Black White Photography - UK (2019-11)

40
B+W

COMMENT

When automated captioning to pictures comes into play, it raises some

interesting questions, as Vicki Painting discovers...

REFLECTIVE PRACTICE

@vickipaintingphoto

I

recently uploaded some photographs

onto a web page and each time I did

this, I found that every image had what

appeared to be an annoying caption

attached to it. This wasn’t something I

had knowingly activated but it was being

offered as a helpful suggestion should I wish

to include some text to explain my picture.

Initially I found this a rather odd suggestion,

as writing a simple description of the image

would seem to negate it. In addition, none

of the text offered was remotely helpful

and most was so way off the mark as to

be positively misleading. For example, the

text suggestion for a row of pollarded trees

growing in front of some buildings was

‘traffic lights on a busy road’. A few of the

descriptions covered the basics, yet seemed

to me to be only stating the obvious.

This was my introduction to ‘alt text’, or

alternative to the image text, the purpose

of which is to explain the content of

visual information and convey meaning in

instructional and communications materials.

Picture 1 Picture 2

‘...AN IMAGE IS SOMETHING WE VIEW AS HAVING

MANY LAYERS OF INFORMATION THAT CAN OFFER

A NUMBER OF EXPLANATIONS AND INTERPRETATIONS.’

we can all summarise the essence of it,

but there may perhaps be other information

embedded within the image that could be

interesting but would distract us from what

we need to know which, in this case, is the

most basic information that will form a

short explanation of the flower.

A

lt text doesn’t like complications

or nuance, it needs to be

compliant. Explanations offered

by alt text are only ever going

to be as good as their originators and this

is where alt text stumbles. Search engine

recognition technology has improved greatly

over the years, but basic search engine

explanations can only scratch the surface

of an image, and success clearly depends on

how the image might be classified – after

all, an image is something we view as having

many layers of information that can offer a

number of explanations and interpretations.

The upshot of this means that we, as

photographers, need to write our own alt

text when we upload our images in order to

ensure that we incorporate the information

that we want to convey about them, and to

also ensure that our images are properly

indexed and ranked on search engines.

Many questions arose as I continued to

experiment uploading. I was interested

to see how alt text would deal with subtle

differences around a similar theme. Working

with two pictures of the same woman, alt

text wasn’t able to recognise that it was the

same person, and began each explanation

as ‘a picture of ’ – a major pitfall. I was

amused by the fact that alt text appears to

conform to gender stereotypes as picture 1

is described as ‘a woman’ and picture 2 as

‘a person’, perhaps picking up on the more

traditionally masculine details such as the

bow tie. The alt text to picture 3 tells us

that these arms belong to a man, which is

also incorrect.

Of course, not all images require alt

text – the more fine art or conceptual the

image, the more difficult it is going to be to

nail it with a spot-on explanation. Alt text

dismisses such images as ‘decorative’ and

therefore not needing explanation. Does

this matter? Everyone is entitled to accurate

information, though that may be a minefield

in itself in these times of alternative facts.

With a limit of 125 characters, tell it like it is.

vickipainting.com

Of course, what I hadn’t grasped, coming

at this from the standpoint of someone

looking at an image, is that there has to

be an alternative to seeing. I had failed to

appreciate, in my sighted world, the necessity

for a textual alternative to visual information,

including photographs, which could then be

produced as alternative content.

This is in line with the principles of

Universal Design for Learning which

states that all information should be

understandable in multiple ways in order

that its meaning be available to people who,

for whatever reason, cannot see it, whether

due to technological limitations, such as poor

internet connection, or visual impairment.

Nobody is going to argue against this,

and for information in the form of graphs

and charts this makes perfect sense, but

alternative text for an image is surely a

different prospect and becomes slightly

more problematic. If basic information is

all that is required – for example, if we were

to write a description to explain a sunflower,