Advertisement

You're blocking ads, which pay for BlenderNation. Read about other ways to support us.

Brecht Returns to Blender Institute, Works on new Renderer

306

Blenderheads, rejoice! After a brief detour as a developer for Octane Renderer, Brecht van Lommel has decided to return to the Blender Foundation to work on Blender development.

He has already started working on a complete overhaul of Blender's internal renderer. He comments:

There is also the problem that the internal render engine is showing its age. Before and during the Sintel project, I have worked on refactoring and improving it, but found it simply too hard to get it where I think a modern render engine should be.

The preliminary work is nothing short of amazing (did we expect otherwise from him?). Watch and drool:

Links

About the Author

Avatar image for Bart Veldhuizen
Bart Veldhuizen

I have a LONG history with Blender - I wrote some of the earliest Blender tutorials, worked for Not a Number and helped run the crowdfunding campaign that open sourced Blender (the first one on the internet!). I founded BlenderNation in 2006 and have been editing it every single day since then ;-) I also run the Blender Artists forum and I'm Head of Community at Sketchfab.

306 Comments

  1. Is this really coming to Blender? Freaking awesome! rooling all over ... have to clean up my floor now;-)
    Thanks a lot Brecht!

  2. I was in awe struck! Brecht your awesome i am thrilled to see the instant results in blender i will be waiting for this development to come out so that i can try it on my NVIDIA GTX 460 which i am using now. Its simply awesome to see such results.

    All the best Brecht and thank you.

  3. I was in awe struck! Brecht your awesome i am thrilled to see the instant results in blender i will be waiting for this development to come out so that i can try it on my NVIDIA GTX 460 which i am using now. Its simply awesome to see such results.

    All the best Brecht and thank you.

  4. So will this function like luxrender where when you hit render it will go on rendering forever, or will it have a point where it finishes? Thats one thing I like about the internal renderer, it has a starting point, and an ending point.

  5. Il en sera de cette fonction comme luxrender où quand vous frappez la rendre ira sur le rendu pour toujours, ou sera-il un point où elle se termine? Thats une chose que j'aime dans le moteur de rendu interne, il a un point de départ et un point d'arrivée.

  6. I was only thinking to myself the other day "If only Blender had a renderer like octane built in" you know like maya and max have mental ray. This is a happy day for all Blender fans.

  7. Je ne pensais qu'à me l'autre jour: «Si seulement un moteur de rendu de Blender a comme indice d'octane construite en" vous savez comme Maya et Max ont mental ray. C'est un jour heureux pour tous les fans de Blender.

  8. stratégie intelligente d'attendre jusqu'à ce que vous aviez quelque chose à voir. C'est génial! J'attends avec impatience plus de nouvelles. Merci, Brecht.

  9. This was the one area in which I was disappointed with Blender, so I just fainted. How much faster and when might we see this stunning development ready for download?!!!!!

  10. Ce fut le seul domaine dans lequel j'ai été déçu par Blender, alors je me suis évanouie. Quel gain de temps et quand pourrait-on voir cette étonnante prête pour le téléchargement ?!!!!!

  11. Poisson d'avril???? Allons, ne faites pas rire comme, mon petit cœur ne le supportent pas. Vraiment, "Da Man" est de retour?

  12. This is badass. One of the sore points of using blender in the past has been the renderer. But now personal knowledge, technology and software sophistication within the blender domain is making this much more of a level playing field. I dig having the added capability for visualizing.
    Now I won't have to port out meshes to C4D for better visualization. And with rigify, C4D is starting to gather dust.

  13. C'est dur à cuire. Un des points sensibles du mélangeur à l'aide dans le passé a été le moteur de rendu. Mais maintenant, une connaissance personnelle, la technologie et la sophistication des logiciels dans le domaine mélangeur est davantage ce bien d'un pied d'égalité. Je creuse ayant la capacité supplémentaire pour la visualisation. Maintenant, je n'aurai pas à mailles port de sortie pour C4D pour une meilleure visualisation. Et avec rigify, C4D commence à accumuler de la poussière.

  14. Sounds great. I'm not too in the loop with the renderer as I'm more into games design but it sounds like good news.

    How come its grainy and goes to smooth rather than filling out the screen in blocks like the old one? Is that how renderers work now? Also, like Adam said, will this have an end point? I don't know how you'd use it for animations otherwise...

  15. Sonne bien. Je ne suis pas trop dans la boucle avec le moteur de rendu que je suis plus dans la conception des jeux, mais il semble que de bonnes nouvelles. Comment se fait de son aspect granuleux et va en douceur plutôt que de remplir l'écran en blocs comme l'ancien? Est-ce ainsi que les moteurs de rendu de travail maintenant? En outre, comme Adam dit, cela aura une fin en soi? Je ne sais pas comment vous pourriez l'utiliser pour les animations autrement ...

  16. Gentleman German on

    I think I just wept a little bit...I've been waiting for an update to the renderer like this for so long. I consistently tout the advantages of Blender to other 3d artists and even my 3d teachers at college and this will just give me another thing to add to the list. After this I really won't be able to understand why production studios won't use Blender in a professional capacity. It'd cut down their production costs from having to purchase multiple software licenses to not having to purchase anything at all.

    Truly amazing stuff, I can't believe Blender has come this far in such a relatively short time and I can't wait to see where it is even a few years from now.

  17. Gentleman German on

    Je crois que je viens de pleurer un peu ... J'ai été en attente d'une mise à jour du moteur de rendu comme ça depuis si longtemps. J'ai toujours vanter les avantages de Blender à d'autres artistes 3D et même mes professeurs au collège 3d et ce sera juste me donner une autre chose à ajouter à la liste. Après cela, je ne sera pas vraiment en mesure de comprendre pourquoi les studios de production ne seront pas utiliser Blender à titre professionnel. Il faudrait réduire leurs coûts de production d'avoir à acheter plusieurs licences logicielles à ne pas avoir à acheter quoi que ce soit. Vraiment des choses incroyables, je ne peux pas croire Blender est venu aussi loin dans un tel laps de temps relativement court et je ne peux pas attendre de voir où il est encore quelques années à partir de maintenant.

