Sample interior scene

[ecximer]

Are you comparing exactly the same scene?

A4 result - from "A4 benchmark"
Build timestamp: Sep  8 2014 result - from this thread. I opened file and click render butoon.

I don't know precisely, than two scenes can differ. But absolutely not incorrect global illumination, a noisiness and, as a result, the increased time - me confused.
If it is only temporal, I will wait and trust =)

I began to panic because to release not and long it was necessary to wait. And "regress" as it seemed to me, was strongly expressed.

[Polymax]

Build timestamp: Aug  6 2014 - working fine, only 09/07 - have a bug with SecGI(but! working fine with HDRI lighting).
https://www.dropbox.com/sh/g4qz1fh98mycjg6/AADYLXbvXE6w6trSU2lSPnaga?dl=0

Testing: same scene ("max" edition) in all versions, no reset, no change any parameters

[Ondra]

ok, there was a weird bug. In some of builds it did nothing, in others it made the scene brighter, in others darker (like the one I released as daily build). It is fixed now.

[polbo]

OCTANE and CORONA comparison


Hi! I take the scene and converted to Octane Render, so I can compare the quality and speed of the two renderers.




They are some differences but they are because the values for specular/ glossiness  and reflection are very different in Octane.
I had some problem with the portal, because with the version of Octane I have can't change the environment when use the Sun.
In render time it's 3 time faster with a gtx 590 in Octane than a i7 2600k in Corona to get similar image quality with the same amount of noise.
In Octane I used Pathtracing with a maxdepth of 1024 to be totally unbiased, but i can get the same quality with faster render times with a maxdepth of 8 or 10.

[juang3d]

Can you post your render times and do the Octane test with a 21 ray depth?

Cheers.

[polbo]

Can you post your render times and do the Octane test with a 21 ray depth?

Cheers.

Why 21 ray depth? Where you get that number? In my experience with Octane go further than 16 doesn't make any difference. I had renderer water and fruits with 16 max depth and don't have any clamp on the reflections.. So, 21 doesn't make any sense for me... If I want to be sure I go for 1024 that take the same time to render than 21.

[Ludvik Koutny]

Yeah, matching ray depth is completely irrelevant, as long as ray depth is larger than what Corona uses. And even if it wasn't, anything above 8 does not make very big difference.

Anyway, sure Octane is fast, but you are still comparing very expensive GPU with monstrous power consumption and quite restricting memory limitation to quite cheap CPU with moderate power consumption.

[polbo]

Yeah, matching ray depth is completely irrelevant, as long as ray depth is larger than what Corona uses. And even if it wasn't, anything above 8 does not make very big difference.

Anyway, sure Octane is fast, but you are still comparing very expensive GPU with monstrous power consumption and quite restricting memory limitation to quite cheap CPU with moderate power consumption.

The intention of my post wasn't to compare which renderer was the fastest, but to compare the render quality.. I think they are pretty close.
I don't know why you talk about the cost and the power consumption. Who cares? You are not buying a Lamborghini... It's only a few bucks less or more.
I like Corona Render very much.
But as an art director, Actually I prefer Octane because have instant preview and can make a lot of changes very fast to get the style I'm looking for. And that time is much more valuable for me than the final render time.
I can spend a lot of money on a computer or send my Corona renders to a render farm.. but I will not get instant preview anyway.

[Ludvik Koutny]

...but I will not get instant preview anyway.

You will...

[andrerosh]

[sas]

My test
How long should I wait to get the results in your gallery?

[Black Sun]

Do I use the Obj File to load in Max. Obj is the only one that makes sense to me.

Thanks

[geesve]

For maybe someone like me, who interested with some old AMD CPU speed with this...

Phenom II X6 1100T
Time: 0:12:26

[agentdark45]

Just tried this out in Max 2016 on my 12 core Xeon v3 using the UHD cache for the secondary bounches (as opposed to the old HD Cache). Result is 3 minutes 17 seconds.

[bud85]

Just tried this out in Max 2016 on my 12 core Xeon v3 using the UHD cache for the secondary bounches (as opposed to the old HD Cache). Result is 3 minutes 17 seconds.

Same max, same Xeon = time 03:03