I thought of this while trying to think of a good way of making an Animatic/Leica reel in Blender.
Before Blender, I used Lightwave, so I’m one of those people you won’t hear complaining about Blender’s interface (to me, they’re similar in the sense that they’re built for speed[/i]). Lightwave is considered “defunct” nowadays, sure, but there are still things we can learn from it.
In particular, LW had this little neat OGL rendering mode. Just like Blender can switch between internal and Yafray, LW could choose the built-in renderer, a plug-in renderer, and the OGL renderer. This was, of course, ideal for previsualization…
As it is, I can speed up animatics with a number of tricks (the light and mat fields in “Render Layers”, for instance) and lo-res meshes or cut-outs from designa of course, but nothing would match the speed of a plain-Gourad OGL renderer for animation. Wish Ctrl-F3 gave me animated output.
Thoughs? Suggestions? am I missing something absurdly obvious, or some feature I could take advantage of?
In the header of any 3D window, completely to the right, past the padlock, there is a button which if clicked will give you an OpenGL stil; if Ctrl+LMB are used you get an animation.
The framing is a bit maddening though: the image is rendered to the settings in the Scene buttons (10) and so, for it to reflect what can be seen in the 3D view this one must be set at the same ratio. Quite doable in a Camera view, although once I got it right I sure recorded a Screen just as to be able to recall it without having to fiddle with the size of the 3D view anymore.
I am not so sure that the business model of Blender aims to World domination
So there’s the gap in our strategy slaps forehead Dang! I knew we were doing something wrong. That’s why Autodesk is so far ahead in world domination.
Heh. Seriously, though, I must have seen that button a billion times (I believed it was synonymous with Ctrl-F3), so I was looking for some key combination or something.
It still may not be such a bad idea to include this in the rendering engine list, specially considering GLSL shaders are underway. It’s really wonderful thing to have, and IMHO it deserves a bit more exposure.
If the nebula script refuses to run, don’t hesitate to drop a line. I haven’t used it myself for some time.
Back to the topic :
what’s bugging me on the ogl preview is that it takes camera zoom level from screen, which results you can never see the real proportions of the final sequence in the preview. Otherwise it works quite fine
The parameters Lens (focal length really…) ClipSta and ClipEnd of the OpenGL camera (one per 3D view) can be set in the panel View > View Properties of each 3D window. By setting ‘Lens’ the same as for the rendering camera the problem is solved.
The parameters Lens (focal length really…) ClipSta and ClipEnd of the OpenGL camera (one per 3D view) can be set in the panel View > View Properties of each 3D window. By setting ‘Lens’ the same as for the rendering camera the problem is solved.
That is one useful tip I’ll be using extensively from now on! Thank you!
No, that doesn’t solve the problem. try it yourself.
place an object, place camera.
set the properties etc.
render the view
hit the OGL preview button.
result is different!
trust me, I’m using the preview quite often, so I know how it works.
No, that doesn’t solve the problem. try it yourself.
place an object, place camera.
set the properties etc.
render the view
hit the OGL preview button.
result is different!
trust me, I’m using the preview quite often, so I know how it works.
Quite right you are.
When it came round to doing an OpenGL render, I remembered this thread and tried it. No change. The issue isn’t the lens setting, it is that the OpenGL view point isn’t actually tied to the camera location even in camera mode- you can still zoom out and pan revealing areas that the camera wouldn’t see.
So I’m back the the solution I was using the past month - try and center the camera borders in the 3D view (in camera view) and try zooming till the borders fill the window. It is the best I can do (btw, I know there is a way to zoom smoothly but as I use the mouse wheel I have forgotten it! Can you remind me?)
I certainly want to trust you but I can’t not trust my eyes.
Is your 3D view a camera view?
Is you 3D window width/height ratio close enough to that of the render camera?
<EDIT>
OK, I see where the confusion lies. There are two cases.
1- When we are working in a view that is perspective but not a (render) camera view, matching window size ratio and lens parameter is important and useful. It becomes then possible to use the 3D view to find the best camera angle. When satisfied we can just use Ctrl+Alt+0 to position the rendering camera quite close to what was found in the 3D view.
2- For those who work directly in the camera view then the View > View Properties parameters are of no importance, aren’t they? Otherwise It couldn’t be a rendering camera view…
</EDIT>
@Koba: CTRL + Middle Mouse Button and move mouse up or down.
I’ve been thinking lately how awesome it would be to have the ability to do pre-rendered stuff with OpenGL that wouldn’t be dependent on Framerate (ie, you could take a few seconds to render each frame, since you don’t need it to be realtime) and have it keep the lighting / shadow / etc settings that are set up (motion blur and all that crap). Like an OpenGL-based ray-tracing renderer, hehe.
