I think his point was that the impact habitat photos have on the CV is negligible, not nonexistent.
2 Likes
Yeah, I’m not calling people liars. I’m just saying this is a red herring that the forum community has fixated on and it shouldn’t form the basis for policy decisions without better evidence.
Growth is the norm for all things human and all things digital; that’s never going away. Limiting our vision of potential scope because of current storage limits is foolhardy-- after all, in a just few years Moore’s Law will make those redlines an afterthought. Teenagers today have no idea why 1.44 MB used to be a dealbreaker file size, or why someone might be keeping their file names under 9 characters. Our dreams still have to work in the today, of course. But I don’t see habitat photos being the threat to iNat’s servers-- they’re a tiny minority right now.
That said, if I were king for a day picking a priority, I’d probably focus on making the Apple app functional, not habitat photos.
2 Likes
About being able to find that photo, of habitat for example
https://forum.inaturalist.org/t/add-a-dqa-choice-for-survey-photos/190/5
https://forum.inaturalist.org/t/sort-by-type-angle-of-photo-for-id/20031/3
1 Like
You are saying that the narrative expressed by the people responsible for building and maintaining the site is untrue. There are few explanations that square with that assertion and none are complimentary. They have asked that the community refrain from doing certain things for reasons that include this concern.
Pretty sure the folks running the show get that, given that they have a Features Requests section and all. Regardless, there are constraints on all growth that isn’t cancerous. The constraints here are pretty gentle and pretty reasonable. The notion that a requirement for observations be limited to recordings, drawings or images of the species observed is arbitrary unless data are presented to justify it is a bit much, especially since this topic is a question about what’s allowed, not what should be allowed.
I wonder if that has something to do with them not being permitted. What would happen if they were permitted or actually encouraged is not difficult to anticipate. There are already plenty of compromises like those on drawing boundaries that are sometimes inconvenient and limits were recently placed on users defining places in order to limit computation and bandwidth consumption. These are just facts of life in a world where resources are limited.
Amen. Although there are plenty of other things that would also be nice.
The folks running this place do a good job and offer a great tool to the community that continues to improve, bit by bit, as time passes. Playing by their rules is not too much to ask.
2 Likes
After reading through this thread, it seems we really to need two things:
- The current discussion, which will help come to an agreement on the most useful ways to add a variety if media to an observation.
- A discussion that can lead to a request for enhancement of the the database, to the software running “in the background” on that database, and to the web app and mobile app interfaces we use to enter the media. The most important part of that is what is needed. That should lead, later, to how the interface should look and feel.
I am a beginner, but I’d like to begin discussion #2 by asserting that the current interface does not satisfy the needs of the iNaturalist community, including the needs of those using it for rigorous scientific work, and that some changes to the “behind the scenes stuff” will likely be needed to support the data as well as the web and mobile app changes we would like to see.
Media we need to be able to upload (and differentiate between)
a) photos of the organism, which can potentially be used to train the AI on the organism’s appearance
b) photos of other visual evidence of the organism (tracks, scat, nest, web, claw/scratch marks, shed fur or feathers, shed skin or exoskeleton, etc), which could, at some time in the future, potentially also be used as separate AI training sets
c) audio recordings of the organism (which could also potentially be used, at some time in the future, in AI training sets)
d) photos of habitat
e) audio recordings of habitat.
Anything else?
This is true, unfortunately, and adding functionality for annotating individual photos has been on the list for some time. Other things have taken priority, however.
There isn’t a hard and fast rule banning habitat photos from observations, but currently, we say:
Photos attached to observations should include evidence of the actual organism at the time of the observation, observed by the user who is uploading the observation.
So we ask observers to only include photos that depict the actual organism or trace evidence of said organism. Is it the end of the world if some people include a habitat shot every so often? No. But it’s not great, and I would discourage people from doing so.
8 Likes
Why wouldn’t the sound recording be the evidence that you ask for?
1 Like
The sound recording is. It’s the habitat shot that isn’t.
1 Like
I think theyask why would you need habitat shot at all, but it’s very helpful to see where you found the thing.
2 Likes
Yes!!! Indeed!
So, I recorded an animal vocalization and a habitat shot with it to establish and auto upload the meta data for location and date/time. Besides more precise location/date/time information, that habitat shot shows some additional information about the organism.
Uploading a sound on it’s own does not bring in any meta data, so that is an (unnecessary) loss of precision in the location/date/time fields.
I don’t intend to buck the advice above, but I do feel it is an unnecessary loss in data quality.
3 Likes
What I do is upload the photo (taken at the same place and time as the audio recording) and let iNat read its metadata, remove the photo from the observation upload-in-progress, then add the audio file. It retains the metadata from the photo even after it’s deleted.
