Does SC 1.2.1 require the alternative for time-based media for video-only to be visible?

[kengdoj]

Hello,

This question originates from ACT Rules Issue 1815: Should the text alternative for a video without audio always be visually present?

For reference:

SC 1.2.1: Prerecorded Video-only
Either an alternative for time-based media or an audio track is provided that presents equivalent information for prerecorded video-only content.

alternative for time-based media
document including correctly sequenced text descriptions of time-based visual and auditory information and providing a means for achieving the outcomes of any time-based interaction

[dav-idc]

Some personal thought that might help? Potentially not.

The alternative for time-based media, when it's in the form of a text transcript is sometimes offered on a separate page or as a separate document, and linked to near the video.

In that case, the alternative is definitely not 'visible' near the video if visibility is interpreted as "can be observed visually".

But, once the transcript is accessed, wherever it lives, I would think that it needs to be available visibly.

So:

• no, the video alternative doesn't need to be physically visible on the same page as the video
• yes, wherever the video alternative lives it should be available visibly

[patrickhlauke]

I'd say, more generally: since for images / non-text content in general, having something like alt which is only exposed to AT is an acceptable way to pass 1.1.1, it would logically follow that for video, having an alternative that is only exposed to AT (i.e. not visible in the page, or linked out, or otherwise accessible separately) is a valid way to pass 1.2.1 as well

[rscano]

I'd say, more generally: since for images / non-text content in general, having something like alt which is only exposed to AT is an acceptable way to pass 1.1.1, it would logically follow that for video, having an alternative that is only exposed to AT (i.e. not visible in the page, or linked out, or otherwise accessible separately) is a valid way to pass 1.2.1 as well

Agree but think also that a transcription can be useful also for other AT, e.g. AT for cognitive disabilities that can simplify languages. So due that for alternative for image there is a specific attribute (that can be catched by AT), for transcript or description there are no specific attributes so IMHO shall be percepible not only by people that cannot see video.

[mraccess77]

The same argument to put things off-screen could be applied to other criteria as well but it's not the intent of the criteria in my opinion.

Putting it another way- for 1.2.1 - an audio file could have off-screen text as the transcript used by people who are deaf or hard of hearing? That wouldn't be acceptable? The flip side is why would low vision folks be excluded from having visible alternative for video only content?

[patrickhlauke]

The flip side is why would low vision folks be excluded from having visible alternative for video only content?

why should they be excluded from having visible alternative for <img> elements? and yet that's kosher under 1.1.1

[mraccess77]

I think the different between an image with alt and a media alternative is the length of it. A media alternative is defined as a document.

[dav-idc]

So due that for alternative for image there is a specific attribute (that can be catched by AT), for transcript or description there are no specific attributes so IMHO shall be percepible not only by people that cannot see video.

I definitely agree that image text alternatives aren't a one-to-one comparison with multimedia text alternatives.

Text alternative for multimedia (like a transcript) doesn't match well to something like alt text, due to the potential length of the transcript. Having a transcript read out the same way as alt text gets read out by screen readers would be pretty disastrous, as alt text is read out all at once. For an hour long piece of multimedia, there's the potential for several hundred lines of transcript material.

[rscano]

Text alternative for multimedia (like a transcript) doesn't match well to something like alt text, due to the potential length of the transcript. Having a transcript read out the same way as alt text gets read out by screen readers would be pretty disastrous, as alt text is read out all at once. For an hour long piece of multimedia, there's the potential for several hundred lines of transcript material.

We are talking about video-only, not multimedia contents. For multimedia there are usually captions and audio-descriptions.

[dav-idc]

We are talking about video-only, not multimedia contents.

Thanks for the note, I definitely missed that distinction. My bad there!

[bruce-usab]

Trying to summarize this thread:

• Yes, text alternative for video-only must be visible to satisfy 1.2.1.
• Caveat is that the text alternative might not be on the same page, but then the UIC for the text alternative would need to be visible.

Please thumbs up, or thumbs down with comment to continue the discussion.

[dbjorge]

Please thumbs up, or thumbs down with comment to continue the discussion.

I agree conceptually, but thumbs-downed only because I think we need to be careful about the terminology; I think for that summary to be accurate, you'd need to replace the term "text alternative" with "alternative for time-based media".

Normatively, "text alternative" (as used by 1.1.1) and "alternative for time-based media" (as used by 1.2.1) are not (necessarily) the same thing.

