First RC of the 2.70.
Awesome improvement of Cycles, as usual. Volumetrics are there 🙂
And a lot of new features, everything here : http://wiki.blender.org/index.php/Dev:Ref/Release_Notes/2.70
Smoke and motion blur in Cycles, awesome, as usual!
Big news for Cycles: Brecht will work on Arnold renderer for Solid Angle leaving the Blender Institute. I’m really curious of how Cycles is going to evolve. (but I’m also quite curious because we’re using Arnold at work and Brecht can bring something to us too!)
That’s a quite sad news for Blender, Floss and 3D. But that’s also the strongness of free softwares that are not made by just one person/company. I’m sure it will well evolve anyway.
My first reaction was the same : oh no, why?!
But actually, I don’t know developers of other softwares. Do you know at least one developer in the Maya team? I personally don’t even if I use it more than Blender at work. And I don’t believe that the big studios who are working with Maya really care too. Blender has something special for a software: the coders are more famous than the users! Weird but this is cool. Brecht has done a lot of Blender features, including Cycles, thanks to him, awesome job. But he’s not alone. Thomas Dinges, Dalai Felinto, Sergey, all the volunteers. They all are good.
Steve Jobs died and Apple is still alive. I don’t want to think that when Ton will leave the Blender Institute, Blender will stop too. Same thing for Linus Torvalds and Richard Stallman.
Perfect moment to learn C++ and start to code core feature! (I’m joking !!! :D)
(And Brecht is still welcome at the BBUG events :D)
Brecht will be missed, but I think you are right. One or more others will step up and continue his work, hopefully with the same dedication.
Talking about learning C++: This afternoon I’m following the first class of a series of lectures on General purpose GPU programming and we will use both CUDA and openGL in C. I am very curious about this.
I have always wondered how hard coding in the Blender core, or integrating something like cycles actually is?
The biggest hurdle to take in Blender core development is really starting off with it. The code consists of millions of lines of code, and they’re not very thoroughly commented. Luckily its all kind of structured into different source files in different directories, but its still a literal haystack to be looking around in. And then theres the actual code: of course things are organized in a certain way (DNA, rna,…) and its worthwhile reading some articles on the Blender site and Wiki to get at least some starting idea of how Ton put it all together in the beginning. When I started off with core, I was also still not very experienced in C also, so it mustve been harder for me than for someone already good at it, but even then hours and hours were spent trying to gain “first access” to the code, mainly scouring many source files for the small part that was doing the things I was trying to change. My advice is to persevere even when things seem all too crazy, cause its this “looking around” that shapes your broader knowledge of the core code structure.
When you do (persevere) things will grow easier soon, especially in the areas of code you already explored, and I was able to progress to coding more intricate things soon enough by growing into the code, but even then eg. coding the preselection highlighting drawing code I made, took many iterations, cause each time while coding one solution, I discovered Blender structure really demanded some other solution and the one I was trying didnt really fit in so I had to change it all.
So when integrating something like Cycles (not even touching on the intricacies of constructing a renderer itself) you first need a lot of experience with the Blender codebase, since it has repercussions in many parts of Blender that need changes, and often the already existing code was not made with the presumption of someone wanting to do the things you want to do.
Also debugging can be a nightmare, mainly because some bugs hide in a part of Blender you didnt explore before and are very difficult and slow to track down.
Of course theres always the IRC freenode #blendercoders channel where you can ask questions when you really get stuck, but people there are quite busy also and most of the problem solving will be up to yourself.
Still, glad I all did it; the experience I gained in reading and working with other peoples code alone is invaluable.
This is very promising: node material and compositing for real time GLSL viewport.
Nice as always but OMG
Pie menu arrived just a while ago and we must already suffer from clusterfuck pie menu worst than Maya 🙁
Ton, please find some funds to pay for usability engineering. My eyes are bleeding already…
Live & Learn
Maybe try to do your own pie? there is so many uses of Blender that makes impossible to have pie menus that fits to everyone.
I think that was the goal to add pies in the python API and it seems really easy to do.
Nope. It has nothing to do with me or my personal preferences.
The core ideas that were the selling points of the pie menu system are :
- Muscle memory
- Clicking in a general direction around the mouse cursor instead of precise click on a far way button
The clusterfuck above doesn’t fit either of those two paradigms.
Live & Learn
The new multi edit object of the 2.75 release is finally there and is working for most of the properties, even custom ones from python and modifiers!
@georgesl, this is for you!
Download : http://download.blender.org/release/Blender2.75/
Release Notes : http://wiki.blender.org/index.php/Dev:Ref/Release_Notes/2.75
Not yet and not full nodal, it will never be full nodal if I understood it well but that will open many doors anyway.
No idea for when it will be in the trunk and available in a release. Maybe for 2.8?
You must be logged in to reply to this topic.