What software do you all use to render the 3D images?

Thanks

Thanks,

James

The rolling ball loop has been rendered as movie plus mask by our artists and we converted it to 192 transparent pngs offline. We then created 3 atlases of 64 frames each loaded as textures of the size 1024×1024. Using thumbnailimageAtTime was just an idea you might follow to do this online – directly in your app.

To apply masks with OpenGL ES 1.1 (this applies to cocos2D) you can use FBOs to change the alpha channel in a second step after you had rendered to them. Cocos2d supports this with CCRenderTexture directly.

Cheers,
Heiko

Thanks very much for your reply. You summarised what I would like to do exactly.

So if I understand correctly, to achieve your affect of having the water flow in front of the sphere, you took random frames of the playing movie (using thumbnailImageAtTime:timeOption) which returned a UIImage; which you then convert to a mask and put in a UIView. You then convert the UIView to a EAGLView? This is a very useful tip!

We are using Cocos2D because we also have sprites to draw in another layer. Do you add the mask layer as a node to the sprite layer? How do you blend the two layers to mask the sprites? I am sorry for the rudimentary questions – I am new to OpenGL programming.

Thank you again for your invaluable help.

Best regards,
Ash

I am glad I could inspire you with my blog post about video composition. In general, you can place any UIView on top of the rendering view of a MPMoviePlayerController as you can read on different online tutorials. If your particular UIView contains transparent areas the content of the movie shines through.

What you want to achieve is the opposite, though. Parts of your movie should be rendered transparently based on its content (chroma keying). I share your opinion that this is currently not possible. Neither can you place a UIView behind the movie (until SDK 3.2) nor can do you have live access to the frames. What you could do though, is storing the individual frames as sprites if this fits in memory (you can render thumbnails of random frames to UIImages, since 3.2).

Hope this helps.

Cheers,
Heiko

Thank you very much!

The MacDeveloper Magazine (German) will publish an article about the production of screencasts for the iPhone I wrote last week. Aroud the 27th I complement that text with a short blog post to provide links to the tools I wrote about. Hopefully, this will help you.

Please let me know which further details -apart from code- you need to understand the liquid effect.

I’m impressed with the effect but understand that you can’t open source the app. I’m really interested to replicate a similar effect and would appreciate further details as far as you are comfortable to release.

I’m also interested in the screencast production. It looks really good!

Cheers

First of all I want to apologize because of my bad English. Thank you so much for this amazing tool, it is being very useful for me.
I have the same problem than Robert. It works but wthout texture, and I am working with the banana example and GLGravity project to be sure that I am not making mistakes, and nothing. Could you help me to solve that please? Thank you Heiko.

Could you possibly share how you captured teh video and how did you rotate the iPhone simulator so smoothly (I can only rotate it 90 degrees at a time…).

Thanks!

unfortunately, the script currently neither supports materials nor separate groups of faces. Both would be required to colorize your models. Sorry for that.

If you come up with a workaround or a patched version of the script, feel free to post it here!

Peter Friese and Heiko Behrens have been blogging (check this and this links) about using MDD techniques to efficiently build native iPhone applications with the help of domain-specific languages….

I was just writing back again in regards to my last post, to give you more specifics as to what would be useful to include in the script.

The .obj file that was produced for my 3D model contains color shading information such as the one below since I did not use an actual texture:

g pSphere1
usemtl initialShadingGroup
f 1//1 2//2 12//3 11//4
f 2//5 3//6 13//7 12//8
f 3//9 4//10 14//11 13//12
f 4//13 5//14 15//15 14//16
f 5//17 6//18 16//19 15//20
……

So I was wondering if there would be any way to parse these values into the equivalent in OpenGL ES that the function glColorPointer(..) would be able to take.

Thanks,

Carlos.

I do have one question though. If an object (3D object that is) was painted with some colors instead of using textures, is there any way to obtain the color data in the obj file and then be parsed to the C style arrays like your script does it?
The color array that I am talking about would be used in the function: glColorPointer(…) of the OpenGL ES API’s.

Please let me know either by email or directly in this blog.

Thanks in advance.

great to see that someone even tries to use this for VS! Unfortunately, I cannot assist you without further information. Did you try to narrow this down already? How does a simple header file (e.g. the examples) behave? Are you able to visualize anything loaded via glVertexPointer? How do these arrays differ from those generated by obj2opengl?

