My question is regarding lower quality media (such as blurry photos or recordings with background noise). I did check that there are few forum posts on this topic but i feel this is different enough to start a new forum post. If not please let me know !
For the past couple of months, I’ve truly started uploading observations and making broad IDs in the ‘Unknowns’ . When I notice that an observer is new to the platform, I’ll sometimes leave a comment explaining how to add an ID, along with a link to the relevant instructions. Similarly, when I come across duplicates, I’ll link to the existing observation and suggest that the observations be merged. I’ve also seen other users doing this, so I assumed this was generally considered a normal and helpful way to contribute.
Recently, I’ve received several comments on some of my observations suggesting that they should be deleted because the media quality is not considered good enough. One example was a bird recording. The bird stopped calling while I was recording (I swear they know when youre recording and stop) , but I did manage to capture a few seconds with some longer pauses in between. One user commented that I should delete the observation directly translated “observation to be deleted”. No comment on how to improve, just to delete. I replied that the call is audible and included timestamps to make it easier to find. They responded that it just isn’t good and a shame to upload it.
I’m completely fine with an observation being marked as casual, remaining unidentified, or simply not reaching research grade. I understand that not every observation will have enough information for research grade, and I’m happy for them to remain at whatever level is appropriate.
What I find discouraging is being told that an observation should be deleted simply because the photo or recording isn’t particularly good. For me, that takes some of the enjoyment out of contributing. Sometimes a partial recording or blurry photo is the best one can do and I would rather leave it available for someone who might still find it useful than simply remove it.
So I’d really appreciate some clarification on the general guidelines: When should an observation with imperfect but potentially usable media be left as is, marked casual and when would deletion actually be appropriate?
And related to that: Is it generally appropriate for users to tell other observers to delete their observations because the media is poorer quality or would it be better to suggest leaving the observation unidentified unless there is another reason to delete it?
I think most people would argue imperfect media shouldn’t automatically be considered a reason to delete an observation. A blurry photo or short recording may still contain useful information, especially if the subject can be identified or the observation documents that a particular species was present. Even if it doesn’t deleting a legitimate observation could discourage a new user. Helping them understand what to do instead is what I would do.
In general, I’d usually advocate to leave an observation online. Especially if it represents a genuine encounter and the media is relevant, even if it remains Casual or unidentified. Deletion in my view is for things such as accidental uploads, duplicates that should be merged, test observations, non-wild or captive organisms where applicable, or media with no meaningful connection to the observation. Even with some of those it is a gray line.
It’s reasonable to explain that better media would improve the observation, but simply telling someone to delete it because it isn’t high quality can be discouraging and may remove potentially useful evidence that you didn’t notice. It would be helpful to suggest adding context, timestamps, habitat details, or an identification while allowing the observer to decide whether to keep the observation.
Don’t neglect that partial recordings and blurry images can still be valuable. They may help someone recognize a call, narrow down an identification, or provide supporting evidence for future observations. Not every observation needs to reach Research Grade to be worth keeping as you said. Some low quality observations could be people making their first attempts to enjoy nature and naturalizing. See this:
I don’t think it’s ever acceptable to order a user to delete an observation like that purely because of low fidelity, but I will also say having listened to the observation in question it is really quite clear, trimmed to the subject, and has no background noise. I have seen much much worse bird audio observations get identified to species.
You’d be amazed how blurry something can be and still be recognisable if it’s very distinctive, so one person’s crappy photo may be fine for someone else.
I think it’d usually be better to move on and identify something else rather than spend time demanding deletions. Sometimes I might say ‘oh next time you could take the back of the flower too to work out the species’ or something similar - I figure that might actually be useful and it often gets a good reception
There are a few curt people on here, here and there, but my advice is to ignore them and to just enjoy yourself being in nature and recording what you see :)
In my opinion, there are some observations that should be deleted. They are just so unidentifiable that they are nothing but a waste of space. Nevertheless, although I have been tempted to do so, I have never told anyone to remove an observation or even hinted that they should. I did dare to say as much on the forum once and was slapped down for it. The general opinion seems to be that nothing is too blurry, too small or too dark.
I would generally just ignore comments like this in the situation you described. If an observer doesn’t want to ID, they can skip/move on. If the commenter is being rude and ordering you to delete, you could raise a flag. You can also ask them to stop leaving such comments, mute, or block them if the behavior continues.
In general, I don’t think that there’s any need to delete an observation like this unless maybe it is not possible to see/hear any organism at all. Otherwise, just IDing it to a general level is fine. You can always tick “as good as can be” on your own observations.
