PDA

View Full Version : Does Lightwave have a next gen renderer...or is it a bit outdated?



tyrot
09-20-2012, 10:21 AM
reading any constructive feedback would be cool.

I feel like using Octane GPU renderer over Lightwave in any archviz project that i do. I cannot compete with Vrays or other renderers with native Lightwave renderer in terms of speed and quality. It just cannot cut it..

Is it me or is it really F9 got older..and nobody confessing it?

OFF
09-20-2012, 10:36 AM
maxwell-like would be more interesting ->
Maxwell Render plug-in for LightWave: Maxwell Fire
http://youtu.be/MPkuIuGCxBg

realgray
09-20-2012, 10:54 AM
There's a few things missing. Micro poly disp. Modern Pass system. But it is my belief these things are coming.

hrgiger
09-20-2012, 10:57 AM
It shouldn't be outdated after all, its been the focus for the last several versions of LightWave....

But as far as complaints go, these seem to be common ones that I hear... no render pass system, no bucket rendering, no micropoly displacement... My main focus is not rendering though so I will admit to not being an expert on rendering systems.

I know though from my brief stint with using Mental Ray in XSI that LW rendering is a hell of a lot easier to setup.

Netvudu
09-20-2012, 11:16 AM
I use Mantra (Houdiniīs renderer) along with LWīs native renderer for some projects. Mantra is very well considered in the industry and increasingly so, for its flexibility and speed. Itīs being actively used in many feature films. I can testify that Mantra is very nice but it takes much longer to setup a scene compared to what it takes in LW and itīs tough to get LWīs quality. Only micropoly displacements are clearly superior to what stock LW has.
The only thing I think LWīs renderer should be starting to take into account is PBR. With more and more physically based renderers appearing into the market, some slow, gradual steps into this should be taken. (BTW, Mantra is doing it)
So, no, itīs not outdated at all. Itīs much better than most others. Itīs not a specialized render but way more flexible than most.

Obviously, if you are working in a specialized sector such as arch-viz, then itīs possible to find better engines just for that task (i.e. Kray, or the slow but beautiful Maxwel)

3dWannabe
09-20-2012, 01:01 PM
no render pass system,

I doubt LW will add this as Janus and other tools already exist - better to focus on what we can't purchase.

Janus is the best - and it is designed for Lightwave

http://faulknermano.com/janus/index.html

DonJMyers
09-20-2012, 01:26 PM
Lightwave has always been known for quality renders. Don't forget for a very long time Maya didn't have a native render engine!

The show Terra Nova, for instance, animated in Maya and rendered in Lightwave. Obviously tests showed LW did it better with lighting/texturing/rendering.

LW_Will
09-20-2012, 01:41 PM
You know, the Lightwave Renderer isn't the same old renderer. It has been totally re-engineered since 11.X... so they are updating the system.

I was shocked to learn this, partially because the system will still work in the old manner (included wasted render time). Look to Dave Jerrard and Jennifer Hachigian you tube posts on unified sampling in the new renderer.

tyrot
09-20-2012, 02:05 PM
thanks mates - sure i ll check those explanations... from jenn and dave..

monovich
09-20-2012, 03:00 PM
Yeah the unified sampling thing is definitely worth reading up on and wathching. I'm still getting up to speed on proper usage.

chco2
09-20-2012, 05:22 PM
You cannot compete with vray's renderer against LightWave's renderer? And not only Vray, but other renderers as well ?
All I can say is.. compare the quality against the time it takes. I have not found one renderer as versatile and fast as LightWave Native renderer providing the same quality.


reading any constructive feedback would be cool.

I feel like using Octane GPU renderer over Lightwave in any archviz project that i do. I cannot compete with Vrays or other renderers with native Lightwave renderer in terms of speed and quality. It just cannot cut it..

Is it me or is it really F9 got older..and nobody confessing it?

Ernest
09-21-2012, 12:55 AM
maxwell-like would be more interesting ->
Maxwell Render plug-in for LightWave: Maxwell Fire
*WHACK*
I don't have 12 hours for a ****ing render!

Have you seen that commercial? It's hilarious.

Use Fresnel on everything!:ohmy:

Iain
09-21-2012, 01:45 AM
Maxwell is a fabulous render engine. The little touches you get that you don't even expect can lift a render way above the acceptable norm. Is it worth the render times though? After years of experimenting, I've decided not. With a tiny amount of post processing, any decent engine can get very close, including LW.

Lightwave does still hold its own but, as said above, if you want realism, especially interiors, it is not on par with specialist engines like VRay or even Kray. But then I wouldn't expect it to be at its price point and as a built in feature.

scratch33
09-21-2012, 01:48 AM
otoyo is looking for someone to developp octane integration in lw...
I think a biased renderer with the speed and the quality of octane at a so cheap price can be a excellent
solution to integrate in lightwave.

Red_Oddity
09-21-2012, 01:54 AM
LWs renderer has come a long way, but quality vs speed wise it can't even come close to what we can do with VRay.
The project we do now, probably wouldn't even be able render as LW still needs to load EVERYTHING in memory, VRay on the other hand scales really nice, no matter how many billions of polygons and gigabytes of light/irridiance caches we throw at it.

Red_Oddity
09-21-2012, 01:56 AM
Also, people on this forum still don't seem to grasp the differences between a non spectral and spectral renderer, and a biased and unbiased renderer and what those differences mean for workflow and speed vs quality.

chco2
09-21-2012, 02:25 AM
And all white people are trash ?
There are people that keep up with technology, more than you think.

Also, use LightWave if it fits your needs and if you think other packages are better or more suitable use those. Vray does not fit in mine simply because it doesn't produce the results I am looking for in a workflow I'm not happy with.
You cannot simply say this renderer is better than this one. It all depends on what you need for a project. Overall LightWave renderer has not let me down speedwise and quality wise.
[Why do I think your response will be that your projects need the highest quality and LightWave cannot produce that. As if I would settle for less..]


Also, people on this forum still don't seem to grasp the differences between a non spectral and spectral renderer, and a biased and unbiased renderer and what those differences mean for workflow and speed vs quality.

tyrot
09-21-2012, 02:26 AM
red oddity - not just me mate - if you pointing out me - i do know but my clients do not care about any of that.. they throw a render on the table..very cool looking render from a guy who does not know anything about 3d -just using presets - adding stupid highpoly models from archvision models..

Question can i match it with native Lightwave renderer... that is all.

Years ago this was not a huge issue -Lightwave was beating them all - we could have more realistic more faster built in renderer as we were always proud of ...

scratch33
09-21-2012, 02:30 AM
otoyo is looking for someone to developp octane integration in lw...
I think a biased renderer with the speed and the quality of octane at a so cheap price can be a excellent
solution to integrate in lightwave.

geo_n
09-21-2012, 03:16 AM
Its outdated compared to others but its still the best thing that lightwave has to offer and comes free with the package. The quality is pretty good for general purpose rendering. The only competition at this price point is modo which is a superb renderer.
What's outdated?
Speed - AA is still not as fast as the top renderer out there. Gi calculations still slow even with interp radiosity. I was baking gi to textures and had to revert to fprime for speed and better gi.
Bucket rendering - N/A
Distributed rendering - N/A
Micropoly disp - N/A
More buffers
Faster hair shader rendering

Frankly though I don't want NT to work on the renderer for sometime since its good enough for the current lw users.
Nt should work on something else.

Red_Oddity
09-21-2012, 03:40 AM
And all white people are trash ?
There are people that keep up with technology, more than you think.

Also, use LightWave if it fits your needs and if you think other packages are better or more suitable use those. Vray does not fit in mine simply because it doesn't produce the results I am looking for in a workflow I'm not happy with.
You cannot simply say this renderer is better than this one. It all depends on what you need for a project. Overall LightWave renderer has not let me down speedwise and quality wise.

Sorry you took that reply so personally, i wasn't implying anything.

It's just that some people keep crying that LW doesn't match VRay or why their results aren't as fast and good looking as some people show in videos on what they seem manage with Octane (which are horse crap videos mind you, Octane still needs hours to solve a perceptively noise and firefly free image.)
If LW doesn't cut it, use something else, but stop bugging developers to focus on unrealistic goals. Use LW when it fits the need, that's all i'm trying to say here.

Right now we use VRay because with the amount of data and resolutions we require would just not be feasible to do with LW (at least not with the budget and time we have, and let's not forget, the ability to hire freelance talent without having to educate them in a program they have never used, or worse, never even heard of.)


[Why do I think your response will be that your projects need the highest quality and LightWave cannot produce that. As if I would settle for less..]
Doesn't every one else's projects need the highest quality? I sure never heard of a client requesting the mediocre quality option (and no, i'm not implying LW's results are mediocre.)
It just comes down to time and money, and unfortunately, LW draws the shortest straw on that for us here.

Red_Oddity
09-21-2012, 03:54 AM
double post

dulo
09-21-2012, 04:49 AM
Janus is the best, but there is still much room for improvement on the user interface side...
Newtek should add something like a passes Node editor ( a better integrated extension to the things DP is doing with his buffer nodes .. )

sorry this should have been a quote for the janus message some posts ago ..
dont mind it ..

Netvudu
09-21-2012, 06:02 AM
I totally agree with duloīs opinion. It doesnīt matter that Janus is there. That would be like asking for no VPR because FPrime was there, while the truth is itīs vastly more useful to have VPR because of so many reasons (volumetrics, color space, integration, etc....)
We do need urgently a strong passes system to backup such a strong rendering engine.

tyrot
09-21-2012, 06:47 AM
unrealistic goals?

supporting octane natively from Lightwave would open lots of opportunities.
I prefer octane's directlighting (which is crazy fast with a decent nvidia cuda card) for any exterior shots. Do not mislead people please if you use VRAY use it.

All i am saying competing with Octane or VRAY shouldnt be unrealistic goal for Lightwave native renderer. Lightwave's flagship was rendering as far as i remember.. that is all i am saying..

If someone from NT says "guys it is unrealistic we never produce such renderer " ok then we can ask for may be an octane support natively at least .. that is all i m saying..

Red_Oddity
09-21-2012, 11:16 AM
unrealistic goals?

supporting octane natively from Lightwave would open lots of opportunities.
I prefer octane's directlighting (which is crazy fast with a decent nvidia cuda card) for any exterior shots. Do not mislead people please if you use VRAY use it.

All i am saying competing with Octane or VRAY shouldnt be unrealistic goal for Lightwave native renderer. Lightwave's flagship was rendering as far as i remember.. that is all i am saying..

If someone from NT says "guys it is unrealistic we never produce such renderer " ok then we can ask for may be an octane support natively at least .. that is all i m saying..

As an unrealistic goal i meant trying to turn LWs renderer into an unbiased spectral renderer.
As for Octane support, why is that Newtek's job? Shouldn't that be Otoy's job, after all, they sell the product.

Don't forget LWs renderer is very versatile, it is build to be able to handle as much situations and styles as possible (Jack of all trades, master of none.)
Octane is build (and VRay to that extend) to do one thing very well, and that is realism. Try to do stylized or any other look and it becomes difficult real fast (unless you do most of your look in post/compositing)


And just to add to what i think might be realistic goals that benefit not just current users but new ones (or ones that went to slightly differently shaded green grass on the other side) would be things like: bucket rendering, native OpenEXR render element support (sorry Mike), a render pass system, etc.

(And just to add, when i say Jack of all trades, master of none, i don't mean that as a slight or in a derogative manner.)

netstile123
09-21-2012, 02:23 PM
I Agree lightwave render needs work and think it should be noted in the lightwave gallery due to the fact it seems to me a large % of the renders were in kray and not native lightwave. I was working on a space animation and just gave up with lightwave. For realism and the computer I have I would be dead before I completed anything with just lightwave so I gave up and starting something new.

