Alvaro Luna Bautista sent us a detailed article about the state of Yafray. He touches on the struggle of the team to keep development going during the last year and he provides a detailed list of the upcoming features. It all looks very promising - I hope they'll be able to get a stable release out soon!
Dear Yaf dudes:
Last year was a difficult one for YafRay. Original developers did not have more time to collaborate actively in the YafRay development, although Jandro is still collaborating by other (important) means and eeshlo has been involved in the Blender development lately. Blender users were concerned about what YafRay future was going be. At that time, Mathias Wein a.k.a. Lynx3d stepped foward to carry on YafRay development. Nine people out of ten would run away of getting involved in such a disastrous project. A big "thank you" is all we can say.
This article is a "heads up" about what Mathias a.k.a. Lynx is doing 'in the shadows'. I think that his work deserves a place in Blendernation and our recognition, since YafRay is a substantial part of this comunity. The list of features is taking shape and it looks very promising:
Plugin based (as before), with following classes
- Materials (new, partially replaces shaders)
- Cameras (new)
- Integrators (new)
- Point light
- Spot light
- Area light (rectangular, or parallelogramm actually)
- Mesh light (uses a triangle mesh as light source)
- Sphere light
- Directional (with optimal radius)
- Sunlight (basically a directional with incoming direction sampled from a cone, blurred shadows)
- Environment light, with importance sampling for efficient image based lighting (even without GI)
- Basic diffuse w. specular reflection, transparency and translucency, diffuse color, support for shader nodes on diffuse color
- Diffuse+glossy reflective material, blinn microfacet distribution w. fresnel effect (isotropic version of Ashikhmin&Shirley), shader nodes support on diffuse color
- Coated version of the above mentioned, adding specular reflection with fresnel effect
- Basic glass (dielectric) material, with fresnel, filter and absorption and dispersion
- Basic image textures (tga, jpeg, png, exr, hdr), cloud, marble, wood, voronoi, musgrave, distorted noise and "RGB-cube"
- Backgrounds: constant, sunsky, texture
- Perspective camera with DoF, orthographic camera
- basic surface integrators for direct lighting, path tracing and photon mapping with final gather (all under heavy construction!)
- Adaptive (simple color threshold based)
- Variable size box filter (other filters can be added easily though, just laziness)
- Importance sampling, multiple importance sampling etc.
- Multithreaded rendering passes.
- Very basic XML writing and reading (using libXML) for scenes.
More info: http://wiki.yafray.org/bin/view.pl/UserDoc/YafaRay
woooh! yafray is gettin diffused glossy material shaders!!
And its getting the realistic sunlight cone effect!!
I was waiting so long for thoes things to come along :D :D :D
Ahh, a much needed addition to Blender. Can't wait to test it against other professional GI-renderers.
Yeaaah, very nice to see work fresh news on Yaf(a)Ray... I'm glad it's sort of become the official future of YafRay.
Good to see that yafray survied the "personal changes" and that they heading to an next release.
Talking about Yafray one has to talk also about "the dark side of Opensource projects" . It`s a pitty to see, how a "substantial part" of blender comunity only survives because of the enormous efforts of particulaly one, maybe three or four persons. Three years ago we had the Yafray hype, now the Orange hype. This is no glorious chapter for blender organisation.
"Talking about Yafray one has to talk also about â€œthe dark side of Opensource projectsâ€. It`s a pitty to see, how a â€œsubstantial partâ€ of blender comunity only survives because of the enormous efforts of particulaly one, maybe three or four persons. Three years ago we had the Yafray hype, now the Orange hype. This is no glorious chapter for blender organisation."
Any 'one-man' project is immune against the problems that today is facing Yafray. This is true for open source and for propietary software. In fact, Yafray might survive because is open source: perhaps Lynx got interested in a software he could actually 'read', and the other people currently involved are that committed because 'it is open source'.
In fact, YafRay survival is not guaranteed at this point, it still depends on the commitment of too few people. Mathias is just an university student, like Jandro when YafRay saw the light.
YafRay 0.1.0 will be a milestone for this project, that's for sure.
I wouldn't say it's been a disaster, Yafray is still genuinely useful, and the programmers learnt a lot from trying to integrate it with Blender.
The latter is more of a problem, as there's features in Yafray you can't get at in Blender, and recently that appears to have taken a step back (around 2.41 to 2.42 anything with image textures seems to die) whilst Blender's internal advances massively.
The other issue for Yafray is Sunflow, which whilst it doesn't have the integration but has got performance and a perception of more active development.
I only want to point out one more thing. One has responsebility of things, you invited and you are familiar with and I don`t want to belive, that it`s only a Yafray problem. Blenderorganisation -in my opinion- has to do a very critical self analysis of it`s own role in this whole process.