I never intended the photo to add habitat info - although I suppose it could, I’d just never thought of using it like that. If I think the situation/habitat/location, etc are pertinent, I add them in the notes for the observation.
But my point is, the photo’s metadata can be read and retained for the audio file even after the photo is deleted - if the photo doesn’t show the organism thus doesn’t meet iNat’s guidelines.
(adding: I have no dog in this race - I don’t care one way or another what is being argued or ultimately decided)
3 Likes
Which platform do you use? I am, I suppose, a bit frustrated by how many (freaking) steps it takes to deal with audio obs in iOS platform. And, alas, the new capability to record audio with in the iOS mobile app produces a lower grade quality audio than voice memo.
1 Like
To start with, I’m on Android with location turned on. And to (perhaps) complicate it, I use Google. But I’ll walk through the steps I use just on the off chance they spark an idea for you.
(I’m betting there is some easier way but I use what I have and know how to use rather than search out other ways since I was to ‘use’ and not ‘search’… )
I open my photo app on my Android Phone (Nokia 6.1) and take a video. I point the mic toward the sound and try to keep it out of the wind. The app is the one that came with the phone. It’s called < Camera HMD Global > at Google Play and when I just checked on the name, it tells me it’s not compatible with my phone anymore ?? I mean, it came with the phone and it’s been working so, I’ll ignore that for the moment.
After taking the video (I keep it between 20-30 seconds and, if need be, I take more than one. I just don’t want the file size to get too large), I take a photo. This is purely to document the location and time since I let my phone track me.
All my photos are stored in my Google Photos app and when I get home, I sync the photos with my online account. Then I download the photos from Google Online (web based) to my home desktop computer. (I find this the easiest way to transmit phone photos to my home computer)
To extract the audio from the video file, I searched for an online converter. I found lots but a few got persnickity about file size or my ad blocking so I’ve settled on this one:
https://www.fileconverto.com/extract-audio-from-video/
Once I have the audio file extracted and downloaded to my computer, I open it in Audacity (see note about this at end) and do any editing - cutting off dead space at beginning or end or an attempt to reduce noise a little. I’m familiar with this software but I’m not well trained in it. I think you could upload the file without editing it at all and in some communities, it is preferred that you not edit it in any manner.
When I want to upload the finished audio file to iNat, I first upload the photo I took and let iNat read its metadata. Then I delete that photo and add the audio file to the in-progress observation. The CV won’t give you any suggestions so you have to provide your own identification. For me it’s usually ‘birds’ or some branch of ‘insects’ that covers grasshoppers/katydids/cicadas, etc. That’s all I’ve recorded.
It sounds fussy and I suppose, depending on the flow of your uploads, it may be too fussy. But I have developed a routine and rhythm that keeps all this pretty straightforward. I don’t make as many observations as some but I’m not too shabby in volume, either. However, I am retired so… there’s that. :-) I usually leave the audio files to the end of my day’s uploads and often put it off a day or two till I want to deal with them.
I tried some audio recording apps and didn’t find them at all useable. Then I tried this method and it worked so I haven’t looked around for any other.
I’m not sure how much that was helpful. That is, how much you could modify for iPhones or how fussy it all seemed to the point of ‘not an option’.
regarding Audacity: I have been working with this for years and I’m not young or overly technical. But I’m not afraid of technology and software either. How someone else feels about using it will vary with temperament. But if you want to try Audacity, I would suggest something that is pre-Version 3. There was some discussion about its ability to collect data with later versions. I can give you more info if desired about that. And I think there’s been some discussion about audio editing on the forum so a search for ‘Audacity’ might provide some additional info or options.
Here are my audio recordings if you want to listen to quality.
https://www.inaturalist.org/observations?place_id=any&sounds&user_id=mmmiller&verifiable=any
3 Likes
I know this started with the issue of habitat shots without direct evidence of the observed species to go with an audio of the species. That said, here has been a lot of talk here about iNaturalist not wanting “habitat shots”. If iNaturalist doesn’t want habitat shots without the direct evidence of the observed species, I would still encourage users to include views of the subject species that also show the habitat, and any other species around them. This could include shots from further away than needed for the best view of the subject species, without necessarily making the subject species hard to see. I would like to stress that knowing the habitat, including species near the subject species, is often a key to knowing what species would be there, and what the subject species might be.
6 Likes
All my sound files are converted videos, it’s the conversion process that removes the metadata. So I just keep the original video file and copy the metadata manually onto the observation. It’s time-consuming but results in no loss of accuracy. Obviously this won’t work if you use a different process, like recording sounds in non-video form with a device that doesn’t record date/time, but it works for me.
3 Likes
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.