Cageman
09-21-2012, 03:51 PM
I totally agree with duloīs opinion. It doesnīt matter that Janus is there. That would be like asking for no VPR because FPrime was there, while the truth is itīs vastly more useful to have VPR because of so many reasons (volumetrics, color space, integration, etc....)
We do need urgently a strong passes system to backup such a strong rendering engine.

Here is what I think, from what I see in the line of productions we do...

1. Janus+exrTrader combination beats a lot of the competition
2. The lack of MicroPoly displacements is starting to hurt LW quite badly... just because there aren't any realistic alternatives to use. DPont do have such a tool, but it is Volumebased, and as such, takes too much a renderhit.

What I am saying here is that you can purchase tools for LW to deal with renderpass management, and as such, it is one of the best systems I've encountered for many reasons (including Maya and Max). So, with that said, you have a solution that you can use and it works very well. The same can't be said about MicroPoly displacements, and, LWs lack of them is starting to hurt it.

Netvudu
09-21-2012, 05:33 PM
Iīll admit you made a point cage. And as much as I would like to be able to play with native micropoly displacements, just for a moment donīt think on your own pipeline (which looks pretty cool by the way). How many users are suffering the lack of micropoly displacement vs the lack of native render passes. I donīt know there, but around here I know 0 people that might be Janus users. (I have to take a look at it. You guys are praising it a lot).
This being said....I hope both aspects get implemented in Lightwave.

Where are you suffering the lack of MP disp the most...on ZBrush-related work?

LW_Will
09-21-2012, 07:41 PM
there is also the "you also need 5 experts to write shaders, a bunch of people to massage it, oh and it cost 50 times more that your hideously expensive system for ONE frakking node for NO better quality" thing as well...

not that I'm bitter...

- - - Updated - - -

there is also the "you also need 5 experts to write shaders, a bunch of people to massage it, oh and it cost 50 times more that your hideously expensive system for ONE frakking node for NO better quality" thing as well...

not that I'm bitter...

Iain
09-22-2012, 01:58 AM
I Agree lightwave render needs work and think it should be noted in the lightwave gallery due to the fact it seems to me a large % of the renders were in kray and not native lightwave.

What are you basing that on? I know for a fact that a lot of the more realistic stuff in the gallery is 100% LW.
It's irrelevant though-any cg artist should have a few options open to them when it comes to rendering. Some engines deal with certain things better, just as some types of modelling are achieved a million times easier in ZBrush or whatever.

Cageman
09-22-2012, 01:16 PM
This being said....I hope both aspects get implemented in Lightwave.

Yes... At some point I do believe they will.


Where are you suffering the lack of MP disp the most...on ZBrush-related work?

That is the most obvious one, yes, but, in general, anything we do that we need details for closeup shots, be it hard surfaces or soft surfaces, we end up having millions of polygons. We could utilize a mix of displacement, normals, bumps and cavity maps, of course, but that doesn't allways work due to normalmaps sometimes generating artifacts depending on how the light is hitting the surface. Essentially, when you use normalmaps, in particular, to add that extra detail, you'll start to be a lot more concious about how you light, and as such, you might not be able to light something the way you want or need.

prometheus
09-22-2012, 05:15 PM
Yes... At some point I do believe they will.



That is the most obvious one, yes, but, in general, anything we do that we need details for closeup shots, be it hard surfaces or soft surfaces, we end up having millions of polygons. We could utilize a mix of displacement, normals, bumps and cavity maps, of course, but that doesn't allways work due to normalmaps sometimes generating artifacts depending on how the light is hitting the surface. Essentially, when you use normalmaps, in particular, to add that extra detail, you'll start to be a lot more concious about how you light, and as such, you might not be able to light something the way you want or need.

+1
For micropoly displacements then.

Michael

SkiddlyMcZ
09-22-2012, 10:27 PM
Maybe it's just me, but I find LW's render engine to be more than adequate for what I do.
I haven't been using it for a couple years now, and have been using mental ray and V-Ray, but I'm now using the LW 11 trial and being reminded just how nice the renders are and just how easy it is to set up, compared to mental ray.
It's smoking fast on my PC too. I haven't had a chance to really push it yet, but one thing I do know is I can get a render out of LW in a fraction of the time it takes me to set one up in MR, with adequate speed and quality.

Cageman
09-23-2012, 02:46 AM
Oh, and an observation that I've made is that... LightWave is far from limited regarding shading. I would argue that it is one of the more versatile engines nowdayds, when you start to dig into the nodes and using all kinds of techniques to generate data to feed the engine with, and the node-editor is amongst the easier ones to learn from an artist standpoint.

Cageman
09-23-2012, 03:05 AM
Oh, and an observation that I've made is that... LightWave is far from limited regarding shading. I would argue that it is one of the more versatile engines nowdayds, when you start to dig into the nodes and using all kinds of techniques to generate data to feed the engine with, and the node-editor is amongst the easier ones to learn from an artist standpoint.

Red_Oddity
09-23-2012, 04:08 AM
Oh, and an observation that I've made is that... LightWave is far from limited regarding shading. I would argue that it is one of the more versatile engines nowdayds, when you start to dig into the nodes and using all kinds of techniques to generate data to feed the engine with, and the node-editor is amongst the easier ones to learn from an artist standpoint.

That was a big surprise to me too, it shows you what a complete nodal system is capable of.
Also, with Denis' node kits LWs nodal shading capabilities even far outweigh some of the competition (the ease at which you can manipulate the ray tracer paths is quite amazing.)

jasonwestmas
09-23-2012, 07:37 AM
It's the lack of bucket rendering and efficient memory handling and the crappy render buffers that make LW render look old, but that's about it imo.

I totally agree about LWs flexible shading system, it's top notch.

Quote from another thread: "My personal opinion is that Lightwave 11 is still the fastest when it comes to achieving an optical sensibility that the user wants stylistically in a shading context. The nodal shading network in LW has a lot to do with that. Other renderers out there are more powerful in a "numbers crunching game", but that doesn't mean much when the goal is to achieve a specific "flavor" as an artist in your production. In other render engines it seems you are locked into a specific "look" and it's much harder to change that look and feel than it is in lightwave nodal.

So yeah definitely a great time to use lightwave. . . .I think LW12 has a great chance to turn things around completely as a must have toolbox for maya, max, softimage, modo, messiah artists."

SkiddlyMcZ
09-23-2012, 08:10 AM
Speaking of rendering, I think I found a nasty bug.
See this post (http://forums.newtek.com/showthread.php?130750-VPR-F9-render-discrepency-in-shadows)

Since I posted that I've been messing around and I discovered that this happens if I create one type of light, and then change it to another type.
In this particular case, had created a point light, tried out a spherical light, and then changed it back to a point light. After that it never worked right in an F9 render at all.

And I've tried that several times now, with the same effect each time - once you change it., it's broken in F9 but works fine in VPR.

Strangely enough, it doesn't seem to apply to the default light created in a new scene. I've changed that one several times and it always works no matter what, but any subsequent lights created get broken.

I guess maybe I should try it out more in more scenes before calling it a bug, but so far it certainly looks that way.

Amurrell
09-23-2012, 01:15 PM
I have to agree with the memory argument. I'm dealing with a project now with millions of polys and the process of moving everything into memory takes a long time and the render itself a couple of seconds litterally. Overall, I've never be dis-satisfied with the engine, though I do see some areas for improvement.

OFF
09-23-2012, 07:53 PM
:agree:

Cageman
09-25-2012, 04:14 PM
I have to agree with the memory argument. I'm dealing with a project now with millions of polys and the process of moving everything into memory takes a long time and the render itself a couple of seconds litterally. Overall, I've never be dis-satisfied with the engine, though I do see some areas for improvement.

I agree. It is the translationtime that hits the most. Renderingtime isn't affected much by the number of polygons. Fast Subpixel Displacements would arguably make LWs engine a very hard to beat engine, especially when thinking about how versatile it has become with all those nodes regarding shading (and other things).

It would also be nice to see a speedup in GI, especially the brute-force MC, but then again... there are ways to work to rectify that stuff quite easily with compositing...

On that note... It will be very cool to take a look at DPs new Image Filter Node that does Displacements as a post-effect. I'm waiting for the x64 version to pop up! :D

toby
09-30-2012, 01:03 AM
I still have 9.6 - is the nodal system "complete" now? Can you connect from shader to shader, and to other things in the scene? Can you separate the r,g,b, and a channels? Hopefully GI respects different shading models too.

50one
09-30-2012, 03:27 AM
Would be great to have the micropoly disp, bucket rendering, maybe two different method for Gi bounces, scene referencing and easier to use render passes, something that modo has, for instance when you want alpha for specific material, you just drop the alpha output to that shader and voila, easy. But i can already see the changes going in right direction, so i guess we all just need to sit and wait:)

toby
09-30-2012, 04:24 AM
Hopefully any new render pass system will allow you to add arbitrary layers, with different matte shaders allowed for each of them. On Real Steel we had paparazzi camera bulb flashes in the real footage that had to be matched in 3d, in vray we could render a dozen flash bulb lights on constantly and each would go to a separate buffer, to be animated on and off in compositing. That's in addition to being able to make multiple rgb matte passes without having to assign different shaders.

toby
09-30-2012, 05:11 AM
Matte objects not working as normal for reflection or GI just bit me in the *** again last week! That should have been taken care of *years* ago, is it still not in LW11?? It's very simple - camera rays see a solid color, reflection & gi rays see the original shader. If I knew how to program I could write it myself.

Cageman
09-30-2012, 12:39 PM
I still have 9.6 - is the nodal system "complete" now? Can you connect from shader to shader, and to other things in the scene? Can you separate the r,g,b, and a channels? Hopefully GI respects different shading models too.

Complete? No, not by a longshot... but it is quite intruging what you can do with it regarding a lot of things, epecially shading. What you are asking for are major architectural changes and they will take a lot longer than a couple of versions.

You can use Poms nodes for storing and sending shading data across surfaces (so, in essence, tweak one surface and you tweak them all). If you purchase Rebelhills nodal videos, you'll see that there are a lot you can do in that regard. You also have TrueArts Global materials which is pretty handy as well. But Poms node allows you to, for example, transfer what you do in the displacements node-editor and use the results of that in the shading, and vice-versa. You can separate channels if you want to... for example, we had a bunch of textures where R, G, B and A all controlled different things and needed to be mapped as masks/filters.

Nodal in LW11 isn't much different from LW9.6, except some nifty adjustments and tools for search/namefilter the nodes. Oh.. and you don't need to be zoomed in 100% in order to connect nodes. :) The main reason to get LW11 if you are into shading is the much faster and easier to use Unified Sampling implementation, which, in most cases, speeds up rendering. Oh, and the improvements to many nodes, of course... if you are on LW9.6 there should be many improvements to be found.

tburbage
09-30-2012, 01:02 PM
LWs renderer has come a long way, but quality vs speed wise it can't even come close to what we can do with VRay.
The project we do now, probably wouldn't even be able render as LW still needs to load EVERYTHING in memory, VRay on the other hand scales really nice, no matter how many billions of polygons and gigabytes of light/irridiance caches we throw at it.

One complaint I remember hearing in the past -- that there wasn't much optimization to evaluate what assets on a frame by frame basis are really needed in memory to complete the frame. With ray tracers though, this is not so easy to calculate because of things like ray traced reflections where the reflective object is "in camera" but the things its reflects may not be...

I also remember from quite a few years ago a comparison of renderers including Renderman, mentalRay, and LW where speed, texture sampling issues in a motion context (where you would get flickering you couldn't get rid of), and quality of motion blur were cited as LW problem areas. That was a long time ago though.

Cageman
09-30-2012, 01:57 PM
Matte objects not working as normal for reflection or GI just bit me in the *** again last week! That should have been taken care of *years* ago, is it still not in LW11?? It's very simple - camera rays see a solid color, reflection & gi rays see the original shader. If I knew how to program I could write it myself.

