Re: Ocean Simulator: The research continues...

[simonharvey]

Hello everybody
I am sorry for the lack of updates, unfortunately I am just getting back into blender development after some thesis and (time wasting) UT2003 gameplaying...

Anyway, right now the Texture effect compiles, and runs however I am getting all zeros for the output instead of a nice looking intensity map of what I would expect it to look like. I can also render the scene using the texture, however as above there is no visible difference, I will have to look into this further.

I have done some basic GUI updates:


You can see the addition on the 'Linked' button, this it to allow blenders texture code to source the data from a linked Object that has an OceanFFT mesh effect attached to it, This crashes blender in its current implementation, so this to will need to be looked at.

Kind Regards
Simon Harvey

[bullx]

tnx for the update, it sounds promising.
keep it up!

[gabio]

this smells like needing a new objdata structure where you can interconnect all whaht you want.
glad to see your progress.
and yes ut is addicting

[bullx]

you know we are hungry of updates, show us something new please!!

[simonharvey]

Well Hi all

It has been sometime since I have last posted to you, this has been due to masters. house moving and aerospace distractions.

Since the last post I have managed to get some of the texture code up and running to the point where I can texture a mesh, an image is shown below:



In this image the waves would be travelling to the left. Unfortunetly it dosent yet move when animating, so I will have to get that sorted out later.

The Control panels are:


For the texture panel and:



For the material buttons. I have used the texture as a normal map since a straight color map dosent look teribbly special.

Some will notice that it looks somewhat like the marble texture, the texture effect will only really comes out when (1) it is animated and (2) when an artist applies real material settings to it. As I have said in other areas of this forum, I am more of a coder and don't have the creative flare that artists tend to have.

Kind Regards
Simon Harvey

[harkyman]

Oh man, this is great. I was just wondering about your project the other day. Awesome to see it's on its way to usefulness!

[bullx]

thank you for the interesting update.

[mjordan]

Thank you Simon for your great work.
Tools like that are greatly appreciated by users. Personally, I think that Blender has a big need for this specialized tools.
This night I dreamed an L-grammar based plant generator for Blender, some specialized tools for clothing, hair and fur modeling, a complete terrain generator, a nice native integration of MH and a ocean simulator... Glad to see one of this dreams is becoming reality!
Are you releasing some patch from time to time? I'd like to experiment with your code!!!
Did you use some document that inspired you in this work?
Personally I know nothing about waves modeling, so I'm interested in some document explaining the teory behind this.

I'm a Foley/Van Dam/Feiner/Hughes student (well, not properly, just have this book on my shelf and trying to finish it), so you can imagine yourself my level of knowledge in implementing this stuff.

Hope to see more from this great work. Thank you dude.

P.S.: Don't try to play Half Life 2 please

[levon]

This night I dreamed an L-grammar based plant generator for Blender, some specialized tools for clothing, hair and fur modeling, a complete terrain generator, a nice native integration of MH and a ocean simulator...

whats wrong with python? it works just fine for MH, rip sting's fiber, the beast, and blender world forge.

[mjordan]

This night I dreamed an L-grammar based plant generator for Blender, some specialized tools for clothing, hair and fur modeling, a complete terrain generator, a nice native integration of MH and a ocean simulator...

whats wrong with python? it works just fine for MH, rip sting's fiber, the beast, and blender world forge.

Hi Levon.
Nothing against Python. I think Python is the best language for powering up an application to make it scriptable and extensible.
But sometimes I feel that python extensions are somewhat "extraneous" to Blender.
I think MH is a great add on to Blender. Human modeling is one of the most exciting things computer graphics has to offer. I feel that human modeling tools are somewhat "mandatory" for a 3D modeler, so I thought that would be nice having it completely integrated into Blender UI. I feel, sometimes, that python extensions runs like "subprograms" in Blender and I don't feel like using one tool but a miriad of them.

I don't want to cause some flame Just my poor opinion. I like how Blender is going on. When I use it (even if I cannot consider myself an expert, of course) I feel that sensation of "power". Just see the UI, for example. It feels like it can be expanded and improved with no restrictions at all. Take povanim, for examples, or YableX. They are quite nice, but they lacks those "guidelines" to make me feel that they are "part" of Blender. It would be nice if plugins can run like plugins but work like blender does. Something that the user can't recognize if he is using an extension or a Blender feature.

Sorry for Simon, this is somewhat off topic here.

[joeedh]

This night I dreamed an L-grammar based plant generator for Blender, some specialized tools for clothing, hair and fur modeling, a complete terrain generator, a nice native integration of MH and a ocean simulator...

whats wrong with python? it works just fine for MH, rip sting's fiber, the beast, and blender world forge.

