PDA

View Full Version : mental ray



tisoy
03-26-2004, 06:40 PM
Is it possible to integrate Mental Ray into LW?

tokyo drifter
03-26-2004, 07:05 PM
No, LW currently doesn't support third-party renderers.

mkiii
03-26-2004, 07:07 PM
...Apart from FPrime.

Lamont
03-26-2004, 08:20 PM
Originally posted by tokyo drifter
No, LW currently doesn't support third-party renderers. I think he asked "Is it possibe?", not "Can it?". ;)

GruvSyco
03-26-2004, 09:06 PM
for the love of all that is sacred... please NOOOOO Mental Ray. Max, Maya and XSI all use that as the default renderer we don't need to do it just because they do.

If they implement something like that at all, it should be Renderman so, people with money can buy PRMan and people without can use some of the Open Sores Renderman compliant renderers.

tisoy
03-26-2004, 09:50 PM
The reason why I ask is because I've been reading that LW renderer reach its age and I have to agree that its getting old compared to others with built in Mental ray which is competetive agains renderman....And I was reading the updates on LW8 and it stated in increase of speed........Does LW8 support a 3rd party? if so then its possible to integrate other rendering plug ins?

sailor
03-27-2004, 01:00 AM
"for the love of all that is sacred... please NOOOOO Mental Ray. Max, Maya and XSI all use that as the default renderer we don't need to do it just because they do. "

for the love of all that is sacred? what a drama queen :)

personally i wont be that drastic...if you mean better not have anything new on the render side rather than Mental ray i totally disagree with you...why so often you LW fanboys are on a binary system ...like lets have Renderman or die... everything or nothing etc... if LW renderer show its age then ANY way to improve this situation is welcome

my opinion

GruvSyco
03-27-2004, 01:41 AM
wow 2 personal attacks in one post... tnx man :)

tokyo drifter
03-27-2004, 03:00 AM
Originally posted by Lamont
I think he asked "Is it possibe?", not "Can it?". ;) Oh, you're right. Well, anything is possible, so I guess the answer is yes. :)

tisoy
03-27-2004, 03:10 AM
Not that I'm complaining on LW renderer, I love its rendering capabilities. But I think with other programs with high end renderer, LW's renderer will have a hard time to compete with them......Before, other programs like maya, max and so on uses either renderman or MR as there external rendering for movies and etc,,,now MR is built into some programs, And I'm sure some uses the LW rendering for movies too......and as u know long time ago it cost money to render in PRman or MR, I think its time for LW to update theres, not just speed and some extra function but the quality of it......Im sure LW have think of that too. Maybe on LW9....hmmm.......So what u think? Are u content on LW built in renderer?:rolleyes:

DarkLight
03-27-2004, 03:25 AM
I would imagine that is possible to intergrate Mental Ray into Lightwave. You would have the same limitations as fprime does at the moment of not being able to use shaders, hypervoxels, etc.

Transferring the object data itself should be a fairly simple job.

Jaffro
03-27-2004, 03:57 AM
Darklight is right, you can render in 3rd party renderers but you'll have to write your own conversion plugins to get the data out, and then you'll have the limitations of the sdk - which is basically why 'lw doesn't support 3rd party renderers' (because of the sdk). I dont know how limited it is, it maybe that mental ray cannot be supported because of the number of limitations. *shurg*

Mentalray and Renderman are world class renderer's just like Lightwaves is. However Lightwaves is old and doesn't have many features that modern renderers have. There's nothing wrong with using another renderer if it gets the job done, no matter what it is or where it came from!

Mylenium
03-27-2004, 06:18 AM
Well, why would anybody want to use a combo of LW with MR? It just doesn't make sense. Due to the limitations of LW (not node based, different GI method, different basic shaders...) you would only have about 30% access to any features related to MR and even that would be tricky in parts (think of translating LWs Multiply mode to MR). The other 70% could only be used with a very massive and complex converter which simply would be impractical.

Mylenium

Jaffro
03-27-2004, 06:26 AM
Exactly why its said that lw doesn't support 3rd party renderers. Maya has similar issues, not all of mental ray features are supported and its poorly intergrated into the program.

