If you zoom in on all of the pictures that have been created by the interleaving process, you go from here:

to here:

I find that kind of neat because the picture as a whole looks like it’s mostly black and white. Or black and gray, I guess, is more appropriate. There is a little bit of color scattered around the picture, but definitely nothing as bright as what exists, apparently, all throughout the picture.

This seems like a weird sort of pointillism that exists on top of the pointillism of all computer images. That is, pointillism fixates around dots being used to create images, and these dots have a modern analogue in computer pixels, which all digital images are made from. However, in these interleaved photographs, there are distinct dot clusters formed from pixels and which go on to form the top image.

Okay, okay, not ALL digital images are made with pixels. All raster-based images are. Not vector-based. And all digital photographs are raster-based, I believe.

I created a page with some of the images that I liked the most.

You’ll notice that a lot of the pictures are pretty similar. This is because I was just experimenting and didn’t want to be messing with a whole range of photographs. I wanted some sort of constant to see how many ways a single photograph could be altered. However, if I wanted to explore a certain technique a little bit more, I used different pictures to see how different types of photographs (color, black and white, etc) would be affected.

The pictures used are:

The other stuff was kind of neat, but I think it might have been getting a little too abstract. It was also getting away from converting Nietzsche’s work to an image, which is what I want to stick with.

I decided to take an existing image, convert it to text, and then combine it with the text of Beyond Good and Evil. Since the text of Beyond Good and Evil is so much shorter (402,272 vs. 3,686,400 characters), I made sure that the shorter text repeats itself until it makes it all the way through the image.

Not a bad start!

Well, darn. It would have been nice to have a neat image just come right out of PIL for me.

To continue from last post’s experiment – where I converted an existing picture to a string – I’m wondering what happens if I convert an image to a string and back again. Will it be the same as before?

Answer: Yes, it will be the same before, but only if the same dimensions are used. Otherwise, everything goes crazy, as might be expected.

Here, I am using this image, and converting is from a 1280 x 960px RGB picture to one that is 960 x 1280px RGB.

Original – 1280 x 960px RGB

960 x 1280px RGB

The result is pretty weird. The photograph gets horizontally collapsed into thirds and is repeated in each one of those thirds.

800 x 600px RGB

Refinery – 800×600 – CMYK

The last one is the same, but with a different color model – CMYK as opposed to RGB. The streetlights look almost floral.

Refinery – 600×400 – RGB

You get the idea.

The images created might not have been all that cool, but they do kind of provide a springboard for some questions about how the text is converted to images. All the questions you didn’t know you had are answered in a whole bunch of detail on the JPEG Wikipedia page

Failing to understand that Wikipedia page, I guess one way to get some insight is to convert an image into text.

There are 3,686,400 characters in a digital picture (694kb) converted into a string of text.

Here are characters 56,000 – 60,000:

Pretty ugly. In fact, these characters are different from text editor to text editor, so they probably don’t match any sort of character encoding and are more just a representation of the bytes that are actually being stored in the file. But we can definitely see that regular English characters aren’t going to make much of an image – explaining why we get such ugly blobs of pixels with Beyond Good and Evil.

For reference, here is what Beyond Good and Evil looks like when its text is repeated 24 times until it hits about 9.5 million characters.

Enlarged version of Nietzsche’s Beyond Good and Evil compressed to 10 x 10 pixels

Nietzsche’s Beyond Good And Evil compressed to 1280 x 960 pixels

These images are my first tries. They are just Nietzsche’s Beyond Good and Evil converted directly into images. Not very exciting. The first one is converted into a 10 x 10 pixel image, while the second one is 1280 x 960 px.

I got the idea for this project a while ago when I was working with the Python Imaging Library (PIL) for the first time. First of all, PIL is a Python module – meaning that it lays a set of easily accessible capabilities on top of the Python programming language. In this case, the features that are provided are all image-based.

I was looking through the documentation for the PIL and one of the functions struck me as being potentially interesting – the ability to convert text to pictures and vice versa. Considering myself an occasional practitioner of abstract photography, as well as being familiar with some texts, I thought I would try to convert a book into an image. But I never really had the time or an excuse to do so until recently.