[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Public WebGL] invalid framebuffer operation in Firefox



On Thu, Dec 23, 2010 at 5:22 PM, Steve Baker <steve@sjbaker.org> wrote:
> On 12/23/2010 12:03 PM, Vladimir Vukicevic wrote:
>> ----- Original Message -----
>>
>>> after updating DirectX9 (Windows XP) I get "invalid framebuffer
>>> operation" error in the "2D Game of Life" application
>>> http://www.ibiblio.org/e-notes/webgl/ca/life.html
>>> is render to LUMINANCE FBO prohibited now? render to RGBA FBO works
>>> fine
>>> in the "Barkley's model"
>>> http://www.ibiblio.org/e-notes/webgl/waves/barkley_rnd.html
>>>
>> Correct, only the image formats listed in table 4.5 of the GL ES 2.0 spec are color-renderable:  RGBA4, RGB5_A1, RGB565.  It's an error if we weren't enforcing this with the straight GL (non-ANGLE) backend...
>>
> I understand that some cellphones can't do better - but this is an
> insanely crippling limitation.  Doing shadow rendering without
> render-to-depth texture, without floating point destinations, and
> without any kind of 16 bit format...argh!

These limitations are only when you're rendering to an FBO with a
renderbuffer as the color attachment. If you have an RGBA texture as
the color attachment you ought to be able to portably get 8 bits per
channel. (I think...)

We are still investigating real OpenGL ES 2.0 implementations with
GL_OES_texture_float support to see whether they tend to support
rendering to floating-point textures.

-Ken

> I believe my current shadow renderer works with 4/4/4/4 but I'm working
> in a very small view volume (basically a single room).  If anyone wants
> to render an outdoor scene with shadows - it's going to be an absolute
> bitch!
>
> You'd have to use cascading shadow maps - where the precision of the
> depth map got progressively worse further from the camera - but because
> precision is a big issue with shadowmaps, it'll be really tough to get
> artists to maintain separation rules to prevent the shadows from
> flimmering...and the resulting fragment shader to apply the shadows
> would be horrendous.
>
> We really need better options on desktop systems because the qualitative
> clobbering we'll take from cellphone limitations is a pain for people
> who don't particularly care about that market.
>
>  -- Steve
>
> -----------------------------------------------------------
> You are currently subscribed to public_webgl@khronos.org.
> To unsubscribe, send an email to majordomo@khronos.org with
> the following command in the body of your email:
>
>

-----------------------------------------------------------
You are currently subscribed to public_webgl@khronos.org.
To unsubscribe, send an email to majordomo@khronos.org with
the following command in the body of your email: