{ S T A G E   T W O } BA Graphic Design . . .

This is a record of my process and research during Stage Two of the course.


VISIBLE LANGUAGE {04-18/10/16}

For the first project starting the second year we were given the brief 'Resolution, Definition, Distortion' and asked to look at how these words and explore their effects through digital and analogue methods with the format of our final outcome being up to us.

Begining my research I knew I wanted to look at type in terms of distortion and legibility, in my first year I had looked at FF Beowolf a font that uses a digital programme to randomise it's shape in the printing process. The idea was to go against the uniformity of digital type, one of the designer Just van Rossum said'A certain roughness or varying unevenness is quite pleasing to the eye,...For reading, sameness is not necessary: we can read handwritten text, type superimposed on flickering TV images. The sameness of type seems an arbitrary thing..'. For me I found this really inspiring and used this as a refrence for a lot of my work in the past year. Although I still find this an interesting area of design to explore but for this project I wanted to look at another way that I could showcase the differences between handwritten and digital text.



Above is a Ted Talk video about reCAPTCHA a new version of CAPTCHAs. CAPTCHAs a used on websites for security purposes and are used to distinguish real human users from computer programmes. Before they were just random selections of words or numbers but the new system looked to used this system and the time needed to do these CAPTCHAs in a more useful way. Therefore the reCAPTCHA system instead is being used to digitilise books so that they are available to the public online. So now the words that are taken from scans of books, words that a computer programme wasn't able to recognise but that a human wouldn't stuggle with. What I liked about this was, one that something that is being used for security could be put to such a good effect, but also that although computers have come so far there are still things that we as humans can recognise and understand that a computer can't that the words through older printing methods are distorted in a computers eye but still legible to us and this was something that I wanted to explore in my work.


OCR (optical character recognition)

. . . is the recognition of printed or written text characters by a computer. This involves photoscanning of the text character-by-character, analysis of the scanned-in image, and then translation of the character image into character codes, such as ASCII, commonly used in data processing.


From this I decided I wanted to look more at how far computers recognition of printed or handwritten text has come and look more at how I would explore these ideas visually. Above is the definition for OCR which is the general term for this idea and from my research is what I wanted to focus on. On how things we see and can make sense of like a misprinted piece of text or a handwritten note can appear distorted to a computer since it isn't uniform or doesnt follow a certain pattern.

Look at an article on the Guardian website it seemed like computers although have come a long way with printed text still would really sturggle if given a piece of handwritten text especially in cursive lettering. 'OCR works best with high-quality printed materials and worst of all with handwriting, so you’re not starting from the best position. In my experience, you can only get handwriting recognition to work well enough by doing it in real time.' It also talked about a free OCR programme which although may not be the best quality could produce some interesting results if I tried using it in my experiments.


Through this I also found a new programme developed by the University College London, a programme that was able to recreate your handwriting through a small text sample. For me I find handwriting really interesting and how it varies person to person and am impressed that now with technology computers can replicate something so personal with such a high success rate.

'Up until now, the only way to produce computer-generated text that resembles a specific person's handwriting would be to use a relevant font,' Dr Oisin Mac Aodha, a member of the UCL team said. 'The problem with such fonts is that it is often clear that the text has not been penned by hand, which loses the character and personal touch of a handwritten piece of text. What we've developed removes this problem and so could be used in a wide variety of commercial and personal circumstances.'

This also show's how computers could learn to find patterns in handwriting although it still can't yet understand them as words it can pick up individual glyphs and so in theory could be eventually used to decipher handwriting although in this case it has to be a certain amount of text so notes written on different types of paper or less legible text could still be quite difficult for a computer to understand.



    Add comment

    Fields marked by '*' are required.
    Comments are moderated. If you choose to make this comment public, it will not be visible to others until it is approved by the owner.