Description of need:
As more and more phones and cameras have resolution and file size much higher than 20MB, it is much better for the workflow to have iNat accept input, as iNat will rezised to current rules by default. Many will compose images properly in the field, rendering cropping destructive. User then must re-edit images just for iNat to a lower quality (for instance from 12 to 9 or 10 in Photoshop) and then save as a unique copy just for iNat.
I have hundred of images shot with 8256x5504, and yet there are other cameras with far higher than that, which means that I have to spend a considerable amount of time processing all the images just for iNat - lowering quality not dimensions. Describe the iNaturalist community need that your requested feature addresses. Include screenshots, URLs, and other details to help us all understand the issue.
Feature request details:
When uploading images, iNat will accept any sized input, and continue processing them as currently executed. Thus this is merely a UI/UX/workflow improvement rather than a change in the back-end processing mechanisms.
there are lots of programs and utilities that will allow you to resize batches of image files. you could easily save resized versions of hundreds of images in seconds.
there’s no sense in loading huge files to iNat’s servers just to have iNat do the resizing for you on that end.
doing a quick web search should bring lots of suggestions, but if anyone needed recommendations, all they had to do was ask nicely and preferably provide some information on what kind of machine and OS they’re using.
since i’m a Windows user, it would just use the Image Resizer in PowerToys. for Mac folks, it seems like Automator gives you way to resize batches of files.
since the original poster mentioned Photoshop, there’s also a way in that program to basically record set of actions that can be reapplied to a set of photos, and although i probably wouldn’t use that just for resizing (since Image Resizer and Automator are probably easier to use), it might be useful for a Photoshop user just to learn how to do this kind of thing so that they know how to automate other tasks.
On a Mac, the feature is built in, select the images, Right Click (or Ctrl+Click), select Quick Actions, Convert Images. Choose JPEG, Large, and keep preserve metadata checked.
On PC, IrfanView is a great image utility with batch resize and convert features.
You mention using Photoshop. Batch resizing and saving at a different image setting has been a standard built-in feature for a very long time now.
File → Scripts → Image Processor… enter in your settings, load the images (you can load batches directly from the folder, they don’t even have to be open in Photoshop), and Run.
You can combine this with other scripts too, but I never bother doing that.
Just a side note: Since the identification of certain organisms can take considerable effort (e.g. comparing photos and consulting taxonomic literature), perhaps the observers could also spend a little bit of time on cropping and adjusting the size of each photo before upload.
Perhaps I’m understanding this wrong, but I’d much rather be notified if my image is going to be compressed or resized. That was I can crop or otherwise tailor it to preserve as much visual information as possible.
Personally, I’d prefer the keep the warning. Compression can ruin an image and I’d like to be warned that the image may be compromised during the upload.
Since the observer is hoping for an ID, it is a basic courtesy to offer images that are cropped and chosen to show the relevant details.
When you put your identifier hat on … nice landscape, but what are we looking at? Bunch of leaves photobombing the top left corner … Sigh, any more photos of That plant ? I don’t want to scroll across a huge image, I’m not wandering, I AM lost.
This is wrong: maximum dimension on iNat is 2048x2048 pixels. If your image is larger, it will automatically be resized.
So in the end, your image quality will be worsened by two steps - your lowering of the quality, iNat’s lowering of the size.
See also https://www.inaturalist.org/pages/responses#crop
I had a lot of issues with iNat uploads for awhile and it turned out to be because I was trying to upload really large files. I started resizing to the maximum iNat dimensions using Lightroom with bulk exports and that really made my whole workflow much faster as I didn’t have to wait so long to upload overly large photos and then wait more for iNat to process them. It also means iNat doesn’t need to resize, so presumably no loss in quality from what is uploaded.
So, I’d say a warning is a very good thing. Not sure what it says at present. If anything, it should say to resize to the maximum dimensions for the most efficient uploads and image quality.
I certainly don’t disagree. As an observer I crop everything and as an IDer I hate having to zoom in constantly.
But it seems only a minority of observers is doing this. I’m inclined to think it is usually not the lack of courtesy, rather a problem of education and offering integrated tools.
My experience is that the bit I need to see for the ID is very often not the bit that the observer was intending to photograph.
Both resizing and cropping lose information. It’s not clear which one of them loses important information more often.
I think “know what the relevant details are but don’t know the identity of the plant” is a pretty narrow window. Most people are on one side or the other.
My plant window is wide. Not just one artsy picture of a wow flower!
Flower / fruit, leaf, and a wide view for scale and habit and habitat. May still miss THE diagnostic feature, but it has a fair chance of a middling ID, and a comment about the missing feature. (Manulea - we need leaves - says my trusted identifier)
And occasionally we don’t even get a pretty flower to look at
EDIT: I"ve removed a link to an observation of plant roots because it was taken to be negative, which was not the way I intended it but I can see that it could be taken that way.
We’re going to keep the limit at 20mb for a few reasons:
image processing is one of the bigger slowdowns on iNat, and processing larger images would make this even slower.
bigger images mean longer observation upload time, and they might also cause issues on the client side depending on one’s set up.
It’s possible that at some point we can do the image processing on the client side, but there are a lot of variables there, and we’ve had issues with doing it in Android.