No, it isn't fixed in LW11. I have to blame Janus for never having to think about this type of stuff. :D

That said, with DPonts tools, it is quite easy to make all desired objects to behave exactly as you describe. It is done in the Node Pixel Filter, so, turning it off (one click) you are back where all objects are rendered normaly.

I've attached sample-content and two F9 renders (saved as 32-bit PNG so you can download and take a look in the alpha). Also, the very ugly cube in the middle is emitting 500% luminosity, but only visible for GI-rays, just to showcase that DPonts Pixel Filter (and the tools I use in there) are 100% supporting what you are suggesting.

I know that this is something that LW3DG should address, but in the meantime, there are solutions around the corner.... oh... and this setup I did, is done so that I only add the objects I want to render normaly in the list in the Mask node, all other objects that I add to the scene, will be seen as Matte-object by the camera, but still reflect correctly and illuminate for GI correctly.

Have fun!

:)

EDIT: The two images are also in the rar-file I attached.

erikals
10-01-2012, 07:06 AM
thank you Sir :]

wonder, is it FogBugzed? (recently?)

Cageman
10-01-2012, 04:08 PM
I don't think it is a bug.... it is more a design-thing that has not been adressed yet.

EDIT: If anything, it should be fogged as a Feature Request, I think...

:)

toby
10-01-2012, 05:47 PM
No, it isn't fixed in LW11. I have to blame Janus for never having to think about this type of stuff. :D

That said, with DPonts tools, it is quite easy to make all desired objects to behave exactly as you describe.
How do you avoid the issue with Janus? Don't you still need matted objects to relfect?

Yes it can be done with DP nodes, *if they're available on your system*, but wiring up the network, when you should have to do no more than click the matte object button, is not really any better than simply duplicating the object and setting one to unseen by camera and the other to matte. The issue isn't really the workaround, it's just how silly it is that the matte button sets reflection and GI rays to matte, and not just the camera.

toby
10-01-2012, 06:22 PM
duplicate post

geo_n
10-01-2012, 07:06 PM
How do you avoid the issue with Janus? Don't you still need matted objects to relfect?



I think he still needs to render two breakout scenes created by Janus. One for the beauty pass and the other for the matt pass with dpnfe. Still not ideal imho considering the render time.
If he could render both pass at the same time that would be interesting.

Cageman
10-02-2012, 02:55 AM
How do you avoid the issue with Janus? Don't you still need matted objects to relfect?

Through Janus I generate two scenefiles... one that renders with all the glory stuff, but the objects being Unseen by Camera, the other one is a pure matte-pass for the objects. The pure matte-pass is rendering extremely fast, so it doesn't matter.

- - - Updated - - -


Still not ideal imho considering the render time.

Matte-passes usually takes a couple of sec/frame to render in full HD... so I really don't see the issue.

Cageman
10-02-2012, 03:10 AM
The reason why I say I don't have to bother with this stuff, is because I am using a preset for Janus that I've setup and simply used for years now. I don't have to *think* about it. Janus do the things for me.

Also, understand this: I'm not saying that this isn't something that should be fixed. I do agree with the sentiment that it is silly, but I also understand how and why it is the way it is, and for the moment, I've choosen to find a solution that isn't too technical nor timeconsuming. And, just for the sake of argument: Fire up Maya and try to find a solution to this problem using mental ray... I'll be curious to see what you come up with, but, over here it looks like you have to approach this in a similar fashion to how LW does it, but way more complicated since it needs shader-overrides to become Matte-objects, unlike LW where it is a on/off toggle.

:)

geo_n
10-02-2012, 11:00 AM
Through Janus I generate two scenefiles... one that renders with all the glory stuff, but the objects being Unseen by Camera, the other one is a pure matte-pass for the objects. The pure matte-pass is rendering extremely fast, so it doesn't matter.

- - - Updated - - -



Matte-passes usually takes a couple of sec/frame to render in full HD... so I really don't see the issue.

Not an issue to get hourly changes from clients and have to update multiple master scenes and create specific matt/hold pass for those scenes and render them individually a second time? Time is gold in other studios.
Here's a tutorial for c4d.

http://vimeo.com/18932758
The video shows how much control you can have with the renderer that's easy to use and direct to the point just ticking checkbox. At the last part there's also the compositing tag to render out a beaty and matt pass together with no trouble.
Maya and mray. Still not on vray? :D

Cageman
10-02-2012, 12:29 PM
Does C4D have an absolute awesome nodeshader that can do spot-replacements (yes... we have something that is similar to vRoom done just with nodes).

You choose where to fight the battles, so to speak... and I rather have the flexibility of LWs shadingsystem and Janus than something far more limiting, but with "perfect" renderpasses system. When LW3DG will start to adress those areas of LW (renderpasses/layers etc), I think it is going to be an awesome solution...

Oh and regarding rendertimes... first of all... we have ONE masterscene / shot, so that ain't a problem, secondly, rendering mask-passes doesn't bog down the renderfarm. We have comps that can take several minutes / frame to render. In such cases, we use pre-comps for the stuff that takes time to manage renderspeeds for the master-comps... so, a matte/maskpass that takes 3-4 sec/frame doesn't really cost us time.

With well over 100 rendernodes during night, and 40 during daytime, it is very fast to render things.

:)

Oh... and vRay... no... we don't have vRay, nor do I think we will ever need it. We have other engines to use... ;)

Netvudu
10-02-2012, 01:40 PM
(yes... we have something that is similar to vRoom done just with nodes).

erm...could we have an insight on this? pleeeeeeeeeaaaase ;DDD


Oh... and vRay... no... we don't have vRay, nor do I think we will ever need it. We have other engines to use... ;)

Quoted for agreement.

m.d.
10-02-2012, 02:07 PM
yes please elaborate on the node setup....very interested

Cageman
10-02-2012, 02:31 PM
It is a further development of this: http://forums.newtek.com/showthread.php?86107-Fun-with-Nodes-Holo-instancing-Batman!&highlight=

I remember Hagerman told me there is more to investigate with this technique. :)

EDIT: I just have to say it is absolutely wonderfull to work with him here at Massive. I certanly can't get my head around what he is doing, but I am very happy to have him as a co-worker, and he surely can innovate on LWs shadingsystem. I'm very curious to see where he can take this technique... :)

geo_n
10-02-2012, 05:35 PM
You present a problem and a way other programs deal with it and you get how wonderful lw is in A, B aspects. Double Lol everytime you do that in the forums.
In the C4D tutorial its not even about just rendertime btw. The system there is quick and direct. Nodes are powerful but if there's a faster way why not. This is what C4D and 3Dmax have imho.
One master scene? You don't have a masters scene each for say
Scene 1, cameracut01F-098F, cameracut 099F-150F, cameracut150F-500F.
Scene 2, cameracut01F-300F, cameracut 301F-450F, cameracut451F-800F
Because those camera shots per scene could be one master scene each depending on how complex each one is. You have them in one scene? Interesting...

Its good to have other renderers indeed but after actually using vray for some time, its easy to understand why studios switch to this renderer. Maybe Arnold is better, if it has something other than photoreal renders that will be interesting and something to look at. Maya and mray is not a good combo for studios so don't know why the need to bring that up. That's why many switch to vray.
Of course with vray in a studio someone else will be hired that's more knowledgeable with vray and that means the user of the current renderer who doesn't want to learn vray will be bumped out unless that person takes time to learn vray, too. Happened to one of our TD, didn't want to learn vray and sticking to old ways as we were transition our rendering to vray. Got bumped out after more than 7 years working with the old ways.
Vray will have haters just because "haters gonna hate" what's popular. :thumbsup:

Cageman
10-03-2012, 12:47 AM
You present a problem and a way other programs deal with it and you get how wonderful lw is in A, B aspects. Double Lol everytime you do that in the forums.

Please enlighten me on my wrongs. Why does it seems to bother you so much that we use LW for rendering? And that we are very successfull with that? It seems to bother you that LW works so well for us.

You also doesn't seem to get what I am saying about LWs shadingsystem. Janus works very well for what it offers, and it allows us to keep using LW, which is something we want to do, thanks to the shadingsystem. What about that is so hard for you to understand?

Cageman
10-03-2012, 12:58 AM
Maya and mray is not a good combo for studios so don't know why the need to bring that up.

Because it is the renderengine that comes bundled with the package. It is quite silly that you should have to fork out an arm and a leg just to get descent rendering in an application, don't you think?



Vray will have haters just because "haters gonna hate" what's popular.

Ehh... !?

And again, you seem to ignore everything I've talked about regarding the reasons for us to keep using LW. Please...

geo_n
10-03-2012, 02:05 AM
Please enlighten me on my wrongs. Why does it seems to bother you so much that we use LW for rendering? And that we are very successfull with that? It seems to bother you that LW works so well for us.

You also doesn't seem to get what I am saying about LWs shadingsystem. Janus works very well for what it offers, and it allows us to keep using LW, which is something we want to do, thanks to the shadingsystem. What about that is so hard for you to understand?

Because it is the renderengine that comes bundled with the package. It is quite silly that you should have to fork out an arm and a leg just to get descent rendering in an application, don't you think?




Because like the Titanic the fanboys would not look at its flaws and only point at other things that are shiny. Tyrot already said stop your fanboism, too, in a different thread recently.
In another thread(one of many you praise lw too much) there's a comparison between hdinstance and native instancing and the fanboy inside you just wants to say lw is great.
It doesn't bother me that you use lw renderer. I use lw renderer everyday for personal projects. What's bothering is the fanboyism that is detrimental to progress. A lot of expert users who have experience with other software have left and when they bother to give insight how to improve lw, the fanboys just brush it off. Most of these veterans don't post anymore I notice.
A solution or workflow is presented from other software, no need to re-invent the wheel there. Why dismiss it and say shiny things? You can use Janus with lw all you want but a feature that just works in other software is a feature we could use in lw ,without needing to say lw doesn't need it, it has this shiny thing.
Your issue with unification, I know you said its trivial, you've never modelled in camera before for the 8 years you've worked in 3d. Good for you. But other people less than 8 years experience probably modelled in camera. They probably work on different stuff that needs it. Pointing out how maya and mray sucks just to raise the flag for lw over and over is not really helpful or even relevant since many studios adopted maya vray from said pipelines.
That is your wrongs. As I said before being a TD should be open minded. Please...fanboism...stop... nothing personal btw :D

Btw, if you think its silly to fork out an arm and a leg for a renderer then its silly to fork out an arm, a leg, a bladder for motionbuilder just to retarget mocap in lw. Just to retarget mocap!!! Other software has good mocap retargetting since more than 5 years ago.
The renderer cost doesn't seem silly in that respect since its THE part of the software that you use no matter what. No matter any project you will need a renderer unless you render invisibility. So its ok to pay for a good renderer imho. The lw renderer is great but I'm not going to ignore functions that are there on other renderers and not request them to be added.

erikals
10-03-2012, 03:51 AM
 
in my opinion the reason why LW works great for so many projects is because it's great at rendering exteriors.
and after all, about 95% of renders and SFX renders for TV/Film today are exactly that, exteriors.

Lightwave is also splendid for NPR.

the pitfall is shown especially when rendering photo realistic interiors. (like mentioned by so many)

there are new alternatives on the horizon though according to Rob Powers, (www.youtube.com/watch?v=nZb-oX1GPck#t=7m40s)
we'll just have to wait and see which render it will be...

i don't think Cageman is a fanboy, it's more of appreciating of what Lightwave can do.
and after all, if Lightwave renders 95% of your renders beautifully, that says a lot.

the problem with rendering with Lightwave basically emerges when aiming for photo realistic interiors.
+ interior animations


http://www.youtube.com/playlist?list=PL36DAD32935586C03&feature=edit_ok

erikals
10-03-2012, 07:42 AM
once again you harass...

not sure why you're not swamped in LW jobs. maybe it's more of an attitude problem.

erikals
10-03-2012, 07:58 AM
well, good to know why you're hanging out here.

erikals
10-03-2012, 08:09 AM
cool.

probiner
10-03-2012, 08:49 AM
Ai ai ai...
Erikals, I think Oliver has the right to question your "95%" figure relating it to experience, though I think it wasn't necessary to stomp your toes...

