Hello,

I have seen long time ago that one can render only a part of an image
(you put some command in the command line). This is really cool to
render some animation, for example:

you have an image with a door, when you click on the door (this is a
program), it open. But just the door move not the whole scene, so you
render only this part of the picture -> gain of render time and gain of
animation size. But suppose that you have an image 320*240, you render a
just a part of it with the size 100*200, the image you ontain is still a
320*240 but with black color around the part of the image: you don't
obtain a 100*200 image. Ok with PSP or corel photo paint it is easy to
get only the relevant part of the image but this take much time because
you must work on each image before construct the animation. It is
possible to render and GET only the relevant part of the image?

Thanks

Fabian.

Post a reply to this message

No.  And since the partial images get put on the first row I don't see any
feasible way of simply chroma-keying the black backgrounds out, which if you
had any black in the scene itself would also get blanked out anyway.  The
non-descript background might be clear if Alpha channeled too but once again
there's the problem with the partial image showing up in the wrong place, so
I see no way as is to compile them together with whole images.
Would need a different method of image production, I think might be the only
potential answer.  Seems like the idea about a whole image remaining for
each subsequent partial render would go nicely with this concept.

Bob

Fabian Brau <fab### [at] umhacbe> wrote in message
news:3819A015.128291F5@umh.ac.be...
> Hello,
>
> I have seen long time ago that one can render only a part of an image
> (you put some command in the command line). This is really cool to
> render some animation, for example:
>
> you have an image with a door, when you click on the door (this is a
> program), it open. But just the door move not the whole scene, so you
> render only this part of the picture -> gain of render time and gain of
> animation size. But suppose that you have an image 320*240, you render a
> just a part of it with the size 100*200, the image you ontain is still a
> 320*240 but with black color around the part of the image: you don't
> obtain a 100*200 image. Ok with PSP or corel photo paint it is easy to
> get only the relevant part of the image but this take much time because
> you must work on each image before construct the animation. It is
> possible to render and GET only the relevant part of the image?
>
> Thanks
>
> Fabian.

Post a reply to this message

Hello omniVERSE,

this is easy to do but this take time. Here is the method:
1) I render completely the first image. I choose the area which I will
render, say the door. I put this into the command line.
2) I begin the animation rendering. So I get a given number (depend on
animation) of partial image which are on the first row, with black color
around.
3) I take each image into corel photo paint. With a special tool I
select the non-black area. I copy and paste into a new document and save
it into bmp (for example). And I have myst first image with good
dimension.
4) I do it for each image (a nightmare :)).
5) I compute the animation with the images.
6) In my program (a game) I load the first image (a complete image).
When I click, say on the door, I load the video which I place at the
right position on my image. The complete image don't disappear (I choose
a correct update method for updating the screen) and the video play at
the correct place. This work perfectly well (I have tested) but take
time to do the animation :).

If someone which know well the code (perhaps Nathan) could look at it
(when one render a piece of an image, the output image is only this
piece without black color around) this could be cool :)

Fabian.

omniVERSE wrote:
> 
> No.  And since the partial images get put on the first row I don't see any
> feasible way of simply chroma-keying the black backgrounds out, which if you
> had any black in the scene itself would also get blanked out anyway.  The
> non-descript background might be clear if Alpha channeled too but once again
> there's the problem with the partial image showing up in the wrong place, so
> I see no way as is to compile them together with whole images.
> Would need a different method of image production, I think might be the only
> potential answer.  Seems like the idea about a whole image remaining for
> each subsequent partial render would go nicely with this concept.
> 
> Bob
> 
> Fabian Brau <fab### [at] umhacbe> wrote in message
> news:3819A015.128291F5@umh.ac.be...
> > Hello,
> >
> > I have seen long time ago that one can render only a part of an image
> > (you put some command in the command line). This is really cool to
> > render some animation, for example:
> >
> > you have an image with a door, when you click on the door (this is a
> > program), it open. But just the door move not the whole scene, so you
> > render only this part of the picture -> gain of render time and gain of
> > animation size. But suppose that you have an image 320*240, you render a
> > just a part of it with the size 100*200, the image you ontain is still a
> > 320*240 but with black color around the part of the image: you don't
> > obtain a 100*200 image. Ok with PSP or corel photo paint it is easy to
> > get only the relevant part of the image but this take much time because
> > you must work on each image before construct the animation. It is
> > possible to render and GET only the relevant part of the image?
> >
> > Thanks
> >
> > Fabian.

