Your right it is a coincidence and it is a rumor that no one can validate at this point. Guess we'll have to wait and see.
David,
Don't be so hard on yourself. If the rumour mongering party (whom I suspect is correct) would release a bitstream or a description of how to recreate the event, well, then, we can all verify it.
Software decoders skip frames anyway, so it may not be surprising that they play the stream with no unusual artifacts. ESST is a programmable chip and may choke on difficult stream segments. This would indeed be a fundamental architectural flaw, and not uncommon in an industry of visionaries prone to excess handwaving (MicroUnity, UltraSparc, 8x8, TriMedia, etc.).
It could be just a bad interpretation of the MPEG syntax, which can be corrected with a new firmware (microcode) update. It could even be an illegal bitstream element which other decoders are more forgiving. Or it could be a critical bug in the hardware, such as a missing or errant table entriy hardwired onto the chip. Even this can be corrected with a new rev.
Give us the recipe. |
