I guess almost everyone has a digital camera these days, and we have all tried to enhance an image in some sort of image editing software. Now those who tried probably found out that editing images is difficult, and that for the most part it cannot be done with a few mouse clicks. From Adobe’s famous image editing software PhotoShop, this art has gotten the name “photoshopping”. If you know what you’re doing, you can substantially alter an image to your tastes. Done right, it can be an art in itself.
So, what’s wrong with photoshopping?
Lets just take an example, okay? Take this image that came through Reuters in 2006. What do you think it shows?
It shows Beirut on fire, right? And heavy black smoke, right?
Readers who know what a “clone brush” is, please look closely at the smoke in the upper left corner. Does that look “cloned” to you? It sure did to somebody.
Now as it turned out, that image wa really a photoshopped version of this image:
That is Beirut with one building on fire and some smoke.
Now, I’m totally not getting into why there was an israelian air strike in the first place, and what would be right or wrong with that. What I’m aiming at, is that we just cannot trust images anymore the way we used to. There have been numerous occasions where photoshopped images were released as “news”, this just happens to be a very prominent example.
This raises quite a few questions:
Q1. Why was the image altered in the first place?
I don’t know much about Adnan Hajj, the photographer, so I can only do guesswork on question 1. A lot of photographers get paid only if their photo gets published. So if you are a photographer, you want to get images that are exciting and news-worthy. The manipulated image sure looks more alarming than the original. With the level of competition on the news market, news agencies have outsourced more and more staff. People who only get their money if they sell a photo will be more likely to forge something to make a living than somebody who gets a regular salery no matter what.
On the other hand, Hajj is lebanese, and might have done that to shift public oppinion. But using the same kind of reasoning, the whole thing could also have been a setup to discredit anybody reporting on that side of the war.
Sorry, I cannot give you the real answer on what happened here, but it turns out not to be as important as you might think. Sure, Reuters could rely more on their own staff, but they cannot do so entirely. And there will always be somebody tampering with images just to get them sold or to make a point.
Q2. Did Reuters know the image was modified?
I’m not sure which answer I should hope for. Either they didn’t, which would be an unbelievable level of neglegance, or they did, which would be a level of bias only few people would expect from them either.
With only very few exceptions, it is possible to detect image manipulation. Even on a large scale. If they had any provisions to check for altered pictures, that would have caught Hajj’s image. Any experienced digital artist can easily see that it was photoshopped – if the Reuters staff didn’t, they’re blind or plain incompetent. There are ways to automatically detect such conduct, or at least raise a flag for further investigation. They could probably collect the RAW camera image files, which are a bit harder to manipulate, and check if the image had been altered.
On the other hand, what if they did know it was manipulated? Is Reuters biased, as some people think? Honestly, I don’t know, but if that were the case, they would be pretty stupid in releasing this particular photo. As said before, its rather easy to notice that the image is photoshopped, even with the low quality copy shown above. If they had wanted to manipulate public opinion, I personally find it hard to believe that they wouldn’t do a better job at it.
In the end, they were either sloppy at manipulating it, or they were sloppy checking for manipulation. That probably says more about how a news agency works than anything else.
(side note: if you don’t believe that this could have been automatically detected, here are two white papers on how it very well can be done: Image Manipulation Detection with Binary Similarity Measures [PDF],Image Manipulation Detection [PDF] )
Q3. What does all that mean to us?
To us, it does not really matter why the image had been altered, since there will a number of reasons to do so, and there would always be reasons to do that.
It does shed a bad light on Reuters, and we have no reason to believe that Reuters is any different from other news agencies.
In the end, it should teach us that we cannot believe something to be true, just because we see a photo of it. Not that that would be all that new – there have been manipulations before digital imaging, and you never know how and why a picture came about. It could have been a set-up, an optical/analog manipulation or a digital one.
A picture sure can say more than a thousand words, but it can also tell more than a thousand lies.
Q4. So why am I blogging about the whole subject, anyways?
Its not new… been there, done that, got the t-shirt, now what?
Any new technology brings new opportunities and new pitfalls. This time its digital image manipulation and a stubling news agency. Next time it will be something else. What we really need is discussion on the technology’s implications.
We need to set moral standards on how a technology may be used, and where it should not.
After all, the pitfall doesn’t suddenly turn the technology into one from hell. It is always people who decide what they make of it – just think “nuclear fission” here. It can be used for nuclear bombs and in power plants. It is up to us to decide which uses we allow, and which we prohibit. Despite how tricky it can be to monitor its use we do need an open and public discussion on the use of technology.
There are a number of pretty new technologies that really do need such a discussion. Photoshopping is just one of them. I’d love to read your comments on this.