Post a reply to this message

Fabian Brau wrote:
> 
> Hello omniVERSE,
> 
> this is easy to do but this take time. Here is the method:
> 1) I render completely the first image. I choose the area which I will
> render, say the door. I put this into the command line.
> 2) I begin the animation rendering. So I get a given number (depend on
> animation) of partial image which are on the first row, with black color
> around.
> 3) I take each image into corel photo paint. With a special tool I
> select the non-black area. I copy and paste into a new document and save
> it into bmp (for example). And I have myst first image with good
> dimension.
> 4) I do it for each image (a nightmare :)).
> 5) I compute the animation with the images.
> 6) In my program (a game) I load the first image (a complete image).
> When I click, say on the door, I load the video which I place at the
> right position on my image. The complete image don't disappear (I choose
> a correct update method for updating the screen) and the video play at
> the correct place. This work perfectly well (I have tested) but take
> time to do the animation :).
> 
> If someone which know well the code (perhaps Nathan) could look at it
> (when one render a piece of an image, the output image is only this
> piece without black color around) this could be cool :)
> 
> Fabian.

Why not simply use multiple camera's and zoom in on the area you want for
a particular animation sequence ? It is not that difficult to position a
camera.

-- 
Ken Tyler -  1100+ Povray, Graphics, 3D Rendering, and Raytracing Links:
http://home.pacbell.net/tylereng/index.html http://www.povray.org/links/

Post a reply to this message

Fabian Brau wrote:
> 
> Hello omniVERSE,
> 
> this is easy to do but this take time. Here is the method:
> 1) I render completely the first image. I choose the area which I will
> render, say the door. I put this into the command line.
> 2) I begin the animation rendering. So I get a given number (depend on
> animation) of partial image which are on the first row, with black color
> around.
> 3) I take each image into corel photo paint. With a special tool I
> select the non-black area. I copy and paste into a new document and save
> it into bmp (for example). And I have myst first image with good
> dimension.
> 4) I do it for each image (a nightmare :)).
> 5) I compute the animation with the images.
> 6) In my program (a game) I load the first image (a complete image).
> When I click, say on the door, I load the video which I place at the
> right position on my image. The complete image don't disappear (I choose
> a correct update method for updating the screen) and the video play at
> the correct place. This work perfectly well (I have tested) but take
> time to do the animation :).
> 
> If someone which know well the code (perhaps Nathan) could look at it
> (when one render a piece of an image, the output image is only this
> piece without black color around) this could be cool :)
> 
> Fabian.
> 
> omniVERSE wrote:
> >

I understand you're using this for a game. Do you program that? If so, it
shouldn't be hard to do what you want programmatically. 
I don't know exactly what you need, but perhaps this may help:
I'm using the same technique you describe for a project I'm working on. For that
I had to make several simple utilities. One of those takes a series of images
and cuts out the relevant part. In my case I need all the cut-outs on a strip so
it pastes them into one image, but it'd be easy to make it so it saves them as
seperate files. In fact I have another utility which will take a strip of
images, cuts it into frames and saves those seperately. If you think this could
help let me know. 

Ciao!

Remco

Post a reply to this message

Well, I think that it is really 'yes'. But tricky.

When I checked, many many moons ago, if you output .TGA files, then POV-Ray did
the right thing. The bits stored start on that first line, but the image header
itself should have an offset. The TGA format has X and Y in addition to Width
and Height. I used to take advantage of this for sprites and compositing.

Anyway,  given image tools that do the proper thing, and recognize the offsets
in TGA headers, then you could use that to save time on renderings... (TGA is
not hard. You can program that yourself...)

EXCEPT...

If you have an animation, and are raytracing, it is very possible that other
portions of the image might change. Reflections, shadows, etc. This would then
throw things off for the look and quality of your animation.

omniVERSE wrote:

> No.  And since the partial images get put on the first row I don't see any
> feasible way of simply chroma-keying the black backgrounds out, which if you
> had any black in the scene itself would also get blanked out anyway.

[SNIP]

> Fabian Brau <fab### [at] umhacbe> wrote in message
> news:3819A015.128291F5@umh.ac.be...
> > Hello,
> >
> > I have seen long time ago that one can render only a part of an image
> > (you put some command in the command line). This is really cool to
> > render some animation, for example:

Post a reply to this message

Hello Remco,

yes, I do all in my game :), I program in Java (not quick language, but
for a point-and-click game it is enough). But i never work with image
file before! If you could give me an utility which can cut a given part
(which I choose), al ways the same, of several (hundreds) images, this
could be great for me :). 
I am using win 98.

