This topic contains 31 replies, has 23 voices, and was last updated by 
nobuyuki
2 years, 1 month ago.
-
Posts
-
I disagree… 3 then 2. and obviously, the 2 engine won’t have the same feature set.
Most people probably won’t take advantage of GLES3-specific features but it’s definitely not going to “cut your userbase in half” like some people are saying. If you’re targeting 3d for a new project then you’re probably not targeting a legacy userbase. If you are, why not backport it yourself? This seems like a special case scenario considering the circumstances. Usage of GLES2-only devices is on the decline.
Also, one thing that hasn’t been mentioned yet is that GLES3 guarantees availability of ETC2 texture compression, whereas on GLES2 you maybe got ETC1 but it wasn’t guaranteed to be available, and it didn’t support alpha channels anyway which often necessitated figuring out what type of hardware compression the device did support (cludge around or have a table of GPUs lying about) and creating duped assets for each of these, then specifying it in the manifest to reach the widest audience.
It could be a mess; PowerVR, nVidia, and AMD all had their own standard, and in AMD’s case, they had two(!). Part of the convenience of GLES3 is you don’t have to mess with any of these; you just use Ericsson 2 compression and you’re good to go.
You must be logged in to reply to this topic.