Huh! I had not heard this before. Well, I crop most of my photos anyway, so I will try to remember to constrain them to a square going forward, as much as practical.
2 Likes
Nope. If nature had intended species to all fit in a square…
As for aesthetics, seriously? Cropping for identification is, or should be, more than an aesthetic goal. There’s also a question of what to include in order to facilitate identification. I see so many lovely bug eye pictures, face to camera that contain almost no additional signs from the body because, well, it’s not as pretty or engaging. Maybe the butt end is the key feature. I post a lot of those because, well–sometimes the thing just won’t turn around no matter how long you wait but I also now that for some species that’s where the keys are.
Another related question: if you’ve got a bug with incredibly long but low-in-features, legs or antennae – is it safe to crop those out in order to be able to post more of the rest of the creature?
As for collage construction, there’s a lot of very great apps that can handle multiple cropping sizes, some of which are built right into the phone (like Samsung’s photo Gallery). And if that doesn’t work for you there are even free tools online (like https://www.fotor.com/collage).
As for trying to optimize cropping to assist the CV, that could be a risky slippery slope. And I’d hate to think that others are NOT posting really interesting, valuable observations because they didn’t take a shot that would work in a 1:1 ratio.
As for ID work (which I don’t do yet – still climbing Mt. Stupid), I think given the volume that needs doing, I can see why a square grid of thumbnails would be helpful in narrowing off the low-hanging fruit. But a lot, if not most, gallery display options I’ve seen in online galleries (like Amazon photos for example) or on mobiles (Samsung Gallery) give you an option of square grid or unique crops, which get resized best to fit. Any possibility that the same could be done with iNat’s IDer views?
1 Like
If we’re talking snakes or flatworms, sometimes a 1:50 ratio is the one that would get the least background around the organism.
6 Likes
Seems improbable. No one is saying only 1:1 pictures are acceptable, rather they tend to work better in the context of the iNat website (especially identify portal).
4 Likes
Yeah, I understand this. It’s just that when it comes to things like technical image standards it doesn’t take much to go from a little suggestion to a misinformed but perceived hard rule. Especially if it’s coming from the IDers.
Squares are easy to understand. And optimal cropping and submission resolutions is another layer many would prefer to not think about. But without more of that understanding, you might end creating more problems in the ID process if more people just think, oh well, I’ll just crop everything to a square.
Is good cropping practice detailed in any of the tutorials?
1 Like
1:1 is useful because the tops or sides aren’t hidden from view when displayed in iNat (very annoying for me as an identifier). Also, there are many good pics of species that would make excellent representative photos for a taxon–but because they’ve been cropped to be very narrow or very tall, they don’t display well in the “compare” mode. I’ve remove many like that and replaced them with more square ones. So something close to, but not necessarily exactly, 1:1 is better for these particular reasons. I don’t think aesthetics should play into the decision.
3 Likes
Just wanted to say I appreciate that this was posted with the tag “mostly-humor.” 
Like others above, I make cropping decisions mostly based on the subject matter and not how the website converts pictures into thumbnails. I often reuse pictures for multiple sites with various optimum image ratios and sizes, so tailoring them to one site in particular doesn’t make too much sense to me. I do crop and resize to specifications for profile/banner pictures and such and carefully choose images that work well in those dimensions. However, everything else is mostly driven by choosing the best crop/ratio for the subject matter. I usually upload multiple images per observation though and if I notice the first image doesn’t display very well, I might change the order to bring a better image to the top.
As an identifier, I always click through and look at all the pictures in an observation before adding an ID. I find a lot of stuff in the ‘unknown’ pool that looks crappy on thumbnails and fails to pull up anything useful in CV but is perfectly identifiable based on the full size image or second or third picture in a series.
3 Likes
But for vertical things (e.g. https://www.inaturalist.org/observations/104542175), that would mean taking a much bigger photo in order to have to crop it down, and thus much lower resolution.
Same; I use my photos for presentations (which typically use 16:9 in order to use all the space) and publications, and for general outreach for which I want the highest resolution possible and semi-standard ratios. I’m not making a duplicate set by hand-cropping for iNat, I already spend way too much time curating photos that few people see 
I’d be happy if all the photos just had the subject organism in focus and at a reasonable size!
5 Likes
I think that the issue should also be addressed from the perspective of whether users use a desktop than a cell phone. From my point of view, an 16:9 image format is visually more comfortable with wide desktop screens.
I’m a 95% desktop user; but I agree as @jnstuart mentioned before, the subject should determines the format.
7 Likes
Good suggestion!
This discourse can’t be had without getting a little technical.
If I get you correctly, by 1:1 you actually mean a life-size (macro-like) , tight crop, and in square 
format.
From experience as a photographer, this doesn’t work for all observation photos. And besides, camera photos are churned out in either landscape or portrait orientation.
Only vintage/film cameras do square format.
Today, in digital photography, square format can only be achieved by cropping with 1:1 aspect ratio in a photo app like Snapseed, PicsArt, Adobe Express, Lightroom, Photoshop, GIMP, and in the iNat app as well.
iNat users simply need to tightly crop their observations (using any of the standard aspect ratios) to macro look, to facilitate speedy ID. And if freehand/custom works, fine.
