Portraits to Text was developed in July 2014 by Thomas Zimmerman as a Sweat Pants Era project. You can comment on this project and interact with other people interested in this project on Facebook at: PortraitsToText Facebook group. This site is funded through advertising, which is to say that it generates almost no money. If you need to interact with us to troubleshoot a problem or you need us to answer a question, please do so on Facebook.
Quick Start Guide
Portraits to Text redraws your photo using words of your choice. The new image consists of a single background color and each font character has a single solid color.
Selecting a Photo
Redrawing a photo using words results in a dramatic decrease in the resolution and density of the image. This process also introduces a lot of "white space" between and around the letters that may be compensated for by selecting an appropriate background color. For best results, your photo should focus on a single subject/person and you should crop out as much of the background as possible before uploading the image (this tool does not have cropping capability). If your photo is of a person, that person should consume most of the area of the photo. You can experiment with text orientation, font face, font colors, and font sizes in the interface.
You can paste any text of your choice. You will need at least a dozen words of varying lengths for the generator to work properly. You should make sure that you provide several very short words that are one to three characters long. Absent these very short words, the generator will attempt to fit large words into these areas and text intersections will occur. The maximum amount of text that you can upload depends on the amount of data that your device will send to our server. On most laptop and desktop computers, textarea contents can contain several million characters.
You can select from one of the several dozen books that we have provided or you can find your own vintage text at: Project Gutenberg or similar web sites.
Why Do my Output Images Look Terrible?
Sorry, a lot of the things that I have made look terrible too. Keep trying and you will get a better feel for how the generator works. If your output is too light, add a darker background color. Try different fonts. Try setting a completely black background and set the text color to black and white. Try changing the format/orientation of the text. Try using more or less of a span between your smallest and your largest fonts. Again, this tool works best when you crop your photo to focus on a single thing or person.
There is a Word in my Output Image that I Don't Like
The generator pulls from the list of words that you send to it. If there are words that you want to remove inside one of our stock books, download the book here. Copy and paste the text into notepad or into Word. Do a search and replace and remove the word that you don't like. Then, select the option to Paste Text to Use in Step 2 and upload your revised book.
How Do I Make My Output Look Less Like a Regular Photo?
If you want your output to look less like a photo, you should feed the tool an image looks less like a photo. You can use filters, tools, or apps available on your computer or device to change your photo prior to uploading it.
In the image above, I took the following steps to get this output.
A Note about Copyright
You should not make artwork from a copyrighted photo. Converting a photo to text does not strip the copyright from the original. We are not certain about whether or not using most frequently occurring words from a copyrighted text is permitted under Fair Use. Therefore, we request that you do not use any photos or text that are copyright. We reserve the right to remove without notice any artwork on our server that we believe violates copyright. We will remove all artwork of any user if a third party reports copyright infringement.