Irrlicht Engine

The rendering problem was confirmed due to some other compatibility issue on my side due to some true type font code that I had adapted. Thanks for the pointers on how to set up for custom rendering on the latest code. With these and changing back to bitmap font fixed the issue. Cheers!

A quick update to the CBillBoardSceneNode, I managed to get the health bar back to red colour by modifying private Material in the constructor of the class:       Material.Lighting = false;     Material.MaterialType = irr::video::EMT_ONETEXTURE_BLEND;     Material.MaterialTypeParam = irr::video::pac...

Hi Nadro, CuteAlien, I have upgraded to 5004. Unfortunately, it still doesn't work. I am supposed to see a button that is magenta in colour, and in stead the whole screen is magenta. I suspect there is some coordinates conversion error but could be due to my own issue, so will investigate a bit on m...

Hi CuteAlien and Nadro,

Thanks for the pointers. I will test it out soon and let u know the outcome

I recently updated my usage of ogles branch from revision 4898 to 5001 and noticed that there are big improvements, especially for OGLES2 driver where the shaders were overhauled (kudos to the effort). However, one side effect it had on me is that it breaks some of my existing code which uses the OG...

Thanks for all the awesome works

I managed to get ogles2 running on iOS simulator again by doing the following. Not sure if I am doing all the right things but it seems to render most things (as opposed to a white screen with my previous hack). Some stuff still not rendered correctly, but may be due to lack of shaders. Please feel ...

Hi Nadro, Thanks for the well thought out reply. Originally, I am trying to delay going to ogles v2 later instead of sooner. There are several 'issues' that I need to overcome to bring v2 on par with v1. Though most of these 'issues' could just be a matter of configuring customer shaders like you sa...

My knowledge with OpenGL is quite basic and a little rusty, hence will prefer to stick with v1 which already works on most aspects, and try to fill in the last piece that I need (skeleton anim in hw). I also have a feeling much of the custom code is also reusable if I decide to go v2 later, so will ...

I am staying with ogles1 cos it is better supported by irrlicht at this point in time? I tried v2 and didn't get it to run properly, hence thought it will be more tricky.

@Nadro thanks for the pointers! I will go in the recommended direction. @hendu i did consider creating a map of {frame => animated mesh} as cache so that each frame is only animated once, but fear it may become another bottleneck further down the road, hence think will stick to the fixed-function sk...

Hi Nadro, thanks for the info. I thought over this for some time. For my usage, I don't think I can get reasonable performance without hardware skinning. Irrlicht is ideal for my usage except for this last case, do you think it is feasible to hack a simple custom scene node that does hardware skinni...

I have a new observation that HardwareSkinning (a private bool) is false in the corresponding CSkinnedMesh instance. I am hopeful enabling this can close the performance gap (please correct me if I am wrong). Any idea if this can be enabled in for ogles1? If yes, how?

Thanks!

Hi CuteAlien, Thanks for offering to help. I did some more test and also looked a bit into irrlicht source code, turns out my original good performance (~15fps) is unrealistic and is a result of me using irrlicht in an unnatural manner, i.e. I loaded many nodes of the same 3d model at startup, and t...

@CuteAlien Thanks for the suggestion. To verify this, I tried node->setFrameLoop(0, , node->getMesh()->getFrameCount()) so that it animates through all the frames, but still face the issue, hence I doubt that's the cause...