My only suggestion for the flip side of this scenario is that I have seen observers get upset when IDers bump back an ID on an observation with lower quality media or tick “as good as can be” when they feel the evidence doesn’t support a more specific ID. If those IDs/votes are made in good faith, that’s ok. Sometimes the evidence doesn’t support a very specific ID, and there’s nothing wrong with that.
Which we all know is untrue. Some observations are so poor that they cannot be IDed to anything. But they are someone else’s observations, not mine, so it’s not my place to tell the observer to delete them.
I don’t think you should mark observations casual when it is not immediately identifiable from a blurry photo or a bad sound recording. Leave as ID required and move on is my motto. I’ve personally ID’d quite a few obs that was clearly made from a moving car - but then I know the area and I know what is there and sometimes I go there myself to check. I can always learn.
That being said, at a sampling rate of 48000hz, your recording is not bad at all and the three second window is sufficient for a suggestion at a rather good 70% confidence rate - if one actually tries. I’ve heard much worse and frankly this does not have bad wind or traffic noise or even people talking in the background. So no need to delete it or even mark it as casual (BTW what makes you think it is captive?)
As for the tone of the comments - translation sometimes do make for a tone that might be perceived as confrontational or rude, but is in fact not. Ignore viz my screen shot.
If you are posting an image that looks like used toilet paper in the fair hope that someone out there will recognize it as the most rare giraffe on the planet, then maybe you are wasting time. But on the other hand, someone might actually place it correctly and it might provide some sort of insight or convince someone to go there and see for themselves. It is data, and rather have something than nothing. We do not all have access to top line equipment and this is citizen science, not lab quality research.
I prefer to say, “this photo is hard to use for indetification, it’s [too blurry/only bark/etc], do you have a more useable photo with [trait]?”
a positive suggestion that still politely informs the user that the photo is… not very good.
if you don’t know with certainty what is missing for ID, and only ‘suspect’ the photo is unusable, you probably don’t know enough to make the call to make it casual anyway.
as for telling people to delete: don’t. better they stay and improve over time. many will go back to clean up old unusuitable observations on their own.
This. People listen, if you don’t like something that doesn’t mean it is rule breaking. I wouldn’t make an observation like this but if some handicapped user who for some reason cannot move or photograph well still enjoys to make observations even if they are unidentifiable, I’m super happy for them!
Do not tell people that their observations aren’t good enough or need to be removed. Just don’t. That’s rude and it can alienate people who would otherwise go on to do good things here.
Do we have really bad, useless photos on iNaturalist? Yes. (Some may be mine.) That’s one of the costs of open, welcoming citizen’s science projects.
There is never a ‘too low quality media’ for observations. All quality media have the chance to be identified at some point and therefore asking to delete their media is pointless and often inapporiate.
I agree wholeheartedly. There is also eventually going to be a line where opinions of the quality or usefulness of an observation will differ. Who’s to judge? I saw this a lot on BugGuide, on which moderators can delete poor quality pictures. Can’t say I always agreed with their choices.
In general I would not demand someone to delete a photo, no matter how poorly taken it was. The exception would be if a photo was obviously posted with malicious intentions, the key word being “obviously”.
The others have already given good reasons why even poor photos are acceptable in iNat. I would add that for me, it could be that the person is using iNat as a log of personal observations rather than with the intention of contributing to science, and for me that’s a perfectly valid.
I don’t think I’d ever feel it was my place to tell someone to delete an observation, but I’m not sure I can agree with your statement. When you get fuzzy ‘bug’ pictures where it’s literally impossible to tell whether the thing they took a photo of is dirt, insect, rock, spider, distant bird or something else… I can’t see the value in that.
It might have value to the observer, as a reminder of that cool insect they noticed, even if they didn’t have the proper camera equipment for a good photo.
It should probably be in casual though, and that gets tricky. If two people agree on “insects” or even “life” it can be marked “as good as can be.” But if there’s no community taxon because it could literally be a speck of dirt, then I don’t know if there’s an acceptable way of making it casual. I’m not sure “no evidence of organism” would apply, since there’s evidence of something that easily could be an organism. I think we just have to put up with them, and be happy the observer noticed an insect and had the impulse to photograph it instead of squishing it.
I agree with the general sentiment on this page.
If something is so blurry that I can’t confirm the ID I leave a broader ID, if I feel like it, and also a comment of “I’m not sure I can ID past this with what I can see here, if you have any other photos to add that might help. etc.”
If it is something I know more about then I can add that. e.g. Tamarix species ID it challenging for the experts, it requires closeup photos of the nectar disk inside the flower, and often also requires genetic testing… and some others (Datura, and lichens are notorious for this.)
Now for my own, I’ve never been told to delete, I have been told “with current photographic data we are unable to proceed with ID.”
I think telling other people what to do with their data is not fair, mark it as cannot ID further and move on. Or just mark it as reviewed and move on.
