Decode BasisU textures to BCn format instead of RGBA8 #474
Conversation
lilleyse
requested review from
weegeekps,
r-veenstra,
corybarr and
mattelser
as code owners
|
I confirmed that the texture memory is lower because of this change. See the
|
src/core/src/Context.cpp
Outdated
| @@ -637,6 +641,15 @@ const pxr::TfToken& Context::getCesiumMdlPathToken() const { | |||
| return _cesiumMdlPathToken; | |||
| } | |||
|
|
|||
| bool Context::isCompressedTexturesSupported() const { | |||
| // Compressed textures do not work in Kit 105.0 or 105.0.1 | |||
| if (_kitVersion.rfind("105.0", 0) == 0) { | |||
There was a problem hiding this comment.
Is this even necessary? We just released version 0.11.0 which claims that versions less than 105.1 are unsupported.
There was a problem hiding this comment.
Oh good point... I'll remove it
There was a problem hiding this comment.
I think I worked on this PR before realizing that 0.11.0 wouldn't be backwards compatible with Kit 105.0
There was a problem hiding this comment.
In that case do we even need any of the code to pass the kit version up to the native layer?
There was a problem hiding this comment.
Nope, I just removed it all
… on kit version" This reverts commit d31637e.
Fixes #266
Previously, compressed textures were decoded to RGBA8 instead of the native BCn format due to a limitation in Kit 105.0. Now with Kit 105.1 we can decode to BCn and keep the data compressed on the GPU. This involves a bit of a strange workaround with the
strideparameter (see comments in the code).Test data:
agi-ktx2-explicit.zip
agi-ktx.usda.zip