This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
It is reasonable to suppose that tracks designed to support people with disabilities may be distributed separately from the primary media object for which they would represent a track element. But it appears impossible, in a case where the main content uses a CDM, to allow for a different CDM to apply to the tracks. This would significantly inhibit the production and distribution of accessibility enhancements by third parties, and therefore potentially the use of content by people with disabilities.
So, if I understand your description, one has the following options: (1) a/v tracks are encrypted and some CDM-1 would decrypt (2) t/t tracks may be: a. included in the container containing a/v tracks, and those t/t are also encrypted according to CDM-1 b. included in the container containing a/v tracks, but those t/t are not encrypted c. not included in the a/v container, and encrypted using an encryption according to CDM-2 d. not included in the a/v container, and not encrypted You are expressing concern about support for (2)c, yes? How likely is it that (2)c would apply instead of 2(d)? I would not expect producers and distributors of accessibility enhancements to use encryption at all, especially when they are a 3rd party, but I agree it can't be ruled out. Is my interpretation correct?
This was discussed in the F2F. We discussed that site authors could use the WebCrypto APIs to provide an equivalent level of protection. Resolving WORKSFORME.
(In reply to Adrian Bateman [MSFT] from comment #2) > This was discussed in the F2F. We discussed that site authors could use the > WebCrypto APIs to provide an equivalent level of protection. Resolving > WORKSFORME. WebCrypto might resolve the caption (and other text-file) issue, but what would happen when the supplemental content was picture-in-picture sign-language (which has previously been identified as an accessibility requirement - http://www.w3.org/WAI/PF/media-a11y-reqs/#sign-translation)? Recognizing the limitations of hardware (processing overhead), will user-agents be able to decrypt streams that have been encrypted using 2 (or more) CDMs (aka Glen's 2c scenario)? The current WORKSFORME response does not address this question.
I don't believe it is a requirement to support multiple video/audio tracks using different CDMs in the same media element. I don't think this is a common scenario and not something that implementations would be likely to support. This is a design constraint of EME that simplifies deployment. If you have examples of where content has been delivered with different encryption mechanisms please share.