Rich
03-27-2004, 07:56 AM
I'll have to agree with Jaffro about Mental Ray and Maya. I went to a school for 2 years learning Maya. We used Maya 5 which has Mental Ray integrated. It gave me many problems when trying to use Sub-D objects. Mental Ray would refuse to render certain Sub-D objects. It was a real pain. There is a lot of hype that surrounds Mental Ray and most who use it when compairing it to other aps like LW won't mention its limitations so they can brag that they use the best renderer around. The only complaint I have about LW renderer is speed but at least it will render all of my objects.

Titus
03-27-2004, 08:34 AM
I'm programmer and I don't know what the SDK limitations are, but yes, it takes a couple of good programmers and some time to write and exporter for mental ray. But NewTek then need to pay for every mr license since you can't buy it independently.

Mental ray won't make renders faster, only more phisycally acurate. Do you know how to program RenderMan/mental ray shaders? just wondering.

UnCommonGrafx
03-27-2004, 08:46 AM
Another sales post...

Titus, well said.

wacom
03-27-2004, 09:30 AM
I'd rather have other renders than MR- and hell lets not forget the only other renderer for LW that's in developement- anybody remember FPrime? Worley is going to be working with Newtek...can't say that the other renderers are getting that kind of support so....

279 bucks was such a good investment...

Jaffro
03-27-2004, 09:52 AM
wacom, not from nt they're not no. lets not forget that softimage's native render is mental ray and im sure plenty of development is going into that. Not to mention the other renderer's out there that dont have many of the worries that a whole 3d package development has.

fprime is good, no its great, and will develop to be something as amazing like sas is im sure. but the real development needs to go into lw's own renderer and im sure as worley and nt are helping eachother now that both are going to come a long way in the next few years.

tisoy
03-27-2004, 05:38 PM
that would be great if they can improve LW's renderer, not just speed but more advance option to it and quality similar to high end renderers..........Lw's renderer is awesome but its aging now compaired to other highend programs even if it has some limitations............if fprime can cut it then that would be great....I cant find the review comparison between fprime and lw renderer and it only shows small differences and overall LW renderer still the best.....if worley can make fprime as good as MR or PRMAn then Fprime is good to integrate it to LW..........small difference still not good enough....

jamesl
03-27-2004, 05:51 PM
How about a decent RIB exporter with support for custom coordinate systems for texturing like Mtor? There are several cheap/free renderman compliant renderers out there, even free shader building programs. Anyone want to write LtoR? BTW, the sdk wouldn't be an issue for something like this, as you would ignore all of LW's rendering functions, concentrating on modeling and animation only. It's just geometry in worldspace, with UVs and projection coordinates. Entirely possible.

j

Lightwolf
03-27-2004, 05:51 PM
O.k., I'm jumping in now :)

We bought three FPrime licenses, and I have to say, it really changed the way we work, instantly. The difference is huge for most scenes we work on, in terms of speed and quality. Oh, and there is the instant feedback as well :)

Areas where I think LWs renderer needs to improve:

Speed - raytracing is extremly slow and not well optimized
Adaptive displacement and tesselation - for displacement maps and SDS
GI - more options, degradation of evaluations with higher bounces
texture filtering - nuff said, I keep going on about it anyhow :)
volumetrics - could use a speed boost as well.
DOF and motion blur - well...
Antialiasing - needs to be improved in general.

None of these are really advanced requests, but this is what I would expect from a state of the art commercial renderer nowadays. As a special bonus I'd take distributed single frame rendering across a farm :)

In general I'd like to have more options to tweak, to optimally adjust rendering times.

