iNat compresses any uploaded images to 4 Megapixels, and while that seems reasonably enough, it definitely eliminates a lot of finer details that are often crucial for ID, such as number of hairs, finer sculptures, etc.
People often say: “just crop that important part and upload as a separate image” but we just don’t have time to crop every photos we’ve taken. It’s a heck load of extra work especially when you post like 1000 observations at once. If I didn’t have to do that, I had days of extra time to work on IDing people’s obs.
Also, cropping reduces the quality of the image significantly as the image becomes more pixelated, which is fairly unpleasant.
I know that iNat isn’t really a place for posting high quality photos, but still it’s definitely useful and pleasing to have beautiful photos so people can realize how fascinating the species is.
While our photos are compressed to 4 MP, we can upload 15 images in an observation. That is A LOT of data being processed.
Therefore, my suggestion is, why don’t we make an option to reduce the number of the images in an observation for posting fewer higher quality (at least 8MP, which is 4K) files.
If this makes iNat server tired (which I doubt, because less images are required at the end), it’s definitely possible to restrict this feature to people who can donate to iNat.
Majority of the cameras out there are around 20 MP, and there is a reason why.
4MP is not always enough for seeing finer details in a photo.
Especially when we think as iNat observations something humans use later in the future, it’s best if we could keep the photographic data in a better condition.
Images are reduced to 2048 x 2048 (if square).
There’s an existing feature request related to this here:
which seems like it would be a good place for this discussion, since many of the issues likely to be brought up are laid out there. @invertebratist is there a strong reason to leave this as a separate thread?
You could also consider working this up as a full feature request and submitting.
Another related discussion is
And there may be more.
My apologies, that’s right. Edited the post, it’s 4MP (2000x2000=4 million)
@cthawley Yes I know those older threads, but I think it is worth having a separate discussion because iNat admins haven’t updated on this for about a year.
I think it’s one of the biggest issues we have on iNaturalist. So much data is lost. I’ve heard of dozens of people not using iNat because of the huge image compression.
Since iNat is such a great platform I really hope this will be improved.
This can have sense if technically feasible.
Anyway, I think you have noticed that in most observations that are not, how to say, not so satisfying as far as the quality of photos is concerned, there is not an issue of “fine details”.
Definitively, in parallel, I would also try to raise the awareness of certain users making them realize that they should evaluate what they are going to post and ask themselves if the photo could allow someone to recognize that organism.
How huge the compression is depends on where one starts, of course.
If people are wanting a website for archiving original high-resolution photo files, there are other more suitable options like flickr. In my experience, 2048x2048 pixels is plenty to preserve details needed for observation and identification of most organisms, which is what iNaturalist is all about. In instances where critical minute details were captured at original resolution but might be lost in re-sampling, these can be cropped below the 2048px re-sampling threshold and added to the observation.
Rather than cropping I usually take photos of plants both from far away (showing the whole plant) and close up photos of different features. If it’s a bird or insect, I get as close as I can then crop before uploading so I don’t lose resolution.
I don’t think people use all 15 images (or even close to that) very often.
Sometimes people don’t realize their photo is reduced and delete the original. It would be good to make it more clear to users when uploading images that the resolution will be reduced. That “original” on iNat is not actually the same as the file they uploaded.
I don’t think making larger images is a good idea unless there will be more levels of zoom for people who have a slow connection.
Exactly, that’s why I wrote that we need it as “an option” for people who spend a lot of effort to capture images with good details. Those details really matter for identification.
Yes, iNat is all about observations and identifications of most organisms.
I’m not suggesting this for archiving high-resolution photo files, which isn’t the purpose of iNat. 4K is not really a high-resolution anymore- 8k is, I’d say.
In my experience, 2048x2048 pixels is not enough for many cases where fine sculptures matter.
While we can still add IDs from 4MP images, it’s just so much less reliable. For example, looking at surface sculptures of gastropods helps improving accuracy a lot.
Yes, that’s what most of us do. The point is, even if we photograph the organism from all the necessary angles, quite frequently we need to magnify the image to see finer sculptures later. It’s not always the case that we know which part to crop and preserve the resolution.
I really do agree with that clarifying the compression is a good idea. The word “original” is not quite accurate.
I think cropping already solves this issue. I checked your photos and they are beautiful. I can imagine that bigger photos are more beautiful, but I am afraid many uncropped photos will be uploaded which only increases the load of the servers… Most people make - I guess - photos with the phone camera and spent not to much time on preparing photos before upload.
Be carefull … SPIDERS!
I do not like them, but you can make nice pics of the hairs on the legs…
Focus stacked images.
One option may be to put the original high-resolution images on Flickr, and then import into iNaturalist through the existing feature that allows to link accounts for easy import. This will create a link to the Flickr original on the iNat photo pages. Of course, this also is more work and may only work as long as you are a paying Flickr Pro user.
I have one with 16, presumably from before when this limit was implemented? There are hours of work captured in that one observation. It was a fun way to spend time during pandemic lockdown but certainly not something I do very often.
We hear your frustration with low-resolution when high-resolution detail is what you want for identifying. We would not favor the option of people buying more features on inat because we worry who gets left out of that economic model. It does not meet our needs for inclusion or consideration. Are you willing to support only solutions that are available to everyone?
In terms of inclusion and consideration, I don’t think there wouldn’t be any problems. If it happens, if should be available for everyone to pay for any extra features.
Higher resolution files will affect iNat server, so it makes perfect sense for the observer to pay for that. It’s quite unrealistic to expect everything to be free.
And it’s wouldn’t be really needed by majority of iNat users who uses compact cameras or smartphone cameras anyway.
However, it’s definitely helps to have some higher resolution images available on iNat for more accurate identification and comparison, etc.