Developer Meeting Notes: April 6, 2014
My personal favorite this week: Tamito Kajiyama starts working for the Blender Development Fund on Freestyle development!
Hi all,
Here are the notes from today’s meeting in irc.freenode.net #blendercoders.
1) Blender 2.70a update
- Check the long list of fixes already.
- Proposal is to do the ‘a’ release mid this week. Sergey Sharybin and Campbell Barton will coordinate it.
2) Release targets for 2.71
- Updated the project page (see 2.71 targets list below).
- We should move to “BCon2” now, but several developers have to confirm whether their proposed projects for 2.71 will make it (Compositor, Multiview, Baking, …).
- Martijn Berger tested new OpenCL drivers (private beta). Even though it’s a big step forward, it won’t justify enabling OpenCL in Blender for the next release.
3) Other projects
- Per April, Tamito Kajiyama works for the Blender Development Fund on Freestyle targets. Everyone who’s interested in his work can join the NPR mailing list.
Thanks,
-Ton-
2.71 targets
Release Notes for features already added.
- Python 3.4 (fixes annoying windows issue).
- Split Normals, stage I.
- F-Curve Easing patch.
- FBX binary exporter (7.4). Already in master.
- soc-2013-paint
Candidate targets
- Cycles Baking
- Branch & Patch migration list
- Cycles SSS on GPU, AO node
- Multi View (stereo 3D)
- Compositor cleanup and restructured node conversion
- Freestyle: Memory consumption improvements
- Freestyle: Textured strokes
I use the Ambient Occlusion shader in cycles a lot.
It would be very helpful to have 2 inputs to control the dark and light balance.
I once used the AO pass in the render layer and had that pass adjusted via contrast nodes in compositor to enhance the dark areas. Playing with compositor nodes may do the trick without having to redesign the AO shader.
I’ve been wanting Cycles Baking for so long! I really hope it makes it into the next branch!
It’s already available to testing at BlenderArtist, with early builds available at GraphicAll.
Cycles SSS on GPU is the holy grail for me! Fingers crossed they can make it Nippy!
Oooh, Multiview, I hope it makes it into this release finally!
I wonder when fully integrated rigid body physics is going to come around… current system just doesn’t cut it for anything but the most simple falling animations with mostly kinematic bodies doing the motion.
How exactly does it not cut it? I’m curious.
1) Abuses the time-stepping which causes constraint stiffness to vary with Blender frame-rate
2) Absolutely no way to implement motors (which are half the reason anyone uses Bullet in the first place)
3) No programmable drivers (only curves for kinematic objects, no way to implement something like a non-linear oscilatory system with discontinuities)
4) No option for other solvers including multithreaded and MLCP
5) No easy way to give motor targets based on time (largely because “time” is very convoluted in Blender and nobody can figure out what they mean by it, second because it’s basically impossible to import a simple csv without knowing the ins and outs of blender python)
6) No easy way to export results
I’ve sent them suggestions on fixing 1 (not too difficult to do, but they like ignoring it), but 2-4 require rewriting the entire C interface (to properly work) and 5-6 requires functionality external to bullet side that blender doesn’t currently offer.
Tamito Kajiyamaさん, 期待してますよぉ！！
よろしくお願いします！！！
I hope Tamito Kajiyama will do great things with Freestyle. NPR is often undervalued.
GANBARE, Kajiyamasan!
Cycles baking :) FTW !!
Very disappointed to not see any UI targets in 2.71. Was also disappointed by lacking ui improvement in 2.7. New version has only added more hotkeys onto the gigantic pile of them, so problem is only getting worse.
Most topical way I can describe how to address Blender UI woes:
Think about how to redesign the UI so that it could be used on a tablet (which is where all development is heading in the next 5 years, btw). On tablet, you have limited space and mostly only button access to features (Can’t rely on users having keyboards except popup one, but that takes 2/3rds – 1/2 screen, so not really feasible to use constantly while working).
Within these two constraints, there are a few almost inevitable conclusions.
-You have to decouple Blender from hotkeys
-Plan single click solutions to every option.
-Adopt some iconography, because words either take too much space or are too small to read.
Given the limited screen space, you should also consider how to make all menus tuck away. You have to maximize workspace (something that none of the packages do well right now).
Wait, why would someone who is seriously using blender or any other 3d software limit themselves to a tablet?? I mean, you get the right tools for the right job. Like if you were hired to dig a hole and you came out with a spoon and started on it, people would probably laugh at you and generally think you aren’t very serious. After all, blender uses the keyboard for more than just hotkeys. If you are keeping organized, you are naming things a lot as well. Also, using hotkeys gives a great speed increase in your workflow. All that said, I do see a place for tablets like the cintiq, but that would be for just the sculpting part, or texture painting. That would still be connected to a desktop. I can definitely see the standard computer users going to tablets, but a professional cg artist??