Thank you very much!!

Fabian.


> I understand you're using this for a game. Do you program that? If so, it
> shouldn't be hard to do what you want programmatically.
> I don't know exactly what you need, but perhaps this may help:
> I'm using the same technique you describe for a project I'm working on. For that
> I had to make several simple utilities. One of those takes a series of images
> and cuts out the relevant part. In my case I need all the cut-outs on a strip so
> it pastes them into one image, but it'd be easy to make it so it saves them as
> seperate files. In fact I have another utility which will take a strip of
> images, cuts it into frames and saves those seperately. If you think this could
> help let me know.
> 
> Ciao!
> 
> Remco

Post a reply to this message

On Wed, 03 Nov 1999 10:34:49 +0100, Fabian Brau wrote:
>If you could give me an utility which can cut a given part
>(which I choose), al ways the same, of several (hundreds) images, this
>could be great for me :). 
>I am using win 98.

The netpbm utilities can do this, among other things.  For a win32 version,
go to ftp://ftp.uci.agh.edu.pl/pub/tex/systems/win32/web2c/ .  These are
command-line utilities, so they're perfectly suited to scripting with a 
batch file or your favorite other scripting language.  The documentation is
in Unix man format, but you can find an HTML version at 
http://www.ctyme.com/linuxdoc.htm .  The programs you're likely to be 
interested in are tgatoppm, pnmcut, and ppmtotga.  All are in section 1 of
the manpages.

-- 
These are my opinions.  I do NOT speak for the POV-Team.
The superpatch: http://www2.fwi.com/~parkerr/superpatch/
My other stuff: http://www2.fwi.com/~parkerr/traces.html

Post a reply to this message

Thanks Ron,

but bmptoppm don't work for me :(

Fabian.

Ron Parker wrote:
> 
> On Wed, 03 Nov 1999 10:34:49 +0100, Fabian Brau wrote:
> >If you could give me an utility which can cut a given part
> >(which I choose), al ways the same, of several (hundreds) images, this
> >could be great for me :).
> >I am using win 98.
> 
> The netpbm utilities can do this, among other things.  For a win32 version,
> go to ftp://ftp.uci.agh.edu.pl/pub/tex/systems/win32/web2c/ .  These are
> command-line utilities, so they're perfectly suited to scripting with a
> batch file or your favorite other scripting language.  The documentation is
> in Unix man format, but you can find an HTML version at
> http://www.ctyme.com/linuxdoc.htm .  The programs you're likely to be
> interested in are tgatoppm, pnmcut, and ppmtotga.  All are in section 1 of
> the manpages.
> 
> --
> These are my opinions.  I do NOT speak for the POV-Team.
> The superpatch: http://www2.fwi.com/~parkerr/superpatch/
> My other stuff: http://www2.fwi.com/~parkerr/traces.html

Post a reply to this message

And tgatoppm seems work but no output file?? I have search in my hard
disk, nothing!!

Fabian.

Fabian Brau wrote:
> 
> Thanks Ron,
> 
> but bmptoppm don't work for me :(
> 
> Fabian.
> 
> Ron Parker wrote:
> >
> > On Wed, 03 Nov 1999 10:34:49 +0100, Fabian Brau wrote:
> > >If you could give me an utility which can cut a given part
> > >(which I choose), al ways the same, of several (hundreds) images, this
> > >could be great for me :).
> > >I am using win 98.
> >
> > The netpbm utilities can do this, among other things.  For a win32 version,
> > go to ftp://ftp.uci.agh.edu.pl/pub/tex/systems/win32/web2c/ .  These are
> > command-line utilities, so they're perfectly suited to scripting with a
> > batch file or your favorite other scripting language.  The documentation is
> > in Unix man format, but you can find an HTML version at
> > http://www.ctyme.com/linuxdoc.htm .  The programs you're likely to be
> > interested in are tgatoppm, pnmcut, and ppmtotga.  All are in section 1 of
> > the manpages.
> >
> > --
> > These are my opinions.  I do NOT speak for the POV-Team.
> > The superpatch: http://www2.fwi.com/~parkerr/superpatch/
> > My other stuff: http://www2.fwi.com/~parkerr/traces.html

Post a reply to this message