Jump to content

Glass Test


Devin Johnston
 Share

Recommended Posts

My client wanted to see what the exterior of his building was going to look like so I decided to give Maxwell a try, this is the result. I'm pretty happy with it and I've learned a little more about the material editor. All the glass is AGS, the original image was rendered at 3000x2810. I didn’t have the guts to render it on a single PC so I used cooperative rendering, let it run for 13 hours and the result was an almost noiseless image, no post processing has been done.

Link to comment
Share on other sites

Are you kiddin' me? My god, how can you make money waiting for a render like that especially when you are tying up more than one machine. You must have clients that have bottomless pockets or you are eating peanut butter sandwiches all the time!

Link to comment
Share on other sites

Hmm, ... well I can understand that you are happy with it, but I guess

this is just because of all the RC/V1 weirdness. Do you remember fusos

facade ... that convinces me more. The metal material you've used is

one of the culprits ... it is to even and the reflection/lighting is a bit odd.

The glass, ... well it's AGS ... :p:D;)

 

Anyway, I know it's not that easy to obtain really good materials with the

new material system. And btw. what SL was the rendering on each machine?

I'm wondering 'cause I can see those ugly artifacts in the sky ... you know

horizontal stripes. :/

 

How many nodes where rendering this pic and how long took the merging?

 

ps: Sorry for the betatalk. ;o)

 

edit: And have a look at what kinda box he was rendering ... oO

 

 

take care

Oleg

Link to comment
Share on other sites

Are you kiddin' me? My god, how can you make money waiting for a render like that especially when you are tying up more than one machine. You must have clients that have bottomless pockets or you are eating peanut butter sandwiches all the time!

 

My office has about 120 work stations, at night I can use as many as I need. Right now I only use Maxwell on projects that don’t have super tight deadlines so the speed issue isn’t as big a factor. In reality the work stations are there regardless of whether or not I send anything to them to render so if one image requires 40 computers over a 13 hour period it’s no big deal as long as the client gets the image.

 

. And btw. what SL was the rendering on each machine?

The SL numbers were all over the place, some machines were at .5 others reached 11 but they all were pretty fast machines so I have a feeling that something is still not functioning correctly in cooperative rendering.

I'm wondering 'cause I can see those ugly artifacts in the sky ... you know

horizontal stripes. :/

Yes I saw that and wasn’t sure what to make of it since Maxwell isn’t supposed to generate any artifacts. This could be the result of the MXI merging process; I’ll post it on the Maxwell forum and see what people have to say.

How many nodes where rendering this pic and how long took the merging?

I got back 30 usable MXI files from 30 PC’s, each file was about 900 MB and it took a good 20 or 30 minutes to merge them all manually.

Link to comment
Share on other sites

Hmm, ... if the SL is "to low" you'll have a problem, ... but I thought you

knew this already --> klick

 

Did you use HDRI for the reflections on the glass or some kind of reflection

plane? Just wondering ... ;o)

 

 

take care

Oleg

Link to comment
Share on other sites

I know the problems associated with low SL levels but seriously how do you control this when you only have x amount of time to render something, I can only do so much and the rest is up to the computer. I did put a map in one of the BDSF layers but you can't even really see it why would that matter?

Link to comment
Share on other sites

Miahi just posted on this issue, he says that this is most likely noise that would disappear at a higher sample level. He said that machines that don't reach higher sample levels won't contribute anything to resolving the noise; only higher sample levels can fix this.

 

This is opposite of how I understood cooperative rendering worked, I was under the impression that it didn't matter what SL level a machine reached that even a low SL would contribute to reducing the noise pattern. If what Mihai said is true then your image can only be as clear as the highest SL a single computer can generate.

Link to comment
Share on other sites

If what Mihai said is true then your image can only be as clear as the highest SL a single computer can generate.

 

I don't have Maxwell...never used it...nor do I want to, but that is a bunch of crap if you ask me.

Link to comment
Share on other sites

He just posted another comment and said that each machine no matter what SL level it reaches contributes to the overall resolution of noise; it's just that low SL levels don't contribute very much.

 

Wow, from the first post you cited it sounded like he was saying "cooperative rendering is completely useless." This noise pattern bugs me - it's gone from random noise that can pass as film grain in many situations to something an inkjet printer would do when low on ink. They've effectively made Maxwell much slower for a lot of people who now have to get that noise out.

Link to comment
Share on other sites

I don't think Mihai knows what mihell he's talking about in this situation.

 

Just for kicks, you might want to try rendering that scene overnight on one of your fastest machines to see how it turns out. If it doesn't have that pattern, then I'd think it'd prove that your lower SL mxi's are contributing the pattern... which wouldn't be a ringing endorsement for cooperative rendering.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share

×
×
  • Create New...