Moore’s law for speed increase, and then factor that a lot of tech from the towers is quickly being migrated and made efficient into the mobile space. Tablets will be more than enough to handle any 3D applications in five years. Already in my industry, games animation, we have many contractors moving to laptops, so that they can enjoy open air. More and more laptops are moving to tablet laptop design. There is a growing swell of art programs blossoming on the tablet space. Wacom is even creating a tablet pen to support them.
Well.. The trend should be apparent.
The reason is that people like to be mobile, and they don’t want to carry around bulky machines. If I can sit in my living room and tinker on my tablet on creature animation curves, I am happier and enduring less relationship issues than the person tethered to the heavy tower.
The portability is definitely a good point, but I wasn’t talking so much about them being limited in speed so much as limited in physical buttons. Buttons that are always there and do the same thing. Maybe it’s just me, but I find physical buttons much faster to get to than digital ones. Like, a volume knob instead of using the slider in the task bar. Part of the reason it’s faster is you can train your muscles to find it in an instant. Without having to even think about where it is. Of course it’s probably possible to come up with a interface that is just as fast for touch only using gestures or something similar, but that is so different from where blender is now, it would be a completely different program. Also, I’m pretty sure you can do most everything from the UI that you can do with hotkeys. Maybe I’m wrong.
You’re right. I believe everything is technically accessible through the interface. However, the organization is such that no one does. Tutorials about blender always teach the hotkeys. People always use the hotkeys. The interface is scattered and non-intuitive, so programmers and users use hotkeys to get around it. I’m not against hotkeys, but the core experience should be accessible and understandable through the interface. Addressing Blender to work on tablets would force the development team to tackle this glaring issue with the program.
As for it not being blender anymore, is the program defined by its interface or by the feature sets and developers? I think as long as it’s an open source program under the guidance of the Blender foundation, there will be something undeniably Blender about it. I think the question is if it will address its interface to become a boon to 3D industries?
The day Blender Foundation takes UI seriously, will send Autodesk and a tonne of other other software companies in a panic.
No, not really. Let’s examine exactly why:
1) While Blender has its own many strengths, it as well has many weaknesses, and a UI change is just the face of its weaknesses. You can’t even begin to address an improved UI until you first deal with an improved internal structure for Blender, which in itself will necessarily require an improved style of developing Blender altogether.
The further we develop all ad-hoc, anything-goes, tack-new-feature-here, stuff-it-in-the-full-closet style with Blender, the more its UI will reflect as much. A new UI at this point would just be a shortsighted attempt to mask a deeper and growing issue. Even if you change the UI completely, Blender still behaves like Blender, in many respects. Some for better, but some for worse.
2) Contrary to popular belief around here, nobody’s constantly eying Blender.
Some other companies might check around Blender at times, and might borrow an idea every now and then (which, by the way, Blender borrows just as many ideas from the others), and some may even take a smart jab at Blender at conventions once in a blue moon.
But generally, unless Blender just stops its development aims and starting conforming to the standards that the entire industry is using, which hasn’t and likely won’t happen, generally no other company’s going to “panic” over what Blender does.
Raise an eye curiously at Blender’s ambitious moments, sure, but they’re not worrying–they’re too busy being industry-standard tools, introducing the next generation of standards. There’s no reason to panic about a swordsman when you’ve got a gun (well, unless if you’re a bad shot or out of ammo, of course).
Open-source nature of Blender isn’t threatening major business for these software companies, largely because the effort isn’t timely matching the standards, and also simply because most of its users aren’t even businesspeople. Despite its large user base, Blender isn’t even close when it comes to the ratio of professionals using other software more comfortably enjoy.
The majority of the Blender user base are hobbyists, enthusiasts, and students of 3D graphics. Those who do use Blender for a professional living are a minority among the user base. And unless you work freelance or own a small business, you’re going to be using industry-standard tools provided at work, regardless–and it won’t be Blender.
3) There are things about Blender’s open nature that will always make it tough for becoming some sort of industry-standard tool.
In particular, we find difficulty just adopting the technologies that everyone else freely enjoys, such as full FBX support without worries of license issues and workarounds, and direct seamless interoperability with other commercial software like ZBrush and Mari.
