Hey there,
stupid question: I arranged a scene with 3 different renderlayers. I rendered them as openexr. In the compositor I combined them together. Now my stupid question is - how do I export this as a video / png-sequence? Do I have to rerender it or is there a way to use the sequencer? If I add the imagesequence there it doesn´t use the compositor nodes…
Many thanks!!
There is a checkbox for rendering the Compositor (Compositung) in the scene rendering settings. After choosing this option it is possible to use the Compositor result in another scene’s VSE (add > scene) without rendering it but it is very possible that this will slow down the editing drastically.
Lähetetty minun SM-A300FU laitteesta Tapatalkilla
Thanks for your response!
I already tried this. But somehow Blender doesn´t use the nodesetup from the compositor in the playback window of the VSE. Any ideas??
thx
One more thing - in another thread (see link below) I read, that I have to use at least a new scene to export the project via the VSE. Is that right?
https://blenderartists.org/forum/archive/index.php/t-288126.html
okay, I try to be a little more concrete …
I have an animation of three different render layers: For one of them I baked the textures to save rendertime. This one I would like to render in the final render with 1000 samples. On the second layer I have another object, which I will render as well with 1000 samples (together with the first layer). The third layer contains only background-stuff, which I like to render with 300 samples.
For compositing purposes (and if blender crashes) I am working with OpenEXR-Multilayer-Files. So I render all three layers in different renderings and combine them with other VFX in the compositor. But then - how do I get the OpenEXR-sequences out as a video or png-sequence without rendering it again?
Like I said - in the render tab I checked the fields for “Compositing” and “Sequencer”. In the VSE I add the scene before I start rendering. But somehow I only see the animation as a solid animation, without textures, VFX etc. And if I press the animation-button in the render-tab again, it will render the whole animation like in the first place…
Does anyone have any ideas or does someone probably know a good tutorial on this workflow? surprisingly I couldn´t find one …
Many thanks,
t.
Umm. I’m not sure if I’m now talking about the same thing but to my knowledge the relation between Compositor and VSE works so that Compositor effects won’t show up if the scene is added in its own VSE. Instead you need to create another scene and add the Compositor-fx-scene to this new scene’s VSE. At least this way I have gotten it to work.
Lähetetty minun SM-A300FU laitteesta Tapatalkilla
sounds complicated… haven´t worked so far with different scenes. Does this mean, that I have import the openEXR-files in the compositor of the new scene and rebuilt the whole node-setup? Or is there a way just to add the nodesetup? Or is it possible to somehow import the complete “old” scene into the new one?
in any case - many thanks for your tip!!
I’ll now try to answer your longer explanation.
First of all, have you imported the openEXR-files into an empty scene? I think you should do that. There shouldn’t be a need to activate the ‘Sequence’-checkbox in the scene’s render settings but you definitely have to activate the ‘Compositor’-checkbox. I bet you have already done it but double check that you have to have an output (not the viewer) node last in the node chain/network.
You also don’t have to add the scene to any VSE timelines if you just want to render the frames. You only have to use VSE if you want to edit the Compositor-sequence together with other sequences.
Lähetetty minun SM-A300FU laitteesta Tapatalkilla
You can copy the whole scene.
Lähetetty minun SM-A300FU laitteesta Tapatalkilla
Okay, meanwhile I created a little testproject to try different things. In there I now created a new scene, after rendered the first scene into 2 OpenEXR (which are composited together in the first scene). Afterwards I created a new scene within the project with the option “full copy”. But again, if I switch to the VSE I only do see the testanimtion in solid mode.
And just to understand it: If the animation would be there as wished - what would be the next step to save it as a video oder png-sequence? going back to the render-tab in the default layout, changing the output-format and hit “Animation”? Isn´t rendering again the hole project as before? And sorry, that it takes me that long to understand the workflow. Don´t know why …
Many thanks again,
t.
