Overlaying images on canvas

[atiro]

As a museum that carries out scientific and conservation work on our items that generate images from different parts of the light spectrum

I want to be able to overlay those images onto the canvas

So that users can choose which parts of the spectrum to see

[atiro]

Providing similar functionality to that Mirador provides with the layers tab.

[edsilv]

some discussion on image choice here: UniversalViewer/universalviewer#119

[edsilv]

Here's a mock-up: https://docs.google.com/presentation/d/1VTgKn6ZCl1xFeUa1cvm1WhTsv5h6gYl9i6MFn3lSDIo/edit?usp=sharing

[tomcrane]

Mirador's layers implementation isn't quite right; it treats everything as a choice even when it's not. If the motivation is painting the content should be painted onto the canvas by default, unless the content is provided inside the oa:Choice construction for when the user should decide between alternatives.

For example. in the Biblissima example at http://projectmirador.org/demo/advanced_features.html (Grandes Chroniques de France (Châteauroux, BM, ms 5)), the illuminations are shown only when the user switches them on, but that isn't what's being modelled in the manifest, where the main image and the illumination are both direct parts of the content.

If you viewed the Chateauroux example with the logic used by the current AV canvas component (which you can't quite do unfortunately) you'd see both images by default, they are both equally valid parts of the Canvas.

I'm not saying that the UV shouldn't (like Mirador) give users the ability to toggle the visibility of the Canvas's content - in fact this is desirable as more content (like text annotations) is rendered. It's just that the user experience offered by Mirador isn't quite what's being specified in the manifest.

In your case - multispectral imaging - the correct construction would be oa:Choice, with each band labelled; that gives control of the visibility to users. Mirador handles that correctly.

There are two distinct scenarios that the UX needs to take into account. The first is an explicit choice, as in the mock up @edsilv linked to linked above. The second is the ability for the user to see that the view is composed of separate parts (it might not be obvious) and possibly give the user the ability to manipulate their visibility. The Choice construction allows the publisher to say that it's OK for the user to do this, but offering UI to do this all the time might not produce the desired outcome. Again, in the AV Fire example the Canvas has many distinct pieces of content, and you don't want to interrupt the intended user experience with dropdowns etc by asking the user to control their visibility.

This is a subset of #4 as well.

[edsilv]

We should add this to the UVCON agenda to discuss on the UX day.

[tomcrane]

More generally, for UVCON, this issue is UV's support for Advanced Association Features of IIIF:

http://iiif.io/api/presentation/2.1/#advanced-association-features:

6.1. Segments
6.2. Embedded Content
6.3. Choice of Alternative Resources
6.4. Non Rectangular Segments
6.5. Style
6.6. Rotation
6.7. Comment Annotations
6.8. Hotspot Linking

Perhaps a quick vote on these, the extent to which they are required by stakeholders (some are far more commonly encountered than others)

Check this one out:
https://iiif.durham.ac.uk/index.html?manifest=t1mz029p473h&canvas=t1th128nf28z

Issues - I know for fact that many Wellcome manifests exhibit images that are really alternatives (fold-outs, etc) - the problem is in capturing the fact that two images are representations of the same page/view during digitisation.

Presentation 3.0 introduces Canvas-on-Canvas too.