Page 1 of 4 123 ... LastLast
Results 1 to 15 of 54

Thread: Maxwell Render in Lightwave test.

  1. #1
    The Real Homestar. Intuition's Avatar
    Join Date
    Feb 2003
    Location
    Venice, CA
    Posts
    2,632

    Maxwell Render in Lightwave test.



    Well, I was sucked in by the Maxwell gallery and broke down. Purchased the Alpha/Beta (whatever it is right now) of Maxwell Render and ran a test.

    This Renderer is IMHO Awesome!

    Does it make F-Prime less useful. NOPE. F-Prime still ownz in the fast visualization pipeline hands down.

    What does Maxwell do? It handles light so well I almost feel like I'm looking at photographs instead of 3d.

    It uses Polygon Normals as light emitters or you can use a global illumination or sky illumination (similar to skytracer).

    You have to use tripled polys or subdivided nurbs objects right now and you can only use one image map per surface (I'm sure it will be fixed in full release in July).

    Pretty nifty.
    -----------------
    KC

  2. #2
    looks nice..how fast is the renderer though?

    in comparison to lightwave, vray, final render, brazil or mental ray?

    speed is a huge consideration in a renderer as "slow but nice" will mean still images only...how good is it for animation..does it have 3d mo blur or render artifacts when used for animated characters?

    also does it only "shine" when using GI rendering..what about standard lights?
    how does it look and what speed has it for those?

    would be VEry interested to here your overal comments on the renderer as you become more familiar with it.

    cheers

    steve g
    stee+cat
    real name: steve gilbert
    http://www.cresshead.com/

    Q - How many polys?
    A - All of them!

  3. #3
    The Real Homestar. Intuition's Avatar
    Join Date
    Feb 2003
    Location
    Venice, CA
    Posts
    2,632
    Speed is an issue with Maxwell.

    This render is 20 minutes on a single 3.2 GHZ 2GB ram.

    Not bad if you have a renderfarm like I do for animation needs.

    But on a single machine it would be a huge stretch.

    Right now you can't use Lightwave's standard lights.

    You use objects with polygon (normals) to light the scenes (using luminosity values) OR you can use one of thier preset globals or skys.

    It can't keep up with F-Prime as far as speed radiosity solutions but what it does do photoreal way ahead of the other render engines right now.

    Even ugly low poly objects look real when rendered in MW.

    The full version will have greater control over surfaces and interface.

    NO shaders as of now. No Hypervoxels yet, but it is made by NEXT LIMIT which did realflow so I am sure thier solution to fluid/particle renders will be awesome.

    I think it shines no matter what lighting you use, thats what they claim for fame.

    Caustics are way better then anything I've seen, it even calculates caustic motion blurring.

    Reflections/refractions look real as it gets. In thier forum there is a Lightwave test with a glass on bamboo and its real nice.

    MW's main catch is that its rendertimes are very slow except on the very latest computers. Hopefully this will be one of the render engines that takes advantage of 64bit processing as its potential is great.

    I'm not sure if Lightwave's native radiosity would make as nice an image but I'll do a test today. Also will do a F-Prime test as well and post here.
    Last edited by Intuition; 03-30-2005 at 12:18 PM.
    -----------------
    KC

  4. #4
    Super Member Captain Obvious's Avatar
    Join Date
    Dec 2004
    Location
    London
    Posts
    4,502
    MW's main catch is that its rendertimes are very slow except on the very latest computers. Hopefully this will be one of the render engines that takes advantage of 64bit processing as its potential is great.
    Most renderers use floating points instead of integers (ie, fractions instead of just whole numbers). This gives them a higher dynamic range and numerous other advantages. It also means they'll see a zero speedup with 64-bit processing. All processors since the early 90s have been able to do 64-bit floats.

    Of course, it can use more RAM if it's 64-bit, but it won't make it faster, as such.

  5. #5
    The Real Homestar. Intuition's Avatar
    Join Date
    Feb 2003
    Location
    Venice, CA
    Posts
    2,632
    Hmm... Good point Captain.

    Here is a Lightwave Native Radiosity. 2 hours and 20 minutes.

    Problem is I had reduce noise on so it removed the subtle hints.

    Also, the surface settings are not matched very well, The specularity/glossiness was modified for results in Maxwell which ar enot stright comparisons.

    What I am focusing on here is more light diffusion/scattering and less on surface.
    Attached Images Attached Images
    -----------------
    KC

  6. #6
    The Real Homestar. Intuition's Avatar
    Join Date
    Feb 2003
    Location
    Venice, CA
    Posts
    2,632
    F-Prime test 45 minutes. 400 Passes.

    Not bad at all.

    Surfaces need to be better matched again.

    But still nice.
    Attached Images Attached Images
    -----------------
    KC

  7. #7
    Dreamer Ztreem's Avatar
    Join Date
    Jun 2003
    Location
    Sweden
    Posts
    4,102
    I like the maxwell render best. Most because the light to the left is yellow in the reflection and not white. In LW and Fprime renders the light looks too bright(white) compared to the light hitting the skull. BTW if it renders at 20 min it's quite fast compared to Fprime and LW.