  18. ...
    ...
    ...

    (I find no words to say how amazing it is... Thank you Brecht !!!!!!!!!!!!!!!!!!!!!!!)

    :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :)

    :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :)

    :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :) :)

  19. ... ... ... (Je trouve pas de mots pour dire à quel point il est formidable ... Je vous remercie Brecht !!!!!!!!!!!!!!!!!!!!!!!):):):):): ):):):):):):):):):):):):):):):):):):):):):):):):): ):):):):):):):):):):):):):):):):):):):):):):):):): ):):):):):):):):):):):):):):):):):):):):):):):):): ):):):):):):):):):):):):):):):):):):):):):):):):): ):):):):):):):):):):):):):):):):):):):):):)

  20. Fantastic news! Great to see Brecht is back and working on open-source again!!!

    I'm absolutely stunned by the progress he's already making. Welcome back Brecht!!

  21. nouvelles fantastiques! Formidable de voir que Brecht est de retour et de travail sur les open-source à nouveau! Je suis absolument abasourdi par les progrès qu'il a déjà prise. Bienvenue Brecht!

  22. So was Brecht working as a double agent while at Octane? (j/k) Very nice addition. Is this an unbiased render option, utilizing parallel processing or is it mostly GPU powered? If it's unbiased, how will you render animations?

  23. Il en était de Brecht travaille comme agent double tout en Octane? (J / k) addition très gentille. Est-ce une option de rendu sans biais, en utilisant le traitement parallèle ou est-ce principalement alimenté GPU? Si elle est impartiale, comment allez-vous rendre des animations?

  24. Excellent news and outstanding work. However, it looks suspiciously 'Octanely'. No chance it will raise issues on technology theft or software patents or whatever?

    It's always good to have one of the Gurus back in our team.

  25. Excellentes nouvelles et le travail exceptionnel. Toutefois, il ressemble étrangement 'Octanely. Pas de chance il soulever des questions sur le vol de technologie ou les brevets logiciels ou autre chose? Il est toujours bon d'avoir l'un des gourous de retour dans notre équipe.

  26. YYYYYEEEEEEEESSSSSSSSSS!!!!!!!!!!!!!!!!!!!!
    YYYYYEEEEEEEESSSSSSSSSS!!!!!!!!!!!!!!!!!!!!YYYYYEEEEEEEESSSSSSSSSS!!!!!!!!!!!!!!!!!!!!YYYYYEEEEEEEESSSSSSSSSS!!!!!!!!!!!!!!!!!!!!YYYYYEEEEEEEESSSSSSSSSS!!!!!!!!!!!!!!!!!!!!YYYYYEEEEEEEESSSSSSSSSS!!!!!!!!!!!!!!!!!!!!YYYYYEEEEEEEESSSSSSSSSS!!!!!!!!!!!!!!!!!!!!YYYYYEEEEEEEESSSSSSSSSS!!!!!!!!!!!!!!!!!!!!YYYYYEEEEEEEESSSSSSSSSS!!!!!!!!!!!!!!!!!!!!YYYYYEEEEEEEESSSSSSSSSS!!!!!!!!!!!!!!!!!!!!YYYYYEEEEEEEESSSSSSSSSS!!!!!!!!!!!!!!!!!!!!YYYYYEEEEEEEESSSSSSSSSS!!!!!!!!!!!!!!!!!!!!YYYYYEEEEEEEESSSSSSSSSS!!!!!!!!!!!!!!!!!!!!YYYYYEEEEEEEESSSSSSSSSS!!!!!!!!!!!!!!!!!!!!YYYYYEEEEEEEESSSSSSSSSS!!!!!!!!!!!!!!!!!!!!YYYYYEEEEEEEESSSSSSSSSS!!!!!!!!!!!!!!!!!!!!YYYYYEEEEEEEESSSSSSSSSS!!!!!!!!!!!!!!!!!!!!YYYYYEEEEEEEESSSSSSSSSS!!!!!!!!!!!!!!!!!!!!YYYYYEEEEEEEESSSSSSSSSS!!!!!!!!!!!!!!!!!!!!

  27. YYYYYEEEEEEEESSSSSSSSSS !!!!!!!!!!!!!!!!!!!! YYYYYEEEEEEEESSSSSSSSSS !!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! YYYYYEEEEEEEESSSSSSSSSS YYYYYEEEEEEEESSSSSSSSSS !!!!!!! !!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! YYYYYEEEEEEEESSSSSSSSSS YYYYYEEEEEEEESSSSSSSSSS !!!!!!!!!!!!!!! YYYYYEEEEEEEESSSSSSSSSS !!!!! !!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! YYYYYEEEEEEEESSSSSSSSSS YYYYYEEEEEEEESSSSSSSSSS! !!!!!!!!!!!!!!!!!! YYYYYEEEEEEEESSSSSSSSSS !!!!!!!!!!!!!!!!!!!! YYYYYEEEEEEEESSSSSSSSSS !!!!!!!!!! !!!!!!!!!! YYYYYEEEEEEEESSSSSSSSSS !!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!! YYYYYEEEEEEEESSSSSSSSSS ! YYYYYEEEEEEEESSSSSSSSSS !!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! YYYYYEEEEEEEESSSSSSSSSS YYYYYEEEEEEEESSSSSSSSSS !!!!! !!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! YYYYYEEEEEEEESSSSSSSSSS YYYYYEEEEEEEESSSSSSSSSS !!!!!!!!!!!!! !!!!!!! YYYYYEEEEEEEESSSSSSSSSS !!!!!!!!!!!!!!!!!!!!

  28. Speachless. Simply speachless. The time I spent last year to finally learn Blender was extremely well spent.

    No "thank you" is big enough.

