Roku Developer Program

[RokuMarkn]

Hm. You can't possibly really care whether OpenGL is present since there is no API to access it -- I think what you really want to know is whether the graphics performance exceeds a certain level. If some future Roku doesn't have OpenGL but has blindingly fast 2D graphics performance, you'd lump it in with the OpenGL models, correct? But what do you do if you find that OpenGL is not present?

For the memory size, are you using that to determine how many bitmaps to cache?

--Mark

[squirreltown]

"RokuMarkn" wrote:
Hm. You can't possibly really care whether OpenGL is present since there is no API to access it -- I think what you really want to know is whether the graphics performance exceeds a certain level. If some future Roku doesn't have OpenGL but has blindingly fast 2D graphics performance, you'd lump it in with the OpenGL models, correct? But what do you do if you find that OpenGL is not present?

For the memory size, are you using that to determine how many bitmaps to cache?

--Mark

Well there are only 5 platforms on which you can rotate an object an arbritary amount- the ones with OpenGL. I surely do care about that. And yes the amount of memory matters when you are dealing with big images or whether you can cross dissolve between screens or plenty of other things.

[EnTerr]

"squirreltown" wrote:
Well there are only 5 platforms on which you can rotate an object an arbritary amount- the ones with OpenGL. I surely do care about that. And yes the amount of memory matters when you are dealing with big images or whether you can cross dissolve between screens or plenty of other things.

Aha!
That is a good one, i thought about that too - the ifDraw2D.DrawRotatedObject(), which un-documented-ly rotates to non-right angles on OGLES models.

I see said method returns a Boolean - can that be used to test? I mean, ii remember on DirectFB models it does not draw the object at all if theta mod 90 <> 0 - but does it return False there? Can someone check pls... i am getting tired of testing things today.

[TheEndless]

"EnTerr" wrote:
I mean, ii remember on DirectFB models it does not draw the object at all if theta mod 90 <> 0 - but does it return False there?

It does not, but I suspect, deep down inside, you already knew that.. 😉

"RokuMarkn" wrote:
I think what you really want to know is whether the graphics performance exceeds a certain level.

Is there a way to test that?

[EnTerr]

"TheEndless" wrote:

"EnTerr" wrote:
I mean, ii remember on DirectFB models it does not draw the object at all if theta mod 90 <> 0 - but does it return False there?

It does not, but I suspect, deep down inside, you already knew that.. 😉

No, that's not it. I honestly felt it very likely to return False when it could not draw it due to "unacceptable" angle. After all, why else does it return Boolean ?

C'mon now... you think me that unwholesome?
"I am not a pessimist. Just a well-informed optimist" 8-)

"RokuMarkn" wrote:
I think what you really want to know is whether the graphics performance exceeds a certain level.

Is there a way to test that?

+1
If not .hasFeature(), can we at least get the bug fixed in DrawRotatedObject(): make it return False when it could not draw at non-90 rotation?

[TheEndless]

"EnTerr" wrote:
I honestly felt it very likely to return False when it could not draw it due to "unacceptable" angle.

That's the thing, it does draw it, just at a 0 degree rotation (doesn't even round it to the nearest 90).

[EnTerr]

"TheEndless" wrote:

"EnTerr" wrote:
I honestly felt it very likely to return False when it could not draw it due to "unacceptable" angle.

That's the thing, it does draw it, just at a 0 degree rotation (doesn't even round it to the nearest 90).

This topic popped when i was searching on our prior discussions of ifDraw rotation angles on OGL vs non-OGL models... very much a current subject, given the RSG. So anyway, i was asked to do a check how is that performing now and apparently it behaves differently now!

Specifically, the DirectFB models these days "round up" the non-multiple-of-90 angles to the closest 90 degree notch and do rotate to that angle. With a cut-off at 45 degrees (45 degrees going towards the next click for some raisin). A behavior that makes more sense.
More details at http://chat.stackoverflow.com/rooms/125 ... les-models

[Rcosta]

"RokuShawnS" wrote:
The letter definitions at the end of model numbers were posted on the General forums a while back:

R means retail
X means online
EU mean Europe (UK, Ireland)

Thus, they can be generally ignored.

C. Shawn Smith

Shawn, As I know that there is a need and desire for a knowledge base. Wouldn't a sticky to define the model #'s be advantageous for the users of this forum? I know that there is a Wikipedia page that describes the models with other information such as CPU and hardware features. The Wikipedia page does not define software features between the models such as "on-the-fly" DD+ to DD transcoding that is supported in the Ultra, but not in the Roku4. That would be a helpful matrix as a purchasing decision could be made to define specific feature sets between the models.