Last night at about 4AM, I think I figured this ‘bug’ out. Turns out it is not a ‘bug’ at all, and has nothing to do with the obj2opengl script.

If you manipulate a mesh in Silo and try to export it as a .obj, Silo will export it as Quads, not Triangles. Before you export your model from Silo, you must select it and choose “Modify” -> “Triangulate” or [Control + Shift + T]. Once you do this, you can export your model from Silo, run obj2opengl and the iphone will render the .h properly.

Thanks to Jeff LaMarche for confirming the solution. Hope this post saves someone a couple days.

Thanks again Heiko, for such a great little script.

best wishes to all,
ty

Thanks for getting back to me. Great little utility! I am really interested in figuring this out and appreciate you taking the time to look at it. I posted both a cube and sphere here: http://gist.github.com/333101

Both models have the same visual artifact.

cube.obj: http://gist.github.com/raw/333101/b24c888179d871c4355cbd54a21ad616b4afa3d1/cube.obj

sphere.obj: http://gist.github.com/raw/333101/e2b4dc69df4f25bd0196817fc3f5d7ee5aec2b9a/sphere.obj

I have tried dozens of models, all have the same issue. For what it is worth, Blender can’t open Silo .obj files, so perhaps Silo went off standard.

Anyway, thanks again…. Wow, “Blender makes my head heart.” Apparently Blender makes me funny too.

please provide the .obj file of the most simple version with errors you are converting from (e.g. a cube). I think, the format of the .obj file is slighty different from what the script expects.

I seem to be experiencing some problems converting Silo 2 .obj files to .h files with obj2opengl. The conversion appears to work and I am able to load the .h file on an iphone using GLGravity. But the iphone only seems to render the model partially, in a triangle pattern as indicated here: http://imgur.com/fGN8f.jpg

I found another script, for blender, which enabled me to export solid models:
http://iphonedevelopment.blogspot.com/2009/07/improved-blender-export.html

But I would prefer to model in Silo. Blender makes my head heart. Any idea what’s going on here?

first of all , thanks for your presentation, I like it !!
actually , i’m trying to use xtext in my project , so i need to see some examples first to get ready !
So , can you make the chess example available for us?
thank you for answering

Best regards

What is “Center” that script saying ?
Center:

I have two models and i have no success draw it correctly in space.
I’ve tried glTranslatef to “Center” and glScalef by “Scale by” before rendering – but my models located in space at the wrong coordinates.

How can I draw 2 models on their coordinates same as in 3dMax?

thanks,

Thanks for answer. Do you have a plan make script working with some textures ?

and my code to draw it:

glPushMatrix();
glBindTexture(GL_TEXTURE_2D, textures[13]);
glVertexPointer(3, GL_FLOAT, 0, cubeVerts);
glTexCoordPointer(2, GL_FLOAT, 0, cubeTexCoords);
glDrawArrays(GL_TRIANGLES, 0, cubeNumVerts);
glPopMatrix();

I have been having problems getting the UV texture coordinates to work.
Here is a screen shot of what it looks like in Maya: http://imgur.com/z2Ica.png
and what it looks like on the iPhone: http://imgur.com/A1alQ.png

Does anyone know how this could be fixed?
Thanks in advance for any help

But what if the object 2 or more textures?
thanks.

I made a small change to this so that it interleaves the vertex data in one contiguous array, according to http://developer.apple.com/iphone/library/documentation/3DDrawing/Conceptual/OpenGLES_ProgrammingGuide/Performance/Performance.html

obj2opengl.pl:
http://docs.google.com/leaf?id=0B3MMs33O0RexZDUzZDU1ZjItMjRlZi00OGQwLThkOTktN2RkYWM4MzNiMDM3&hl=en

Example output (cube.h):
http://docs.google.com/leaf?id=0B3MMs33O0RexMGY0YTI0NjAtY2E1MC00MGYwLWIxM2EtMGRmYzExMzE0ODUz&hl=en

Example draw (note the required stride constant):

glVertexPointer(3, GL_FLOAT, cubeStride, cubeVerts);

glNormalPointer(GL_FLOAT, cubeStride, cubeNormals);

glDrawArrays(GL_TRIANGLES, 0, cubeNumVerts);

Http://www.hundertliegestuetze.de

should be:
print OUTFILE “glTexCoordPointer(2, GL_FLOAT, 0, “.$object.”TexCoords);\n”
if $numTexture > 0;

It stops glTexCoordPointer appearing in comment.

Nice script though!