  29. Speachless. Il suffit de speachless. Le temps que j'ai passé l'année dernière pour finalement apprendre Blender a été extrêmement bien passé. N «merci» est assez grand.

  30. FInally from direct raytrace to PATHTRACE!!!!
    GI is totally needed in internal.... Best news for a long time ago....

  31. Enfin, de lancer de rayons directs à PATHTRACE!! GI est totalement nécessaire internes .... Meilleures nouvelles pour longtemps ....

  32. Looking unbelievable. THANK YOU!

    Blender should keep the old internal too imo. For animations it's sometimes better to use a simple and noise-free renderer that has an ending point, rather than photon-based with GI.

    For very fast rendering blender has an awesome openGL-export, but why not make it a "renderer" too? I've made animations that don't need any reflections and normal map bump maps are enough. But then when I've wanted a DOF, I couldn't do it, because OpenGL export doesn't go into compositing nodes. Would be easy to fix by making the openGL-export an actual renderer. So in the future blender would have 3 renderers: OpenGL for very fast, old internal for simple rendering, and this new Cycles for hyper-realistic stuff.

  33. Vous cherchez incroyable. MERCI! Blender doit de garder les vieux internes trop imo. Pour les animations, il est parfois préférable d'utiliser un moteur de rendu simple et sans bruit qui a un point de fin, plutôt que photoniques avec IG. Pour mélangeur rendu très rapide a un impressionnant openGL-export, mais pourquoi ne pas en faire un "rendu" trop? J'ai fait des animations qui n'ont pas besoin de réflexions et normal "bump maps" carte sont suffisants. Mais quand j'ai voulu un DOF, je ne pouvais pas le faire, parce que l'exportation OpenGL ne va pas dans la composition des nœuds. Serait facile à corriger en faisant l'openGL à l'exportation un rendu réel. Ainsi, dans le mélangeur avenir aurait 3 moteurs de rendu: OpenGL pour très vite, ancien interne pour rendre simple, et ce de nouveaux cycles pour des trucs hyper-réaliste.

  34. Brecht,

    God Bless you my friend:) You have my permission to start over entirely with blender's material system. Nodes/shader base system would be so much quicker and easier to manage.

  35. Brecht, Dieu vous bénisse mon ami:) Vous avez ma permission de recommencer entièrement avec le système de matériel de Blender. Nodes / système de base shader serait tellement plus rapide et plus facile à gérer.

  36. This is awesome and all but...

    how does a cycle powered renderer handle animations and you are kind of ditching everybody who has ati/amd video cards(me included) by going with cuda. Whats wrong with openCL? Shouldn't an open software use an open GPGPU library? (I know openCL is not truly open but it is less proprietary then cuda)

  37. C'est génial et tous, mais ... Comment fonctionne un moteur de rendu cycle alimenté manipuler des animations et vous êtes bon de tout le monde amerrissage forcé qui a ati / amd cartes vidéo (moi y compris) en allant avec CUDA. Quoi de mal à OpenCL? Si pas un logiciel libre utiliser une bibliothèque ouverte GPGPU? (Je sais OpenCL n'est pas vraiment ouvert mais il est moins exclusive puis cuda)

  38. OH MY GEEEEEEEEEEEEEEEEEEEEEEEEZ"

    thank you, this is PERFECT

    also, will this have an endpoint?

    I exspecially love being able to move things IN THE RENDER, that will make setting up for a shot SOOOOO MUCH EASIER!!!!!!!!!!!!!!!!!!!

  39. OH MY GEEEEEEEEEEEEEEEEEEEEEEEEZ "merci, c'est parfait aussi, cela aura un point de terminaison? Je exspecially heureuse de pouvoir faire bouger les choses dans le rendu, qui fera la mise en place pour faciliter un coup sooooo beaucoup !!!!!!!!! !!!!!!!!!!

  40. un travail très impressionnant que jamais. J'espère qu'il n'y aura pas une question de droit avec le logiciel de réfraction

  41. goodmorningburger on

    Pretty sweet. I hope this does good motion blur. GI is great, but for animators motion blur is at least as important, maybe more. Blender internal has fairly primitive motion blur, and Yafaray doesn't have it at all. Lux does, but Lux is too slow to use for long animations.

  42. goodmorningburger on

    Assez doux. J'espère que cela ne bonne motion blur. GI est grand, mais pour le mouvement animateurs flou est au moins aussi plus important, peut-être. internes Blender a assez flou mouvement primitif, et Yafaray ne l'ai pas du tout. Lux fait, mais Lux est trop lent à utiliser pour les animations de long.

  43. Guys, stop that panic about an "octanish" view. First - I'm sure that Brecht will not violate any contracts and agreements. Also - just think why other renderers (for well known commercial soft) looks "suspiciously alike"?
    Don't need to go for long explanations - path tracing was known before the Octane renderer. So it's just a matter of how to connect any renderer to an animation system.
    Alike algorythms of a visualisation will give alike results.

  44. Les gars, arrêtez de dire que la panique autour d'un "octanish" point de vue. Première - Je suis sûr que Brecht ne violera pas les contrats et accords. De plus - il suffit de penser pourquoi autres moteurs de rendu (pour bien connue commerciale soft) ressemble "étrangement similaire"? Vous n'avez pas besoin d'aller pour des explications longues - chemin tracé a été connu avant le rendu d'Octane. Donc, c'est juste une question de savoir comment communiquer toute rendu à un système d'animation. algorythmes identique d'une visualisation donnera des résultats semblables.

  45. OMG!!!!!!!

    Blendersome, Blendersome, Blendersome hahahha

    :*)

    finally!! >:) we have somethin more to shut up Maya and 3DMax
    later Zbrush.

    With the glad and emotion I don't understand well XD
    ¿This is the future of the Blender Internal Render? or ¿The future of a external render integration?