Due to its open nature, Blender is the fastest software in development in one regard, but the slowest of all the widely-known 3D packages to adopt industry-standard and emerging-standard formats. Do we have Alembic support yet? How about Ptex? How
about OpenSubdiv? Full PSD support? Vector displacement
maps? Substance Designer substances? No? Then what’s the worry for
the other guys?
5) What works well for many in the Blender user base doesn’t necessary work for other people in relevant industries, even though some people have made it work with themselves with effort. Just because Blender is free and open-source, many around here think that’s strong enough reason why it’d be someone’s preference.
These were rather harsh, I say all of this as a realist who loves Blender. I love Blender and its community, but I wish they both could use a deeper look at ourselves before championing Blender as what’s “threatening” the commercial software industry.
Just addressing your claim of tower technology migrating towards mobile, I’ll have to disagree.
Moore’s law of increasing rates with technology isn’t relevant to any growth of tablets’ adoption over
desktops–it’s just relevant to the increasing tablet speeds, portability and affordability. And most
of the tech going into mobile devices are significantly scaled down, and
made so efficiently for the mobile platform, it mimics the results of
stronger devices.
Tablets are mobile, which is useful for many things, but not ideal for the most demanding jobs. Tablets will likely never be ideal for high-end work, just from simple physics alone. High-end rendering and processing takes horsepower, which creates high amounts of heat, which is much more detrimental to smaller units like mobile devices.
Mobile units are also a lot less adaptable as machines. You usually have to buy a new tablet if you want the latest tablet tech. Mobile devices can only handle so much. And mobile devices are fragile.to heavy-duty use–even sturdier notebook computers will wear out sooner than a desktop, from frequent use.
People also commonly mistake the high adoption rate of tablets as it “progressing” passed more stationary desktop. Tablets haven’t surpassed the usefulness or matched the capacity of a desktop–more people just replace tablets far more frequently than they need to replace their powerful desktops, which lasts for years without them needing a new one or being upgraded.
They tend to use their mobile devices more for the more trivial things at a high frequent, one that’d be a pain to do quickly on more stationary computers. Most mobile devices are used for things like checking the weather, watching a show while away, catching the news for a minute, or checking train schedules–stuff you wouldn’t bother to do on a computer, if you don’t have to. But for everything else–high-end work and high-end entertainment–computers can’t be topped.
A powerful desktop (such as a workstation) offers things tablets can’t begin to offer, such as a wide variety of cooling systems, maximized expandability, and long-term sturdiness with use. They offer nearly-infinitely more potential in terms of heavy-duty work and adaptability.
Workstations in particularly are built to work long hours doing hard work, with all the power and expandability you need to work better. The likes of the stationary computer alone carries the straining growth of technology, while mobile rides on its giant back. Mobile devices have made people more flexible, but stationed computers have made people stable.
Though, I do find it interesting that you’re able to work this way. In fact, i’m rather impressed that you can work so exclusively on a tablet. Though, just as to what you said about tablet’s importance, I just wanted to chip in why I don’t think “people” will be replacing desktop stations for tablets anytime soon.
When it comes to technology, I have a motto:
Sometimes the best option in life is options.™
I disagree to everything you say. Mobile market is accelerating faster than PC, because of the huge market forces towards mobile. We’re coming out of an age where mobile used pretty ugly standard hardware, and into an era where people do want the best experience. There’s a lot of knowledge in the x86 market, and it’s just taking development time to bring it to mobile market under less power consumptive ARM architecture. Watch NVidia for this. They’re doing fantastic things. You’ll see PC worthy tablets in the next year or two if half of their claims come true.
As for physics, ARM is a less power consumptive technology, and power is heat. Not to mention we’re just starting the era of 3 dimensional processors (NVidia), and we haven’t seen optical processors yet, which would cut heat from resistance to nothing.
Mobile is leaping in bounds and hurdles. It’s actually catching up to the PC market. You’d be foolish to believe tablets couldn’t ever handle what PCs have been doing excessively well for 15 years.
[Weird, my previous comment disappeared. Try again…]
It would be a farce to try to have a productive Blender session on a tablet. Show me a company that asks its designers to use AutoCAD, Maya, SolidWorks, Pro E, or any Adobe application on a tablet for 8 hours a day. In all these programs, and Blender too, the precision required for complex selections and adjustment of multiple variables in parallel makes a tablet interface unusable. 80% of your time would be spent swiping and pinching just to get the required view and/or controls to appear. However, very large super high-res touchscreens might just be usable, say a 27″ or larger 4k screen.