Either that or i need to figure out a way to do vertex lighting in Blender’s internal render. Low-poly work would be so much easier to light in Blender if you could get something like that working. The artifacts on Low res models from the lighting and shadowing is one of the hurdles i’ve got to overcome eventually, so I’ve either got to figure out vertex lighting or some similar method or i’ve got to come up with some kind of really clever node setup to simulate it…
/* EDIT - When i say vertex lighting, what i really meant to say was “Gouraud” shading, the type of lighting used in OpenGL (the game engine uses it, but it’s not available as a diffuse shader for the pre-rendered stuff…) */
The use of Ctrl + Alt + 0 is interesting am good for animated scenes with static cameras but not very useful if your camera is animated. Most of my animated scenes have animated cameras as well.
2- For those who work directly in the camera view then the View > View Properties parameters are of no importance, aren’t they? Otherwise It couldn’t be a rendering camera view…
Unless I’ve misunderstood, the view controls apply here as well because if you have an unusual camera lens setting, matching the lens setting in “View Properties” would help make the OpenGL render reflect the final result. If only there was an “inverse” to Ctrl + Alt + 0 ie - making the view match the camera borders.
Well, if the fly mode or if the Gkey/Gkey+MMB/Rkey/Rkey+MMB shortcuts were as comfortable as the navigation in the 3D window I would agree with you but they are both rather awkward.
I find that, for example, if at frame 1 I find my camera position in the 3D view, use Ctrl+Alt+0 and Ikey, then move to the next frame where I’ll need a key and repeat, etc… I can animate my camera quite fast and efficiently.
Unless I’ve misunderstood, the view controls apply here as well because if you have an unusual camera lens setting, matching the lens setting in “View Properties” would help make the OpenGL render reflect the final result.
They matter for matching any perspective 3D view but the camera view: for the latter they are overridden to match those of the active rendering camera.
If only there was an “inverse” to Ctrl + Alt + 0 ie - making the view match the camera borders.
Koba
It certainly would be a much more ergonomic way to find our camera keys. May the coding Gods hear you!
They matter for matching any perspective 3D view but the camera view: for the latter they are overridden to match those of the active rendering camera.
You beat me to it! I was going to correct my mistake as I’ve just rendered a scene with a lens IPO and yes the change is reflected in the OpenGL camera view. My mistake.
Well, if the fly mode or if the Gkey/Gkey+MMB/Rkey/Rkey+MMB shortcuts were as comfortable as the navigation in the 3D window I would agree with you but they are both rather awkward.
Unless I’ve misunderstood everything entirely, the flyby mode and those shortcuts (rotation, translation etc) aren’t keyable in themselves without a camera so if rendering out to OpenGL (ie making an avi), you can’t have view motion without a camera defined. Which is the thing at stake here.
I find that, for example, if at frame 1 I find my camera position in the 3D view, use Ctrl+Alt+0 and Ikey, then move to the next frame where I’ll need a key and repeat, etc… I can animate my camera quite fast and efficiently.
Ooh…that reminds me of one huge gripe I have with Blender. Have you ever found the perfect camera position in the 3D viewport and wanted to put a camera there without modifying the scene’s currently assigned active camera? In detail…
I find a perfect position.
I create a camera.
Now how do I make the camera active without switching the view to it and losing this position?
I could always use the active camera but if it keyframed, I must be careful not to change frame!
I suppose I could use the active camera, duplicate it and then delete any object, lens IPOs etc it might have. Not a very neat solution though!
No, not neat at all I’d say…
Shouldn’t this work though ?
1- create a new camera;
2- make it active using Ctrl+Numpad0
3- position it relative to your newly found ideal view wth Ctrl+Alt+Numpad0
That should do it.
Nope, Ctrl NumPad 0 also sets the view to that camera.
You could do it by just duplicating the original camera and leaving it in place. Then snapping the camera to your view will move the original camera, but not its duplicate. Then you have to go and rename things to make sense. It works, but a pretty ugly workflow, and hardly intuitive, as shown by the example of this thread…
Mmmmm,
should have tried it first, shouldn’t I?
Tsk tsk
Nevertheless, with just a small change to the work flow:
1- Find the view needed
2- Create a second camera
3- Split the 3D area in two: the same OpenGL camera angle will be reflected in both.
4- in one view apply Ctrl+Numpad0 to make the new camera filming
5- in the other view apply Ctrl+Alt+Numpad0 to position the new camera…
Still a bit of a pain but workable if one must have it.
Nevertheless, with just a small change to the work flow:
1- Find the view needed
2- Create a second camera
3- Split the 3D area in two: the same OpenGL camera angle will be reflected in both.
4- in one view apply Ctrl+Numpad0 to make the new camera filming
5- in the other view apply Ctrl+Alt+Numpad0 to position the new camera…
Not bad…not bad at all. I’ll try that next time I need it.
What does this all have to do with the incorrect size of the OGL preview? sorry, I probably lost the storyline
edit:
I’d really like to see this in Blender:
-if view is not in camera, do it as it is now
-otherwise to the preview with the correct camera settings/dimensions.
It’s quite Important when it comes to composition.
but it is doing great.
