The WAVE presentation starts with the earliest video and audio sample

[yanj-github]

The WAVE presentation starts with the earliest video and audio sample that corresponds to the same presentation time as the earliest video sample.

This observation aims to verify that the audio and video start with the same media time. The comparison is made against the JS starting time, focusing on the following:

• The media time of the first presented video frame.
• The media time of the first audio segment.

If we allow the possibility of the starting audio being missed (or video missing), how should we define and implement a tolerance for this observation?

[jpiesing]

This is a very good question.
I've tried several times to come up with a technical solution.
I lean towards a more structural solution, which is to disable the starting audio being missed / starting video being missed tolerance in video+audio observations for whichever is the later to start and instead use a new tolerance for this observation.
EDIT: I did think about using the same tolerance for missing audio at the start in an audio only test as for missing audio at the start in video+audio tests. After more thought, I think separate tolerances is better even if they're the same value.

[yanj-github]

I have added earliest_sample_alignment_tolerance default to 60ms.
Here is the updated results: