Matias N. Goldberg

> to prevent a division by 0, I guess; which can NEVER happen since roughness lower bound is set to 0.001 Mmm this is a very good point! The extra...

OK I took a fresh look at you results. HOLY CRAP THAT IS LARGE DIFFERENCE. AND R BECOMES SMALLER INSTEAD OF BIGGER! You're right the 1e-6f needs to go, and...

OK removed. Although I don't know what will happen with numerical stability, since at low roughness G can get really big. We also need to monitor what happens with HDR....

We went over this one before. Our formula (I will ignore the epsilon and PI): ``` k = pixelData.roughness * pixelData.roughness float gL = NdotL * (1-k) + k; float...

> If doubles are not supported for such data, the RenderSystem should probably assert or throw when receiving such data as input Yes, you're right. D3D11 does not support doubles...

> Sorry, I think I misunderstood the issue. Indeed. One thing is to have Real typedef to double (which affects CPU calculations), and another one is to request VET_DOUBLEx (which...

I fixed all the `FIXME is not defined` + `MovableObject::calculateCameraDistance`. I also just fixed `FIXME ScheduledTasks`

1. Looks good. Thanks! 2. Good point. I didn't realize we're still pointing to Mercurial. All platforms are in the same state. On Linux & Mac it's clear to use...

Is there a small/quick way to repro this? (you could try exporting the scene via SceneFormatExporter and sending it to me) I would like to understand what's going on, because...

Hi! We're sorry you're experiencing these issues. The error you're having almost always can be traced to Ogre not being built with rapidjson support, which is required to parse the...