Hi Guys
After reading through a couple of samples in the PSPSDK, and modifying the Cube example to be a pool table, I need to get a better understanding about 3d graphics, as i last study this about 8 years ago.
Anyways since various people have done 3d graphics, I'm wondering if anybody can give me a recommendation as to which books to read.
I think i really need to go back to basics, obviously since i was able to modify the cude example , I was able to work out the modeling in 3d dimensional space, but that being said, I really need to understand the pricipals again (normals, planes, intersection, projection... as so forth)
Note clipping seems to be a big issue, as when I get to close to the table, various polygon disappear making it look completely crap.
If there are any significant 3d examples out there, aside from quake, that would be great to.
Hope you guys can help.
3d graphic books recommendations
-
white rabbit
- Posts: 60
- Joined: Wed Jul 06, 2005 7:03 pm
3D Game Engine Design by David Eberly is very good and covers pretty much everything - by the end of it you could write your own software graphics & 3d engine.
He uses opengl in the source code that comes with it, and his own software engine.
There's a few bits in there that aren't correct, or not done fully, but it will cover more than you need in a decent depth.
The OpenGL red book is also good, but it is all based on the OpenGL API.
He uses opengl in the source code that comes with it, and his own software engine.
There's a few bits in there that aren't correct, or not done fully, but it will cover more than you need in a decent depth.
The OpenGL red book is also good, but it is all based on the OpenGL API.
yes that is a good book by david eberly and highly
focuses on opengl (which youll need to learn beforehand :P
and psp gpu is very similar in functions to opengl so
learning opengl with 3d engine book should get you
probably ahead of most people concerning learning
gpu and 3d on psp ....and might i add READ THE SAMPLES
:P they are the best so far for gpu learning
focuses on opengl (which youll need to learn beforehand :P
and psp gpu is very similar in functions to opengl so
learning opengl with 3d engine book should get you
probably ahead of most people concerning learning
gpu and 3d on psp ....and might i add READ THE SAMPLES
:P they are the best so far for gpu learning
10011011 00101010 11010111 10001001 10111010
another vote for the red book, the fact that it's about OpenGL only makes a difference when you look at the code samples (although the PSP's gl is very similar anyway). It's a really good source for the theory behind everything, the Graphics Gems books are good too, the older ones I & II, have more basic theory for rendering shapes, again interesting and a good base to build from
Hi Guys
Thanks for the recommendation, I have borrowed a couple of books from some guys I work with that have just finished uni.
Anyways I'm going to go down and try to get the redbook, since I can always write off the books for work anyways.
I tried some of the online sites, and as good as they were, they just don;t go into enough detail in the maths... and leaving me more confused. I still have a basic understanding, so I'll outline it as to developing on PSP is hugely different from my past experence.
*) Objects are defined with in there own world space
*) All objects are then translated into a single world space
*) World space in then translated into camera view.
This all seem pretty simple, but my main question arises from messing around with the PSP examples. I'm assumes the vertixes I would pass off to the 'gu' would be for the single world space, as obvious the gu would know nothing about each object, just the vertexies.
Using just a single world space (yes I'm lazy in not knowing how to translate all objects into a single world space) is cool I can make objects and render them well enough. Problem arise when the camera gets too close. Obvious a vertex is behind the camera and therefore any triangle that has a vertex behind the camera disappears. The maths I'm looking for is to determine when a vertex gets behind the camera, and calculating the new polygon (I'm not worried about the triangles as I wrote a piece of code that breaks up a polygon into triangles).
So yes I need the maths, although the hardware does the majority of the work, realistic clipping is impossible if I don't know the maths.
Thanks for the recommendation, I have borrowed a couple of books from some guys I work with that have just finished uni.
Anyways I'm going to go down and try to get the redbook, since I can always write off the books for work anyways.
I tried some of the online sites, and as good as they were, they just don;t go into enough detail in the maths... and leaving me more confused. I still have a basic understanding, so I'll outline it as to developing on PSP is hugely different from my past experence.