It's great that people in the forums are using other things and take time to talk about them.
Even though there were really troll characters around we must admit that most people criticizing LW here, do it so because they want it to be much more and even take time to explain it and present references. Though at times one wonders if pushing it too much will bring the whole house down or not.
But it makes everybody richer to acknowledge other solutions (and roadblocks) that go around out there.
Unfortunetly many left but it's kind of perceptible that some, have an underlying whim that one day, they can work with LW once again, but on their terms, which doesn't happen now.

As for the topic. I meddle very little with rendering and many times in an NPR approach, but would be interesting to know where rendering workflows are going. Is this a hint?:
https://www.youtube.com/watch?v=op-Dw1Ai91I&t=2m41s

Cheers

erikals
10-03-2012, 09:09 AM
not the first time, and i'm talking LW and TV/Film work. not Maya etc and 3D/Film work.

erikals
10-03-2012, 10:09 AM
not Lightwave, but take a look at "Snow White and the Huntsman" where it's close to a 100%

-NPR
-Exterior
-Exterior
-SFX Exterior
-SFX
-SFX Exterior
-SFX Exterior
-SFX Exterior
-SFX Exterior
-SFX
-SFX
-SFX Exterior
-SFX

that's half of the movie, didn't bother going through the second half, might be an interior render in there. (somewhere)
using Lightwave for this one would be quite do-able.

though not saying all movies are like this, maybe i should have said 90% instead, not 95...

jwiede
10-03-2012, 11:09 AM

about 95% of renders and SFX renders for TV/Film today are exactly that, exteriors.
Do you have any quantifiable evidence or data to support the stated premise, though? You know as well as anyone here that citing one empirical example and concluding anything from it has near-zero validity. Obviously, just because something looks like an exterior shot doesn't mean it is one.

3dWannabe
10-03-2012, 11:17 AM
Do you have any evidence or data to support the stated premise?

For 'Sanctuary' (not sure what software they used, maybe not LW as some of their creatures were a bit lacking), that figure might be reversed, but ... for for Tera Nova (which did use a lot of LW), that sounds about right?

jwiede
10-03-2012, 12:07 PM
For 'Sanctuary' (not sure what software they used, maybe not LW as some of their creatures were a bit lacking), that figure might be reversed, but ... for for Tera Nova (which did use a lot of LW), that sounds about right?
His citation was for industry in general, not any specific show, though. If you're going to make statements like that and expect them to be taken credibly, the backing needs to be more solid than empirical evidence.

Cageman
10-03-2012, 02:23 PM
Because like the Titanic the fanboys would not look at its flaws and only point at other things that are shiny. Tyrot already said stop your fanboism, too, in a different thread recently.

Lets just say that you do not know the whole story about Tyrot and me... we PMed eachother as well... :)



In another thread(one of many you praise lw too much) there's a comparison between hdinstance and native instancing and the fanboy inside you just wants to say lw is great.

Hmm... I said that in my experience, LW11 native instances are faster to render compared to the volumetric instances of HDI. And when I tested the scene provided, the native Instances were faster, even if it was just around 10 seconds or something in that magnitude.


It doesn't bother me that you use lw renderer. I use lw renderer everyday for personal projects. What's bothering is the fanboyism that is detrimental to progress. A lot of expert users who have experience with other software have left and when they bother to give insight how to improve lw, the fanboys just brush it off. Most of these veterans don't post anymore I notice.
A solution or workflow is presented from other software, no need to re-invent the wheel there. Why dismiss it and say shiny things? You can use Janus with lw all you want but a feature that just works in other software is a feature we could use in lw ,without needing to say lw doesn't need it, it has this shiny thing.

Re-read what I have written in this thread. I have never written anything like that. Where did I write that LW doesn't need a native solution? What I have written though, is that as it is right now, today, LW has a very functional plugin to deal with Renderpass management. Some people have said that they had to quit using LW because of lack of any renderpass/layer management, and not even bothered to look at what is a very usefull workflow until LW3DG can get something done natively. As such, if I could choose what to come first for advancing LWs renderpipeline further, I would not choose renderpass management, but instead things LW can't do, even with plugins, one example that is very dear to me is Micropoly Displacements (DPs one works, but is Volumetric and as such, very slow)... Ultimately though, LW3DG will do things in the order that works best for them, with insigts into the developments and where code are at different stages.


Your issue with unification, I know you said its trivial, you've never modelled in camera before for the 8 years you've worked in 3d. Good for you. But other people less than 8 years experience probably modelled in camera. They probably work on different stuff that needs it.

I would argue that in production, the need to adjust models in camera, rather than model them in camera, is a good start and something that should be quite trivial to implement, yes (it could be implemented as a vertex deformer). This alone would be much better to have ASAP and have that while we wait for in-camera modeling features (or the unification of the two programs).

I do base this on what I see in production here, where many different tools are used for modeling... a lot of the detailing today is done in apps that I don't think can import a camera. And, again, based on the experience I have with directors, art-directors and customers in general is that, in 99 out of 100 times, they want to change the animation, the camera, change lights, do this, do that (sometimes even adjust the positions and orientations of objects to make the best frame). In the end of the day, the built from cameraview model is either thrown a way or completely remodeled to work with the changes that happens. At this point, it will take longer to model this same model everytime things changes, than it is to model it properly from the beginning.

That doesn't mean I don't see the usefullness of in camera modeling (it just doesn't work that well for most of what we do), but, and this is a big but.... I would argue that all LW-users would be quite happy to be able to adjust their models in camera, in Layout, at least as a first itteration.


Pointing out how maya and mray sucks just to raise the flag for lw over and over is not really helpful or even relevant since many studios adopted maya vray from said pipelines.

This is a good argument, I think, because when you purchase Maya, you get 8 or 10 mental ray renderlics. You don't get vRay. So, it becomes an economical question... you also have to factor in what type of work is done, and how much of it is done as well. If we do 1 cinematic pre-render / year, the rest of the work is In-game cutscenes, or maybe one or two logos or whatever... would that justify the cost of 100 vRay lics? Especially when we are doing so well with what we allready have...

And, another argument is still related to all the shadingcontroll we have in LWs nodal interface that is quite atonishing and very flexible, and quite frankly, very artistfriendly as well. When you factor these things together, it becomes quite clear that vRay is not a necessary thing for us to use, and, furthermore, we have other engines apart from LW and MR to use, if need be.



Btw, if you think its silly to fork out an arm and a leg for a renderer then its silly to fork out an arm, a leg, a bladder for motionbuilder just to retarget mocap in lw. Just to retarget mocap!!! Other software has good mocap retargetting since more than 5 years ago.

I'm working with an animator who has been in the industry for 17 years. No matter what the Features says on the box of Maya, 3DSMax or XSI, MotionBuilder is still leaps and bounds more powerfull than those regarding Motioncapture. It is a fenomenal software for that stuff, because it is specialized on that area. As an example, we had an FBX containing well over 10.000 individual models (we can say it was a proceduraly modeled city). After 30 minutes of waiting for it to import into Maya 2012, I hit abort. Loaded up MotionBuilder 2012 and after two minutes, everything was loaded. I could then pick out the 100 or so objects I needed, save them out to a new FBX and bring into Maya. Imagine the speed it operates with lots of characters... it beats the competition by miles. And animators want snappy, fast, no jerk motion. So, MotionBuilder isn't just an expensive retargeting application, it is a very good animation application as well.



The renderer cost doesn't seem silly in that respect since its THE part of the software that you use no matter what. No matter any project you will need a renderer unless you render invisibility. So its ok to pay for a good renderer imho. The lw renderer is great but I'm not going to ignore functions that are there on other renderers and not request them to be added.

And neither do I.... the difference, I think, between what you and I say in this thread, is that your solution to LWs problems right now seems to be about not using it at all, but instead get vRay. This is an LW-forum, after all, and I suggest to people to get familiar with Janus. In the end though, I want everything native and nicely integrated, but that will probably take quite some time, so, again, my suggestions and praise of Janus, is because it exists now, and allows LW to be used quite comfortably with lots of renderpasses and layers.... and, it allows us to use the nodal shading that has become extremely good over the years.

:)

erikals
10-03-2012, 02:55 PM
"His citation was for industry in general"
no, though the industry in general for TVfilms / Films comes close.

i'm a bit surprised, i thought it was common knowledge.

toby
10-03-2012, 05:18 PM
we must admit that most people criticizing LW here, do it so because they want it to be much more and even take time to explain it and present references.
*Thank you*
I couldn't have said it better myself.

toby
10-03-2012, 05:47 PM
His citation was for industry in general, not any specific show, though. If you're going to make statements like that and expect them to be taken credibly, the backing needs to be more solid than empirical evidence.
I doubt he meant exactly 95%. He probably meant the vast majority, and I think he's probably right.

But interiors aren't lw's biggest limitation - the same studio that did Snow White also did Life of Pi, with a cg tiger splashing around in cg water. Just try doing hundreds of shots like that in lw - and really there's no reason to use it for it can do, when other renderers can do it all. It's main issues are inflexibility, memory management and narrow list of what it can do, imho. Render speed is another problem.

toby
10-03-2012, 05:51 PM
double post
jesus what is up with this forum

Netvudu
10-04-2012, 10:06 AM
Lets just say that you do not know the whole story about Tyrot and me... we PMed eachother as well... :)

This is a good argument, I think, because when you purchase Maya, you get 8 or 10 mental ray renderlics. You don't get vRay. So, it becomes an economical question... you also have to factor in what type of work is done, and how much of it is done as well. If we do 1 cinematic pre-render / year, the rest of the work is In-game cutscenes, or maybe one or two logos or whatever... would that justify the cost of 100 vRay lics? Especially when we are doing so well with what we allready have...

And, another argument is still related to all the shadingcontroll we have in LWs nodal interface that is quite atonishing and very flexible, and quite frankly, very artistfriendly as well. When you factor these things together, it becomes quite clear that vRay is not a necessary thing for us to use, and, furthermore, we have other engines apart from LW and MR to use, if need be.



Not to mention when Maya had no vray option, Autodesk marketing machine pushed and pushed Mental Ray whenever they could and many people now praising Vray were back then praising Mental Ray, and stating how wonderful it was because their SSS and skin and mia shaders that we couldnīt touch...now that Vray is around all the stuff is proven to be just marketing bulls**t, and itīs the turn for VRay to be praised.

In a couple of years Arnold will be the thing to use and VRay will be crapped all around...in the meanwhile weīll still be happy LW users, with a strong all-round renderer that keeps evolving and getting better through the years.

Obviously, when you donīt agree with geo_n you receive personal attacks just because he holds the absolute truth (i.e. when he states big studios are dropping Houdini..lol). In my case I get the Houdini fanboy tag (mind you, he could also tag me a LW fanboy ), cageman gets the LW fanboy tag (dang it, I want the tag too), while he mentions vRay in every single thread for the last three years (seriously, itīs 3 years http://forums.newtek.com/showthread.php?72492-We-need-VRay-for-LW!!!!!!!!!!!!!!!!/page23 ), but of course he is NOT a Vray fanboy...itīs us, begin wrong...go figure...

m.d.
10-04-2012, 01:27 PM
Renderers are an evolving tech....right now Arnold seems to be coming on strong....
I am a casual Houdini user...but have not dived too deeply into Mantra....
Whats your impression Netvudu....I know it is totally capable...phonton mapping, physically based and Reyes (i think)....
it was just the speed seemed a little lacking to me...

