Messages

121

[Max] Daily Builds / Re: Daily Builds 1.7

« on: 2017-10-07, 09:49:00 »

a 6000px image took 8-9min of parsing time on my Dual Xeon render node and is only rendering with around 3mil ray/s actual. This scene only took a couple of mins to parse and is rendering with over 6mil rays on my single Xeon workstation. So I still think something is off when it comes to performance on dual xeons.

Running 1.7RC6 on max 2016.

Follow up after trying the RC7 Build:

Parsing is better, but still performance is way worse than I remember the performance of the first 1.7builds.

Image being rendered is 6000x4000px.

Render node is a dual Xeon E5-2670:
Parsing 4min (down from 8-9min, so that is atleast something).
Actualy ray/s: hovering just below 2mill (so worse than prev build I guess?)

Workstation is a single Xeon E5-2696:
Parsing = 9sec(might have remembered the RC6 stats wrong last time, so screenshots now)
Actual ray/s: around 6.2-6.5mil (sorry I had stopped the render when taking the screenshot on this one).

122

[Max] Daily Builds / Re: Daily Builds 1.7

« on: 2017-10-05, 23:08:41 »

a 6000px image took 8-9min of parsing time on my Dual Xeon render node and is only rendering with around 3mil ray/s actual. This scene only took a couple of mins to parse and is rendering with over 6mil rays on my single Xeon workstation. So I still think something is off when it comes to performance on dual xeons.

Running 1.7RC6 on max 2016.

123

[Max] Daily Builds / Re: Daily Builds 1.7

« on: 2017-10-03, 11:32:55 »

Hey guys! Is the speed issue with dual socket systems i reported, solved with RC5?

Not yet. We're still trying to figure out the best way to optimize this which would have the least chance of introducing additional bugs

Is this only related to dual socket systems or is it # of cores dependent?

124

[Max] Daily Builds / Re: Daily Builds 1.7

« on: 2017-09-30, 18:11:29 »

I'm noticing slower performance on RC4 too. Especially very choppy performance when orbiting even the simplest objects/scenes in IR mode. This is on max 2016 and a single xeon 2696v4 cpu. It's as if the update to boost performance on many threads has done the opposite :)

125

[Max] Daily Builds / Re: Daily Builds 1.7

« on: 2017-09-17, 18:00:04 »

Very impressed with the size of the new library! Great work guys.

A request would be to have a "settings button" or tab where we could change the asset path. I.e. put the material folder on a network drive and link to that so it's shareable on the network and nodes can access maps. Just like with Siger Shaders.

126

[Max] Daily Builds / Re: Daily Builds 1.7

« on: 2017-09-12, 19:06:06 »

Connecter ! :- )

...And could even IES if Corona makes it droppable slot)

+1

If not, is there a way to change the default path that pops up when clicking new light's IES slot? But seriously, drag 'n drop would be lovely for IES as Connecter handles them previews like a baws.

127

Gallery / Re: Impremta Garden

« on: 2017-09-12, 12:09:37 »

Thanks Juraj

Also, I'm pretty sure Mr. Cop is trolling ;)

128

Gallery / Re: Impremta Garden

« on: 2017-09-12, 09:46:22 »

When I notice I'm sitting talking to myself about the images, the work is usually extraordinary - or I'm just a bit loony. Amazing work yet again guys. Real stunner!

Are you still using that special framebuffer plugin for CC tweaks or how does you workflow look nowadays once the render finishes? Feel free to point to a different thread if you've answered this recently ;)

129

[Max] I need help! / Re: Corona Image Editor - output gamma?

« on: 2017-09-03, 18:30:41 »

Small bump, I forgot about this problem and here we are again. What am I doing wrong here?

When I send a job to my render node, I save my img output with a gamma overide of 2.2 and save as a .cexr

When I then load the .cexr on my workstation in the CIE everything looks great and I tweak the lightmix as I want. If I now save to jpg, bmp or png it burns in the gamma of 2.2 nicely, but I only get an 8bit image to play with in PS.

If I save as a normal .exr the gamma goes back to 1 and I have to retweak in PS. Is there a way to save the exr with a gammaoveride like from max?

130

General CG Discussion / Re: "adding" more sky to a composition without changing camera / render twice?

« on: 2017-09-02, 19:31:37 »

I perhaps absolutely don't understand the question properly...but isn't 'Camera shift (vertical)' doing exactly that ?

You change resolution composition to more vertical (just like PS canvas) and then shift the camera up to remove the unnecessary bottom part. Nothing regarding the perspective changes, the net end result is simply added upper part on top of original composition.
The only difference to PS Canvas is the slight order of things, as Canvas lets you shift directly at resize, but 3dMas imitates tilt-shift lens process, results are absolutely identical though.

hmm, I feel slightly silly now. That does indeed seem to do the job. Thanks Juraj, hadn't thought of utilizing the lenshift and output size like that. Apparently I'm not the only one :)

Love these "aha" moments. Cheers man!

131

General CG Discussion / Re: "adding" more sky to a composition without changing camera / render twice?

« on: 2017-09-02, 09:19:07 »

Hey mate did you find any solution? Sometime ago i wanted to render a little wide, but as your problem, its really hard to do.

Sorry no, still haven't found a good solution to this.

132

General CG Discussion / Re: "adding" more sky to a composition without changing camera / render twice?

« on: 2017-08-15, 22:36:41 »

hehe no worries fraine, still appreciated ;)

133

[Max] Resolved Feature Requests / Re: Corona Camera functions

« on: 2017-08-13, 21:37:40 »

Bormax suggested me this script in a dif thread. I would love to see this function in the coming corona cam:
http://www.scriptspot.com/3ds-max/scripts/camera-resolution-mod

(a per cam render output/aspect res).

134

General CG Discussion / Re: "adding" more sky to a composition without changing camera / render twice?

« on: 2017-08-13, 21:35:33 »

Thank you Bormax, it's not quite what I was looking for, but definitely useful for my workflow as well! Maybe we could suggest this script function was incorporated with the new corona camera. Kind of like you can force dof off on a pr camera basis.

135

[Max] Resolved Feature Requests / Re: Instancing with Forest Pack and Rail Clone - status?

« on: 2017-08-08, 11:17:18 »

Recap of this thread:
Railclone:
1) Instancing on - works as expected, parsing is faster, there is an issue with randomizing.
2) Instancing off- works as expected
3) Randomizing does not work when instancing is enabled - using Corona Multi Map

Jens, could you contact us about this at support@corona-renderer.com with a very simple scene demonstrating this? It could be even a scene like in your screenshots, with some boxes scattered around.

If there are some other issues with Railclone or other iToo plugins, please also report them at support@, but one issue per one separate email please.

Done.

It's a general problem that Corona doesn't seem to recognize the way RC "manipulates" the instances with it's "Instancing engine". Neither random mat IDs or UVW offsets nodes work. Corona only sees the scattered object as the exact same instance, hence no randomization will occur. So the new Corona UVW randomizer doesn't work either. - Unless we turn the instancing engine off, but that comes with the drawbacks listed above.