*) Objects are defined with in there own world space
*) All objects are then translated into a single world space
*) World space in then translated into camera view.
This all seem pretty simple, but my main question arises from messing around with the PSP examples. I'm assumes the vertixes I would pass off to the 'gu' would be for the single world space, as obvious the gu would know nothing about each object, just the vertexies.
Using just a single world space (yes I'm lazy in not knowing how to translate all objects into a single world space) is cool I can make objects and render them well enough. Problem arise when the camera gets too close. Obvious a vertex is behind the camera and therefore any triangle that has a vertex behind the camera disappears. The maths I'm looking for is to determine when a vertex gets behind the camera, and calculating the new polygon (I'm not worried about the triangles as I wrote a piece of code that breaks up a polygon into triangles).
So yes I need the maths, although the hardware does the majority of the work, realistic clipping is impossible if I don't know the maths.
-
white rabbit
- Posts: 60
- Joined: Wed Jul 06, 2005 7:03 pm
I remeber some early discussion about clipping close to the camera not being good. Not sure what the outcome was (it was about 6 months ago), other than to use small triangles in the mesh.
As for the maths, the David Eberly book has a lot of proofs and maths. I'm not sure where in the world you're at, but I'd say it's end of 1st year maths degree in the UK: It's all pretty basic stuff, but you have to know your stuff to get to grips with it. He also misses the point or glosses over things he doesn't fully understand sometimes.
If you are serious about 3D and you can write off the expense, I'd say get it, it covers 99% of stuff and to a good level of detail too.
Edit to say - it also has all the maths to detect when to and how to clip triangles into component triangles for the situation you talk about - I will get my PA to OCR that bit for you if you want? PM me if you're interested.
As for the maths, the David Eberly book has a lot of proofs and maths. I'm not sure where in the world you're at, but I'd say it's end of 1st year maths degree in the UK: It's all pretty basic stuff, but you have to know your stuff to get to grips with it. He also misses the point or glosses over things he doesn't fully understand sometimes.
If you are serious about 3D and you can write off the expense, I'd say get it, it covers 99% of stuff and to a good level of detail too.
Edit to say - it also has all the maths to detect when to and how to clip triangles into component triangles for the situation you talk about - I will get my PA to OCR that bit for you if you want? PM me if you're interested.
Sent you a pm. It's actualy pretty easy to figure out by checking out examples on rejecting object clusters by a view frustum. It's the same concept except you then do a line test against each plane and clip the lines at the point of intersection with the view frustum.ipsp wrote:Hi Guys
Thanks for the recommendation, I have borrowed a couple of books from some guys I work with that have just finished uni.
Anyways I'm going to go down and try to get the redbook, since I can always write off the books for work anyways.
I tried some of the online sites, and as good as they were, they just don;t go into enough detail in the maths... and leaving me more confused. I still have a basic understanding, so I'll outline it as to developing on PSP is hugely different from my past experence.
*) Objects are defined with in there own world space
*) All objects are then translated into a single world space
*) World space in then translated into camera view.
This all seem pretty simple, but my main question arises from messing around with the PSP examples. I'm assumes the vertixes I would pass off to the 'gu' would be for the single world space, as obvious the gu would know nothing about each object, just the vertexies.
Using just a single world space (yes I'm lazy in not knowing how to translate all objects into a single world space) is cool I can make objects and render them well enough. Problem arise when the camera gets too close. Obvious a vertex is behind the camera and therefore any triangle that has a vertex behind the camera disappears. The maths I'm looking for is to determine when a vertex gets behind the camera, and calculating the new polygon (I'm not worried about the triangles as I wrote a piece of code that breaks up a polygon into triangles).
So yes I need the maths, although the hardware does the majority of the work, realistic clipping is impossible if I don't know the maths.
http://www.amazon.ca/exec/obidos/ASIN/0 ... 55-0309914
All the watt/watt/policarpo books are great for the theoretical background behind computer imaging.
Jim
All the watt/watt/policarpo books are great for the theoretical background behind computer imaging.
Jim