Search Results

Found 1 results for "5ed1b0d2f7e8d4ae7546fe2c26bd86e3" across all boards searching md5.

Anonymous /e/2967700#2968009
5/27/2025, 3:40:16 AM
Trying to find the error, not even sure which effect it's in. Slightly rotating and raising the grass controller has the same effect of fixing the bug. With NormalDepth-Core.fxsub:

Changing pixel shader output, from "IN.vView.z" to "IN.vView.z + IN.vView.y": changes position of flowers, but doesn't affect whether they are bugged and don't show up. Changing IN.vNormal to "IN.vNormal + 0.001" appears to fix the bug.

The model's normals are unremarkable. Following the includes to MMxCompatible.h, the input also looks fine, with correct semantic. Don't know what "float Index : PSIZE15;" is, but it isn't used by this effect.

Changing to "s_mtxView = s_mtxBaseView;", to avoid any use of s_mtxAdjust: breaks it. Using s_mtxBase in place of s_mtxBaseView: also breaks it.

Using s_mtxBase[0], [1], and [2] for the first three rows of s_mtxBaseView: no change (when rotation is ~0, reverses tilt of control bone). Changing the last line to float4(0,0,0, 1.0 ): cancels movements of control bone, but doesn't break or fix anything. Processing here appears to be fine, with no bugs from normalizing a zero-sized vector or something.

Change output from "IN.vNormal", to "-IN.vNormal": is blue, meaning non-black. It was just negative. Suggests error, which is fixed by "IN.vNormal +0.001", is in processing of "Grass-R-Core.fxsub".