1) I never argued that the mobile market isn’t accelerating faster than PC. I told you why the tablet mobile sees greater sales and demand than the PC market: tablets need replacing much more frequently than PCs do. You’ll likely buy two generations of tablets during the lifespan of one desktop before needing a new one or to upgrade its parts. However, I thought we were comparing technology between tablet and workstations, not their respective markets…
2) We’ve already entered an age of options, where mobile technology gives people flexibility, while powerful PCs give people stability.
3) You keep on bringing up lower power consumption. You fail to understand that the efficiency is for the tablet, not technology overall. You can use less electricity with tablets, but they’ll never match the output of workstations, which uses its own relative rate of power consumption.
If it takes a workstation cutting-edge horsepower to achieve the likes of real-time physically-accurate visuals, or powerful processor-heavy operations for voxel sculpting, or million, billions, or now even trillions of polygon for visual FX nowadays, without taking shortcuts in fidelity, what makes you think a measly tablets can handle such soon?
Better tablet efficiency caters to extending its own fragile battery life and keep its own fragile system cool. A tablet running better doesn’t mean it’s running on equal footing to a computer–it’s just beating its personal best.
4) Nothing in their mobile line compares to the likes of the Titan Z, which will find home in workstations worldwide. And don’t hold your breath on NVidia’s hype–they’re no stranger to hyperbole. They’re businessmen, selling a product. To sell products, they must first sell promises. Stockholders come first, consumers second.
5) About the ARM processor tech taking less heat, again, less heat doesn’t necessarily mean a thing when you’re comparing it to mobile tech. The efficiency for mobile doesn’t bode well as a measurement towards matching the output power of workstations. Benchmarks are what you use for such, and in a benchmark test that compares the same load of work across a tablet and a workstation, workstation will come out on top every time.
6) 3D ICs generate even more heat than standard ICs, which requires even more focus on cooling, which is harder to factor in with a mobile device’s slim design. Optical processors are theoretical at best, and nobody is sure of its promises yet, as there hasn’t yet been any practical example available to test, compare and judge its resulting limitations.
7) Mobile is leaping in bounds and hurdles…for mobile. It’s catching up to the PC market–not its rate of technological output. Comparing markets and comparing technologies are two different things.
8) Never claim what I’d be foolish to think, esp. when you keep jumping topics and you’ve yet to defeat my main point.
My point isn’t that tablets couldn’t ever handle what PCs can do. My point is that tablets will never be sufficient for high-end, demanding work. It’ll never match the quality of what a workstation produces. And it’ll never be sufficient for heavy-duty usage, esp. all while trying to keep to its mobile design.
What’s the point of touting along an detachable keyboard, mouse and extra speakers for your tablet when a notebook (or ultrabook) can handle the same thing, usually for a lower price? What’s there to gain in trying to make tablet something they’re simply not? If anything, what you’re seeing is tablets trying to become PCs, losing
more of its mobility aspect by just becoming essentially what notebooks
already are.
If you need that much performance out of a tablet, that you need to turn it into a notebook, you might as well go with something that was built for performance to begin with–not some mere emulation of performance. Tablets will get stronger for themselves, but they’ll always be scaled imitations of the power and adaptability of workstations.
They need to a complete redesign from the bottom up, the sooner they realise this the better. At the moment they are just papering over the cracks.
I don’t know why you say this. Do you use Adobe CS applications? They have far nastier, cramped little button panels (non scalable too!), and plenty of totally obscure button combinations. And if I ever upgrade, I have to spend ages working out where to copy all my workspace preferences files (scattered all over the disk), or face hours resetting everything. In Blender, it’s just one click on the splash screen and I’m done!
The difference is Adobe CS applications are industry standard learn them and you can get a job, so it make sense to put up with Adobe CS idiocies, there plenty of them, while very few people use Blender in a professional settings and so it doesn’t make sense to put with its idiocies when there are software which are much more widely use by industry and in my experience easier to learn.
Absorption for GPU please! Had a project already, where I had to switch to cpu, which is about 7-10x slower in my case.
just keeps gettin better and better. mabye i missed this, but everything is updated now, as far as the UV’s and properties window with node editor, and painting with texture’s, sculpting texture’,UV image’s and properties window. all updating better than ever.it use to be i would make a texture for a brush or whatever and it would be hit or miss if the information would show up across all window’s.Now all information show’s up where it should. hats off to all the developers and contributors and anyone who work’s to make blender what it’s become, a f’n great piece of software.sure has done ton’s for me. long live the blender nation.:)))))))))
Multiview! Multi-View! Moo-Tea-Vue! ;) make it happen, Felinto! trust the force!