Bart, I see some grammatical errors in the article. I'm sorry, I've studying english for five years, still I make awful mistakes when writting. Can you check it up a bit, please?
need Final Gathering... Please....Â
That said Sunflow is looking to be the new happy until this new Yafray is here... Go Render power Go !
I always wondered why Yafray technology isn't directly included into the Blender renderer.
I know Yafray can also be used by other 3d apps to render their things, but instead Yafray it would be the Blender renderer.
Don't get me wrong, Yafray is great. It only sometimes looks to me as people are working on islands... A great example is the sculpttool... which first was an external application, and now is integrated into Blender 2.43.
Thanks for your comments.
First of all, I've used the 'disastrous' adjective in a light-hearted way. I wish YafRay never lost its aura of informality and unexpected succes. In fact, I have had a lot of fun these last months writting yafray docs, and I have learnt an awful lot of things about raytracing, lighting and shaders, which I hope its helps to improve the quality of my work in the future. I suppose that Lynx is working in his code for the same reason, he is 'having' fun and learning a lot of things.
Remember, it is yet another free raytracer. Therefore, I think I can speak for the entire yafray team (present and missing ones) when saying that we are not interested in any any kind of raytracers competition. Fortunatelly, there are great options available apart from YafRay and you should always include them in your considerations about what engine to choose. Regards.
hey alvaro, we had been using blender+yafray with our interiors and stuff (until we had to switch to max+vray cuz we took on a few partners) working with max has reallly helped me appreciate blender!
we want to have yafray creep back into some of our high-end work again...any news about general speed/detail of GI in 0.1.0?
thanks for all your hard work person-to-person in the forums and to the masses via ur docs!
Hmmm intresting comments here... i agree with the comment of Toni Grappa - but i see that opensource
is a very dynamic pool of different people ( ..or one ) working on a projekt. How larger this "personal pool", how larger the possibility that the projekt survive - also if someone ( .. or group ) who leave. Mayby thats the risk
of opensource.... but .... no risk no fun. And by the way: As i was the first time really intrested in blender,
i found the information that blender is gone - the company crashed. Hmmm... for a dead projekt blender
looks fantastic!! ;)
is not Yafray and Yaf(a)ray two sepearte programs?
I wonder how the DoF will work like. Yey! I'm so excited. Keep going guys! Good luck. *thumbsup*
Reynante M. Martinez: If nothing else helps, you can always render the depth pass in Blender and then add DOF to your image by using the Defocus node.
I think it's unfair to speak about "the dark side" of the Open Source. You have to remember the people work for free and on their own time. Everything you get, you get out of generosity. Hence it is unfair to put enormous demands on the developers of the Open Source software. You cannot expect to get everything you want unless you ultimately code it yourself or hire someone to do it for you. Blaming the developers about lack of development is about the worst thing you can do.
I wish the best future to the Yafray project.
originalsurfmex, AFAIK yafRay will remain a biased raytracer, we will keep features to trade accuracy for render times, such as photons, irradiance cachÃ©, etc. In fact, Lynx is more or less doing the core rewritting planned by Jandro, and with that rewritting, Lynx is probably able to fix the more obvius yafray shortcomings.
In fact, as suggested by ccherrett, the 'old' YafRay and the new 'yafray 0.1.0' based on yaf(a)ray are two different programs, under the same name.
@Tony Grappa: How would you like to see the yafray project beeing held? Could you give us an idea?
From my distant viewpoint, I can say that the organization of implementing Blender work pretty good at this point. We see features getting into it fast - they are stable and they just work. The only example I have from the industry is the XSI-team, Softimage Co. They formed a strategic development relationship with Microsoft to develop their XSI founded. They could do it with a lot of financial support - Blender is doing it without! Well, I know, who's my winner in terms of management here ;)
I hope, that webmaster of yafray.org reads this comments too, because it is bothering to click twice at cancel button, when you want to visit yafray wiki :-(
BeBraw said on Feb 7th, 2007 at 8:10 am:
> I think it's unfair to speak about "the dark side" of the Open Source.
> You have to remember the people work for free and on their own time.
> Everything you get, you get out of generosity. Hence it is unfair to put
> enormous demands on the developers of the Open Source software.
> You cannot expect to get everything you want unless you ultimately
> code it yourself or hire someone to do it for you. Blaming the
> developers about lack of development is about the worst thing you
> can do.
While looking at all the different Open Source projects that are aimed to do pretty much the same thing (and I'm not talking just about rendering here), I often keep wondering how powerful application could be created if all the developers worked together, instead of all of them starting their own project. In a way I think it's sad that all the talent is scattered across so many different projects, but then again I understand that not everybody wants to develop using the same programming language, or structure the program in the same way. Still, I keep wishing that they could just work together...