4441
[Max] General Discussion / Re: Best way to optimise large terrain scenes?
« on: 2018-03-01, 19:24:30 »
A Corona AO in multipass? Or some native C4D AO? Not sure if the second works (I never use it :) )
2024-06-04, 05:40:08
News:
Show Posts
This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.
4442
[C4D] Resolved Bugs / Re: C4D Multi-Pass Velocity Map Black ?
« on: 2018-03-01, 19:22:38 »
Looks like a bug to me - the Velocity pass should work without Motion Blur being enabled in the render settings (but it is likely that it wouldn't work in IR even then, it doesn't in Max... but in Max it does work without MoBlur being enabled for the Beauty in a full render, which is what would be expected since the idea of the pass is to apply the MoBlur in post :) )
I can confirm the same thing happens to me in a simple test scene.
I can confirm the same thing happens to me in a simple test scene.
4443
[Max] General Discussion / Re: Best way to optimise large terrain scenes?
« on: 2018-03-01, 17:43:59 »
Scenes can be sent via the private uploader, then only the Corona team gets them and only for testing purposes :)
https://corona-renderer.com/upload
https://corona-renderer.com/upload
4444
[Max] Bug Reporting / Re: Multisub and lightmix?
« on: 2018-03-01, 17:16:50 »
May need the scene here, or at least more information about the scene itself.
I set up a Blend, replaced one material with the CoronaLightMtl, applied it to 2 boxes, and used the auto create LightMix, and only got on LightSelect per box. So it may depend on what the material is being assigned to. Also, if you just assign the CoronaLightMtl directly in place of the Blend, do you still get the 60 layers (which would again suggest it is not the Blend, but what the material is assigned to)?
I set up a Blend, replaced one material with the CoronaLightMtl, applied it to 2 boxes, and used the auto create LightMix, and only got on LightSelect per box. So it may depend on what the material is being assigned to. Also, if you just assign the CoronaLightMtl directly in place of the Blend, do you still get the 60 layers (which would again suggest it is not the Blend, but what the material is assigned to)?
4445
[Max] General Discussion / Re: Best way to optimise large terrain scenes?
« on: 2018-03-01, 17:10:25 »
I notice in your screenshot that your memory is pretty much maxed out - this will cause significant reduction in render speed. Looks like there is a lot of geometry here with the landscape objects, so that could be what is pushing the memory up so high, and would likely result in slowdowns (and why the slowdowns show up depending on the objects you add).
Other notes
- Denoising amount always has an effect, it's the amount of blending between the un-denoised version of the image and the denoised version, and applies whether Denoising is set to Full or Fireflies only. See https://help.c4d.corona-renderer.com/support/solutions/articles/12000021554-how-to-use-denoising-c4d-
- Denoising won't significantly affect render time. It may affect memory requirements though, especially depending on image resolution (the larger the image, the more memory required), and passes in MultiPass (the more passes that have Denoising applied, the more memory required). Depending on your scene (would still be good if you could send the scene as Marcin requested :) ) there could be memory savings to be found here, e.g. lower resolution, less multipasses, etc.
- With a pass limit of 10 and noise limit of 3, it is most likely stopping at the pass limit of 10 (usually it would take more than 10 passes to get down to a noise level of 3%). Remember that the renderer stops whenever any ONE of the limits is reached (time, passes or noise level). Helpdesk articles like https://help.c4d.corona-renderer.com/support/solutions/articles/5000518064-how-many-passes-is-enough- explore things like passes, GI vs AA balance, etc.
But, I think memory is most likely the issue in this instance looking at the screen grab.
Other notes
- Denoising amount always has an effect, it's the amount of blending between the un-denoised version of the image and the denoised version, and applies whether Denoising is set to Full or Fireflies only. See https://help.c4d.corona-renderer.com/support/solutions/articles/12000021554-how-to-use-denoising-c4d-
- Denoising won't significantly affect render time. It may affect memory requirements though, especially depending on image resolution (the larger the image, the more memory required), and passes in MultiPass (the more passes that have Denoising applied, the more memory required). Depending on your scene (would still be good if you could send the scene as Marcin requested :) ) there could be memory savings to be found here, e.g. lower resolution, less multipasses, etc.
- With a pass limit of 10 and noise limit of 3, it is most likely stopping at the pass limit of 10 (usually it would take more than 10 passes to get down to a noise level of 3%). Remember that the renderer stops whenever any ONE of the limits is reached (time, passes or noise level). Helpdesk articles like https://help.c4d.corona-renderer.com/support/solutions/articles/5000518064-how-many-passes-is-enough- explore things like passes, GI vs AA balance, etc.
But, I think memory is most likely the issue in this instance looking at the screen grab.
4446
[Max] I need help! / Re: Buy Corona video tutorials?
« on: 2018-03-01, 17:00:20 »
As soon as someone gets Certified over here in the US, they could start to offer in-person here so that US folks don't need to travel so far - who will be first :)
I would suspect the lack of paid video tutorials is perhaps because other engines needed detailed explanations to get any results out of them at all, while Corona is much more "just hit render with the defaults", making demand for tutorials less as people were able to get the results they wanted without extensive training?