For prerecorded video-only time-based media that:

• Doesn't have any associated audio track,
• Isn't a clearly-labelled media alternative for text, and
• Doesn't meet 1.1.1's "Decoration, Formatting, Invisible" exception

... I would interpret the combination of the two SCs like this:

• Authors must provide both a "text alternative" (per 1.1.1) and an "alternative for time-based media" (per 1.2.1).
• Per 1.2.1, the "alternative for time-based media":
  • Must be "provided" somehow (without any qualification, I'd interpret this as meaning "for all users")
    • This implies that it must be visible in order for it to be considered as "provided" to visual users
  • Does not need to be on the same page (I think this is implied by the definition calling it a "document")
    • ...but if it is not, it still must be "provided" (for all users), which would require a visible mechanism providing it
  • Does not need to meet the "programmatically associated" requirements from the "text alternative" definition
    • ...but must be "provided" to AT users somehow. I think "programmatically associated" would be best-practice, but I think so long as the alternative is somehow accessible to an AT user, it'd meet the letter of the requirement.
  • Must provide "equivalent information" to the media
  • Must meet the completeness requirements specified by the "alternative for time-based media" definition (correct sequencing, etc)
• Per 1.1.1, the "text alternative":
  • Must meet the usual requirements from the text alternative definition (especially with regard to being programmatically associated)
    • Per the usual text alternative requirements, does not need to be visible (eg, an aria-label would be an acceptable means of providing a text alternative for a <video> element)
  • Must "at least provide descriptive identification of the media"
  • Does not need to provide "equivalent information" to the media (but would be allowed to do so)
• The two types of alternative may be implemented by the same mechanism, provided it meets all the requirements of both SCs/definitions.
  • eg, an aria-labelledby on a <video> element which points to a visible text description that provides "equivalent information" to the video
  • Although this is normatively allowed, it is only recommended when the "alternative for time-based media" is short and simple. If more than a sentence or two is required, best practice would be to use a separate shorter text alternative in addition to a more complete alternative for time-based media.

[bruce-usab]

Good catch @dbjorge as the context is 1.2.1 and alternative for time-based media. With the context of the original question, I do not agree that 1.2.1 is even applicable. I would prefer a better real-world example of invisible text used as the alternative for time time-based media, as I am skeptical that it happens.

[dbjorge]

With the context of the act-rules/act-rules.github.io#1815 (comment), I do not agree that 1.2.1 is even applicable.

Can you explain why you think 1.2.1 wouldn't be applicable to that example? You talked in this comment about the possibility of that example being a "pure decoration", but even if that was the case for the example in question, that would only exempt it from 1.1.1, not 1.2.1.

(edit: and even for 1.1.1, I think the normative text there is actually a bit ambiguous about what the requirement is for something that meets multiple of the special scenarios simultaneously, like a purely decorative video would - I presume the intent is for the "decorative" requirement to take priority, but I don't think that's actually well-specified in the SC)

[WilcoFiers]

This one surprised me. I very much thought this was required, but thinking through it carefully, I think there is one pretty glaring problem with this; Animated gifs. If 1.2.1 required visible alternatives, then 1. either animated gifs need a visible alternative, or 2. something about animated gifs would make not quality as "video-only" content.

1. It seems generally acceptable that animated gifs can pass WCAG with an alt attribute. I also can't really think of what the benefit for someone with a disability would be. Granted there is the case where some people struggle to process things like social cues from images, but that's as true for still images as it is for short video clips, and it isn't required for still images. So that doesn't seem right.

2. Maybe animated gifs aren't "video-only" content. Looking at the definition of video-only I can't think of an argument for why that might be true. They are a "time-based presentation that contains only video", with video being "the technology of moving or sequenced pictures or images".

[dbjorge]

I very much thought this was required, but thinking through it carefully, I think there is one pretty glaring problem with this; Animated gifs.

I agree - I think normatively 1.2.1 requires visually presented descriptions of animated gifs, which surprises me and strikes me as possibly-unintended.

[WilcoFiers]

@dbjorge I don't think I agree with your conclusion that "provided" means "to all users". 1.2.2 says captions need to be provided. I don't think that means captions need to be accessible to screen readers.

[dbjorge]

@dbjorge I don't think I agree with your conclusion that "provided" means "to all users". 1.2.2 says captions need to be provided. I don't think that means captions need to be accessible to screen readers.

In the case of captions specifically, the "captions" definition explicitly says "visual and/or text alternative" - I think that's the mechanism by which captions are allowed to be visual without a text alternative.

I looked around for other cases where we use "provides" or "provided" in SC text without that sort of qualification and clarify the scope/interpretation in an understanding doc. Some cases are as ambiguous as this one is (eg, 3.3.3), but 3.3.2 is a good example where the SC uses similar language ("Labels or instructions are provided when content requires user input") and its understanding doc clarifies explicitly that our interpretation is that this wording means labels and instructions must be "presented to all users, not just those using assistive technologies." (see 4th paragraph of Intent section)

[mraccess77]

@dbjorge I don't think I agree with your conclusion that "provided" means "to all users". 1.2.2 says captions need to be provided. I don't think that means captions need to be accessible to screen readers.

Since WCAG allows for open captions this may be the case - but captions that are not accessible to screen readers will not be available to people who are deafblind and need them in braille - so this is something should be addressed in WCAG 3.

[bruce-usab]

I realize that an appeal to consequences is a logical fallacy, but it cannot be the case that WCAG conformance requires that all animated gifs have an alternative for time based media comparable to what is required for silent movies and slide shows.

From 1.1.1, emphasis added:

If non-text content is time-based media, then text alternatives at least provide descriptive identification of the non-text content. (Refer to Guideline 1.2 for additional requirements for media.)

If the text alternative provided to meet 1.1.1 serves the equivalent purpose (as the non-text content) it follows that this text is sufficient to also meet 1.2.1.

