar( c++ < 54 ? b: b + 16 & 240 | b + 1 & 15 ) );
}


Compile and run with gcc -o color colorgolf.c && cat a.bmp | ./color > b.bmp

This code supports works with bitmaps. To convert png files to bmp, I used following command: convert -flatten -alpha off png.png a.bmp

Code assumes, that bmp header is 54 bytes long - in this case it works, but I'm not sure if I'm not discreetly breaking something.

Also, this is the rainbow: # Java 142 bytes

public BufferedImage translateColor(BufferedImage image){
for(int i=-1;++i<image.getWidth();)
for(int j=-1;++<image.getHeight();)
image.setRGB(i,j,image.getRGB(i,j)+1118481);
return image;
}


Golfed:

BufferedImage t(BufferedImage i){for(int x=-1;++x<i.getWidth();)for(int y=-1;++y<i.getHeight();)i.setRGB(x,y,i.getRGB(x,y)+1118481);return i;}

• First of all welcome to PPCG! Your code-golfed code uses the i both as parameter and in the for-loop, so I would change the parameter to something else like a. Also, you can golf it some more by removing the int  in front of the j and add it to the int i. So like this: BufferedImage t(BufferedImage a){for(int i=-1,j;++i<a.getWidth();)for(j=-1;++j<a.getHeight();)a.setRGB(i,j,a.getRGB(i,j)+1118481);return a;}. Also, have a look at this post: Tips for golfing in Java. :) Aug 2, 2016 at 9:04
• You're not setting the colors correctly. By adding 0x111111 to the RGB values, you have the possibility of overflow from one hexadecimal place to the next. For example, #49de5f is supposed to become #5aef60, not #5aef70. Aug 2, 2016 at 18:55

# R, 302 bytes

Far from perfect, but here goes:

a<-as.raster(imager::load.image(file.choose()));l=letters;n=as.numeric;d=a;f=function(x){if(suppressWarnings(is.na(n(x)))){if(x=="F")"0" else{l[match(x,toupper(l))+1]}}else{if(x==9)"A" else as.character(n(x)+1)}};for(o in 1:90){for(g in 1:200){for(h in 2:7){substr(d[o,g],h,h)<-f(substr(a[o,g],h,h))}}}


Explanation:

a<-as.raster(imager::load.image(file.choose()));  //chooses file
l=letters;n=as.numeric;d=a;                      //shortens some names and creates
duplicate variable to modify

f=function(x){                                //creates a function that takes in a
if(suppressWarnings(is.na(n(x)))){            character type and checks whether it is
if(x=="F")"0"                              a number or letter.
else{l[match(x,toupper(l))+1]}}            Returns the converted letter or number.
else{
if(x==9)"A"
else as.character(n(x)+1)}
};

for(o in 1:90){                       //loops through every single hex digit and applies
for(g in 1:200){                     the converting function, modifying the duplicate
for(h in 2:7){                    variable. Now you have 2 images =D

substr(d[o,g],h,h)<-f(substr(a[o,g],h,h))

}
}
} 