Rendering software comparison

jenujacob

comparing render engines are very hard actually...
unless u have a set of rules where u can define as to what all features u would like to test i think its kinda fruitless...
i mean take KT for example.. the above render is using pm+Fg.. (biased render)which gives out completely different results from the unbiased render options like MLT or BiPT or PPT...

Podium i think has only unbiased redner solutions.. so its fine in that aspect here...
but when u compare Maxwell or fry render.. they only deal with unbiased rendering... so thats a different benchmark all together...

Vray on the other hand has a mix of both biased and unbiased rendering algorithms...

U could on the other hand... try to compare the render engines on the basis of these :-

based on one touch render solution... without any manipulation of materials in the external render engine..
set up time...
3)ease of use....
4)speed... which u have mentioned here..

i hope i am clear in what i am tryin to explain here... not to step on anyones shoes here..

kwistenbiebel can talk more bout it here.. hes an expert on all things rendering!!!

Stinkie

Valid points, Jenujacob. I must add, however, that since 1.5 Podium also uses MLT.

solo

I have seen crappy Max renders, brilliant Indigo renders...etc.
The render engine is only part of a great render, knowing how to use it is the other part.

I also agree Kwistenbiebel would be the best person to make comparisons, however I see no conclusion arising from it, just like diets cannot be compared as the only one that works is the one you keep to, render engines are the same as the best renders come from understanding, learning and talent.

The top comparison reeks of bias as the wasted time for export and GUI maneuvering in KT was not mentioned, or the conditions used.

Stinkie

Let's also not forget: different tools for different jobs. In my opinion, unbiased yields the best results. However, unbiased renderers seem somewhat unpractical to use in a production environment. Unless, of course, you own a whole fleet of 'manycores'.

Oh yeah: the price counts too, obviously.

Stinkie

Sorry, my bad. Or actually Bigstick's. Search the Podium forum, you'll see.

Your point about comparing apples and pears seems valid to me. To some extent these kind of comparisons seems to derive from the need to find the 'ultimate renderer' -a thing that does not exist.

kwistenbiebel

@unknownuser said:

To some extent these kind of comparisons seems to derive from the need to find the 'ultimate renderer' -a thing that does not exist.

That would be my conclusion as well
...It all depends on the kind of scene you want to see rendered.
What seems to be the ultimate render engine for scene A, is often the worst solution for scene B and vice versa.

kwistenbiebel

Correction Stinkie,
Podium doesn't support MLT (Metropolitan Light Transport), but it does support PT (Path Tracing). Both are unbiased methods. MLT is effective on scenes with a lot of indirect light, while PT is master in scenes that contain a lot of direct lit objects (e.g sun lit exteriors).

@Jenujacob,
Indeed I like to play with different render engines, although that doesn't make me a master.

@Rob D,
As Jenujacob and Solo pointed out, comparisons are difficult to do correctly.
In fact, I tried it before , but in the end, I found out that in some way, the comparison always fails as each render engine uses a different approach (on materials, physical sky,camera,tonemapping, etc..), making it comparing apples and pears.

Instead, I like to render a particular scene using different render apps, and just post the results for all to see, together with some info on the settings used (render time etc...).
I have posted gallery threads like that earlier on sketchucation gallery section (the old forum), but I stopped doing that because the scenes contained too much geometry (.3ds imports) that wasn't native Sketchup.

Stinkie

On the positive side, a side by side comparison could help people to decide what renderer they might get for different tasks.

Also, it might shed at least some light on what your getting for your buck (prices matter!). Visiting the respective forums would of course be a must.

Rob D

Thanks everyone for the tips. I'll try to incorporate your views so please keep 'em coming!

For me, the useful thing would be to have a benchmark for a "passable" render on several different programs. The majority of us just want to stick our SU model into a renderer, tweak a few materials, hit "render", and get something a little more realistic than SU output. So we might as well have that as the benchmark. Anything more than that starts to become a test of the users skill level, rather than the software, as a few of you have said.

sepo

Hmmm.... I can show you scene where Podium will render many times faster than Kerky. It all depends on presets.

Stinkie

I haven't made any comparative renders, but I have tried KT at maxed out settings on my octo core. Took all night for the render to get at about 40%. Obviously this is in part due to my limited experience with the app (which in itself looks cool), but I was nevertheless a tad surprised.

