While I was cropping some images I got to thinking about Inat and its associated hosting fees. By cropping my images in many case it takes up to 50+% less space. Every time someone clicks that photo it is taking less bandwidth now which in turn should add up to savings. If everyone made it a practice to crop their images better I imagine it could make a sizeable difference.
Similarly, I like to sometimes make compilation photos like photo #2 from this observation: https://www.inaturalist.org/observations/69842881
(As I understand these are ok as long as it’s all the same organism and not used for the main photo?)
Provided it isn’t against policy that makes an even larger difference in hosting space/bandwidth, as otherwise that would have taken six separate images. It also uses up less bandwidth of the end user, potentially also saving them money if they pay for bandwidth. Win, win.
I now try to ask myself if my observations really need more then one photo, or if certain photos are showing any new detail or not.
I did notice there is a feature request for various types of cropping, and that the app has this feature. It would seem like a no-brainer to implement on the surface, but I understand matters could be more complex. If there were a non-mobile crop available would the resources needed to simply do the cropping outweigh the bandwidth savings? I could see this being the case, but I have no knowledge of such matters. Maybe some of the more tech-orientated people could weigh in?
Also if you are cropping, I saw many suggest to make sure you crop your images to a square so they display well as a thumbnail.
Just a thought. Thanks for reading.
Somewhat relevant: iNat automatically scales down anything you upload to a maximum dimension of 2048px.
Learning what are helpful photos and what are not is part of the learning process for new taxa… and especially for spiders, and I guess all other small or very distant subjects, we often lament the lack of detail or just the right angle, so we don’t really want to deter additional photos in general, just those cases where one is fairly certain already of the species.
Just a brief note is that this image in your example is now 1536 pixels high. So you loose information from the original photo to the one now stored on iNat. While in this example this is fine and an advantage, because the background has no information. In compilation photos of plants I too often found that the information I really need is too blurred to make out. I need to be able to see if there are 3 or 5 or more stamens and if it is fused with the style or not. An additional issue is that the background can be used to judge the size of a feature. A difference of 1 mm can be the difference of two species. In spiders one needs to count the eyes, which is too often impossible.
I’m so new here that I can’t comment on ways to save iNat money, don’t yet know enough how the platform is funded.
Most of my observations have been with small to very small organisms … fungi … and I know from previous experience, in for example FB groups dedicated to fungi, that compiled pics prevent a viewer from ‘clicking through’ to an enlarged closer-up view of one of the original photos. One thing I really enjoy here on iNat is the possibility of posting photos ‘rich’ in pixels and seeing an increase in detail that I haven’t even been able to get on my own computer.
I’ve thought about cropping but have delayed it for the same reason as above. The loss of pixels/detail makes a huge difference to the eventual quality of being able to ‘see’ a small subject. A cropped photo will usually only be able to be enlarged once while an un-cropped photo can sometimes be enlarged three times and even though there is a loss of definition, as can be expected, important details can sometimes still be identified.
“Cropping” means you select the part of the image that contains information that you want to share. In other words, cropping is the act of cutting out the area around the subject that offers no useful information. No loss of “pixels/detail” (=resolution) of the subject, as you suggest.
“Resizing” does cause a loss of resolution, because it “thins out” the number of pixels in the image while keeping the full extent of the image. If you resize 50%, every second row and every second column of pixels are discarded, leaving you with an image that contains 25% of the original pixel count.
“Compression” is another thing. It keeps all the pixels but saves them in a kind of shorthand to save space on disk. Loss of detail depends on just how much you compress.
Cropping, resizing and compression are separate operations.
All three are useful and should be used where appropriate. Not only to save upload time, and space on the server but also to improve the user experience.
In the iNat context, I think cropping is the most important technique, and it is courteous. No one is interested in the rest of the brick wall.
Do resize images that are so big that they are unwieldy (like 5MB+), or don’t because iNat will do it anyway.
Don’t compress too much when saving images to your disk, because iNat compresses them again, anyway. If you shoot .jpg, the images are compressed when your camera saves them to its memory card. When you process and save them to disk, the jpg’s are re-compressed. So, by this time they had been compressed, inflated and compressed again. When iNat saves them, they are re-inflated and then compressed for a third time. You have control over the first two compressions. Reduce the level of compression by as much as you’re comfortable with. It will have no effect on the final image size after iNat has compressed it.
