What's the accepted way to do translations of Alt Text on Images. On a single language site we'd use an attribute of the Media Item for the alt-text, but that's resulting in us seeing English Alt Text on the "other languages". I'm wondering how people go about implementing this.

That’s a really interesting question, posting really so I can follow this! You could add a block list on the media item itself, choose language, enter alt text but that feels rather clunky, must be a better way than that surly.

that sounds horrible

When I have been asking about adding alt text to images in the image library, the reason I was given it is not currently possible was because the image library did not support variants. I have been assuming this would be one of the improvements moving to the new back office.

Yeah the media library hasn't changed much in 15 years 🙂

Theres a package called Multi-lingual text box that was created for adding multi-lignual capable text fields to media items. That would be a starting point. But alt text should be "context aware" really. So yes having a fallback on the main media item is okay, but it's not the best solution. Having a field located at the same point as the media picker, that is also aware of the crops available would be another option. Ideally you also need to be able to flag an image as "decorative" that wouldn't have an alt description rendered. Finally, consider dynamically generating them where possible - e.g. in a logo lister where you could set it as "Logo for 'company name'" is another option.

100% and was one of the reasons we created Wholething Text Fallback. We always intended to use the image library as a fallback being 1 step better than the node name. And of course Alt text per crop would be ideal. I have seen some AI offerings as well but as you say @Nik context is really important for Alt text and I can't see AI being the solution.

@Paul Marden have you seen this package ; https://marketplace.umbraco.com/package/our.umbraco.multilanguagetextbox No V14 version yet

Thankyou sir. Callum pointed me to that in the MVP slack. It was exactly what i was after

Just getting a client to name a node intelligently and not just blanket upload images that create nodes with img-xxxx.jpg from camera files is hard enough.. now we need multilanguage alt text fallbacks and context aware.. the mind boggles! (the legacy mediapicker used to ask for alttext and caption, but that didn't progress to the latest mediaPicker3...) Had a client recently who also need to attribute the source of every image for copyright purposes!