Cheers,
Mike - off to get some sleep.... :)

Titus
03-27-2004, 07:02 PM
Originally posted by jamesl
How about a decent RIB exporter with support for custom coordinate systems for texturing like Mtor?
j

I started my own translator Light-R (http://www.garagepost.tv/renderman.htm) but I need some free time and support to continue, with this prototype I realized that an exporter is in some way really straghtforward to do. I also had some talks with Nicolas Yue to convince him to release his own exporter too.

jamesl
03-28-2004, 03:00 AM
Lightwolf... wouldn't it be easier to not try to dictate how you want your renderer to work, rather, why not just open up the api to allow for renderers that are already proven? Lightwave is analogous to the one printer manufacturer who, in the late 80's, early 90's, decided that Postscript was a passing fad, and then faded into oblivion. Come on, gentlelmen... it's all RIBs these days. LtoR would be the Maya killler.

j

Jaffro
03-28-2004, 05:04 AM
I dont think there's any problem with the object data. I think its the volumetric stuff that's the problem in the sdk. *shrug*

Lightwolf
03-28-2004, 05:09 AM
jamesl:
Ah, it looks like I forgot one of my other long term requests ;) :
I was thinking... what if not only LW would be .rib compatible (i.e. write them out), but could also _read_ them, including the renderer of course.
That way NT could
a) market their new renderer on its own if they wanted to
b) It would be easier to interface to other renderers.

I don't think NT should give up on the renderer though, no way. After all, what would be left if they dropped it, or you _had_ to buy one to get decent results (in, let's say, two or three years time, which you have to give them for development).

Why not use another package then? What would the strengths of LW be then? I like the fact that I currently get unlimited free rendernodes of a decent renderer with the package. It is the unified, "out of the box" approach that makes LW interesting for me.

Cheers,
Mike

Jaffro
03-28-2004, 05:18 AM
It is the unified, "out of the box" approach that makes LW interesting for me.


Yup same here. Probably the best thing about lightwave is that its out of the box all in one.

Lightwolf
03-28-2004, 05:38 AM
Originally posted by Jaffro
Yup same here. Probably the best thing about lightwave is that its out of the box all in one.
Absolutely, we're a small studio, and couldn't really afford 5 network licences for RM or whatever... Max and XSI are out of the question, to steep pricewise, Maya might be an option, but the cheaper version is a tad limited for my taste.
The only real contender for me (as a target group), is C4D. Not that much more expensive in the studio bundle if you think about what you get for the money (like, a decent network rendering environment ;) ).
Mind you, one thing that wouldn't bother me would be maintenance, 300$ per license per year and minor upgrades every two months or so would be rather neat.
Cheers,
Mike

Jaffro
03-28-2004, 06:55 AM
I dunno, i dont like the idea of maintenance. I'm not even full time freelancer so i'd never be able to afford it, and hobbiests would have a nightmare.

I still find it hard to believe that they can basically charge if you want to render over a network. Its not like an extra seat. Maybe i've got it wrong hehe :)

Lightwolf
03-28-2004, 07:43 AM
Jaffro:
It wouldn't make much of a difference moneywise. Whether you hand down 300$ a year, or 500$ every 18 months or so, same thing.
The good thing about maintenance would be:
* NT would have a more steady cash flow (I personally think that the eraly bird LW8 specials only came about to bind customers, and to get more cash for development).
* you get major releases without having to explicitly ordering them
* and you (should) get more point upgrades.
I like the idea of having more small bugs fixed, and more releases in between major ones.
This must not neccessarily be based on a subscription, but that would probably make it more interesting for NT.
Eyeon for example tend to release tons of point releases between majors, each one with a full list of bugs fixed an features added (and they don't even have a subscription system, but a relatively high initial entrance fee). That is the kind of service I consider when deciding on a platfrom to work on.

As for network rendering: Of course it is no extra seat, but it does cost money to develop. Mind you though, vendors that charge for network rendering usually have a complete solution, including stuff like a network controller (eyeon, C4D), or distributed single frame rendering (hey, on topic again: mental ray :) ).
Max being an exception, since it provides free network rendering, and a very good network controller (but a crappy renderer :) ).

Cheers,
Mike

Jaffro
03-28-2004, 08:01 AM
True true. What worries me is that this method might pressure nt into releasing updates when they may not be fully ready. I dunno, dont really know enough about how other packages handel maintenance upgrades.


With network rendering thing, dont they charge per render node or whatever its called? Wouldn't it be a better idea to just sell a network render controller as an addon or modual perhaps?