50one
10-04-2012, 01:46 PM
After using Vray for a while I started loving mental ray hehe - but that's n 3dsmax - might be a bit slow tho, same with vray - yes the output is great but really hard to find good ratio between render times & noise.

The whole Arnold situation is so just damn funny, I fully understand Solid Angle, but C'mon we all know that some builds are available here & there for people that want them to try, what's the point of making it sooo limited?
Why not just sell it "as is" to the folks that want to try it, create a forum and let the community gather the bugs and solve user issues?
This whole things make me scratch my head, looks like it's more a marketing gimmick "It so awesome - that is illegal in 9 countries, really hard to get...so you know it must be true!"

50one
10-04-2012, 01:47 PM
Deja vu

toby
10-04-2012, 02:02 PM
Not to mention when Maya had no vray option, Autodesk marketing machine pushed and pushed Mental Ray whenever they could and many people now praising Vray were back then praising Mental Ray, and stating how wonderful it was because their SSS and skin and mia shaders that we couldnīt touch...now that Vray is around all the stuff is proven to be just marketing bulls**t, and itīs the turn for VRay to be praised.
Sorry, but that's rubbish! Just because people liked MR then dumped it in favor of Vray doesn't mean it's all marketing BS. Lw is also not an option for maya / xsi users, so why would they push lw? Vray IS a lot better than MR, and may be the best renderer I've ever used. If I recall correctly MR had sss long before LW, and with maya it's more flexible. Vray had a *working* gi cache long before lw had a partially functioning one - is it still slower to render with the cache in lw? It's always been faster in vray, I was using it for production 6yrs ago, when a dual-core was the fastest machine you could get.

Cageman
10-04-2012, 02:37 PM
Sorry, but that's rubbish! Just because people liked MR then dumped it in favor of Vray doesn't mean it's all marketing BS. Lw is also not an option for maya / xsi users, so why would they push lw? Vray IS a lot better than MR, and may be the best renderer I've ever used. If I recall correctly MR had sss long before LW, and with maya it's more flexible. Vray had a *working* gi cache long before lw had a partially functioning one - is it still slower to render with the cache in lw? It's always been faster in vray, I was using it for production 6yrs ago, when a dual-core was the fastest machine you could get.

I know people who swear they can replicate what vRay does in MR... so... I think it is not just the renderengine it is about; the artists and their proficiency with any given renderer makes a difference as well. I'm curious to know what it is with MRs SSS that is more flexible nowdays compared to LW. It would be cool if you want to elaborate more about that.

Regarding LW and caching.... if you have static objects, LW is flying with a rocket-jet through the rendering. If you have animated objects, not so much. But, what I have noticed with LW11 when not using Use Gradients, is that a new sampling system is used (I think it is in the docs as well). This sampling system does a much better job of reducing flicker in Interpolated GI. I've had some interresting situations where I just used MC GI with interpolation, no cache, only 256 primary rays, 2 bounces, 64 secondary bounce rays, rendered on a renderfarm, and no flicker. That doesn't say that there are no room for a lot of improvements, and I guess I was either lucky, or the content used was just working very well, in a sort of magical way.

However... with that said, and this does indeed look like something has changed for the better; it isn't just sheer luck...

There are some tricks that I have revisited since LW9.6 days (if you remember my tutorial on LW regarding Animated Deforming Characters), and have noticed that one of those tricks, in particular, so far works extremely well. I'll need more testing to be done on that, but it is something that seems to be a lot more consitent nowdays, compared to LW9.6 days, no matter what characters I load into an environment.

I guess when the time is right, I'll wip up a videotutorial on that matter, and hopefully I will be able to use some productioncontent from work... we'll see.

:)

lwanmtr
10-04-2012, 02:46 PM
For my uses, Lightwave's native renderer work great. I've tried other solutions, but usually end up going back...not from a quality standpoint, but because it usually take me more time to set them up. I'm sure there are people who are now going to say, they can set them up just as fast, and thats fine...just not part of my own workflow.
From a render-time standpoint, LW renders pretty fast for me...fast enough I dont mind doing test renders on my 7 year old Macbook Pro...lol. (but, I would never say it couldnt be faster..lol)

That being said....

Render Passes really needs to be added. I often need to render passes for various materials and composting
Micro-Poly displacement sounds really cool too (showing the surface of the moon and zooming out to reveal Edward James Olmos for instance)..

But the biggest I would like to see is GI improvements. Specially eliminating GI crawl in animations. I fight that one alot.

Another that I havent seen mentioned much reguards Photometric lights....They are great...but how about including some presets for them? I know they are freely availble online, but unless you know what you're looking for, they are hard to find. I have a bunch of them...and honestly dont know what kind of light most them are because rarely have the ones i've found been categorized.

toby
10-04-2012, 03:18 PM
I know people who swear they can replicate what vRay does in MR
Well there's always one isn't there? MR motion blur is incapable of matching vray's, no matter how good the artist is. MR doesn't take rotation into account for motion blur. If you were to rotate a tire 180 in half a frame, all it's points would be blurred in a straight line across the center of the tire, resulting in a very small, blurry tire.



what I have noticed with LW11 when not using Use Gradients, is that a new sampling system is used (I think it is in the docs as well). This sampling system does a much better job of reducing flicker in Interpolated GI. I've had some interresting situations where I just used MC GI with interpolation, no cache, only 256 primary rays, 2 bounces, 64 secondary bounce rays, rendered on a renderfarm, and no flicker.
I don't turn off Use Gradients, you lose a lot of shape, it needs to be on to be as accurate as other gi renderers.

Cageman
10-04-2012, 03:24 PM
I don't turn off Use Gradients, you lose a lot of shape, it needs to be on to be as accurate as other gi renderers.

But you are on LW9.6, right?

toby
10-04-2012, 04:01 PM
But you are on LW9.6, right?
Yes, but I can't imagine it's changed that much, if it looks just as good now, why keep the option? You want to do some example renders?

erikals
10-04-2012, 06:41 PM
MC GI with interpolation, no cache, only 256 primary rays, 2 bounces, 64 secondary bounce rays, rendered on a renderfarm, and no flicker.

> sure sounds like something to check out though... (!)

Netvudu
10-04-2012, 11:54 PM
Yes, but I can't imagine it's changed that much, if it looks just as good now, why keep the option? You want to do some example renders?

Wrong! Youīve clearly not updated your info regarding the renderer. It has changed TWICE since 9.6 and the new Unified sampling affects A LOT the way it renders, and not just for AA, but also regarding GI. If you consider purchasing 11 you might want to purchase as well Rebel Hillīs cmiVFX GI tutorial where he explains why "Use Gradients" is not needed anymore and suggests to actively NOT use that option for faster less-flickering renders.

How can you be updated on Vray features but not in LW native rendererīs?

By the way, you are rendering an average of 20-50% slower by using 9.6 instead of 11 (it might be for economy reasons of course, just saying).

Netvudu
10-05-2012, 12:00 AM
Renderers are an evolving tech....right now Arnold seems to be coming on strong....
I am a casual Houdini user...but have not dived too deeply into Mantra....
Whats your impression Netvudu....I know it is totally capable...phonton mapping, physically based and Reyes (i think)....
it was just the speed seemed a little lacking to me...

Well, coming from LW I also find it a tad slow even if it allows for amazing tricks that almost no other renderer does (veeeery fast smooth motion blur).
Mantra is defined as a soft-look renderer which scales very nicely when you go up to 2K or above resolutions, so pretty much focused on feature films res.
For the LW user the "hard" part ainīt rendering, but shading, which is indeed slower than our handy layers/nodes setup.

I still canīt get the same quality from Mantra that I get from LWīs renderer, but Iīm getting quite close lately. In the end, itīs a matter of knowledge.

geo_n
10-05-2012, 04:46 AM
Lets just say that you do not know the whole story about Tyrot and me... we PMed eachother as well... :)



Tyrot and others have a point though. Sometimes too much fanboism when something is pointed out about newtek and lightwave in general, even the website forum, and you feel the need to defend it in one way or the other, is just regressive.



Hmm... I said that in my experience, LW11 native instances are faster to render compared to the volumetric instances of HDI. And when I tested the scene provided, the native Instances were faster, even if it was just around 10 seconds or something in that magnitude.


No. Your first post was "Build a jungle and turn on GI... lets see which one is faster."
Its not constructive and belittles the effort of the OP. The OP just said after "Not sure how you meant the remark.. Not quite trivial scenes we are applying instances to...."
Do you not see? Kind of egotistical about knowledge with lw.




Re-read what I have written in this thread. I have never written anything like that. Where did I write that LW doesn't need a native solution? What I have written though, is that as it is right now, today, LW has a very functional plugin to deal with Renderpass management. Some people have said that they had to quit using LW because of lack of any renderpass/layer management, and not even bothered to look at what is a very usefull workflow until LW3DG can get something done natively. As such, if I could choose what to come first for advancing LWs renderpipeline further, I would not choose renderpass management, but instead things LW can't do, even with plugins, one example that is very dear to me is Micropoly Displacements (DPs one works, but is Volumetric and as such, very slow)... Ultimately though, LW3DG will do things in the order that works best for them, with insigts into the developments and where code are at different stages.


That's why I said I don't have a problem with Janus.
I only pointed out that other software can deal with it better(by not having to re-render twice) and posted a video tutorial that does it and lw devs can adapt it to lw. Lw devs don't really have time to tinker with other software when they're busy coding. It should be pointed out to them.
We both agree that lw needs a render pass management but priority is different. Its not so for you because of Janus. That's ok. So we don't agree on is the road to get that lw render pass management. You go for Janus, I will go for native development.




I would argue that in production, the need to adjust models in camera, rather than model them in camera, is a good start and something that should be quite trivial to implement, yes (it could be implemented as a vertex deformer). This alone would be much better to have ASAP and have that while we wait for in-camera modeling features (or the unification of the two programs).

I do base this on what I see in production here, where many different tools are used for modeling... a lot of the detailing today is done in apps that I don't think can import a camera. And, again, based on the experience I have with directors, art-directors and customers in general is that, in 99 out of 100 times, they want to change the animation, the camera, change lights, do this, do that (sometimes even adjust the positions and orientations of objects to make the best frame). In the end of the day, the built from cameraview model is either thrown a way or completely remodeled to work with the changes that happens. At this point, it will take longer to model this same model everytime things changes, than it is to model it properly from the beginning.

That doesn't mean I don't see the usefullness of in camera modeling (it just doesn't work that well for most of what we do), but, and this is a big but.... I would argue that all LW-users would be quite happy to be able to adjust their models in camera, in Layout, at least as a first itteration.


That's a better way to word it instead of what you posted before calling unification, modelling in cam trivial in your 8 years of cg experience. Btw, I have modelled in camera for set extension and everytime I do it I avoid lightwave. I didn't expect this feature before but I will expect it for future versions of lw. Its just foolish not to.




This is a good argument, I think, because when you purchase Maya, you get 8 or 10 mental ray renderlics. You don't get vRay. So, it becomes an economical question... you also have to factor in what type of work is done, and how much of it is done as well. If we do 1 cinematic pre-render / year, the rest of the work is In-game cutscenes, or maybe one or two logos or whatever... would that justify the cost of 100 vRay lics? Especially when we are doing so well with what we allready have...

And, another argument is still related to all the shadingcontroll we have in LWs nodal interface that is quite atonishing and very flexible, and quite frankly, very artistfriendly as well. When you factor these things together, it becomes quite clear that vRay is not a necessary thing for us to use, and, furthermore, we have other engines apart from LW and MR to use, if need be.



