Mirrors/overte
3
0
Fork 0
mirror of https://github.com/overte-org/overte.git synced 2025-04-23 19:53:30 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

Improve blendshape precision for small models

Browse source 
For avatars authored in meters, the max component length would often be less then 1.
In that case, the blendshape offset was not normalized before quantization, resulting in loss of precision.
This change will normalize the offset for all cases, except when the max component length is 0.
This commit is contained in:
Anthony Thibault 2019-03-26 11:21:13 -07:00
parent 754b281bbf
commit 08c6acdf99
1 changed files with 1 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
libraries/render-utils/src/Model.cpp
View file

@ -1648,7 +1648,7 @@ using packBlendshapeOffsetTo = void(glm::uvec4& packed, const BlendshapeOffsetUn
void packBlendshapeOffsetTo_Pos_F32_3xSN10_Nor_3xSN10_Tan_3xSN10(glm::uvec4& packed, const BlendshapeOffsetUnpacked& unpacked) {
float len = glm::compMax(glm::abs(unpacked.positionOffset));
glm::vec3 normalizedPos(unpacked.positionOffset);
if (len > 1.0f) {
if (len > 0.0f) {
normalizedPos /= len;
} else {
len = 1.0f;