    Greetings

    PD: coff... coff... whatever I'm happy :)

  46. OMG !!!!!!! Blendersome, Blendersome, hahahha Blendersome: *) enfin! >:), Nous avons quelque chose de plus se taire Maya et Zbrush 3DMax tard. Avec la joie et émotion que je ne comprends pas bien XD ¿C'est l'avenir de l'Rendu Blender interne? ou ¿L'avenir d'une intégration rendre externe? Salutations PD: coff ... coff ... tout ce que je suis content:)

  47. GPU powered rendering, YES, that is what i like also :)
    But please think to the people that have an AMD GPU and AMD CPU...
    ... don't optimize everythink only for NVIDIA GPU and Intel CPU.

  48. Looks great, though I hope much of the internal renderer stays too. Approximate AO and indirect lighting is one of the best things about rendering in Blender, IMHO, and most of us Maya users wish we had it for the awesome speed!

  49. can't wait to test this beat, any builds there?? :D
    it looks like an OpenCL thing .....please feed us with more specs,test builds...

    I'm really excited about how blender is developing, and it looks really nice that I can't believe it
    keep up the good work guys

  50. ne peut pas attendre de tester ce rythme, tout y bâtit? : D on dirait une chose OpenCL ..... s'il vous plaît nous nourrir avec plus de spécifications, test construit ... Je suis vraiment excité au sujet de la façon dont le mélangeur est en développement, et il semble vraiment sympa que je ne peux pas croire qu'il tient les gars bon travail

  51. Alright :-) this looks like it will be just what blender needs :-) I have always wanted the internal render engine to produce a more realistic image like that of yafaray

  52. :-) Bon cela ressemble, il sera juste ce mélangeur besoins :-) J'ai toujours voulu le moteur de rendu interne pour produire une image plus réaliste comme celle de yafaray

  53. I agree with Olli Äkräs, the work done in this is terrific and certainty Blender needs this cause people have been needing and wanting it for quite a while so this is a definitive plus.

    However I think that the old internal renderer should be kept also for several reasons.

    One reason is that for animations it is also very good to have a more defined starting and ending point than an estimated one. Another one is because it is the renderer that people already know a thrust and many people's old scenes are setup and/or rigged with light rigs that would require a lot of adjustments with a new renderer so it should be kept for compatibility. Another one is that you don't always want a ultra photorealistic style for everything, for some illustrations the unrealistic look of older renderers is sometimes desired. It's better to have a choice of built in renderers that would give Blender quite a lot of power. Let people have a choice.

    And one other thing that I would like to see is a biased renderer running on a GPU cause with all this talk about unbiased renderers running much more faster on GPUs a lot of people have forgotten how much faster a biased renderer would go on GPUs and if unbiased renderers were so slow before GPU rendering in rendering a complex clean scene and they run so much faster on GPUs then imagine how much faster a biased renderer will run on GPUs.

    And we have been adjusting and optimizing our scenes with older renderers to speed things up so if we do that with a biased renderer that runs on GPUs then I presume that we could obtain some blazingly fast rendering times possibly even at pretty high resolutions with enough quality to satisfy a lot of people and that is an additional functionality I would like to see in a renderer and one that I think a lot of people could put to very good and really, really practical use.

    So we could have two options, to render in very photorealistic quality but still waiting a while for a frame to clean up (which is excellent for stills and simpler animations) and a very, very fast non-photorealistic renderer that could gives us decent enough results with tremendous speed. That I think would a more ideal setup and an a better array of choices.

    Now I know that he says that more info will be available later but I'm wondering about some other capabilities of the new renderer like if it does in fact caustics, radiosity and other things. One thing that I know that lot of people have asked for is the ability to convert any mesh (a mesh surface) into a light source cause at the moment the only way to do that with the old renderer is by faking and rigging it and it is cumbersome and the results sometimes can be so so unless you spend quite a while setting up a scene for that. You can do it with the indirect lighting function but I know that these only work in dark scenes that have glowing objects but as far as I know an object cannot illuminate a scene completely like a regular light can so this is different (unless there is something I'm doing wrong and if I am please somebody tell me how to do this correctly) but as far as I know this is not the same capability.

    I would like to see these capabilities in the new renderer even though I'm not asking for it to have all at the get go. Anyway all of this is very promising and I hope that it is ready soon so Blender users can finally have this other capability that I know they have been waiting for for quite some time. Terrific stuff actually, just terrific.

  54. It's an awesome addition to Blender in the future, that's for sure but, as someone stated I would like to see a OpenCL GPU assisted rendering instead of a Nvidia's CUDA.
    This way not only Nvidia's GPUs but AMD's and possibly Intel's GPU would benefit from it.

  55. Well, I think that the expressions of people in these responses say pretty well what the reaction of most of the Blender community will be to these news. Just look at Cjd post up there, I think that we are going to see a whole lot more of that. LOL :)

  56. I'm already addicted. This will make me officially have no life. Sitting in my office Blending forever. Woe to my social life.......

    This is very bad.

    DJ

  57. SebastianErler on

    This news is awesome! Improving the internal renderer is badly needed. Knowing brecht doing that job is very promising. Thanks Brecht! ;)

  58. It's often easier to start a task than to finish one. It's easier to demo an exciting feature with nice background music than it is to completely finish it. These are lessons learned and taught from BMesh and Unlimited Clay. But somehow with Brecht, I have a magical feeling that this new render will be done soon.

  59. """And one other thing that I would like to see is a biased renderer running on a GPU cause with all this talk about unbiased renderers running much more faster on GPUs a lot of people have forgotten how much faster a biased renderer would go on GPUs and if unbiased renderers were so slow before GPU rendering in rendering a complex clean scene and they run so much faster on GPUs then imagine how much faster a biased renderer will run on GPUs."""

    gpus aren't some magical solution that are 100 times faster than a cpu at everything you throw at them. current gpus only are suited for certain mostly brute force algorithms. i think that's why nobody has achieved a useable biased gpu renderer yet.

