Update 10.additional.tables.md

[luctrudeau]

If my understanding is correct, the only valid splits for BLOCK_8X8 are PARTITION_NONE, PARTITION_SPLIT, PARTITION_HORZ, PARTITION_VERT see Default_Partition_W8_Cdf

I understand this will also requires a change in the code https://aomedia.googlesource.com/aom/+/master/av1/common/common_data.h#71

[peterderivaz]

I think you are quite right that these cases are not covered (in Argon Streams coverage we explicitly mark these 4 8x8 cases as impossible to reach), however I think this is just an editorial matter because I believe making this change has no normative effect on the decoding process (do you agree?).

So I agree that this would be a nice-to-have modification to avoid confusion (especially when people are testing coverage), but I believe it would not be a normative change to either code or spec.

[luctrudeau]

I agree that this is not a normative change, it makes the table more consistent with the video format. The reason I point out the code change is that this table is shared between the aom code base and the spec. I'm assuming the intent is that they match.

[luctrudeau]

Any news on this?

[peterderivaz]

I haven't heard anything myself, but in general there is a strong reluctance to change either the code or the specification because people have taped out hardware for AV1 now and get very stressed whenever they see changes being made. If people do feel strongly (either way), then please feel free to comment.

[PDX-Runner]

issue #313 was also filed to track this particular change

[imoccaga]

My understanding is that there is nothing broken. And, I see from past comments that this requires change in both spec and code, so I would assume that currently there is no mismatch between code and spec.

If my understanding is correct, it seems this is a low level change that is not needed.