Check the video So, what do you guys think? is it real? or fake? It looks real, since is odd if not impossible that someone could make such an elaborate hoax, specially in low poly. But it looks fake, mostly due to the fact that the DS is so underpowered (ie: no FPU) that is hard to believe they can make it run something like that.
Hmmm thats interesting considering that it looks like something the Nintendo DS could pull off. Isn't this just connectivity with the Wii and DS? I thought this was a planed function I maybe wrong but if this does turn out to be a completely different version running on the DS itself then I can imagine it since I remember reading a while back that Mario 128 (Now Mario Galaxy) could come out for either Wii or DS.
It's fan made.. EDIT: http://www.dsfanboy.com/2007/12/11/fan-made-mario-galaxy-ds-trailer-almost-fools-us/
Hmm that world thats in the wii version looks like the one that was in the e3 demo, i recognise those scrolly arches.
Few things that seem fake about it 1. Characters are too detailed for the DS's polygon limit. 2. I swore I saw skinning on some models (DS can't do skinning in real time). 3. ! no polygon shaking ! Most DS games have some form of polygon shaking, the more detailed the more often something shakes. There was none in this video which is just not going to happen when camera moves that close or far away. FF4 does a good job of keep stuff from shaking (but few things pop up) yet say the latest Tales of for DS has massive shaking when player models are up close.
Heh, they did a pretty good job on it. It definitely wasn't official. All that math and stuff just wasn't professional at all. It didn't seem like a Nintedo ad at all. Pretty interesting though.
What exactly is the cause of polygon bending/shaking? Is it some sort of programming shortcut that decreases the resource use by decreasing the accuracy of the frame buffer or something like that? Sorta like the way emulators are able to run faster by using Dynarec and stuff like that?
Polygon shaking is caused by the fact the DS has no floating point unit. So programmers have to use fix point math which then its up to the programmers how much space of a varible is dedicated for accuracy. As well me personaly I find fix point math to be far less accurate then working in float. Ps1 had no floating point unit either (N64, PSP, PS2, .... all have FPUs).
Even though they were done by the DSP in the SCU, mostly, I'm pretty sure. Its registers might be bigger, too. And I'd like to add to subbie's post, that even though the PSX doesn't have a FPU, it does have a coprocessor for 3D math, the GTE. It uses fixed point, too. It'd be interesting to see how much 3D the PSX could crank using solely the CPU, though
I always thought Z-buffer, or lack thereof, was the cause for polygon perspective errors and juddering.
Z-Buffer is the cause for depth issues but not shaking. Z-buffers do suffer from issues called z fighting which is where you see polygons not rendering with proper depth in the distance. Jiddering issues or polygon warping is more an issue with poor accuracy on rendering. Stuff calculated with FLOPs tend to have less of this issue (to the point it's not noticable by the end user). Fix point ends up being less accurate because it's up to the programmer to determin how much accuracy he wishes to store and at some point the accuracy is lost when the value is converted to a non fix point value. It's not an easy thing to notice off hand but this trailer for tales of innocence shows it from time to time. http://youtube.com/watch?v=ycMq9krJmpg&feature=related
Trailer sadly does not show it as well as a previous one I saw (probably youtube encoding). You'll notice things like player models twiching when not animating and practicaly no camera panning. I just got to look to find a better video.