A text alternative that serves the equivalent purpose is more than what 1.2.1 requires, which is merely text descriptions of time-based visual … information. Text equivalents are better than text descriptions.

Does SC 1.2.1 require the alternative for time-based media for video-only to be visible?

No. The adjective of visible is not part of that portion of the SC, and not part of the definitions for text, text alternative, nor alternative for time-based media.

[GreggVan]

BRUCE, I see the problem you're trying to solve. But I don't follow your logic. According to this, I could just provide a descriptive identification for a movie and it would automatically meet 1.2.1.

There is nothing about the statement in 1.1.1 that says that providing a descriptive identifier"serves equivalent purpose". It only says that you at least need to do that for 1.1.1 knowing that 1.2.1 will be following and require more. And meeting One success criteria of course doesn't mean that all other related success criteria are met.

Having said that, I see the problem you're trying to solve, and, we should try to figure out how to do it. But I don't see how your current logic works without opening up a Pandora's box of unintended consequences for longer clips. Unfortunately, we don't have anything here about the length of the material and I think that is what you're trying to address

Comment back if I'm missing something...

[bruce-usab]

According to this, I could just provide a descriptive identification for a movie and it would automatically meet 1.2.1.

No, descriptive identification is the minimum requirement for not providing a text alternative that serves the equivalent purpose of the the non-text content.

There is nothing about the statement in 1.1.1 that says that providing a descriptive identifier "serves equivalent purpose".

I concur completely that descriptive identification is not serving the the equivalent purpose.

Unfortunately, we don't have anything here about the length of the material and I think that is what you're trying to address

@mraccess77 also suggested length of the video might be the trigger for requiring something like a screen play versus something like alt text. I don't disagree with the practicality of the approach, but it's not supported by the normative phrasing. It is also possible for a few seconds of video-only media to convey a significant amount of information! This is why I am making the case to assess the quality of the text alternative provided for meeting 1.1.1. If that text alternative is only providing descriptive identification, that satisfies1.1.1 but not 1.2.1. If that text alternative serves the equivalent purpose, that satisfies both 1.1.1 and 1.2.1.

[bruce-usab]

Getting back to a couple of @dbjorge points:

....but even if that was the case for the example in question, that would only exempt it from 1.1.1, not 1.2.1.

I am sure you agree that pure decoration is written so that it can be applied to non-text content other than images. I recognize it is not an exemption within normative phrasing of 1.2, but there are other situations where auditors have to make inference and use common sense. I am optimistic we might arrive at something that can be added to Understanding without contradicting normative phrasing.

even for 1.1.1, I think the normative text there is actually a bit ambiguous about what the requirement is for something that meets multiple of the special scenarios simultaneously, like a purely decorative video would - I presume the intent is for the "decorative" requirement to take priority, but I don't think that's actually well-specified in the SC.

Agreed on both points. When I teach 1.1.1, I focus on descriptive identification versus text that serves the equivalent purpose. Pure decoration is interesting since null alt is the equivalent purpose. There are also many situations were descriptive identification is the text equivalent. But that kind of nuance is not part of "wcag 101"!

[bruce-usab]

Official Working Group Response...

From SC 1.2.1:

Prerecorded Video-only: Either an alternative for time-based media or an audio track is provided that presents equivalent information for prerecorded video-only content.

The definition for alternative for time-based media (emphasis added):

document including correctly sequenced text descriptions of time-based visual and auditory information and providing a means for achieving the outcomes of any time-based interaction

I am glossing over "document" to focus on "correctly sequenced text descriptions" which is just qualifying that the words of text are in a sensible order and sufficiently descriptive.

The definition for text:

sequence of characters that can be programmatically determined, where the sequence is expressing something in human language

Nothing there specifying "visible". I note that this is consistent with alt text not (typically) being visible.

No, SC 1.2.1 does not require alternative for time-based media for video-only to be visible.

[dbjorge]

Nothing there specifying "visible". I note that this is consistent with alt text not (typically) being visible.

No, SC 1.2.1 does not require alternative for time-based media for video-only to be visible.

I agree that this probably has to be our interpretation assuming we want to avoid visual alternatives being required for animated gifs, which I think we're in agreement on; I don't think we have any normative mechanism to say "sometimes it needs to be visual and sometimes not".

When we update the understanding doc to clarify this position, it'd be nice to make sure we also clarify with an example that it isn't required for the text alternative and the alternative for text based media to be the same thing, and that for videos that are particularly long or complex, it might aid usability to separate them.

[kengdoj]

No, SC 1.2.1 does not require alternative for time-based media for video-only to be visible.

Can ACT move forward this decision?

[bruce-usab]

Discussed on TF call 4/12 and, we have consensus on the answer to the substantive questions. Per our process, TF facilitators won't close this issue until AG WG gets a chance to review. Official Working Group sent for WG approval 4/16:

---

On the original question:

Does SC 1.2.1 require the alternative for time-based media for video-only to be visible?

No, it does not.

Can ACT move forward this decision?

Yes please.

[mbgower]

Closed as a result of the Official Working Group Response being reviewed and adoped