Should PixelLayout be required?

[palemieux]

The PixelLayout item is currently required in ST 2067-40 2ED, and duplicates the value of the J2CLayout item.

The PixelLayout item is ignored in ST 2067-21 and the value of the J2CLayout item is used instead.

This divergence is not ideal.

[palemieux]

20200727 telecon

App 4 should match App 2 since both use J2K and PixelLayout duplicates J2CLayout. Recommendation would/could be different in applications that use a difference codec.

[hnlocher]

Just one question after yesterdays meeting, and before modififying 6.1.3.2.3 PixelLayout and 6.1.3.3.2 J2CLayout subclauses.

The situation here is that the in the RGBA Picture Essence Descriptor duplicates the PixelLayout field already present in J2CLayout. (The language used is in the other direction, btw).

I agree that duplication is error prone and can create inconsistencies, but what would happen if later one other codec is accepted into app#4 (non compressed essence, for instance) ? Would we bring the PixelLayout in RGBA Picture Essence Descriptor back ?

[hnlocher]

In the meantime, correction proposed :

• Remove section 6.1.3.2.3 and use the clause wording to define content J2CLayout in 6.1.3.3.2
• Add PixelLayout as "shall not be present" in table 11 : RGBA Essence Descriptor items

35PM-CD-ST-2067-40-revision2018-2020-07-28-11h53(redlineFrom2020-07-22)

[palemieux]

Add PixelLayout as "shall not be present" in table 11 : RGBA Essence Descriptor items

Unfortunately PixelLayout is required in the RGBA Essence Descriptor, shall not be present is not possible.

[hnlocher]

So according to 2067-21, the wording could be :

PixelLayout : "Shall be ignored."

If no other comment, I will modify the document this way.

[hnlocher]

According to a comment made by Erik Dobberkau, the reference to the suppressed section in the table has been removed as well.

[hnlocher]

Changes reflected in pull request #21