Keep rendering ? + Storing renders ?

[Sam]

First request:
Is it possible for Octane to keep rendering when you change the max samples ?
For example you are at 1800/2048 samples and you are sure that 2048 not gonna be enough.
So you setup the max samples to 4096 but Octane reset the samples at 0/4096 and don't resume at 1800/4096.

Same thing for a completed render
You got a 12min render at 4000 samples but there's still a little bit of noise you want to clean up.
You increase the samples to 6000 and Octane render the samples from 4000 to 6000 and doesn't reset to 0

[Sam]

Another related request

I was doing a render at 4096 samples and I was at 4000 samples
I touched the camera node to change the exposure and the samples reset to 0
That made me lost 15min of rendering...

Why not making a sort of memory of the various renders ? (only for still images not animation)
There's a system like that in Cinema 4D version 11.5

http://www.maxon.net/products/new/cinem ... iewer.html

Wouldn't it be great if there was an easy way to compare test renders so that you can see which changes to lighting, materials, render settings or other elements work best?

The new Picture Viewer lets you do this with ease. Your renderings are stored automatically to the Picture Viewer's History list where they can be viewed with a single click. The render time is stored for each test render, helping you to find out which render settings work best for your project.

What if you want to take a really close look at the differences between two of your test renders? This can be done with the new AB Compare tool, which lets you you choose any two of your test renders in the History list and blend between them by simply dragging a slider left or using the Difference mode.

The thing would be that, each time the sample get to 0 (aka each time a new render is set, either because the user want it or either because the user touched a node) a new entry is made in memory and the render is stored.

What would be even more powerfull is that the samples of that render are stored on disk and also the whole node setup. That would allow getting "back in time" in the scene !!

I think this could be usefull for a lot of things
Also in Cinema 4D Picture Viewer you can browse the old renders and they are stored in 32bit, that mean you can get to an old render and click save-as and save that render.

[James]

Something like this would be useful, especially for things like a side by side comparison of response curves.

[gristle]

Have you seen Shaderlight? It does some interesting things; you can change material, env, lighting and texture without re-rendering.

http://www.artvps.com/

Must be some buffer trickery.

[simmsimaging]

Being able to tweak the sample values up would be hugely helpful. A render history (of completed images) is also very handy, but easy enough to workaround.

b

[Sam]

Have you seen Shaderlight? It does some interesting things; you can change material, env, lighting and texture without re-rendering.

Yeah sure, Im very aware of the CG news as I read like 20 bookmarks each morning
Kinda the same thing than Fryrender SWAP wich use the buffer to change mats.

But its not what I want to see in Octane, its more oriented for product design and Octane is not just about that. It could conflict with animation rendering or it could make the software much more complex by adding too much features.

Being able to tweak the sample values up would be hugely helpful. A render history (of completed images) is also very handy, but easy enough to workaround

Yes tweaking the sample value could be great, specially when you need more samples on a 30min render and you don't want to wait 30min again

The render history is mainly for backup, what I mean is that you don't have to save every test renders to disk, you get an automatic backup of your renders for the current scene. Also it could be used to compare images.

And if it store nodes + samples you could "restore" an old render into a new scene or overwrite the current scene.

Its like auto-save but transparent for the user.

UI wise it should be a node called something like "picture viewer" and you would connect the "rendertarget" to that "picture viewer"

Once its done, everything going inside the "rendertarget" node would be automatically saved in the "picture viewer" node

The "picture viewer" node would have settings like
- how many history levels you want to store
- do you want to store samples (yes/no)
- do you want to store the nodegraph (yes/no)
- choose a custom path to save the history (allow to save on another drive like an external drive or network drive)

Note that the "picture viewer" only save something when the sample ammount get back to 0 (everytime a new render is launched)

And in the settings of the "picture viewer" node there would be a button that open a new floating window where you see all the previous renders, where you can compare renders and where you can restore renders or save renders to disk (because the picture viewer would store internally the 32bit data you could save the render to any file format)

The good thing if "picture viewer" is a node is that its only optional, if the node is not here, there's no history. In Cinema 4D its always here, you always have that history and its not always necessary.

And also you could have multiple "rendertarget" nodes (one per camera) and then you could have also multiple "picture viewer" nodes (one per rendertarget node) That way you can have history of different camera or point of view

I think this could be a very powerfull "time machine" and a nice feature, that way we can even forget to save every render to a image file and have that history always there as a backup.