^ It's not that sophisticated. I suppose, potentially it could be, but there are lots of things in HDMI that aren't used in the grandiose manner conceived when they were originally bunged into the spec.
The device that wants to add audio delay has the option of inquiring of the next device in the chain how much audio delay it thinks would be useful to compensate for the time it takes processing the video. In most cases the value returned is a constant -- independent of user settings that might very well alter the actual video processing time.
That's it. There's nothing more sophisticated going on here, and the desired audio delay meta-data value passed back is often bogus. The entire HDMI chain is not surveyed and there's certainly nothing as sophisticated as trying to track markers in the content to see what amount of audio delay might REALY be needed.
In particular, understand that there is nothing in HDMI auto lip-sync correction which will detect or correct for lip sync errors which are already built into the content.
Auto lip-sync meta data is, as far as I know, only ever returned by TVs. And for the most part it is only ever polled for and used by AVRs (i.e., getting the value from the TV they are connected to). I've seen some evidence that some set top boxes are also willing to poll for this value, but if so, it will only work if the set top box is DIRECTLY connected to a TV that offers that value. I've never heard that auto sync delay actually works if the set top box is supposed to be applying the audio delay.
The OPPO players do not attempt to poll for audio delay meta-data.