Lightwolf
03-28-2004, 08:07 AM
Jaffro:
Don't you think NT feel pressured right now as well? Actually, II think this will take some pressure out, since you will not have a definite release date (may be for major releases), but the minor ones pop up when they're ready. It could also just mean releasing a bunch of plugins in a newer version.

Well, they charge per node since you buy licenses for every computer, obviously. It makes sense form a business point of view. You render faster, you can satisfy your customers, you pay more.

Well, we currently are in a situtation where the cheapest part of being in the biz is buying a box to run the software on. :)

Cheers,
Mike

Jaffro
03-28-2004, 08:24 AM
haha true true.

From a business point of view it does make a lot of sense. Still dont like it :) hehe from my point of view, If i've just spent whatever ammount on getting the hardware to make renders go faster, i dont want to have to pay out for software to use the hardware - espically when i've already brought it once. I guess my point is that the hardware makes the render faster, not the software that i already have. (if that makes sense?)

I think nt are under a huge ammount of pressure right now. But i think there maybe more complains about maintainance not being worth it if its just for point updates. We dont pay for that at the moment, we only pay for the major upgrades. I think the pressure would come from people paying in advance for something that may not be in development yet.

Titus
03-28-2004, 08:26 AM
Originally posted by Lightwolf

Mind you, one thing that wouldn't bother me would be maintenance, 300$ per license per year and minor upgrades every two months or so would be rather neat.
Mike

I don't like maintenance upgrades. I remember some years ago when my former studio upgraded Maya from 3.X to 4, there was no significant benefits, the change in the software was minor adjusments (no new big features) and for the studio was really expensive. Why NewTek should want to innovate if they have captive clients?

Lightwolf
03-28-2004, 08:28 AM
Yeah, but they did that last year as well, pay in advance, for a couple of protons videos and a Siggraph stream...
Heck, with every upgrade you pay for, you pay for the point upgrades as well, same thing, _except_ that NT would get a more steady cash flow, and wouldn't have to shoot out pre-release offers like it does right now.
There is nothing like a good reliable stream of cash coming in to plan your future development and ressources, especially for a small co like NT.
Or, looking at it another way, we, the customers, would basically rent the developers for another year ;)
Cheers,
Mike

Lightwolf
03-28-2004, 08:35 AM
Originally posted by Titus
I don't like maintenance upgrades. I remember some years ago when my former studio upgraded Maya from 3.X to 4, there was no significant benefits, the change in the software was minor adjusments (no new big features) and for the studio was really expensive. Why NewTek should want to innovate if they have captive clients?
Because if they have a good reputation for point upgrades, it will help with new sales as well.
You don't have to pay your subscribtion, that might mean though that you don't get a point upgrade every two months, but only every 6 months or so, and have to shell out the current 500$ for the next major release.
I on the other hand would like to see more smaller upgrades with bugfixes, and I'd like NT to be able to calculate their cash more accurately, because that will mean that they can hire their developers more reliably.
Actually, this is how I'd prefer to get my jobs as well :)
Cheers,
Mike

Jaffro
03-28-2004, 08:38 AM
No, i done see that. I paid in advance for a major upgrade (lw8), im not going to have to pay again til the next major update. Its a fairly long time between major updates, so my upgrade to lw8 has paid for well over a year of upgrades but probably at the same price as a year maintainance would. I still think its cheaper this way, plus its very dependent on when the software is upgraded, minor or major. and how often - which is what i think will cause the pressure on development and give problems in long term r&d maybe? (i'm way over my head in this conversation, i'm just saying it the way i see and understand things)

my brains just shut down i dont understand anything anymore! lol

Lightwolf
03-28-2004, 08:49 AM
Ah, I see. I didn't buy the upgrade yet. :)
Well, of course it will put up the pressure a bit, since it also means that development would have to be restructured. On the other hand, we'd probably have those Mac fixes out by now, and would be a 7.6 with a couple of features added that 8.0 will have.
You prefer to pay in advance for something you haven't seen yet, I prefer to pay in advance for the reassurance to get more point upgrades, and not having to worry about updating at a certain point in time :) It does make it easier to plan expenses as well. Its just a part of your fixed costs then, just like rent, car insurance etc...

Cheers,
Mike

Jaffro
03-28-2004, 09:09 AM
Thats what im saying, with point upgrades your not guaranteed on your money's worth. I guess your not guaranteed either way, but if you pay for a particular update then you know what your getting. For example If i decided to wait until now to buy the upgrade it would still cost the same but i know what im getting. Whereas maintainance i dont think you dont get that choice, do you?

I waited til after i saw enough features to satisify the money being spent. If it was maintainance i would've already spent the money before i knew what was in the updates no?

Your right i think we would've seen a 7.6 if maintainance system was in place. But its a risk to pay for maintainance if your not sure if your going to use or need the point updates covered by it. I think im developing this view that maintainance is like expecting there to be bugs with the features you want and paying for them to be fixed, obviously you get new features - but you might not need/want them.

[edit] - I think i got my brain back :) hehe

Lightwolf
03-28-2004, 09:31 AM
Lol :)
Well, my experience is that there were plenty of bugs in the past releases, and I would have rather spent my upgrade money as a maintenance fee in return for quicker bugfixes.
Some of these can be real showstoppers if they hit you in the middle of prodution.
Anyhow, currently you are paying for point upgrades already, it is just called upgrade fee and is 500$ every now and then (rouglhy every 18 months), so the difference isn't that big from a users perspective.
I'd rather pay 300$ now and 300$ in another 12 months time :)
My main point is basically more upgrades for us, and a more static flow of cash to NT, which would put them on firmer feet. Sometimes they do feel a bit wobbly :)
Cheers,
Mike

Jaffro
03-28-2004, 09:39 AM
Yea it does seem a better way of doing things. Just dont like the thought of paying for something if i dont know exactly what im going to get from it :) LW's always going to be cool whatever way they do it.

Lightwolf
03-28-2004, 09:41 AM
Originally posted by Jaffro
Yea it does seem a better way of doing things. Just dont like the thought of paying for something if i dont know exactly what im going to get from it :)
You never go out to eat in a restaurant, do you ;) :p
We are getting seriously off topic here though...
Cheers,
Mike

Jaffro
03-28-2004, 09:51 AM
Your right actually i really cant handle going to new restaurants if i dont know that i'll like something they do :) hehe

True we are off topic. o well ! haha :D

- Jaff

Nemoid
03-29-2004, 12:56 AM
Originally posted by Lightwolf
Lol :)
Well, my experience is that there were plenty of bugs in the past releases, and I would have rather spent my upgrade money as a maintenance fee in return for quicker bugfixes.
Some of these can be real showstoppers if they hit you in the middle of prodution.
Anyhow, currently you are paying for point upgrades already, it is just called upgrade fee and is 500$ every now and then (rouglhy every 18 months), so the difference isn't that big from a users perspective.
I'd rather pay 300$ now and 300$ in another 12 months time :)
My main point is basically more upgrades for us, and a more static flow of cash to NT, which would put them on firmer feet. Sometimes they do feel a bit wobbly :)
Cheers,
Mike

Not a big fan of maintenances here. it's only a way to earn more money upon quite nothing. instead, I'd like Nt do one big release every year (payed) while a monthly bug fix update(free) with even one bug fix, every month.
Its way better , gives the users the good impression that Nt is working well, letting users free and the opportunity for Nt to earn money every year with a .5, or .0 .

Also, you will know for sure that every year a major release will be out just like it happens with Maya and other apps. :)

Lightwolf
03-29-2004, 02:14 AM
Ain't it funny how must of you prefer to pay more and get less in return ;)
Aynhow, just because maintenacne/subscription is screwed up with companies that have bad service anyhow, that doesn't mean that it is generally a bad thing.
I can see that it puts off hobbyist users though.
As a side note, NT Europe actually offers maintenance and support for a fee, but since you don't get anything extra that you'd normally get anyhow, there is no point really.
Cheers,
Mike

tisoy
04-07-2004, 12:53 AM
k,,here we go,,,I've been searching for answers,,,,,and I found this,,,,maybe this is not the best explanation but its worth knowing other opinions,,,,,,Check this out maybe it helps....Choosing ur renderer (http://www.zaon.com/company/articles/3d_rendering.php)