If an Ogg index appears valid, but points to an incorrect time, we'll still use it to direct seeks and attempt to decode forward from the index-suggested time to the seek target. For example if the Ogg index mapped every Theora time stamp to the first Theora page, we'd end up downloading and decoding from the start up to the seek target frame, which could take a while. I'm not sure if WebM has a similar problem. One solution is to should check the timestamp when we finish the keyframe assisted seek, and if it's not within the max keyframe interval of the target, then fallback to a bisection search seek.
(In reply to comment #0) > I'm not sure if WebM has a similar problem. Yep. If the Cues point to an invalid (from the point of view of the parser) location, the seek will fail. If the Cues point to a valid (but incorrect) location, the seek will appear to succeed. If the result is before the desired seek target, DecodeToTarget will eventually reach the correct point. If the result is after the desired target, I'm not sure what happens.
Component: Audio/Video → Audio/Video: Playback
You need to log in before you can comment on or make changes to this bug.