I'm working with an animator who has been in the industry for 17 years. No matter what the Features says on the box of Maya, 3DSMax or XSI, MotionBuilder is still leaps and bounds more powerfull than those regarding Motioncapture. It is a fenomenal software for that stuff, because it is specialized on that area. As an example, we had an FBX containing well over 10.000 individual models (we can say it was a proceduraly modeled city). After 30 minutes of waiting for it to import into Maya 2012, I hit abort. Loaded up MotionBuilder 2012 and after two minutes, everything was loaded. I could then pick out the 100 or so objects I needed, save them out to a new FBX and bring into Maya. Imagine the speed it operates with lots of characters... it beats the competition by miles. And animators want snappy, fast, no jerk motion. So, MotionBuilder isn't just an expensive retargeting application, it is a very good animation application as well.


This same reasoning came be brought up with any other companion that lw lacks imho. Its an econimical question.
You buy lightwave, it is up to you to buy a rendercontroller, rhiggit, exrtrader, kray, motionbuilder, janus, turbulence, lwcad, etc.
My point is its not ridiculous to pay for a renderer because its something you use no matter what. Buy the best, fastest renderer for your studios needs. Buy motionbuilder if you need more than what the native mocap retargetting is available. So the point that its ridicilous to buy a renderer is case by case like any other software since it is an economical question. Is it ridiculous to buy vray as you said? A big chunk of C-AAA caliber studios might argue against that.




And neither do I.... the difference, I think, between what you and I say in this thread, is that your solution to LWs problems right now seems to be about not using it at all, but instead get vRay. This is an LW-forum, after all, and I suggest to people to get familiar with Janus. In the end though, I want everything native and nicely integrated, but that will probably take quite some time, so, again, my suggestions and praise of Janus, is because it exists now, and allows LW to be used quite comfortably with lots of renderpasses and layers.... and, it allows us to use the nodal shading that has become extremely good over the years.
:)


I only posted that its not ideal to render a simple matt again making it twice for every scene and every scene iteration that needs it. You will be wasting effort making unique matt pass scenes and rendering matt passes for hundreds of scenes and its iteration when it could be done as shown on the video tutorial. The render buffer panel in lw 11.03 is under developed and is open to concepts, functions found on that video tutorial.
I don't recall posting not to use lw. I even said I use lw for personal project everyday.
You brought up maya and mray being poor compared to lw and Janus. That's not really a fair comparison or up to date. Ofcourse lw would look good because maya and mray is troublesome that's why studios used other easy to use renderers including lw, fprime, modo, vray, etc.
I brought up vray because it is how most maya 3dmax C-AAA studios are set up now. Make comparisons from there. If Arnold was being used extensively with C-AAA studios then we make comparisons there. Not maya/mray. :D

geo_n
10-05-2012, 05:25 AM
Netvudu, you're very funny but I think we need to take english classes together. vamos amiga
You facepalm me FIRST for posting a hypothetical theory on independent 3d software that's remaining posted by celshader, and which one might sellout(your beloved houdini) given that Lux, as successful as it was, imho anway, still sold out to the Foundry. We can continue that conversation there. But I'm not interested in conversing with houdini fanboys in a lw forum even though I mentioned often that I like houdini as a software.

Netvudu
10-05-2012, 06:08 AM
Before questioning my english understanding, you have to write 646 posts in pseudo-proper spanish. Then we might be even. By the way, itīs spelled "anyway", not "anway".

It would be interesting to know why you keep on pushing ideas as opposed to debating them which looks like the civilized route IMHO.

probiner
10-05-2012, 08:11 AM
Guys, guys, you're losing focus on the topic starting to name call (fanboy with a tone) and hiting low (spellling).

Beyond the software itself a question that might hurt more is if the next generation CG artists will be LW users...

Cheers

Hail
10-05-2012, 08:39 AM
Sbowie, where art thou..?? :D

Hieron
10-05-2012, 09:09 AM
....

But, what I have noticed with LW11 when not using Use Gradients, is that a new sampling system is used (I think it is in the docs as well). This sampling system does a much better job of reducing flicker in Interpolated GI.



Ah, that explains why HDI now tends to crash sometimes when gradients is turned off.. we just figured that bug out and I remembered your post. HDI needs gradients on...

Imho it gives more detail in a GI solution too.. and more detail/contrast would surely lead to more flicker.. So it depends on the situation..


Well there's always one isn't there? MR motion blur is incapable of matching vray's, no matter how good the artist is. MR doesn't take rotation into account for motion blur. If you were to rotate a tire 180 in half a frame, all it's points would be blurred in a straight line across the center of the tire, resulting in a very small, blurry tire.


I don't turn off Use Gradients, you lose a lot of shape, it needs to be on to be as accurate as other gi renderers.

It does give more detail here as well... would need to test things further I guess.
Reducing contrast/detail from a GI solution would surely help to reduce splotches.

Would need testing I guess, but I had my share of (not quite appreciated) tests for the week.

Cageman
10-05-2012, 03:43 PM
Tyrot and others have a point though. Sometimes too much fanboism when something is pointed out about newtek and lightwave in general, even the website forum, and you feel the need to defend it in one way or the other, is just regressive.

Another description would be tolerant and understanding. But you can of course call me a fanboy, because personal attacks and quotes of what other people are doing is the only thing you seem to have as an argument these days. We are who we are, and not jumping on bandwagons makes us stand out, I get that.. so, please, continue to call me a fanboy if you want....



No. Your first post was "Build a jungle and turn on GI... lets see which one is faster."
Its not constructive and belittles the effort of the OP. The OP just said after "Not sure how you meant the remark.. Not quite trivial scenes we are applying instances to...."
Do you not see? Kind of egotistical about knowledge with lw.

Correct... I was very short on that answer, but I think I made up for it later on in that thread. I'm also quite convinced that the OP (Hieron) and I in that thread are using Instancing very differently, and as such, we have seen different aspect of HDI vs Native, which both have their advantages and disadvantages that goes both ways.

What are you doing for the LW community though?



That's why I said I don't have a problem with Janus.
I only pointed out that other software can deal with it better(by not having to re-render twice) and posted a video tutorial that does it and lw devs can adapt it to lw. Lw devs don't really have time to tinker with other software when they're busy coding. It should be pointed out to them.
We both agree that lw needs a render pass management but priority is different. Its not so for you because of Janus. That's ok. So we don't agree on is the road to get that lw render pass management. You go for Janus, I will go for native development.

Correct... as it stands right now, there are things LW needs improvments on that isn't related to passmanagement. With that said, LW3DG do what they need to do and in the order they need to do it. I can't change that. Whatever they do, though, I think it will be welcomed!



That's a better way to word it instead of what you posted before calling unification, modelling in cam trivial in your 8 years of cg experience. Btw, I have modelled in camera for set extension and everytime I do it I avoid lightwave. I didn't expect this feature before but I will expect it for future versions of lw. Its just foolish not to.

In my 8 year cg-experience in commercials and Pre-rendered cinematics, I've not seen a lot of "in camera modeling" that worked out. You can't compare Set-extensions with the regular modeling tasks though. With camera-projections, set-extensions can in most cases be done with primitives by adjusting the verts. The more elaborate set extensions usually mean properly modeled models, and a properly modeled model, usually means the whole kit, that is then trimmed down for the actual camera angle.

One of my co-workers developed a quite cool technique using LW+Nodal displacements+Projections+PS. He painted the whole set from a single image, and that images was then projected onto an object that was 1 milimeter larger (Normal-displacement for the win) than the object that needed raytraycing. So, pretty much everything was painted in (details and even lights), and that co-existed with the reflective stuff that needed raytracing. It looked like it was vRay on steroids, but it was only a couple of min/frame in 720p. He didn't even need to adjust things in Layout, thanks to the normal displacment for the projection geometry. And in the process, he optimized his render in a ******* awesome way. The point I need to make with this example though, is that it wouldn't have worked without a proper model in the first place.



This same reasoning came be brought up with any other companion that lw lacks imho. Its an econimical question.
You buy lightwave, it is up to you to buy a rendercontroller, rhiggit, exrtrader, kray, motionbuilder, janus, turbulence, lwcad, etc.


We don't rig in LW. We don't do effects in LW (no lic of Turbulence here), we don't use kray.

What we do in LW though, is to use the very unapprichiated (by you it seems) Nodal shadingsystem, and it seems that in every response you give, you avoid this subject.


Buy the best, fastest renderer for your studios needs.

Currently, no engine fits us better than LW for polygonal rendering. That is thanks to the Nodal shading system primarely. Without Janus and exrTrader, it would have been a hell of a lot tougher.


Buy motionbuilder if you need more than what the native mocap retargetting is available.

The animators here do prefer mbuilder over Maya for several reasons, even for handkey stuff. Not in all situations of course... so, MBuilder to us is not just a mocap retargeting system; it is a character animation tool as well.



So the point that its ridicilous to buy a renderer is case by case like any other software since it is an economical question. Is it ridiculous to buy vray as you said? A big chunk of C-AAA caliber studios might argue against that.

I am speaking from what we do here where we work... I don't speak for the studios you are refering to. I don't think it is redicilous for them, because they apparently need it. We do not. Get over it, please.

Arnold, on the other hand, would fit us much better than vRay, as an example...



I only posted that its not ideal to render a simple matt again making it twice for every scene and every scene iteration that needs it. You will be wasting effort making unique matt pass scenes and rendering matt passes for hundreds of scenes and its iteration when it could be done as shown on the video tutorial.

Over the years I've wittnessed these things and how they work in practice. Even with a native implementation of renderpass management/renderlayers etc in LW, my number 1 feature-request would be to have the option to export these renderpasses to separate scenefiles. There are many benefits to this, especially when used in conjunction with object referencing (which is part of LWs native workflow allready). Hopefully, I'll be showcasing more Janus-stuff regarding this, and how it actually makes us more flexible, compared to dumping one single file on the farm and hope that it doesn't crash or that some passes gets screwed up.

There is also the communication with the compositing artist that makes things even more efficient. Far from everything we render needs any mask-passes at all. We also make concious desicions on this, and many things regarding masks, normal and point position passes (for relighting in Nuke) are done on a case by case basis.

I really whish that I could invite you here to Massive so that you could see all the automation we have regarding these things, and how tightly integrated things have become, and, as such, LW is fitting extremely well into all this. Not just LW, but Maya, Max, Motionbuilder and Modo... and Nuke, of course.



I don't recall posting not to use lw. I even said I use lw for personal project everyday.

Your touting of vRay and your "...I use LW for personal projects only..." does indeed give the vibe that everyone should get vRay because it is soo much better. You have to understand that I know perfectly well what vRay does, but it isn't a very versatile engine when it comes to developing looks. It is an engine that should be used if you need to go photoreal, or need GI fast, but that is far from the only looks we need where I work... and there are many people in this thread who have said the same thing.

Arnold, again, looks like a much better engine in that regard compared to vRay. Even mental ray is a better engine for creating a lot of very different looks. And then again, LW has that nodal shaiding system that I keep comming back to...


You brought up maya and mray being poor compared to lw and Janus. That's not really a fair comparison or up to date.

As long as mental ray is the renderengine that comes with the package, it is as up to date as it can be. I would never, ever use kray as an excuse to get LW (even if it is a fantastic engine). I rather try to focus on what you get out of the box, and then start talking about alternatives to what you should get as plugins. But the sole reason for us using LW is because of the shading system, easy to use, comparably fast, tons of very usefull third party nodes that can do things I havn't seen in other apps, or at least as easly done. That and the integrated renderengine... then you start looking at ways to simplify and streamline the process... exrTrader (instant purchase), Janus (instant purchase) and then everything from DP (free).

I would also argue that LW as a hostpackage for a renderengine is complex enough to feed the engine with all kinds of goodies, but also provides a very small window for errors. Maya, as an example, is a fantasticly open software, but it adds to the complexity and when that happens, a third party engine will suffer from user-errors on a much larger scale compared to a less expanded product like LW. Max is a much better application to use for vRay compared to Maya, but that is a result of Max being, simply put, a more simplistic software in both function and architecture. Max, in comparsion to LW, is more clunky compared to LW, because it has more complexity in it, but safe to say is that it is, in most cases, a better choice over Maya, at least in terms of off-the-shelf tools for FX-work and rendering.


