forums.ps2dev.org

Thank you TyRaNiD for the quick fix. Works like a charm!

Hi all, I hope somebody can shed some light on this. I am fooling around with adhoc stuff and had previously been running in kernel mode and doing my own stubs and fixups for adhoc functions. This all works wonderful. I would however like to use the adhoc lib instead, partly in the hopes that I coul...

I have used PSP link for some time now and all is great. I now wanted to convert one of my apps that uses some network code. That code previously ran in kernel mode as it needed to load the appropriate PRX files and patch them. Now we can only use pure user mode in PSPLink so I added the appropriate...

First of all I want to thank Tyranid for making PSPLink. An awesome tool! Thank you!! I little info about my setup.. I am using the latest PSPSDK running under Cygwin on a WinXP box. I have PSPLink working fine using the USB method. I then wanted to try out Insight+GDB, I downloaded the sources and ...

Just going off memory here since I do not have that sample within reach. 1/ The sample uses X,Y,Z and Normals as floats. To increase performance they could be cut to bytes which would yeild a savings of 18 bytes per vertex. 2/ The sample furthermore does not do any form of tri-stripping. Doing tri-s...

For vertex normals the GU_NORMAL_32BITF indicates a floating point value for the nomals which makes perfect sense.

Now for GU_NORMAL_16BIT does that use 8.8 fixed point or 1.15 fixed point or something else entirely?

Same question goes for GU_NORMAL_8BIT.

Thanks in advance for any answers.

Does anybody happen to know the performance impact of feeding the hardware 8 weights versus 2 weights?

Most any skinned model will use more than 8 bones. You will have to split up your mesh into several sub-meshes based on what polygons are influenced by which bones and then render them seperately.

You can get around the 8 bone hardware limit this way but it's a bit of a pain the ass.

You are correct, a resonably decent model slicer would not take that long.

However for my purposes "resonably decent" is just not good enough, making an optimal one is a little more complicated.

Thanks, that was exactly the answer I did not want to hear :-)

I'll just start crying now...

It seems the PSP hardware supports 8 weights per vertex, does anyone know the maximum number of bones supported by the hardware?

In other words, what is the maximum index supported by sceGuBoneMatrix(Index, Matrix);

Thanks in advance...

The textures we used were either exported from Photoshop using the NVidia plugin or from a DOS command line using the NVidia nvdxt.exe converter.

According to our testing the above code needs the following modifications to work correctly: void PSPconvert_dxt35(unsigned char *data, unsigned int size) { unsigned short *src = (unsigned short *) data; for(int j = 0; size >= 16; size -= 16, j++) { unsigned short converted[8]; // copy alpha convert...

I found the answer to my own question so I guess I'll post it here in case anybody else wanted to know. It seems that when using GU_TRANSFORM_2D nothing is done to any of the vertice values, including the Z value. This makes you unable to just specify a Z value like you would for a 3D object. Instea...

yeah, that would be an option but that should not be needed at all. I should be able to do this without mucking with the Z-buffer state.

I seem to be having a problem when using GU_TRANSFORM_2D to do my 2D polygons. The polygons draw just fine but they always appear behind any 3D objects that I draw. I have tried feeding the 2D polygons various Z values in the vertices. I would assume that giving it a Z value slightly higher than my ...