  60. What?!! That is unbelievable! This is what get if you go on a short vacation, turn your back to blender and when you take a look back you see this?! After a mere 3 days? A renderer revolution?! Wow!!
    Jaw: dropped. Mind: blown.
    Best regards and a deep bow from Sebastian Koenig.
    Thank you!

  61. I mean, its nessesary to develop OpenCL client, because here s too much ATI users... /and we like ours ATI cards/ :)

  62. Matt Heimlich on

    For people questioning it, OpenCL is simply not as robust at the moment as CUDA is. And in the professional realm, you see MANY MANY more nVidia cards (Quadros) than you do AMDs. Maybe openCL is in the pipeline, but from the standpoint of getting it out to the majority market, CUDA is still the way to go.

  63. That's why I love Blender and people working with it so much... - Constant jaw-dropping surprises :) Great work Brecht. Congrat's to Blender Team !

  64. Can it get any better?? :) Brecht back and with a realtime GPU powered unbiased (+ little biased too?) renderer.

    About OpenCL, check that it says: No OpenCL "yet".

  65. Got bless Brecht!!!
    I've ATI (to old to support OpenCL), but I'm ready to buy nvidia, as CUDA seems to be much more mature and efficient in comparison to OpenCL.

  66. Does that mean blender's development is such an amazing thing that one just can't keep being off the project ?
    That's real good news and in my opinion, it comes from the heart. Good choice for us Brecht, the community needs your skills I guess.

  67. omfhrtajxbyrejsdfg!!!

    If that won't give you a huge Blender nerd-on, I don't know what will! Hail to the King, baby. Hail to the King.

    Just please do make sure that us ATI users get a lookin on this delectable, rendering gold. :)

  68. Brecht joined the octane team and resigned after one month,
    probably taking the saurce code with him.
    and now this ? looks strange to me...

  69. woww.............ive been using octane for a while. and this seems like an octane renderer, amazing!!! keep it up guyz!!

  70. The only thing competition between Blender and Octane will do is force the developers of Octane to improve it out of sight. Then we will have a great free rendering option for Blender, and Octane will remain that little bit extra that you have to pay for.

  71. The code is open source and will be available for everyone to see. This is the future and these guys are setting a trend that will benefit the entire community. A million artist instead of the privileged few. I think we are going to see a lot of super talented artist, a lot of growth, independant studios and some amazing works of art.

    Its going to be a great ride.

    DJ

  72. Matt Heimlich on

    I sincerely doubt that the Blender Institute would steal code from Octane. Seriously bad PR move right there. Honestly, most unbiased renderers look similar in function, and while I'm sure Brecht is using ideas that he learned during his time at Octane, I would be incredibly surprised if he outright copy pasted source code.

    On a slightly related note, is there any news on why Brecht left/came back?

  73. OMG! Yes yes yes!! I use Blender exclusively now for all my professional projects and I have no interest in external renderers. I need and want to use the full power of Blender's compositor.

    This is the best news I have heard yet!!

  74. "On a slightly related note, is there any news on why Brecht left/came back?"

    I quess Ton is just paying more than Octane did..

    Check the "Announcement by Ton Roosendaal" on top of this page

    Exciting times ahead!!

  75. """I quess Ton is just paying more than Octane did."""

    not necessarily. from the user experiences i read at other forums the octane guy seems to be a bit difficult character. maybe that was the reason? :)

  76. WTF, guys?!

    What stolen code? If it looks "like Octane" then it's not mean that it was "stolen"!
    You didn't see Brecht's contract by the way so please STOP these bad talks.

    You didn't SEE the code of Cycles.

    So don't be cretins please! You must be idiots if you think that Open Source code will be looking the same as for Octane in the state of violating the contract.

    If you want to do some trolling action - please, go f***ng away from here. You must thank him for this, not blame (if you are not those loosers from AD).

  77. Nah, I'm with Moolah - you'd have to be stupid or trolling to think Brecht would be unethical enough (or dumb enough to think he could get away with it) to pinch code from Octane. You know, what with Blender having it's source open and accessible to Octane's developer and all...

  78. CUDA is the way to go? yes a vendor lock in is what blender needs because looking at small lux gpu opencl is nowhere near capable enough.

  79. @toto64 "Brecht joined the octane team and resigned after one month, probably taking the saurce code with him. and now this ? looks strange to me…"

    RefractiveSoftware (Octane) did NOT invented Path tracing, they did NOT invented Real Time Rendering, they did NOT invented GPU rendering, they did NOT invented Unbiased Rendering, they did NOT invented Node Interfaces ... there are a lot of similar render engines using the SAME algorithms for years. Keyshot and Bunkspeed are way better than Octane but Octane is cheap. Modo have Interactive Real time render, and so Maxwell, Arion (Fryrender).... .
    Radiance (developer of Octane) was part of Luxrender when he decided to start his own project. Maybe you are a pissed off user of Octane trolling to discredit Brecht because he left RefractiveSoftware or you are just a troll... In any case I suggest you to make a little of research before make "arguments from ignorance"

  80. You can tell alot about individuals by their post. So lets just get back to the positive.

    This is the best news for the blender community since 2.5 and it has stirred our emotions. Silly talk is to be expected as is false gossip. So now that weve gotten it out of our system, lets support the brillant coders that make our dreams a possibilty. Support them financially too.

    Stop breaking down buildings before they are built and just enjoy the ride.

    DJ

  81. There is nothing wrong about worring with potential legal issues, especially when talking about such a successfull open source software as blender. It would be really sad to have judiciary issues put blender off track.

    Refractive software must have been really dumb not to add a non-competition clause into Becht's contract. That doesn't mean that having him work directly on a new renderer and making sush progresses in a small timespan will not raise questions. You don't need to steal some code to benefit from octane's experience...

    In my opinion, B. should be working on anything else (physics? fluids?) for something like a year before working on the renderer (but i am no lawyer).