Ofcourse lw would look good because maya and mray is troublesome that's why studios used other easy to use renderers including lw, fprime, modo, vray, etc.
I brought up vray because it is how most maya 3dmax C-AAA studios are set up now. Make comparisons from there. If Arnold was being used extensively with C-AAA studios then we make comparisons there. Not maya/mray. :D

I ventured into this in the previous quote... but... yes... vRay is much better in Maya compared to mental ray, but that doesn't mean that you will end up troubleshooting the renderengine for an hour before you find out that it was something in Maya that got screwed up.

The most stable and predictable rendering I've ever experienced in Maya was when we used Maya Software (6 years ago now... no... 8 years... ****!), when we did a musicvideo (all CG btw) because it worked so much better with the tools and I could predict it. Just like I can predict LW.... and that is something to think about when choosing a tool for rendering, no matter what fancy features there are.

:)

toby
10-05-2012, 05:32 PM
Wrong! Youīve clearly not updated your info regarding the renderer. It has changed TWICE since 9.6 and the new Unified sampling affects A LOT the way it renders, and not just for AA, but also regarding GI. If you consider purchasing 11 you might want to purchase as well Rebel Hillīs cmiVFX GI tutorial where he explains why "Use Gradients" is not needed anymore

Example renders? Maybe I can dl a trial.


How can you be updated on Vray features but not in LW native rendererīs?
?
Probably the same way you're updated on LW but not Renderman...

toby
10-05-2012, 06:56 PM
Ok just tried the LW11 Demo version, and turning off Use Gradients has the same effect as before. You might not notice it if you have a lot of cg (non-gi) lighting like exteriors, large flat surfaces like buildings, or if your interpolation settings are low-res already. My MPS was 0.5 for example, and rendering human faces with backdrop gi & no cg lights.

3dWannabe
10-05-2012, 07:46 PM
Hopefully, I'll be showcasing more Janus-stuff regarding this, and how it actually makes us more flexible, compared to dumping one single file on the farm and hope that it doesn't crash or that some passes gets screwed up.


Yes, Yes!!! I live for your video tutorials!!!

geo_n
10-05-2012, 08:35 PM
Before questioning my english understanding, you have to write 646 posts in pseudo-proper spanish. Then we might be even. By the way, itīs spelled "anyway", not "anway".

It would be interesting to know why you keep on pushing ideas as opposed to debating them which looks like the civilized route IMHO.

And your civilized route is to facepalm me just because I posted something negative about houdini. Please...amiga..lets go to english class together. You improve my spelling already. :D
You must read about my other post about houdini being superior in modelling due to its procedural/parametric nature and how a veteran lw'ver,creator of Janus, seems to have switched to it. But it seems you only focus on the negative things said about houdini. You can post how houdini is a "league on its own" in a lw forum, but people can't post something negative (hypothetical even) about it. Lol!
Anyway we are getting off topic. A comparison of lw renderer, prman, mantra renderer, vray renderer, arnold renderer, modo renderer is a more interesting topic. And something to back up facts not hearsay or I know a guy who did this, etc.

geo_n
10-05-2012, 09:15 PM
Another description would be tolerant and understanding. But you can of course call me a fanboy, because personal attacks and quotes of what other people are doing is the only thing you seem to have as an argument these days. We are who we are, and not jumping on bandwagons makes us stand out, I get that.. so, please, continue to call me a fanboy if you want....



Correct... I was very short on that answer, but I think I made up for it later on in that thread.



You can call me a vray fanboy if you want. I won't take offense. But if I become an Arnold fanboy then I'll take offense if you still call me vray fanboy. :D
What is regular modelling task? Its regular for some people doing matt painting and using 2d and 3d software to accomplish that job. Its regular for archiviz people to want modelling in camera. its not trivial as you say it is. Would you call your very example of your use of motionbuilder a regular task? I wouldn't call it trivial if you use mbuilder that way. It works for you.

You were not short on that answer in instancing thread. I don't think people need to read between the lines.

Just because one does video tutorials ...well..I'll stop there..What I have done for lw is "rant" about what could be improved in lw by pointing out workflow and features in other software. So I will be labeled as a basher. Big deal. We are all paying customers lest we forget, all are equal. :D
Some want more modelling improvement, better communication with mbuilder, some want better animation tools, etc. Its free to post requests.
The problem with "some" in the community is that a suggestion or pointing out flaws in lw is bashing.
I don't recall being unappreciative of the nodal shading system. I recall posting a tutorial how other software deal with render elements and a rebuttal how Janus makes it that you don't need to think about this stuff. The concept of direct and indirect or answering a question with a question comes to mind. The fact is even with Janus, you will still need to render it twice for each scene.
If you understand vray then please show work you yourself have been involved using vray. Because you only made comparisons with lw plus Janus against mray and mentalray. Why you didn't mention you use vray so its a fair comparison.

jwiede
10-09-2012, 12:29 PM
The problem with "some" in the community is that a suggestion or pointing out flaws in lw is bashing.
Agreed, I feel like this behavior has become rampant of late here, to the point where it's actually alienating customers and potential customers. When a customer presents a problem that's a "pain point" for them, having other users jumping down their throat arguing their usage isn't "relevant", or hijacking the thread to present alternate, unrelated scenarios "proving LW's side", the message sent to that customer is "your needs don't matter". Customers get gunshy about even discussing problems. They come away with the sense that because Newtek does nothing to stop it, they must agree with and support the responders' premise that their needs and problems aren't significant.

I know many customers who've basically given up and left these forums, saying such feelings and that they felt that posting their real LW issues and problems only led to them being attacked and derided. Many of us here have heard other customers explicitly state such complaints in their "goodbyes". I believe the problem feels worse now than even six months ago. These days, anyone not constantly singing LW's praises is made to feel quite unwelcome here IME, regardless whether they're LW customers stating actual issues.

Cageman
10-09-2012, 01:32 PM
Agreed, I feel like this behavior has become rampant of late here, to the point where it's actually alienating customers and potential customers. When a customer presents a problem that's a "pain point" for them, having other users jumping down their throat arguing their usage isn't "relevant", or hijacking the thread to present alternate, unrelated scenarios "proving LW's side", the message sent to that customer is "your needs don't matter".

So... what you are saying is simply... never mention that there are third party tools to deal with some of the shortcomming now? There are two things to concider here, I think...
1) What can be done now?
2) What should LW3DG do for the future? (regarding implementations, ideas etc).

Simply stating that "in app A this is soo much better, and in LW it is non-existant", is, in some cases, a half-truth. Renderpassmanagement, an outright lie!

This doesn't conflict with what LW3DG should do to implement features. It is simply highlighting that in many cases, you aren't left out in the cold; very much thanks to third party developments.

Lets say that when Recoil was released, and not yet a bundled plugin with modo, and someone told you that modo doesn't have any dynamics engine... would you simply state how great Houdini is, or would you say that Recoil works quite nicely and has a good implementation despite being a plugin.



Customers get gunshy about even discussing problems. They come away with the sense that because Newtek does nothing to stop it, they must agree with and support the responders' premise that their needs and problems aren't significant.

Should the discussion only be one-sided? I mean, in the sense that, if someone has a problem, there is a solution to the problem, but since it isn't native, it shouldn't be worth mentioning?

As much as this thread is about highlighting problems, it should also be about highlighting solutions, right? Because, think about it... if that was the rule for a thread like this, it would suddenly look like LW is completely incapable. If these solutions could become more elegant, or native, or better... that is also an important thing to discuss. I admit that I have voiced my opinion about LW+Janus+exrTrader very strongly, but that is also because it has worked out so f*cking great, for us. If it was total crap, and made our day-to-day rendering on projects a huge hassle, would I promote it so strongly as a solution that can be used today, while waiting for something native comming our way?


I know many customers who've basically given up and left these forums, saying such feelings and that they felt that posting their real LW issues and problems only led to them being attacked and derided. Many of us here have heard other customers explicitly state such complaints in their "goodbyes". I believe the problem feels worse now than even six months ago. These days, anyone not constantly singing LW's praises is made to feel quite unwelcome here IME, regardless whether they're LW customers stating actual issues.

Again, there are issues and then there are issues that has solutions. These solutions might range from inginious to tedious, but they do the job.

Looking at it from a customer standpoint, where a thread would be only pointing out the flaws and not even hinting towards solutions, would that reflect the capabilities of a software? Give or take that some solutions are plugins etc?

As much as your sentiment is that "why can't we discuss the problems here...", I know many users who have stopped participating because they get the fanboy stamp, just because they try to highlight solutions or alternative routes, that still works within the package these pages are about; LW.

Netvudu
10-09-2012, 01:37 PM
Agreed, I feel like this behavior has become rampant of late here, to the point where it's actually alienating customers and potential customers. When a customer presents a problem that's a "pain point" for them, having other users jumping down their throat arguing their usage isn't "relevant", or hijacking the thread to present alternate, unrelated scenarios "proving LW's side", the message sent to that customer is "your needs don't matter". Customers get gunshy about even discussing problems. They come away with the sense that because Newtek does nothing to stop it, they must agree with and support the responders' premise that their needs and problems aren't significant.

I know many customers who've basically given up and left these forums, saying such feelings and that they felt that posting their real LW issues and problems only led to them being attacked and derided. Many of us here have heard other customers explicitly state such complaints in their "goodbyes". I believe the problem feels worse now than even six months ago. These days, anyone not constantly singing LW's praises is made to feel quite unwelcome here IME, regardless whether they're LW customers stating actual issues.

Youīre perfectly entitled to this opinion. I just want to point out some customers experiences are different to others.
Because I work on a training center where we "arenīt married" to a brand (not as easy to find as you might imagine) I can tell you the customer service we recieve from Newtek is first-class and waaaaay better than Autodesk, for instance. Autodesk is in fact almost the worse of the market (the award belongs to Chaos, who has repeteadly made us wait many months for already-paid licenses, severely hurting the students in the process).

For us Newtek is almost faultless in their support. And I mean Newtek USA, not Europe.

erikals
10-09-2012, 02:00 PM
i've noticed that sometimes when i suggest an alternative route to a problem using LW certain people mistakenly read it as,
"LW can do this easily"

which is not the case.

keep asking for features, but if guy X gives you an alternative LW route, don't read it as "LW can do this"

take this thread for example, i'm sure it gets NTs attention

and they are working on something >
http://www.youtube.com/watch?v=nZb-oX1GPck#t=7m40s
(that's a plugin though, afaik)

Cageman
10-09-2012, 02:16 PM
You can call me a vray fanboy if you want. I won't take offense. But if I become an Arnold fanboy then I'll take offense if you still call me vray fanboy. :D


Darn... I had hoped to get that one as well. :D Truthfully though, for the very different types of work we do, Arnold and LW (nodal shading and all that stuff), looks to me to be the easiest to work with to get the looks; Arnold being a super-impressive engine. :)


What is regular modelling task? Its regular for some people doing matt painting and using 2d and 3d software to accomplish that job. Its regular for archiviz people to want modelling in camera. its not trivial as you say it is. Would you call your very example of your use of motionbuilder a regular task? I wouldn't call it trivial if you use mbuilder that way. It works for you.

Building something from camera with depth... without saying too much... but something like that happened today at work... Something that was modeled from a camera-angle got thrown away and a proper model had to be built... guess why? The Director. :D The in camera model did work out as a quick n dirty fix for the time being, before the director stepped in and had a look.

I just have to assure you that I know how neat it is to be able to throw thing together in camera-view, I know! But, for that to happen in LW it will take quite some time, and I am not opposing it, but I would rather see tools that allows artists to adjust their models in Layout (as I have written earlier in this thread), so a matte-painter, for example, easily can select some verts, move them around in Layout, to better fit the mattepainting.

