Another question about image descriptions

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

Duothematic channel. Primary topic is virtual worlds/OpenSim, secondary topic is the Fediverse beyond Mastodon. Follow me from Mastodon or anywhere else in the Fediverse.

Ort:: The Hypergrid

Geschlecht:: Männlich

Homepage:: https://hub.netzgemeinde.eu/profile/jupiter_rowland

Another question about image descriptions

2024-12-26 22:09:36

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

If it's absolutely impossible to see where an image was made, should it be mentioned in the image description nonetheless? CW: long (over 1,000 characters), image description meta

Artikel ansehen

Accessibility Metaverse in general Accessibility Metaverse in general

show all 5 comments

2024-12-26 23:03:00

Dr. Daniel Dizdarevic

daniel@social.dhelonious.de

If it's absolutely impossible to see where an image was made, should it be mentioned in the image description nonetheless? CW: long (over 1,000 characters), image description meta

Artikel ansehen

2024-12-27 08:16:09

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

@Dr. Daniel Dizdarevic I always consider "let them ask if they want to know" bad style for such elementary information. It seems to me like one of these things where Mastodon's good alt-text proponents may criticise you for not mentioning it right away.

That is, I wouldn't put that information into the alt-text. I only have about 900 or 1,000 characters at my disposal for describing an image in alt-text. Mastodon, Misskey and their forks chop alt-texts over 1,500 characters off in posts from outside, and I need the rest of the characters to explain where a longer and more detailed description can be found for as long as there are still instances of Mastodon under 4.4 around.

This is information that would go into said long description. I've always put the long description into the post itself where I technically don't have any character limits. The limit of 100,000 characters above which Mastodon may completely reject posts is not much to worry about either as long as I don't have multiple highly detailed images with little in common to describe.

Leaving out the information where an image is from, unless I have very good reasons to keep the location secret, feels like not giving a long description at all. And not giving the long and detailed description, in the case of my original images, is like omitting the alt-text for "normal" images entirely.

I've asked the above question because I have a series of images which are special cases. If surroundings were visible in the images and not too generic, I would definitely explain where the image was made, not although, but because next to nobody in the Fediverse could tell from looking at the image where it was made because even the sighted users would never have seen anything like it before.

I want to give everyone in the Fediverse the chance to see the image not only like any sighted person sees it, but like I see the original. This is also why I describe details and transcribe text so tiny that they're basically invisible in the image at its given resolution.

But in this special case, the images don't carry any information at all on where they're from. In other words, the information where they're from might be completely useless. Or it might not.

#Long #LongPost #CWLong #CWLongPost #Metaverse #VirtualWorlds #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta #Inclusion #Inclusivity #A11y #Accessibility

2024-12-27 09:06:30

Dr. Daniel Dizdarevic

daniel@social.dhelonious.de

@jupiter_rowland This sounds reasonable, and as if it would benefit more than just the visually impaired. Describing images as you see them, and adding information that is barely visible or not visible because of compression, goes beyond "pure" accessibility. It's about context, and if you're willing to put the time into it, I think it's a good thing.

As for your special series of images: Do what feels right. After all, alt text is there to help people understand what you want to show.

2024-12-27 10:16:07

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

2024-12-27 10:51:31

Dr. Daniel Dizdarevic

daniel@social.dhelonious.de

@jupiter_rowland I see, thank you for explaining that. On the one hand, I think this adds a lot to the images. It's like what Coldmirror did in her Harry Podcast (talking in detail about 5 minutes of a film in an hour-long podcast episode). On the other hand, I understand that this is tedious.

In science, we rely on a method called coarse graining to reduce the complexity of calculations:

https://en.wikipedia.org/wiki/Coarse-grained_modeling

This is basically what you said in the first place, i.e. leaving things out.