I would suspect the lack of paid video tutorials is perhaps because other engines needed detailed explanations to get any results out of them at all, while Corona is much more "just hit render with the defaults", making demand for tutorials less as people were able to get the results they wanted without extensive training?
4447
[C4D] General Discussion / Re: Tree Leaves
« on: 2018-03-01, 15:13:10 »
It's working in the preview so all looks good :) You could enable clipping (won't be a "it works or it doesn't" but will speed up rendering). If it is not working in the scene, can you share what you are getting in the scene, along with screen grabs of the scene set up (or the scene itself). If applied to an object, is the object UV mapped properly?
4448
[C4D] General Discussion / Re: Tree Leaves
« on: 2018-03-01, 14:22:25 »
I'd use a Corona Material to do it - not all features of other materials will necessarily be supported.
4449
[Max] I need help! / Re: I just can't get vector motion blur to work right (RSMB, Blackmagic Fusion)
« on: 2018-02-28, 14:18:44 »
Could be the same issue as this thread https://forum.corona-renderer.com/index.php?topic=16585.0 ?
To quote from that thread:
'
"Corona outputs vector data with 0 velocity as black - so values go into positive and negative direction from there. Reelsmart expects data with a midpoint at grey (from the old days when nobody knew what floating point data is). "
You could inspect the "black" areas to see if there is data there (right click on the VFB to inspect the pixel), negative values in effect that won't show (and is not what RSMB is expecting). I can't test, but adding an offset of 0.5 in the Velocity pass settings should correct for that - anyone with RSMB able to test? Also, just to clarify as you have mentioned Fusion and RSMB, it is RSMB that you are using?
'
Let us know if the offset fixes it. Cheers!
(EDIT unfortunately not able to test as RSMB is not a free plugin, but it solved the issue in that previous thread for the user that was having problems)
To quote from that thread:
'
"Corona outputs vector data with 0 velocity as black - so values go into positive and negative direction from there. Reelsmart expects data with a midpoint at grey (from the old days when nobody knew what floating point data is). "
You could inspect the "black" areas to see if there is data there (right click on the VFB to inspect the pixel), negative values in effect that won't show (and is not what RSMB is expecting). I can't test, but adding an offset of 0.5 in the Velocity pass settings should correct for that - anyone with RSMB able to test? Also, just to clarify as you have mentioned Fusion and RSMB, it is RSMB that you are using?
'
Let us know if the offset fixes it. Cheers!
(EDIT unfortunately not able to test as RSMB is not a free plugin, but it solved the issue in that previous thread for the user that was having problems)
4450
[C4D] General Discussion / Re: Picture viewer and VFB not the same
« on: 2018-02-27, 14:30:28 »
Cheers, and much appreciated!
4451
[C4D] General Discussion / Re: Picture viewer and VFB not the same
« on: 2018-02-27, 14:19:33 »
So it looks like it's the same issue as in the other thread - to avoid scattered reports of the same / similar thing, could you post your examples in that other thread, and we'll use that one to track this particular issue of color differences / gamma differences between VFB and PV? Otherwise, it gets hard to see all reports of the same issue, and hard to reply to the same issue in several places :)
That was https://forum.corona-renderer.com/index.php?topic=10472
Thanks!
Tom
That was https://forum.corona-renderer.com/index.php?topic=10472
Thanks!
Tom
4452
[Max] I need help! / Re: I just can't get vector motion blur to work right (RSMB, Blackmagic Fusion)
« on: 2018-02-27, 13:59:25 »
Another key question would be which version of Corona you are using - a long time ago, the Vector pass did not work properly (maybe back in 1.4 or 1.5) but should have been working since then - as you mention, the YouTube tutorial shows it working as expected.
4453
[C4D] General Discussion / Re: Picture viewer and VFB not the same
« on: 2018-02-26, 20:36:51 »
So in your case Ben, there is a shift in some mapping for the texture, depending on whether you render to Picture Viewer, or render to VFB (and then view it in the Picture Viewer in the A/B comparison there)? I assume that was the difference, one was a "Render to Picture Viewer" render, and one used the Render button in the VFB?
Also, is this what the original poster is referring to, or are they referring to the gamma differences?
Cheers!
Also, is this what the original poster is referring to, or are they referring to the gamma differences?
Cheers!
4454
[C4D] General Discussion / Re: Picture viewer and VFB not the same
« on: 2018-02-26, 19:23:53 »
You'd have to let us know what that difference is - if it's a gamma difference, then there is already a discussion about that under https://forum.corona-renderer.com/index.php?topic=10472, you could add your example images there to that thread. If it is something else, though, some examples showing the differences you are seeing would be needed :)
Thanks!
Thanks!
4455
[C4D] General Discussion / Re: Better detail
« on: 2018-02-26, 18:23:50 »
Yes looks like it is just the Shift that is in the default C4D cam, not the tilt (nor auto tilt). We'll keep it under consideration as an addition and see if it might be possible.