The nuts and bolts of true modeling in Layout (or whatever happens on that front), is something I can wait for, if I get some vertex deformer tools in Layout in the meantime. :)


You were not short on that answer in instancing thread. I don't think people need to read between the lines.

What do you mean?



Just because one does video tutorials ...well..I'll stop there..What I have done for lw is "rant" about what could be improved in lw by pointing out workflow and features in other software. So I will be labeled as a basher. Big deal. We are all paying customers lest we forget, all are equal. :D

I never labeled you as a basher, but you seem to be very un-understanding of what I've tried to say about LWs renderengine, and the solutions that are avaliable now to rectify a lot of the shortcommings that LW has without them. It is a two-edged sword in a thread like this. On one hand you have implementations and tools to be inspired by (for LW3DG), on the other hand, you have tools existing in the LW-universe that can deal with things now (and with quite some power and flexibility).

Neither one should be exluded from a discussion like this, because that would give a very one-sided view of LW and it's powers, give or take third party developments.


Some want more modelling improvement, better communication with mbuilder, some want better animation tools, etc. Its free to post requests.
The problem with "some" in the community is that a suggestion or pointing out flaws in lw is bashing.

Pointing out flaws and then stating that the solutions that are avaliable for LW aren't elegant enough to be used (which you indirectly have done), is one way to do it, of course. I have a very different view on the elegance of things, and while I do agree that there are many things outside of LW that does things fantasticly well, especially for renderers like vRay and Arnold, they are still just inspirational things for LW3DG to take a look at and for us users to request. To that, I also add what I feel is a very strong and flexible workflow, and passmanagement with breakout scenes, are, quite powerfull when used with the architecture that LW consists of, right now. Even if we wanted, we couldn't have used LW for the projects we have used it for, if it wasn't for Janus and exrTrader. Period. And, I think that is an important piece of information for users and visitors of this forum to hear. And, just because I state how great it has been for us, doesn't mean that things can (and should) improve, or that LW3DG shouldn't do their native system for this...


I don't recall being unappreciative of the nodal shading system.

Yet it seems you go amiss of why we want to use LW and why investing in something like Janus has allowed us to continue using it.


I recall posting a tutorial how other software deal with render elements and a rebuttal how Janus makes it that you don't need to think about this stuff.

Yes... bad wording from me... I had to think about it once, then I had a preset that does it for me through Janus... how about that? :) Renderpasses have to be computed though, even if they are thrown to a renderfarmcontroller in a single scenefile or if you throw in several scenefiles.


If you understand vray then please show work you yourself have been involved using vray. Because you only made comparisons with lw plus Janus against mray and mentalray. Why you didn't mention you use vray so its a fair comparison.

I've never said I use/have used vRay, but my understanding of it (and this goes for more people than me who also have used it in the team), is that it isn't a versatile engine to use for the wide work we do. You also have think about the host-package... does Max have as a versatile Nodal-shading system as LW? Will any artist ever learn how to think creativily in Mayas nodal system? As I've said, Arnold looks much more promising for the different types of work that we do compared to vRay, and until then, LW works very well...

:)

geo_n
10-10-2012, 07:36 PM
Agreed, I feel like this behavior has become rampant of late here, to the point where it's actually alienating customers and potential customers. When a customer presents a problem that's a "pain point" for them, having other users jumping down their throat arguing their usage isn't "relevant", or hijacking the thread to present alternate, unrelated scenarios "proving LW's side", the message sent to that customer is "your needs don't matter".

I notice a lot of old time ex-lwvers are not posting so much. Driven away by what cgtalk calls lwvers, "rabid fanboys".
Not all of them are like that but sometimes they're the most persistent to defend lw and post. :D





I've never said I use/have used vRay, but my understanding of it (and this goes for more people than me who also have used it in the team), is that it isn't a versatile engine to use for the wide work we do. You also have think about the host-package... does Max have as a versatile Nodal-shading system as LW? Will any artist ever learn how to think creativily in Mayas nodal system? As I've said, Arnold looks much more promising for the different types of work that we do compared to vRay, and until then, LW works very well...

:)

That's just my point really. I posted a tutorial how other software deals with matt passes. No preset in Janus could replicate that, you would still need to break out scenes in janus and render twice. It has to be shown to NT developers. Do you honestly think they have time to learn other programs and code lw at the same time?

I didn't even mention lw nodal shading.
The only comment I had was maya/mray and lw/janus. It would have been better to compare maya/vray to lw/janus for more common situations.
Max has nodal shading. It actually looks like a copy of lw since max retains the layer system while adding nodal on the side. Versatile is relative since both Lw and max nodal shading depends highly on artist that understand nodes.
The type of work you do, its the same type of work we do. Game cinematics, commercials, anything cg.
Vray is used in all types of projects from photoreal to stylized. Its versatile in our case.
Honestly the studio didn't abandon lw because its renderer was lacking, but because it just can't do mocap workflow well without a lot of help. We just had to look for that renderer, when they switched to max, that could do what lw could do. The answer was vray. We tried finalrender, mray, brazil. None compares to the ease of use, speed, versatility of look and the bonus is unlimited nodes. I am sure this is why many studios have opted to switch to vray. I doubt they would agree with you that a renderer is ridiculous investment. The same could be said about lw since it is now lw strongest selling point to some, the renderer.


To get on topic again. Does lw have a next gen renderer?
Does lightwave support proxies? assemblies in mray.
Does lightwave renderer support tiled exr natively?
micropoly displacement natively
robust renderpass management that doesn't require to render twice :D

Cageman
10-10-2012, 08:28 PM
No preset in Janus could replicate that, you would still need to break out scenes in janus and render twice. It has to be shown to NT developers.

What exactly is rendered twice? Nothing is rendered twice. There is one scenefile that renders the beauty, another one rendering the matte. None of these things are rendering twice. If we talk about computation speed, are you absolutely sure that vRay produces matte-passes without actually rendering them? Or, are you talking about having to submit two files to the renderfarm instead of one? If that is the case, well... I can submitt 10000 renderpasses in one go, if I need to.


I didn't even mention lw nodal shading.

I did, because it is an argument to why I view LW being a very versatile engine to use.



To get on topic again. Does lw have a next gen renderer?

Yes, but it depens on what you want to do. Define "next gen"... is it cellshading, volumetrics, shader-based, spherical, photonbased.... to be totally honest here... the question is kind of moot in the first place. Any renderengine that does for you what you need it to do, will be usefull. Trying to put a renderengine into a classification like that is like asking which type of cup is the best to use when drinking coffie vs tea.


Does lightwave support proxies? assemblies in mray.

No. Does it need to in order to get the results though? Again, is it a requisite for a capable renderer to support such a thing, or is it a convinience feature?


Does lightwave renderer support tiled exr natively?


No.


micropoly displacement natively

No.


robust renderpass management that doesn't require to render twice :D

Yes (I've never had to render any pass twice, if that is what you are asking). How about this...

Does your renderfarm software support these layers so that, for instance, if you quickly need to fix one of the mask-passes, you can do so without opening the scenefile and then re-render just that pass?

Hmm... since you put the emphathies on "native" here... lets turn this question around; does Max do anything natively? :)

geo_n
10-10-2012, 09:20 PM
What exactly is rendered twice? Nothing is rendered twice. There is one scenefile that renders the beauty, another one rendering the matte. None of these things are rendering twice. If we talk about computation speed, are you absolutely sure that vRay produces matte-passes without actually rendering them? Or, are you talking about having to submit two files to the renderfarm instead of one? If that is the case, well... I can submitt 10000 renderpasses in one go, if I need to.



I did, because it is an argument to why I view LW being a very versatile engine to use.



Yes, but it depens on what you want to do. Define "next gen"... is it cellshading, volumetrics, shader-based, spherical, photonbased.... to be totally honest here... the question is kind of moot in the first place. Any renderengine that does for you what you need it to do, will be usefull. Trying to put a renderengine into a classification like that is like asking which type of cup is the best to use when drinking coffie vs tea.



No. Does it need to in order to get the results though? Again, is it a requisite for a capable renderer to support such a thing, or is it a convinience feature?


No.



No.



Yes (I've never had to render any pass twice, if that is what you are asking). How about this...

Does your renderfarm software support these layers so that, for instance, if you quickly need to fix one of the mask-passes, you can do so without opening the scenefile and then re-render just that pass?

Hmm... since you put the emphathies on "native" here... lets turn this question around; does Max do anything natively? :)

Watch the tutorial. Or if you don't want to watch that find a blender tutorial.
You create a beauty scene and render that. Create a matt pass scene and render that. 1+1=2
Lw is versatile. But how is that related to the tutorial vid again? This is like talking to car salesman when asking about the engine the salesman says, but its got shiny alloy wheel. :D

In order to get this result lw probably needs those features and better handling of data sets.
http://www.youtube.com/watch?v=28LU4D8z7Xs
They chose mental ray in maya and not mdd it out to lightwave. I'm not saying mental ray was easy to use but it means lw was incapable to do it for them.

My emphasis is on modern renderers like vray,arnold since we're talking renderers here.
But if you must discuss software itself lightwave doesn't have native mocap retargetting and need to buy a 5k app, Yikes!!! No sane rendercontroller, cad modelling tools, too many plugins to list.
This is pointless though since we are talking renderers and we're only dragging out that lightwave will be as costly or even more compared to other packages.
Now again if you've got nothing to show concrete about renderer performance and features this has no substance.

Celshader
10-10-2012, 10:01 PM
...we're only dragging out that lightwave will be as costly or even more compared to other packages.

That may be your opinion, but I'm not sure any other package could easily do what we're doing for broadcast TV effects for the same price.

geo_n
10-11-2012, 12:02 AM
Yep only opinion. Since my personal cost of buying lightwave and other plugins have cost more than 3k USD. Others probably didn't spend as much.
Don't get me wrong I owe a lot to lw. Got me my job even though the company was slowly moving to max. I'm going to be buying my own house soon just from income from lightwave from freelance work saved up for the past 4 years.
But what would all those things mean if we don't push lw more and point out its weaknesses and post possible solution.
The tutorial showed how c4d handled one of toby's problem with lw and even more concepts that could improve lw.
Coders are coders. They don't have time to learn other software much like 3d users who refuse to learn other software.
I would like to show what vray is doing for us, a studio that doesn't have a complex pipeline, no scripters, but a profitable relatively small studio that lightwave needs to get its attention again, at what speed, what kind of projects, all done in vray.
It didn't even have to be vray if brazil, finalrender, mray was as good, as easy, but it is what it is. But what's the point if lwvers think the renderer has this and that to make up for it?

jwiede
10-12-2012, 11:16 AM
So... what you are saying is simply... never mention that there are third party tools to deal with some of the shortcomming now? There are two things to concider here, I think...
1) What can be done now?
2) What should LW3DG do for the future? (regarding implementations, ideas etc).
Nope, not what I said, not remotely what I meant. An actual example of what I'm referring to: In the Hieron instancing case, you didn't present that there were third-party tools to deal with his problem now, you presented a completely different scenario to show how great instancing was, while repeatedly suggesting the grass use case (Hieron's stated "pain point") was "extreme", etc. None of what you originally stated had anything to do with solving Hieron's problem (you did finally make such suggestions, but only later).

To be clear, you're not the only one making such posts, the example above just happened to be timely. Until you misinterpreted my posting, I didn't think specific call outs were needed. Your misinterpretation kind of made it necessary for me to point at a specific example of what I meant.

I don't have a problem with presenting third-party solutions that allow solving a problem now, such as Janus in the earlier part of this discussion. As long as trying to help address the "pain point" presented, and still respectful of the thread's point, I think it is fine. There is a point (reached when it turns into "LW doesn't need this because the third-party option exists") where it starts to become a case of the type I'm describing, but I did not get the sense that was happening in the Janus discussion.

chikega
04-21-2013, 12:58 PM
I would like for Lightwave to support Vector displacements natively.