  82. Condar: Refractive Software didn't invented any of the thing you've mentioned, but released particular rendering engine and now, the new Blender implementation looks very VERY much thesame (except material workflow) so guys... just accept that some people consider it strange... Brecht left Blender Foundation, joined Refractive Software for a few months, then got back again to BF and voilá - we have gpu renderer working very similar to Octane Render...

    Doesn't matter 99% people want this to look positive, it looks REALLY STRANGE and an official explanation from BF in that case wouldn't do any harm to anybody...

  83. Im a loyal blender user and a proud octane user.
    Not a troll at all, im just speculating.

    Brecht works at BI
    Works at RS
    ---Resigns without notice and doesn't train the new dev [roeland]----
    Rejoins BI and comes with this "Standalone" software that runs on "Cuda"

    Which looks/works "ALOT" >sarcasm<<

  84. All this speculating is bullshit.
    Brecht didnt steal any code!!!

    On the other hand you CAN speculate how Radiance managed to get Octane ready to show in such a short time after he left Lux. I am sure he used Code from Lux for quite some time and time after time he rewrote it to make it his own.

    Come on RS trolls, deal with the fact that Radiance is the bad guy here. No need for further posts. ;-)

  85. The Fatsnacker on

    Jesus boyz and girls are you trying to spoil my chocolate eating easter weekend?

    He's back...good news....end of story.

    Go and discuss you conspiracy theories elsewhere....

    You all play nice now..

    regards

    TFS

  86. @ The Fatsnacker

    Have you realized that BF uses a mirrored number 6 three times to form the famous Blender Logo?
    Mmmm....

    lol

  87. I really dont think it helps to speculate about who might have benefited from exposure to this and that or that you could attribute exclusive rights to anything in particular.
    Just like anything else rendering technology is evolving and cross pollinating.
    The people involved in this are relatively few and tend to crop up in various places together.
    As I remember it @Radiance was considering working on or was experimenting with unbiased on GPU back as far as the later period of his his Indigo association.
    I am sure he derived useful experience he has put to use since from involvement with Indigo and Luxrender.
    Honestly its hard to imagine Brecht would be an industry spy on behalf of the BF.
    I think its a little paranoid and mischievous to suggest so.
    We all know he has done quite a bit of rendering experiments for the BF movie projects and that a rehash of the BI has been mooted for some time.
    This looks like just that incorporating some more up to date ideas.
    As they say in the movies any resemblance to anyone is purely coincidental...

  88. A simpler theory, which draws on my 25 years experience about software development, both as an employee and as a consultant:

    Brecht had some good ideas, his boss had to show that he was the top dog and said no even thought this was a damage to his activity, Brecht reached the royally-pissed-off level and went away, slamming the door and started working on his un-accepted idea, with the supreme gusto of showing his old boss that...

    1) he (Brecht) was right and...
    2) he (the boss) was just another example of shining asshole.

  89. I hear that it was actually Brecht on the grassy knoll who shot JFK and that Radiance was driving the "paparazzi" car that made Princess Diana crash (proof: why do you think you can rearrange the letters in "Princess Dianna" to spell "Radiance n Snips"?! Huh? HUH?!)

    Ton also abducts cattle in his flying, antigravity Blender logo and Campbell is an ancient Illuminati emperor (why do you think he loves PYTHON, the great SERPENT of EVIL!)

    Seriously people. Randiance got Octane up and running quickly because he already *knew* and *understood* the algorithms and code needed from working on Lux. Brecht got hired by Octane because he already *knew* and *understood* the algorithms and code to create it. Neither needed to steal any code. They both knew how to create it already.

    Tim to put the schoolgirl gossip away people. Let's bring on the awesome.™

  90. Guys

    take this as a warning:

    This is a site for Blender news and a place to discuss things.
    However bad talking about other people or calling them a thief is going too far.

    Please keep it civil.

  91. people who talk trash about Radiance should seriously shut the F up. you don't know what the guy has done to Lux. he was basically 24/7 at the forums moderating them, answering people's questions while developing the engine and the website at the same time. the guy is a working bee and deserves all the praise and money he gets from Octane or whatever project he happens to do next.

    think and look for facts before you talk off your asses

  92. shashank singh on

    oh what joys , conspiracy theorists bring to this world :) . .
    anyways
    congrats Brecht , on your work . i will wait to for the code to be released so that i can play with it :)

  93. Posted by me back there:

    "And one other thing that I would like to see is a biased renderer running on a GPU cause with all this talk about unbiased renderers running much more faster on GPUs a lot of people have forgotten how much faster a biased renderer would go on GPUs and if unbiased renderers were so slow before GPU rendering in rendering a complex clean scene and they run so much faster on GPUs then imagine how much faster a biased renderer will run on GPUs."

    Response by horace:

    "gpus aren’t some magical solution that are 100 times faster than a cpu at everything you throw at them. current gpus only are suited for certain mostly brute force algorithms. i think that’s why nobody has achieved a useable biased gpu renderer yet."

    My new answer to that:

    horace I understand that. I understand that vector processors are not magic wands "solve it all" things but I'm not so sure that the issue of biased rendering in GPUs necessarily resides on that cause from what I seem they are pretty capable of doing a whole lot of things faster than CPUs.

    I established a while ago a conversation with the LuxRender people in their forum about this issue in particular and from what I got from them it seems that it could be possible to accelerate a biased renderer with a GPU too.

    The problem may not necessarily be that it cannot be done, I think that the problem could reside in the fact that unbiased renderers are usually much more photorealistic, more physically accurate but it was at the expense of time cause they took a very long time to render some images. Now because of the tremendous acceleration in the newer ones (and I know that there are some that work on CPUs alone and some that work on GPUs alone but some take advantage of all available processing power and I mean both available CPUs and available GPUs) and specially those calculated with the help of GPUs I believe has made the focus shifted to unbiased ones for the moment cause now we can have both in them, quality (their realism) and their much more reasonable speed.