  8. #8
    yeah, maxell 20 minute is not that bad for this quality.. I'm interested a little more now..
    anybody heard from them if the full version will support shaders and hypervoxel and other LW stuff ?

  9. #9
    The Real Homestar. Intuition's Avatar
    Join Date
    Feb 2003
    Location
    Venice, CA
    Posts
    2,632
    The only draw back to MaxWave (Maxwell in Lightwave) is the level of surface control is limited in its beta/alpha stage.

    Alot of us that took the leap are spending alot of time experimenting with surface settings. Lightwave's % values are not used in all surfaces. Glass surfaces, for instance, are taking values only from Maxwell. I believe that when the full version comes out (July 05) Lightwave % values will translate to Maxwave fully.

    Just thought I'd let you know before you leap.

    You get great results with thier presets but its a learning curve if you are a control freak like me.

    Full version should support shaders/gradients but hypervoxels support is a mystery as of now.

    Since Maxwell is a NextLimit product they probably don't want to compete with thier own Realflow products so they may leave it so you have to use realflow to integrate your particle soultions into Maxwave.

    Again I don't know for sure, just guessing based on Forum chatter.
    -----------------
    KC

  10. #10
    Registered User Lude's Avatar
    Join Date
    Feb 2003
    Location
    Ireland
    Posts
    518
    20min for MW - Im very interested. I have stopped using FP and gone back to using Backdrop radiosity in lightwave.

    I've been doing work for a paving company and FP had so much noise on my ground planes (paving) it was obscuring the details so it was back to LW as it produced a much smother less noisy result.

    I look at the FP render here and see alot more noise than on the MW render and it took twice as long.

  11. #11
    Nice!

    Q: The Maxwell render compared to LW and FP doesn’t seem to have any specularity in the image? If it did would it have a similar results on the wood and the wall. For the moment it seems the Maxwell render is missing a lot of features. Though it is the easiest on the eyes.

    Tony

  12. #12
    The Real Homestar. Intuition's Avatar
    Join Date
    Feb 2003
    Location
    Venice, CA
    Posts
    2,632
    I ran some tests last night so that Maxwell would have specular in the wood like the F-Prime example.

    Its looking really nice. Maxwave is going to be a powerhouse.

    I did the pic at home instead of here at work but I'll post it Monday.

    Maxwave is missing alot of features but its still in beta/alpha stage so I'm not worried yet as Herve' Otacon and myself are constantly sending our bugs and issues to the MW staff so they can work on it.
    -----------------
    KC

  13. #13
    Super Member Captain Obvious's Avatar
    Join Date
    Dec 2004
    Location
    London
    Posts
    4,502
    How long does that scene take to render with LW's interpolated radiosity? I assumed you used Monte Carlo.

  14. #14
    Registered User Ramon's Avatar
    Join Date
    Apr 2003
    Location
    Georgia,USA
    Posts
    410
    Quote Originally Posted by Intuition
    I ran some tests last night so that Maxwell would have specular in the wood like the F-Prime example.

    Its looking really nice. Maxwave is going to be a powerhouse.

    I did the pic at home instead of here at work but I'll post it Monday.

    Maxwave is missing alot of features but its still in beta/alpha stage so I'm not worried yet as Herve' Otacon and myself are constantly sending our bugs and issues to the MW staff so they can work on it.
    Cheers to that! On behalf of us all, thanks for letting them know all of the items missing in their LW support! Hopefull that will make for much tighter integration!
    What I see as a really great benefit is that it's render time as stated in your comparison blows away LW and FP!!! And does it with an image quality that is certainly the best and also hardly any noise! I like the FP render but, it's way too noisy and with longer render times
    I then wonder if the animations produced through radiosity in Maxwell will be flickerless, which has been (one) the problem for LWers with radiosity.
    Last edited by Ramon; 04-03-2005 at 02:20 PM.
    My online portfolio:
    http://www.ramonart.com/advertising.html
    May the Lord bless you all.

  15. #15
    Registered User Ramon's Avatar
    Join Date
    Apr 2003
    Location
    Georgia,USA
    Posts
    410
    I don't know if most of you noticed this but, the only thing I don't like about the Maxwell render is the lack of diffused light on the chrome ball. Take a look at the chrome ball on both the LW and FP renders. The render it as I believe would be more true to life. Though as was stated, I don't know if the surface settings in Maxwell remained constant / comparable - more or less gloss / spec. It looks like the lightball is not effecting it at all in the Maxwell render - only the reflection.

    Intuition: Another question I have (because I am very interested in Maxwell) is what were the radiosity settings that you used in LW? Monte Carlo? 4x16? Just wondering because the render times between Maxwell and LW are astonishing to say the least, especially when you factor in the quality comparisions. Did you say that Maxwell's main problem was rendering speed? Am I missing something here?
    I'll tell you what I'm missing, Maxwell!!!
    Last edited by Ramon; 04-03-2005 at 02:34 PM.
    My online portfolio:
    http://www.ramonart.com/advertising.html
    May the Lord bless you all.

Page 1 of 4 123 ... LastLast

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •