A question on how far image descriptions should go

Sat, 19 Oct 2024 19:28:10 +0200

jupiter_rowland@hub.netzgemeinde.eu

Image descriptions: Do I have to describe what a herringbone fabric pattern and full brogue shoes look like? CW: long (over 1,700 characters), image description meta

Artikel ansehen

Accessibility

show all 8 comments

Sat, 19 Oct 2024 19:35:29 +0200

Charlotte Joanne

Lottie@beige.party

Image descriptions: Do I have to describe what a herringbone fabric pattern and full brogue shoes look like? CW: long (over 1,700 characters), image description meta

Artikel ansehen

Sat, 19 Oct 2024 19:38:43 +0200

Patrick H. Lauke

patrick_h_lauke@mastodon.social

Image descriptions: Do I have to describe what a herringbone fabric pattern and full brogue shoes look like? CW: long (over 1,700 characters), image description meta

Artikel ansehen

Sat, 19 Oct 2024 19:40:09 +0200

Andre Louis

FreakyFwoof@universeodon.com

Image descriptions: Do I have to describe what a herringbone fabric pattern and full brogue shoes look like? CW: long (over 1,700 characters), image description meta

Artikel ansehen

Sat, 19 Oct 2024 21:14:40 +0200

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

@Patrick H. Lauke Truth be told, you'd have to enlarge the images in question by the factor of two to five based on their original resolution, depending on the colour, to actually recognise the herringbone pattern on the jacket.

I just don't want to omit details in my descriptions that someone may need to know. And I don't want to be called lazy, neglectful or even ableist because people have to look up information that I could have included in at least one of my two image descriptions myself.

I mean, the image description standards are way higher on Mastodon than anywhere else. People get used to them, and I have to keep up with them, if not even be ahead of them in order to not fall behind.

CC: @Charlotte Joanne @Andre Louis

#Long #LongPost #CWLong #CWLongPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta #Ableism #A11y #Accessibility

Sun, 20 Oct 2024 09:06:50 +0200

scope filter

scopefilter@a11y.social

Image descriptions: Do I have to describe what a herringbone fabric pattern and full brogue shoes look like? CW: long (over 1,700 characters), image description meta

Artikel ansehen

Sun, 20 Oct 2024 10:36:05 +0200

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

@scope filter That's what alt-text guides for static websites, blogs and commercial social media silos tell you. The reason why I was asking in the first place was because Mastodon tends to think differently.

See, these alt-text guides keep preaching to keep alt-texts short. Mastodon, on the other hand, loves long and detailed image descriptions, even in alt-text.

#Long #LongPost #CWLong #CWLongPost #AltText #AltTextMeta #CWAltTextMeta #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta

Mon, 21 Oct 2024 06:27:32 +0200

scope filter

scopefilter@a11y.social

@jupiter_rowland Mastodon has probably given a large space for alt text for people with artwork or complex graphics, it doesn't mean it's a blanket rule that should be applied constantly. I think you should consider the needs of users who need alt text before the preference of any social media platform for SEO or other reasons. If the image can be described in a sentence or two, great. But if it's a complex image or an image of text you need more words to describe it adequately.

Tue, 22 Oct 2024 09:45:24 +0200

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

@scope filter

I think you should consider the needs of users who need alt text before the preference of any social media platform for SEO or other reasons.

That's what I already do.

But first of all, my original images are an edge-case not covered by accessibility rules. They are from very obscure 3-D virtual worlds. Nobody knows anything about them. Non-sighted people don't know what anything in them looks like. But they may want to know, also because I expect this topic to make at least a few people curious.

Sighted people may completely ignore the context and start exploring this new and unknown universe by looking around in the image, especially if it is more detailed. I mean, I'm not talking about simplified, cartoonish worlds like Meta Horizons. Non-sighted people may want to do the same, but they can't, and yet, they've got the right to do it. So they need my help: I have to explain details that don't matter in the context of the post.