    That could be the mayor factor that has deviated the attention momentarily to unbiased renderers and there are other attractive possibilities like the other advantage of them which as you and many other people may now know very well already is that in the newer generations they also got this quasi real-time capability of rendering while one was able to change the viewpoint (or camera location).

    This is also somewhat of a misunderstood by some people by the way cause again as many people already know it is not really real-time rendering because the image really, really start to render at the moment you let go of the viewpoint or camera angle change and then it really takes a while to render a fully smoothed and clean image anyway and this is not real-time and by far (it is more marketing pr than anything else say what rendering companies say and we don't necessarily need that with free open source software) it is more really like interactive rendering but it is still an advantage over the non-interactive rendering method and that has to be acknowledged.

    These could be the factors that have made most rendering companies shift their attention toward unbiased renderers instead of also porting biased ones to them at the moment. What I mean is that this could be only at the moment because who knows if later on they start to move some biased ones to GPUs too.

    From what I've got from the LuxRender people is that it may be possible to accelerate biased ones (but without the interactive view as far as I know) but that right now they didn't have the developers to work on this right now but that maybe the capability could be added to LuxRender in the future. I hope that this does happen of course.

    If anybody out there is more knowledgable in the subjet (perhaps a developer with far more understanding of this) maybe that could help shed some more light here on this and tell us if this is really possible to do much faster in vector processors (GPUs) with OpenCL or the like (Direct Compute, CUDA, Stream) such as it is also done with unbiased ones.

  94. Read most of the post, you guys just gave me a headache.

    I'm sure that Mr Brecth had his reasons to leave Octane and that he followed the rules and left without any legal issues getting involve. Lets take a small examples. The different types of cars we have on our roads do have a lot in common and they comme from so different constructers, why would renderers be different. Lets leave that now.

    But man I'm happy that Mr Brecth is back. Its true that blender was lacking a good renderer. I would be really grateful if in a near future that I could only use Blender and Blender alone even for my rendering.

    If I can post an idea, it would be that after the completion of the new renderer that Mr Brecth make a DVD tutorial on how to use to the new renderer to its full extent and how to use it with the node editors "for people who really are nubs like me in using nodes". That would be really a great resource.

    Thx to Ton for making Mr Bretch comme back:)

    Good day to you all

  95. This is amazing! :o

    Why people things that the code is copyed from octane? Maybe the rendering process looks little bit similar, but i am pretty sure that he did write it all again by the new way.

  96. Actually, this looks more like keyshot / bunkspeed.

    A very nice adittion indeed, can't wait to test it out (caustics "drood").

  97. That's the mutt's nutts right there. Welcome back Brecht, really liked the work on Octane you did and amazed at the Blender work you're putting in :)

    Awesome

    TxRx

  98. Having a public & open project means we need to conduct the utmost ethical practices. That's an intrinsic value. Everything we do is open, and anyone can read the code and techniques to verify if that's original, not something copied from other sources.

    Anyone who has ever met with or worked with Brecht knows that he's a very modest, easy going and ethical person. He's also a brilliant developer, who took a long experience and a lot of 3d coding competences to enthusiastically join Refractive Software. The fact this company let him go after a short while was a shock to me as well. Brecht then needed a while to sort out stuff, and luckily decided to join the Blender team again. We unfortunately can't pay him what commercial 3d companies could afford, so his decision was based on choosing for working in a pleasant, encouraging and rewarding environment first.

    What he worked on, and will work on, are common concepts and techniques where everyone in the industry is heading anyway. Check Modo, Lightwave, Max, Maya for example. It's also well aligned with design proposals and discussions here in the studio and on our public developer mailing list from long before he joined RS. If it wasn't Brecht, similar ideas would be in Blender sooner or later anyway. :)

    -Ton-

  99. Now that is truly killer! It seems to me that it may be easier to use than Octane and a better option for blender users than other external renderers. Thanks man.

  100. Well said Ton!!! :)

    Good luck Brecht! Really really happy you are back working on Blender and I'm looking forward to taking advantage of the new blender shading and rendering system! :)

    Thanks!

  101. Great news, naturally. But, as others have pointed out, the old type renderer should be kept. You can't always afford the CPU power to render each image in a physically accurate renderer, and animations are (at the moment) undoable in a LuxRender-type renderer.

  102. If they go with GPU acceleration I am hoping they choose to use OpenCL as opposed to CUDA. Simply because OpenCL works on both ATI AND Nvidia. ATI fan myself and seeing great speed increases in the newest versions of Luxrender. Would hate to miss out if the internal renderer becomes GPU capable.

  103. The Fatsnacker on

    Ton,

    you mention funds, or possible lack off. Would be willing to make a monthly donation to BF to support the Developers.

    regards

    TFS

  104. Another thing that I would like to know later is if Brecht has plans to make it work with all vailable CPU and GPU power, not only with one PC but with other networked PCs that one may also have as LuxRender will.

    When I talked to the LuxRender people they told me that the new LuxRender will take advantage of all CPU cores and all GPU cores in a PC and that it will also take advantage of all CPU cores and all GPUs present in any available network node too together even to render a single frame.

    I would like to now if Brecht has plans to add this "all power available" capability at the beginning (in a first release) or later on in the development of this new renderer just as they are doing with the new LuxRender.

    As far as I know the capability of using all CPU cores in all available networked PCs to render a single frame (to divide the rendering tiles among different PCs and not just among different CPU cores) is planned with the old Blender internal renderer if I'm not mistaken.

    I know that at the moment the new Blender can distribute the frames of an animation among different networked PCs but not the rendering tiles of a single frame (or at least not with the internal network renderer) . If they do add that capability that would be a big welcome too for those of us that have access to more than one PC at the office or at home.

  105. TwinSnakes007 on

    @PixelOz - Brecht said he doesnt have an OpenCL implementation yet - so no CPU/GPU co-op yet as CUDA doesnt support that.

