Once ACL supports streamable cosmetic joints, we must integrate it into the UE4 plugin.
In order to determine if a joint is cosmetic or not, we can use the LOD information from the skeleton. In the database codec, the settings should be split into two sections: one for the base joints, and one for the cosmetic joints. We must also specify the value to use for cosmetic joints, something like anything above LOD 2 is cosmetic. We should also expose a list of joints that are forced to be base joints. This is required if we need to decompress them individually (as opposed to part of the whole pose). We need to warn, similarly to bind pose stripping which this feature will depend on.
See FSkeletalMeshLODModel::RequiredBones for LOD information.
The visual fidelity level currently has 3 levels: lowest, medium, highest. This must be expanded by 3 new levels.
Lowest will contain only the base joint clips which are always loaded in memory (tier 0).Low will contain base tier 0 and 1Medium will contain base tier 0, 1, and 2MediumHigh will contain all 3 base tiers and cosmetic tier 0High will contain all 3 base tiers and cosmetic tier 0 and 1Highest will contain all 3 base tiers and cosmetic tier 0, 1, and 2
It should be possible to strip cosmetic joints entirely similarly to how we can strip tier 2.
Once ACL supports streamable cosmetic joints, we must integrate it into the UE4 plugin.
In order to determine if a joint is cosmetic or not, we can use the LOD information from the skeleton. In the database codec, the settings should be split into two sections: one for the base joints, and one for the cosmetic joints. We must also specify the value to use for cosmetic joints, something like anything above LOD 2 is cosmetic. We should also expose a list of joints that are forced to be base joints. This is required if we need to decompress them individually (as opposed to part of the whole pose). We need to warn, similarly to bind pose stripping which this feature will depend on.
See
FSkeletalMeshLODModel::RequiredBonesfor LOD information.The visual fidelity level currently has 3 levels:
lowest, medium, highest. This must be expanded by 3 new levels.Lowestwill contain only the base joint clips which are always loaded in memory (tier 0).Lowwill contain base tier 0 and 1Mediumwill contain base tier 0, 1, and 2MediumHighwill contain all 3 base tiers and cosmetic tier 0Highwill contain all 3 base tiers and cosmetic tier 0 and 1Highestwill contain all 3 base tiers and cosmetic tier 0, 1, and 2It should be possible to strip cosmetic joints entirely similarly to how we can strip tier 2.