Hi Levon.
Nothing against Python. I think Python is the best language for powering up an application to make it scriptable and extensible.
But sometimes I feel that python extensions are somewhat "extraneous" to Blender.
I think MH is a great add on to Blender. Human modeling is one of the most exciting things computer graphics has to offer. I feel that human modeling tools are somewhat "mandatory" for a 3D modeler, so I thought that would be nice having it completely integrated into Blender UI. I feel, sometimes, that python extensions runs like "subprograms" in Blender and I don't feel like using one tool but a miriad of them.

I don't want to cause some flame Just my poor opinion. I like how Blender is going on. When I use it (even if I cannot consider myself an expert, of course) I feel that sensation of "power". Just see the UI, for example. It feels like it can be expanded and improved with no restrictions at all. Take povanim, for examples, or YableX. They are quite nice, but they lacks those "guidelines" to make me feel that they are "part" of Blender. It would be nice if plugins can run like plugins but work like blender does. Something that the user can't recognize if he is using an extension or a Blender feature.

Sorry for Simon, this is somewhat off topic here.

Unfortunately, the python API is not really suited to work that way. It is possible, sortof, but it takes a lot of trickery with the Registry module, the OnSave scriptlink, scene redraw scriptlinks, etc.

I actually put a lot of work into making a system where my python scripts could appear as seamlessly integrated into Blender as possible, but that never really went anywhere. My only script that used it ended up implementing functionality that was already inside of Blender . It was a morphing script, geared more for animations then modelling.

joeedh

[simonharvey]

Hi everybody,

Well, I have moved my development computer from university to my home which dosent have an internet connection so alas I cannot send any pictures as updates.

I have been thinking about blenders effect code, both its geometry effects (i.e. the 'effects' tab in Object buttons) and also about application specific textures (in texture buttons) - and I have decided that it would not be a good thing to have an application specific OceanFFT texture in the texture type popup menu, since all textures mentioned are quite general making this quite a break from tradition.

What I think should be implemented is some sort of a graph system where people create 'Graph Effects' whos output can be used to modify the geometry of objects as well as can be outputted to textures so as to allow for fine texture resolution for things such as bumpmaps and colour detail.

This would require reclycling the code from the OOPS schematic area window to generate the graph and a module/block framework so that blocks (repersenting a function) could be linked together at the users will to accomplish an effect to meet the users desire.

This would allow for people to do complicated simulations and have it consistantly interface with the rest of blender.


For example (using my module):
The user would press the spacebar which would bring up an effects toolbox, clicking Add->Simulation->Ocean FFT he adds a OceanFFT block to the schematic, he then adds a mesh displacer by clicking Add->Modifier->Vertex Displacer and then by clicking on the output node of the OceanFFT node he creates a link from its output to the input node of the vertex displacer.

This would be leaps and bounds better that what we have now since what we have dosent really allow the user any real control over the order and the execution of how the effect is to be applied.

I realise the the above example sounds rather vapourwareish however as you can see from my previous posts when I have an idea I go out to implement it and so I should not be considered as some users who attend these forums with all of the ideas but who never really go on to implement them in code. How ever I will not start investing the 10's of hours of coding time when:
1. This willl never make it to tophuhu (or any other development blender tree)
2. This is not really what the core blender development team is wanting.
3. There is another project that is currently active that is promising to deliver the same thing.

This is because I have other commitments and am partially doing this out of charity and also 'giving' i.e. the code that I write with the joy I have in creating has the oppatunity in helping others as well.

So before I go further I would like input from other developers who do have a say on decision making matters since there is no 'Effects and simulation' on the official projects list whos names I can get in order to talk to them about this.

Kind Regards
Simon Harvey

[levon]

i will forward this message to the BF-commiters mailing list. more dev type people will read it there then in these forums.

you can subscribe to the mailing list at http://projects.blender.org/mailman/lis ... committers

[levon]

this is tons reply

Message: 4
Date: Mon, 28 Feb 2005 17:28:49 +0100
From: Ton Roosendaal <ton@blender.org>
Subject: Re: [Bf-committers] Ocean Simulator
To: bf-blender developers <bf-committers@projects.blender.org>
Message-ID: <D37C513E-89A5-11D9-A92C-000D936BAA7A@blender.org>
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed

Hi,

The Oops system was not meant to become a 'node based' (object
oriented) function editor. The only thing possible to add here is what
I did for the logic editor, a GUI based way to edit data relations (the
buttons that allow drawing connections).

In very ancient days (early nineties) an experiment was done in NeoGeo
with a real "oops", a first version was built to create a calculator
(really!). However, back then the method was rejected, because even a
simple setup already easily results in an confusing and hard to
interpret visual result.

The popularity of such 'nodal' or 'oops' editors is not because it's
such a great GUI concept, but because it's such a nice simple thing for
object-oriented coders to maintain.
A reason why it shouldn't be rejected for Blender is that people got
used to it... it's not always bad to make coders life easier, eh!

-Ton-

[mjordan]

What does it mean? The ocean simulator is dead?