Uh. I’m sorry if I confuse you. I realise I’ve written instructions that contradict each other.
I will do a test myself to clear my thoughts and get back to you.
Lähetetty minun SM-A300FU laitteesta Tapatalkilla
cool, thanks a lot! 
Ok. Here’s another try. I try to be clearer. Please, forget what I wrote before. I did my test with Cycles, cause I’m assuming you used it aswell. I think it should work the same with Blender render.
Here goes:
You can use the nodes in the very same scene that has the objects and the animation.
[LIST=1]
So, you don’t have to build the nodetree again.
You don’t have to have any other scenes or to use VSE.
From Properties panel > Post Processing -section: activate Compositing but DON’T activate Sequencer
From Output section: Set up the movie/image sequence you want to export the composition to
In Node Editor
Make sure you’re on Compositing tab
Activate toggle Use Nodes that is found at the bottom of the Node Editor
UseImage sequencenodes to import your exr -files
Make sure the node tree ends with a Composite node
[/LIST]
I hope the sample image clears the node settings for you
Okay, so far so good. But, if I go back to my original problem with the “original project”: So I render layer 1 and layer 2 at 1000 samples as an openEXR. The results I bring back to the compositor. Then I render layer 3 at about 300 samples, again as an OpenEXR. That as well I reimport to the compositor. But then - do I have to make another, final render? Or is it possible from this point of the workflow to save the result with three layers from the compositor?
If not - what´s the point in rendering layers with different samples if I have another, 4th rendering. There won´t be much of a time saving I guess…
Sorry, it is still confusing for me …
You’ll have to do another render but only with the Compositor not with counting all the objects, lights and so on. So this render is way faster. Another important point for post work/composition is that you can do final touches and make your work better than with just rendering the 3D stuff. Doing changes in 3D takes a lot more time than doing some minor adjustments in post work.
If your composition is about making some minor touches it might be all the same to do everything at once. But still I’m pretty positive that taking your scene apart and rendering parts of different importance with different settings saves you time. The time saving might become more apparent if you consider a situation where something goes wrong: the program crashes midst rendering or there is something wrong with some object. While doing each render separately might be in whole slower than doing everything at once. Still, each of the render layers is faster to redo than the whole animation. So, in the long run and considering bigger projects I suggest that you stay on this chosen path and experiment on the results.
If the rendering of the composition seems to take a great deal of time, you might want to try this:
Although, I’m pretty sure this doesn’t affect the render time. Still it won’t hurt to try.
Lähetetty minun SM-A300FU laitteesta Tapatalkilla
Hmm. Maybe you’re facing a problem with terminology. Saving the composition of the three openEXR-sequences is also called rendering although it is not quite the same as rendering 3D objects, lighting and such. You do it by pressing the same Render Animation that you do when you render the 3D world that you have created.
In other words. Also the calculations that are necessary in video editing are called rendering although they might not have anything to do with 3D.
Lähetetty minun SM-A300FU laitteesta Tapatalkilla
hey slowly I start to understand. At least, I hope so … to sum up - I do the two renderings with 1.000 / 300 samples. I place the EXRs in the compositor and then I do another “rendering” with - let´s say: again 1.000 samples - which will be much faster as a normal rendering. This is cause all the 3d-light-falls etc. are already calculated, Right? In addition to that I could do the steps you suggest at post #15…
Am I on the right track now?
In any case - many thanks for your patience and your good explanations!!
ah and the final rendering of the compositor will be - different as before - not in EXR-output but a movieformat or a quicktime etc.
I checked it in blender - which lead me to new questions …
right now, I do have changed the render layers with the OpenEXR layers (with low-samples-rendered files in for adjusting purposes, s. attached screenshot). before I do the final render as described above (first with 1.000 samples, 2. with 300 samples, before rendering out the composite), do I have to switch back to the render-layer-nodes? I think so, right?
I don´t know why, but this seems to me the most complicated part in my blender life so far … :eek:
