Is Shutterstock using AI for QC?

I had a couple of very interesting incidents lately with the Shutterstock quality control. From what I understand Shutterstock need to use AI for the quality control of uploaded images. I suspect that the system can’t see the difference between sand, water and noise. I uploaded a batch of 9 images, 4 of them was rejected because of noise, artifacts or grain.

I could not understand the rejections, the images was from the same shoot. Same ISO, same settings, and the camera was on a tripod. I did a search, and ended up at Alex Rotenberg’s Brutally Honest Guide to Microstock. I saw that I was not the only person having problem with rejections for no reason. So, the tip was just to resubmit the images, and the second time the images was accepted.

It’s not a big deal, but it is frustrating when you know the quality is good, but the automated system fails. It will slow you down, it you need to spend more time on resubmit images that should have been accepted in the first place. Anyway, i hope the AI will improve in the future….

Examples below:

Resubmitted 2 minutes later: 

The same went for all 4 images. Hope Shutterstock will look into this.


Maybe I was too early, or not patient enough?

In 2014 my wife and I visited a photo gallery in Barcelona, and I was very fascinated from the images in the gallery. It was some kind of “multilayered” technique used, and the look and feel from the images really fascinated me. As a PS geek I could clearly se how the images were made, stacking the same images a numerous times on top of each other, twisted and moved using different blendmodes.  We rushed back to the hotel where I had my laptop, and I started to experiment with an image.

My first attempt became “UPSIDE DOWN”

“TWIST” Made from elements in Barcelona

“HARBOUR REFLECTIONS” From the harbour of Tromsø in northern Norway

“STREETS OF BARCELONA” As the title says, a street in Barcelona

I really had a lot of fun making the first image “UPSIDE DOWN”, and this was something new. I could not do analyse excactly how the images in the gallery was made, but I was able to find my own way of creating abstract images with my own touch. My idea was to mix the chaos with som sharp elements in the image.
I developed the technique further, and late 2014 I had an exibition in my hometown Sandefjord showing 20 of my abstract images.

I developed my technique further, and moved in to a different style based on the same idea, stacking images from the same exposure, just repeating the same image over and over again with different blendmodes until I got the look I was happy with. I also did a lot of masking too get som parts of the image in focus. The result became  the style I called “Urban Frequency”





So, to the interesting part. I started using this technique back in 2014, and after 2014 I have seen many other photographers working the same way. The funny part is that many of those photographers win prizes in worldwide competitions, they get “honor and glory”, and the images can be found in many top galleries around the world. So, I guess….. I was maybe to early, or not patient enough 🙂

Some links to award winning photographers with the same technique:

– Alessio Trerotoli (Italy)
– Lars Andreas Dybvik (Norway)
– Stephanie Young (Germany)

No more jpeg’s – I shoot RAW

I recently did one more of my jpeg studys in the post from Salinas in Sicily. I’ve been on and off if I should shoot jpg or RAW, and I’ve been experimenting a lot with the two filetypes. Since I mostly shoot for stock, I thought it would be a good idea to shoot jpeg. I also use Fuji X cameras known for their superb jpg quality. Normally I shoot RAW, and I have been doing that for nearly 10 years. The problem is that RAW files takes up so much space, and my cameras produce 50 mb files. It would be nice to have smaller files with good quality, but my conclusion is that I will stick to shooting RAW.

My decision came after an evening walk at the beach tonight. The light was incredible, and I tried to get most of it right in the camera. Correct exposure, no blown out highlights, definition in the shadows etc. I used all the possibilities I had to adjust these settings in camera. What I noticed was my wife’s hair! She has a “silver” color to her hair, and something strange happened. My X-T20 could not figure out this color when I shot the image in jpg. I got some wierd highlights that was impossible to fix. In RAW it was no problem whatsoever. Take a look at the blownout “stripes” in the image below:

Everything else in this image worked out well, except from the hair. Small problems like this is very annoying and frustrating. Like I said, I shoot mostly for stock, and I want my images to be as good as possible.

So…. I’m going to forget my thoughts about shooting jpg’s once and for all. Nowadays hard drives are so cheap, so space will be no limitation anyway. So, #ishootstock in RAW.
Have a nice one… Finn-b

More abstract beachimages using ICM

I find this technique very fun, and I’m always excited to see the results I get from different kinds of camera movement. This evening we had a stunning sunset here at Rio Jara in spain, and I went down to the beach to take some sunset images, something I normally don’t do. Sunsets are something I don’t photograph very often, but today it was really stunning:

As I stood there, the sun went down below the horizon, and I saw the opportunity to take some more ICM images for my FAA portfolio. I tried different movements to get some variations, and the shoot came out very well:

You can find all the ICM images in my FINE ART SHOP

Correcting colorcast with the LEE bigstopper

Like many photographers I really like from time to time to shoot long exposures. Some years ago I bought a set of LEE filters. I have the LEE Littlestopper, the Bigstopper and I also recently added the Superstopper to my collection. The LEE filters are great, but it’s difficult to get the correct colors. The filters tends to have a blue’ish colorcast that is very hard to correct. The result is that some of the images can look a bit “flat”.

I have tried several ways to correct this problem using graycards, the passport colorchecker etc. etc. without getting the result I was looking for. I also tried a method using a RGB curve adjustment, but still no luck. I was actually looking for a simple way to correct the colorcast. Recently I again did a search for possible ways to correct the problem. I came across a video from Peter Zelinka explaning a method that works really good.

The whole idea is to lock the WB and take a picture against a white wall. This image will be used as the reference image. Next is to add the filters one by one, calculate shutterspeed and use the same WB and settings as the reference image. The result will show the colorcast for the specific filter.

Next step is to take the images into camera raw and adjust WB and tint to get the images taken with filters as close to the reference filter as possible. Adjust the two settings until you are happy with the result. You will then get the correct values for how much you will need to add or subtract on WB and Tint

For my Littlestopper I need to add +950 Kelvin and subtract -3 on the tint. For my Bigstopper i need to add +2100 Kelvin, and add + 3 to he tint. I guess that this values will not be the same depending on the batch from LEE filters, so you will have to do the same calibration for your filters.

So, let’s take a look at the result. Below is the image with no adjustments:

And the corrected one: