Last week I entered the 2014 EPSRC photography competition. It’s specifically for anyone funded on an EPSRC grant (which I am) who wants to share a photograph from their research (which I always do!).

I wouldn’t normally bother with this kind of competition but my colleague suggested it after I took a particularly striking photo graph of a curved fiber with a nice colourful effect.

A spectra separating fiber optic

But as nice as it is, I’m not writing an entire blog article about my unquestionable innately amazing science photography skills. What interests me is the use of photoshop in science photos.

Photoshop for those of you who an have inexplicably been living in a cave for 20 years and just recently re-emerged to read my blog, is a catch-all phrase for digitally altering photographs. This can range from innocuous edits such as minor brightness, contrast and exposure correction, to wholesale re-invention of the photographs.

I’ve been an amateur photographer for some time and I have pretty clear self-imposed rules on photoshop. The only edits I’ll do are to try and better capture what I see with my naked eye. No creating colours that weren’t there and no adding or removing anything in shot. For example, below is a photo that I took before and after photoshop.

Untitled - Version 5

Untitled - Version 2

I could go further, I could for example remove the gap in the star trail (created by a cloud appearing and then moving away). But that I feel is going too far – that’s not what happened and is a kind of re-shaping the photograph to be what I want not what was actually there. If I want a better photograph I’ll have to go back and try again and hope for clearer skies. Although I can’t deny that a quick fix in photoshop is tempting, especially when compared to standing on a cold hillside for 2 hours at 1am. This temptation is even greater in scientific photographs.

Whenever I collect data I always have to fight that part of me that wants to see my theory in what ever data I produce. I try to be impartial but it’s human nature that if you think it’ll be linear you see a linear plot. This is compounded even further when looking at complex photographs. If we really believe we’re going to see the Pepsi logo in some random data then that’s exactly what we’ll see.

So enhancing science photos in photoshop can be a bit of a self-fulfilling prophecy. You choose the tools (of which there are hundreds) with which to edit your photos and there is a temptation to modify and correct your photo in ways that may enhance an artefact that you think is your data but not actually improve the data. Worse, you may accidentally misrepresent the data and add conclusions not actually possible from the original image.

For example, what if I told you that I’d taken some microscope images of some festive decorations. Now the microscope I used only had a black and white camera so to aid the eye I’ve colorised the image so you can clearly see the different festive components. I’ve also included the original B/W version so you can compare.

Red and greem

Cholesterol

As you can probably tell the first image is just randomly coloured in. What I’m trying to highlight is that by modifying images you can mislead without wanting to. It’s very easy to think you’re just helping to clarify things but sometimes it’s hard to impartially draw the line between clarifying and fabricating.

With the photograph I submitted for the photography competition, each shot took 3+ hours of exposure. Trying various camera settings and setups took days of work to get right. Work I could have avoided by simply ‘correcting’ my image. The first photograph I took (below) isn’t hugely different but it’s out of focus and very dark.

2014-11-25 at 08-58-15 - Version 2

When I first saw it one thing I noticed was that there was no blue end of the spectrum after green as you might expect. I figured that this was simply because the photograph was so underexposed (and slightly out of focus) that it wasn’t picking it up. A simple correction would have been to do some photoshop magic, sharpen the picture and add in the ‘expected’ colour. Bust instead I went back to the lab spent another day fiddling with my camera and got a great in-focus better exposed shot, with no blue light. Turns out the blue is just not appearing – I was wrong.

I’m not saying all photoshop is bad but it’s powers can be used for good and evil and I think if you open the image tool box that we should all think very carefully before air brushing data into something it’s not.

Finally, in the interest of openness here is the completely un-photoshoped version of my competition entry so you can see what modifications I chose to make.

A spectra separating fiber optic


0 Comments

Leave a Reply

Your email address will not be published.