More often, actually, the whole image matters because the post is either about the image itself or about everything in it. To repeat myself: Everything in the image is unknown and unfamiliar to both sighted and non-sighted people. But both sighted and non-sighted people may be very curious about everything. And so I have to describe everything and explain it if necessary. In contrast to real-life photographs, I can hardly assume anything to be already known and skip it in my descriptions.

Lastly, as I've already said: Mastodon is different from the rest of the Fediverse, from all other social media and from blogs and static websites. The rest of the Fediverse doesn't care for image descriptions. The rest of the Web is all geared towards alt-text that's as short as possible, also because HTML allows stuff like captions in special places in addition to alt-text and such. If your alt-text is over 200 characters long, you're doing it wrong.

Mastodon, on the other hand, loves long, detailed image descriptions to bits and pieces. It loves, loves, loves them. A real-life photo of one blossom in front of a blurry background, 800 characters of alt-text, and Mastodon applauds. Add another 200 characters of whimsy, and the blind Mastodon users applaud even more.

I'm not even kidding. Let the same image be described by an average human and an AI. The whimsical 750-character alt-text from the AI makes accessibility professionals scream in agony, but it makes blind Mastodon users cheer with joy, regardless of how accurate it is. The 200-character human-written alt-text makes accessibility professionals give thumbs up, but it makes blind Mastodon users yawn and fire up the AI on their phones for something more detailed and more whimsical.

Not few Mastodon users like image descriptions the more, the longer and more detailed they are. I'm still not kidding. And I'm still talking about real-life photos where lots of things are correctly assumed to be familiar, don't matter or both.

As I've already said, my virtual world renderings require even more detailed descriptions because nothing in them is familiar. And so my strategy has to be a little bit different: I describe each one of my original images twice.

The first description is the long description: a full, detailed description that goes into the post itself where I don't have any character limits. It contains visual descriptions of everything. It contains transcripts of all bits of text within the borders of the image, regardless of whether the text is readable at the given image resolution or not. And it contains all necessary explanations, starting with well over 1,000 characters which explain where the image is from. This long description regularly grows several tens of thousands of characters long. All this is actually necessary.

The second description is the "short" one which goes into the alt-text. The alt-text always mentions that there is a long, detailed description in the post, hidden behind a summary of the post plus a content warning due to the length of the post and maybe other things.

But I expect there to be alt-text extremists on Mastodon who absolutely insist in there being a useful image description in the alt-text even if there are already over 76,000 characters worth of description for three images in the post itself (caution: these descriptions are outdated), and even if the alt-text itself mentions these descriptions. And so I have to give this second description.

I have to shorten it dramatically so that it fits into the 1,500-character limit imposed by Mastodon, Misskey and their respective forks. Granted, I have to cut the explanations because explanations do not belong into alt-text because not everyone can access alt-text. But I also have to cut most visual details. I have to cut the mere mentions of many elements in the image. I usually also have to cut the text transcripts, for one, because the visual details that mention where these bits of text are had to be cut, and besides, because all text in the image combined may exceed the 1,500 characters itself already or at least take up way too much space.

As of late, my alt-texts tend to be either exactly 1,500 characters or only missing one or two characters.

Even fairly simple images get lengthy descriptions. In this case, one alt-text measures 1,500 characters, the other one is one character short. The long descriptions are over 20,000 characters combined, i.e. almost 11,000 characters for the common preamble for both images, over 2,800 characters for the first image and almost 6,600 characters for the second image.

My personal record is this fairly complex image: Again, 1,500 characters of alt-text of which a bit over 1,400 characters are actual image description and over 60,000 characters of long image description.

Oh, and by the way, in case you're unfamiliar with the places I've linked to: They're both part of the Fediverse. They're all federated with Mastodon. And yet, they both have no character limits and support for inline images.

#Long #LongPost #CWLong #CWLongPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #AltText #AltTextMeta #CWAltTextMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta #VirtualWorlds