Crop Estimate Test

Some of us in Avian have been emphasizing the value of including information about how much any posted image has been cropped along with the more traditional shooting parameters from the EXIF. Accurate cropping information helps in providing useful feedback about image quality as it can explain loss of detail and excessive noise in an image.

To get an accurate value for any crop simply divide the total number of pixels in the cropped image by the total pixels in the uncropped image; e.g.:

Original: 3000 px wide by 2000 px high = 3000*2000 = 6MP

Crop: 2500 px wide by 1500 px high = 2500*1500 = 3.75MP

So, the cropped image is (3.75/6) = 62.5% of the original.

It can be tempting to skip the calculation and just guess by looking at the crop box overlay in Photoshop. But this is difficult to do accurately and usually leads to gross under estimation of the amount of the crop.

To demonstrate this I've prepared a simple test on my web site in this gallery of nine outlines. Please take a few minutes to look at each image from the gallery on its own page and simply guess at what percentage of the total (blue box) is represented by the white rectangle and post your guesses here (#1 30%, #2 50%, etc.)

Since I'm focusing on the difficultly of the estimate please don't actually calculate a value - just mark down your initial gut feeling.

I'll post the actual values here in this thread after enough people have taken it (or at least looked).

Richard, Please explain more why crop info is deemed valuable?
I judge an image as is, as presented.
Crop info means nothing to me, and I doubt I'll be doing calculations to provide it with my own images.
Thanks -
Sandy

Sandy:

Please see the discussion in this thread in the Birding Forum.

See my second comment (time stamp 7:14 MST) in that thread, especially this:

The reason knowing the amount of cropping that was done serves at least these purposes: it lets the viewer know if the photographer was successful in getting close to the subject; it can explain observable issues with image quality; and it educates novice photographers about reasonable expectations about what it takes to take any sort of bird image and what it takes to maintain acceptable image quality.

See the article on Calculating Crop Percentages on my web site where I write:

Disclosing that a posted image is a substantial crop is simply honesty about how close the photographer actually got to the subject; keeping quiet about such a large crop is dishonest and misleads viewers about the photographer's skills and sets up unreasonable expectations for beginning photographers who haven't mastered field techniques on getting close.

See my comments about the danger of relying on large crops in this thread.

I disagree about being able to up-res an image from a small (e.g.1.8MP image) to something that has the detail of a large image that didn't have to up-res'd to get that big. No matter how advanced the software is it can't invent fine detail for the up-res'd image.

If you rely on large crops (low MP files) you are limiting yourself to web posts. If you build a web site on this basis and don't make that limitation obvious how will you explain to a commercial buyer who wants to buy your image for a book or magazine?

You've also ignored the issue of noise: by working with a severely cropped file you are in essence magnifying the image noise along with the detail you've retained.

Ok, here are my estimates. I did each one without figuring or going back to look at or adjust previous estimates. 1.- 33%, 2 - 85%, 3. - 75%, 4. - 80%, 5. - 45%, 6. - 80%, 7. - 25%, 8. - 75%, 9. - 50%.

I couldn't agree with you more that knowing accurate crop size is important in effective critique of an image (particularly in avian) and it's also a very valuable learning tool for others reading the critiques.

When I first started including crop size in my techs I was only estimating them. Then when I started checking my estimates for accuracy I realized that they were always way off (and nearly always seriously underestimated).

Excellent thread!

Of course, I am on board here too as per the thread I started in the Avian forum. Large crops can look pretty good on-line, but try to print them at a decent size and the fine detail and quality just will not hold up. I believe crop amount should be an integral piece of information included with all the other tech info almost everybody includes (SS, f, ISO). The calculation is easy as looking at the dimensions of your image when you are through processing it. Just need to integrate that into your normal workflow and make it a habit. Here are my guesses and I did not look at Ron D's estimates until after I did mine:

1 = 33, 2 = 90, 3 = 60, 4 = 70, 5 = 40, 6 = 80, 7 = 15, 8 = 75, 9 = 50

Thanks for posting this Rich--I totally agree with your reasons for supplying this info on a post, and have learned a great deal from those who take the time to calculate and post crop %.

I also made my guesses before looking at the others' answers. Here are mine:
1. 25%
2. 90%
3. 56%
4. 64%
5. 36%
6. 81%
7. 20%
8. 75%
9. 50%

1:35%, 2:85%, 3:60%, 4:70%, 5:40%, 6:80% ,7:20% ,8:75% ,9:50%

I'll play: my estimates are - 50%, 95%, 75%, 80%, 40%, 25%, 85%, 60%

I am very interested to see the real answers.

Thanks, Richard for this exercise. I looked at the others' responses after typing m own and see we are all over the place!

And here I was envisioning photography for estimating productivity of corn, soy and tomatoes.... Along the lines of aerial marine mammal surveys.

Seriously, a great idea. I think this is useful beyond the NP site for editors and GA's who specify crops. Especially the pixel specs.

I'll play, but did read (but not study) the other responses.
1 = 30
2 = 95
3 = 60
4= 70
5 = 40
6 = 90
7 = 20
8 = 80
9 = 50

That's still cheating, Lon! For that, you will be banished to South Dakota...

What a mind game! Here's my responses: (did not look at others)

1 - 30%
2 - 90%
3 - 60%
4 - 70%
5 - 40%
6 - 80%
7 - 25%
8 - 75%
9 - 50%

Congratulations Carol for the Perfect Score!

You all did better than I expected, although some of the examples appear to be troubling. I've summarized all the responses, marking correct guesses in green. Note that even the misses are generally close to correct.

Thanks to all who took part.

Richard, good stuff - thanks for posting this!

Richard - I'm guessing you expected people to miss by more, or for more people to miss. (If not, please excuse this post.) I was thinking that one of the reasons you haven't seen that here is that people are actually thinking about it and studying the rectangles more than they normally might, since it was clearly a test. I'd still agree with you somewhat that people probably underestimate how much the crop on average; When guessing crop amounts with no "test" being given, it would seem to be much easier to, consciously or not, shade your answer to be less of a crop.

David:

You are correct; I expected the guesses to be less accurate. I also suspect that people did indeed take more care in their estimates than normal because they wanted to do their best. I also know some of those who posted estimates have been concerned about this very issue and have been carefully calculating their own crop percentages when they post for long enough to have a good feel for the amount.