But it’s vital to first get a hang of the standard aspect ratios, or crop ratios to know what best applies to each photo subject. Same works across different image use cases :)
3 Likes
Overall I agree with the 1:1 to crop for some subjects
I use two cameras - Primarily a DSLR - shooting in RAW, and then opportunistically an anor droid phone camera.
When processing the RAW images I choose between Square (1:1), Goldencut (1:1.68) or for simplicity the original DSLR dimensions 1:1.5
On the android phones it varies between 1:1 crop and the original 1:1.77
Reasons
- Partly time
- Not all subjects can be 1:1 - long subjects like entire plants, lizards, worms, etc are not suitable for 1:1 crop. A 1:1 crop might lead to too much space around the the focus object
I also have multiple pictures in some observations and not all of them are the same ratio. Some are left as it is to show the placement / relative size of the subject, some are to show different angles.
3 Likes
I like to tightly crop just about all my observation photos to 1:1 also. While not always the most aesthetically pleasing from a photographic point of view, it’s the most useful and efficient for the website I think. When the thumbnails in the app and the explore webpage are all 1:1 it just makes sense. Not only are they more useful for a taxon photo, but also when just browsing observations. I often will browse my own old observations when I’m sure I’ve seen something before but can’t recall the name. Having nice tightly cropped thumbnails makes it more like looking through a field guide.
It’s more handy for browsing photos on my computer too. I still use Picasa to organize my photos and do quick edits for iNat. All the square crops make for nice thumbnails to browse within Picasa too. I don’t save these quick crops or edits to the image files though. The edits are stored in hidden .ini files in the image folders. No duplicate images created or wasted drive space. I just export my edited images to a temporary folder on my desktop, upload them to iNat, and then delete the folder. This way I can also just “undo” my crops and edits later in Picasa if I want them done differently for prints, etc.
4 Likes
This would be 1x, and depends on the sensor. It means that the size of the object on the sensor itself (or film in the old days) is the same size as it is in real life. Most macro photos that aren’t taken with a specialized macro lens are less than 1x.
Cropping loses information. Hopefully not relevant information, but if one does not know how to identify a particular organism one does not necessarily know what information is relevant. Why would we discard information to improve identifications?
I can see two ways in which cropping is useful:
-
It reduces how often you’d have to click on an image to look at the big version when identifying observations. This seems perverse, to me. Cropping images takes longer than clicking on them. It shouldn’t take more work to be lazy. Especially when the potential gain–clicking a mouse button a bit less often?–is so inconsequential.
-
Cropping could reduce the loss of information from resampling images to a smaller size. I’m not sure how widely-known this is, but iNaturalist does indeed discard the original image and save only a smaller version. Cropping, in this context, means that you’re making an informed choice about what data to discard rather than having iNaturalist make an automatic, uninformed choice about what data to discard. If our goal is keeping data that might be needed for identification, “toss out data by default, then ask people to mitigate the damage that causes” is, again, perverse. Just keep the original image. Yes, I know file storage and bandwidth aren’t free, but iNaturalist is fundamentally a very specialized file sharing platform. If you manage a furniture store, you can cut costs by not stocking any furniture. But, then, what’s the point?
1 Like
(For what it’s worth, my own observations fall into two categories: cell phone photos for quick, low effort, high volume documentation of the plants I’m seeing; DSLR photos for high resolution, high information, low volume documentation of particularly interesting plants. The DSLR photos go through post-processing so, yup, I crop them. I don’t make them square. I want those photos to look good and square photos don’t. The combination of cell phone + iNaturalist is great because it allows high throughput. The app trims effort down to the bare minimum needed for the task. I bet it’s possible to add cropping to that workflow in a way that is merely annoying and inconvenient rather than entirely non-viable for high throughput. However, my viewpoint here is that downsizing images is simply bad data management on iNaturalist’s part; I wasn’t asked and I can’t change it. I’m interested in working to solve problems that I am empowered to solve.)
3 Likes
It needs to be pointed out that sometimes habitat is relevant ID information.
5 Likes
Quite true!
You know, this is practically impossible?
It all boils down to cropping or tight-cropping.
I’ve explored many genres of photography. The one with the most cropping applied is in wildlife, especially bird photography. This is because subjects are often far away, and it takes a long lens, say 600mm f4 + 1.6x teleconverter (at least) to get a 1x, or life size view in full resolution. The bird has to be in great proximity to get same result on a regular lens range.
That equipment right there costs an arm and leg, thus only 1 out of 1000 bird photogs would own it. The majority simply resort to cropping, because it’s the easier option.
This is the situation with most photos on iNat. We have to crop for effect. Moreso, thoughtful crops for faster IDs.
2 Likes
1x refers to macro photos, such as of insects. For example, this damsel bug is at 1x (closest focus on a Canon 100mm macro lens). It doesn’t have anything to do with cropping; if you crop a photo that only trims the visible area, there’s no way to increase the magnification afterward. A 1x photo of a bird would just be a bit of a feather.
1 Like
Rather commonly in my case, as an identifier of ferns and allies. Also in say, a tricky Gleichenia, where the observer didn’t get the all-important closeup-underside shot, I search all edges of the photo for some upturned pinnules. Cropping could make these harder.
1 Like