Cropping reduces pixels by cutting out the peripheral areas with useless pixels. No loss of resolution in the subject area.
Resizing reduces the pixel count by discarding pixels throughout the image. Resolution loss.
Compression does not reduce the pixel count but it saves disk space.
Loss of detail every time an image is saved while using a high level of compression.
Thanks for your explanation, wynand_uys. I shall experiment with cropping. I do often already use cropped photos on other platforms, but found the richness of the iNaturalist experience very engaging.
True, and I do try to crop out unuseful parts of an image…eg most of my hand, with a tiny moth on it. But when I crop out backbround vegetation, for plant observations, I find I miss that information about context very much, even if it is overlit or blurry, as eg the excess light indicates the proximity to the forest margin, or a lightbreak in vegetation, which are very important features in learning how that species fares in different situations.
However I love the clarity of the closer focus achieved by cropping. It is a difficult balance to achieve.
In fact, because iNat does reduce the size of large images, cropping is a good strategy for MAINTAINING resolution and therefore level of detail. With small subjects, I always try to crop the photo down to a 2000 x 2000 pixel square where possible so that the resolution isn’t reduced when the photo is processed by iNat.
As someone with a slow internet connection I really wish more people would crop their photos. Rather than trying to build that into the functionality of iNaturalist, maybe instead encourage users to do it themselves. For example you could give a warning that photos over 2048px will be resized and may cause loss of detail (and that cropping may be a solution to this).
I think adding a cropping function to the upload observations page would also be a good idea.
Welcome to the Forum! That was helpful information. I almost always crop my images, and was concerned about quality loss after doing so.
I don’t know exactly what compression algorithms iNat uses when resizing and compressing images, but in the interest of efficiency it’s likely not one of the best detail-preserving algorithms available, as these tend to require more processing power.
This makes it advantageous to resize and compress your photos with software on your desktop before uploading, if you’re a user equipped to do so. My workflow in Lightroom ends with exporting all images at a maximum size of 2000x2000 (cropped images often end up smaller) and caps the filesize at 1mb. Because the compression is being performed by a dedicated image-processing software this probably preserves detail and tonality better than iNat’s automatic compression, and takes a load off iNat’s server. Most good image processing/organizing software platforms including freeware like Darktable can do the same.
This also makes uploading way faster. When I had a poor internet connection this was a lifesaver.
This doesn’t entirely help folks who post from their phones- and I have noticed the default filesize settings for my phone camera produce obscenely large files relative to their overall poor detail- but I’d hope that the app is using device CPU to do the resizing rather than doing that processing on the server side.
I’ll be cheeky here - iNat should move to Winnipeg! We have abundant Hydro power that should cost less. And it would probably lower the carbon footprint (I don’t know how your electricity is generated). Plus I could wander down every now and then for a chat!
Previous discussion - when we asked for cropping - was rejected, as there is lots of photo software available.
I do sort and crop my photos before I put them in iNat. Sometimes using a cropped detail for the first photo, and a wider view as a second.
There is an open feature request (i.e. not yet turned down/closed by staff) for adding cropping tools to the website here: https://forum.inaturalist.org/t/incorporate-basic-image-editing-tools-crop-rotate/1223/
This is the main reason to crop any image you upload. If you upload large (like 6x4kpx) image and iNat scale it to 2x1.4kpx, you do loose about 2/3 of all information. And it is often crucial, especially if you make photo of flying bird.
I’ve seen some people post 10-20 virtually identical and often non-cropped photos per observation on a regular basis. No doubt that sort of habit really adds a lot of hosting requirements on iNat’s end.
Yeah, true! I find some peace in realizing some of this is due to it being citizen science. Not everyone is coming at this with the same level of knowledge/time/ability/etc.
And not everyone can go into a state of satisfying flow just by opening up lightroom and cropping every photo in existence- oh wait, sorry, this isn’t the ‘you might be addicted to iNat’ thread, my bad
I’m pretty fussy about my photos. For any image not taken on a mobile device I’m already shooting in RAW, post-processing, cropping to the relevant image (which is the subject plus relevant context, as well as trying to tell a bit of a story with the photo), then downscaling to an online friendly size.