Frederik

@sepo said:

Hmmm.... I can show you scene where Podium will render many times faster than Kerky. It all depends on presets.

I sure would like to see that, Sepo...

With all due respect to the Podium plug-in, which IMHO is very intuitive, easy to use and well made...
But as we both know, the render engine behind Podium is the Kerkythea SDK...!

Regardless of whatever kind of tweaking and optimizations you can claim to have made, the engines are identical (apart from some exceptions available in KT and not in Podium...) and thereby I really can't see how you can claim that Podium will render many times faster than KT...

But if you want to give us (the KT Team) an example, I'm sure that we will be ready to give it a fair test...

@unknownuser said:

The render engine is only part of a great render, knowing how to use it is the other part.

That's SO well said, Pete..!!

sepo

I will send you myscene.xml if you want and you render it and tell me the time

sepo

lol ...OK bare with me... I will find suitable scene and we will give it a whirl. I will by you drink if Kerky renders it faster.

I don't understand what do you mean by "And please also tell us how we can test it in Podium ourself..."

Frederik

We're ready...
And please also tell us how we can test it in Podium ourself...
(That will also enable us to see what the output should look like...

You can send the scene to frederik [at] kerkythea.net

Frederik

@sepo said:

I don't understand what do you mean by "And please also tell us how we can test it in Podium ourself..."

I'm not surprised...
First of all to get an idea about the render time in Podium ourself, and secondly to enable us to see what the output should look like...

sepo

OK as you have better idea maybe you should organise the setup...

Stinkie

Oooh, a li'l contest.

We should really keep Sepo indoors. He just can't resist hassling the neighbours.

I'm pretty curious about the outcome. But shouldn't the setup time count too?

Fletch

Stinkie wrote: "I have tried KT at maxed out settings..."

Well, this is the common mistake of new users with Kerkythea (mine too! ) ... people think (quite naturally) "higher settings must be 'better'" When in actuality it is important to understand what all those settings mean. And if one does not, then they are better off sticking to the 'presets' as their aid/friend/guide.

I have done many, many... and yes - many - test renders with various methods and settings in VIZ scanline, VIZ radiosity, Mental Ray, Vue Infinite 5 and 6, and Kerkythea. I can tell you Kerkythea gives me the best quality with fastest set up and ease of use/material mapping & creation... and I have gotten just as good results in 30mins at 2800px on PM+FG low preset in KT as I got in 24hrs with MLT... it ALL DEPENDS ON THE SCENE AND THE LIGHTING AND MATERIALS.

The key is knowing the 'optimum' settings... a professional workflow is to do test renders small and very low qual. settings, and slowly work your way to bigger renders/higher quality until you know what the settings will do without needing so many test renders.

Comparing speed of Kerkythea and Podium is both silly and pointless IMHO. Podium IS USING Kerkythea, but since it is inside SU is obviously much easier and quicker to use... TBD has done a good job 'optimizing' the settings for you, so you don't have to do all the testing/trials to actually learn the program yourself... you just get the fun of rendering.

If Podium makes you and your clients happy, and you're having fun - well THAT'S WHAT MATTERS!

If you want to have choice to render the exact same scene with biased or unbiased methods, hundreds of advanced options, and soon to be much much more... then perhaps keep Kerkythea handy... but really, they serve two different 'audiences'. Podium is for people who want to just stick with SU and not have to "go-back-and-forth" and who do not want to have to think about all the other info available in Kerkythea.

Trying to compare them is, again, IMHO, both pointless and silly.
1-they both have the same engine - KT
2-you would have to make sure every single setting is the same in both of them... which you cannot be sure of
3-what's the point? oh yeah, being happy rendering... and you can be happy with either of them! there's your answer... they both win.

Love, Fletch!

Stinkie

Thanks for the explanation, Fletch. I did think "higher is better". Anyway, I also did it out of curiosity ("will my Mac Pro catch fire?"). By now, I've tried some of the presets - those work dandy. Fast, yes.

"If Podium makes you and your clients happy, and you're having fun - well THAT'S WHAT MATTERS!"

I'll slurp coffee to that!

"Trying to compare them is, again, IMHO, both pointless and silly. "

Valid point. But I think Sepo and Frederik both are smart enough to know that - them youngsters just wanna have some fun.