  106. YAY!!!!!! Welcome back!!!

    And what an awesome entrance!!! I've been waiting for the this functionality for a while now, so it's awesome to see you working on it.

  107. GPU rendering... YES...
    but please think also to the people that owns and uses AMD GPUs and AMD CPUs,
    and not only to that, that uses NVIDIA GPUs and Intel CPUs...

  108. wait a minute!!!! is he just editing ON the render?????? Shouldn't that be Illegal??? Blender is getting infamous! Thanks Brecht you rock!!

  109. First of all this does look sweet! But the whole argument seems like politics, people demonizing the other side with no substance. Blender people understand its really not unfair to raise a suspicious eyebrow to cycles development happening so soon after Brecht left Octane development. I find it funny how serious both sides are getting. Both seem to have no concrete proof to back it up their statements. I am sure everything is Kosher about Cycles, but really people, don't be sold out to one side or the other that you ignore the facts.

    P.S. I am an Avid Blender user and Octane user. Can't wait until cycles becomes available. I think it will only push forward development at a faster pace for Octane and any other render that wants to stay in the game.

  110. TwinSnakes007:

    "@PixelOz – Brecht said he doesn't have an OpenCL implementation yet – so no CPU/GPU co-op yet as CUDA doesnt support that."

    Notice in my post that I used the word later in the first sentence. That is what I meant. It may not be available yet but that is precisely the thing, I'm asking if he will consider to later do it with OpenCL instead, if he plans to migrate it to OpenCL later.

    Is it only possible to do an unbiased renderer with a combination of GPU - CPU power with any other method? Aren't there programs that use CPUs and GPUs to compute together with any other GPU computing standard like Direct Compute or Stream? or is this only possible through OpenCL at the moment?

    Isn't it possible to do a program that combines CPU and GPU power through any other GPU computing standard maybe with Direct Compute or the like?

    Anyway if not I hopes that it can be converted to OpenCL in the future cause then it will also work with other GPUs other than Nvidia's.

    I have a computer for example that has two Evga GTX 260s in SLI mode so it's not a problem there but my second computer has a HIS (Ati chip) 4670 class card so it will not work there. It would be nice if it could work with all brands but I know that this technology is relatively new and only time will tell which GPU computing standard prevail. Here we go with the battle of VHS vs Beta again.

    I think that using OpenCL is a better choice so it looks odd that a programmer that codes open source software didn't choose that instead cause usually open source programmers embrace other open standards as well and very strongly.

    My guess is that if Brecht used CUDA instead there must be a very powerful reason like at the moment he is much more versed in programing for CUDA or that the current OpenCl version is too limited compared with the current CUDA standard or some other powerful reason(s) like that.

  111. Anyway I found the answers that I was looking for myself by reading a bit around. I had to do it cause I had a lot of doubts about the subject for quite a while and I wanted to clear them once and for all.

    The issue is that at the moment it seems to be easier indeed to implement unbiased renderers in GPUs than it is to implement biased ones although from what I read I also gathered that companies and other groups of programmers are trying to get the biased ones to work with GPUs too.

    There are different things that have to be taken into account. Unbiased renderers do have some problems with some things like caustics, depth of field, mesh based lighting and some other things which are problems that the more complex CPUs are capable of handling more easily but that doesn't mean that there aren't people trying to implement biased renderers in GPUs or trying to implement these features in unbiased renderers too.

    Some of these feature are already starting to appear in the newer generation of unbiased renderers or are about to do so. Bottom line is that it is not necessarily impossible, it just might take more time. There is work along that line and we may see something like that in the near future anyway (next few years).

    Another issue that I found is the memory limitations of graphic cards at the moment cause CPU memory in average is much larger than graphic cards memory and people that need to render very complex scenes that already had surpassed the memory limitations problems when we switched from the 4 GB limit of 32 bit CPUs and with the modern CPU memory modules capacity/price ratio can start to have these issues again.

    And what I mean is that it is relatively easy to have computers with 4, 6, 8, 12 gigabytes of ram or more today but look at current graphic cards memory and in average is still much lower than that, much lower than the memory most people already have for their CPUs.

    Modern high end graphic cards have a lot more ram but in average they are somewhat limited anyway compared again to what most people have for CPUs, it is still much cheaper to add a fairly large amount of ram to a motherboard for the CPU than it is to buy a very high end Quadro graphic card or the like and the more expensives of them have about a maximum of about 6 gigabytes of memory but those are the very high end ones.

    Those usually run in the thousands of dollars and you can buy another computer with a six core CPU and several gigabytes of ram that you can put to render in parallel with your other one (I did that myself as a matter of fact and I can control them both with a single keyboard and mouse with a free program called Input Director as if they were one PC and it is a very powerful and flexible setup) with that kind of money.

    Now that issue is changing cause newer graphic cards are starting to appear with more than 1 gigabyte of ram and I wouldn't be surprised if we start to see graphics cards with 6, 8, 12 gigabytes of memory or more in the near future and not to a small degree precisely because of the accelerated proliferation of GPU computing in the mainstream so we'll see.

    We will have to wait a little more to see were all these relatively new technologies will go but don't despair, it may not take that long before these issues are sorted out.

  112. Before engaging in speculation why not first hear from the man himself. If I am not mistaken these are Brecht's own comments he posted in the comments section of the youtube video

    Brecht says
    "Just to be clear, this does not contain any code from Octane, nor does it use any company secrets. There is plenty of code out there that does GPU ray tracing, see for example SmallLuxGPU, Brigade or Optix. I'd say the techniques are quite well documented by now, there's no magical ingredient needed.

    The raytracing core is based on code released by NVidia, from the paper "Understanding the Efficiency of Ray Traversal on GPUs".

Leave A Reply

To add a profile picture to your message, register your email address with Gravatar.com. To protect your email address, create an account on BlenderNation and log in when posting a message.

Advertisement

×