Weeklies

Weekly #29: March 6, 2015

9 March, 2015 |  24 Comments | by Elysia Brenner

Welcome to weekly recap #29! All of this week’s work went toward finalizing the opening shots of the movie. Watch the full presentation below with all the team’s commentary, and then move on to the highlights to get a closer look at some of the art, animations, and new developments from this week.
 

 

The Highlights

 

This Week’s Guests

We were happy to welcome (back) to the studio this week designer Matthijs de Rijk plus two guests from game development studio Thorworks. Watch them introduce themselves and talk about their work at the beginning of the weekly video above.

 

Getting Animated

The opening scene layout is being replaced by more and more detailed animations, especially after this week. We’ll break down here what each of the animators have been up to, but for the bigger picture of where the film’s first moments are going (and it’s a subtly but significantly different place than last week), watch the latest version of the first 5 minutes at the start of the weekly video above, and download it to watch whenever you like (without the commentary) from Mathieu’s folder on the Cloud. The tighter opening focuses more immediately on the most important interaction: Franck and Victor’s meeting. This means no more (other) talking sheep, but Mathieu made the choice to leave them out to create more uncertainty about whether or not Franck will succeed in his mission to off himself.

 
Hjalti has been adding more kicking, struggle, and other details to the very first shot of the film: Franck’s hanging attempt. Check the video below for 4 versions of this 30-second scene showing how the animation has evolved over the week to “pretty much finished,” including the addition and disappearance of what Hjalti calls “the talking branch effect” — you’ll know it when you see it — and a rocky hoof collision. The trickiest part of this shot is getting the audience to be able to feel the weight of the falling branch without fully seeing it.

And here is the same shot (with some missing frames) after Pablo’s attentions to the setting, lighting, and hair.

 
Manu has been working on lighting the dragging action you saw in the last weekly. See Franck’s big drag in full fur from two angles — although some missing frames make the second (front-view) take look like a scene from Cosmos Nightmare Laundromat.

 
Beorn dedicated his first official week on the Gooseberry project to the next phase of Franck’s branch dragging, following the above shot as Franck and his branch burden reach the cliff. The focus here was on the quadruped body mechanics as Franck stumbles over some cliffside rocks, and Beorn estimates this version is about 50% done.

 
Sarah’s animation takes over from here, showing Franck pushing the log toward the cliff’s edge (and being interrupted by Victor). In this video you can see where she started, with Franck pushing the branch with his nose, to her reference footage using her own head and a studio couch(!), to 2 more versions (including a look at the body animation beneath the fur) leading to the final version of the week!

 
Finally, Hjalti also worked more on Victor’s walk cycle. Specifically on his walk when he leaves Franck behind for the tornado to pick up.

 

Tornado Week (er, Month)

Speaking of the tornado… Andy’s first week of colorful tornado experimentation has come to a close…but there will be much more to follow! The first week started out with a bang (a woosh?) with these spectacular still and moving images where Andy experiments with a watery vortex (after all, it’s supposed to come from a washing machine), spin, smoke effects, and adding color. Some shots were deemed “too alien” and in some the smoke obscures the vortex beneath, but each gives a unique take on this difficult but central figure in the story. See more of these takes (including a couple that are approaching a final concept) in Andy’s folder on the Cloud!

Alien tornado
Smokenado

 

Island Designs

The cassette player is ready for Victor to wear after Manu’s final tweaks. Check out what he’s done right here, and then you can find Hjalti’s simple rig in the Cloud Repository under libs > props > victor_casette_player.

Victor's cassette player

 
Manu has also been working on the model of the island as a whole, seen when the camera pulls back from Franck’s failed hanging. It’s starting to look properly desolate and gloomy, like a place a sheep would not want to go on living…

Gloomy island

 
Meanwhile, Pablo has been painstakingly adding all the details to that opening sequence, making sure everything is right from the lighting of the hooves to the curved hairs (thanks, Lukas!) to the moss on the rocks. You can see where things ended up below, and then all the steps and changes along the way in 14 renders you’ll find in Pablo’s folder on the Cloud.

Opening shot

Franck by the cliff

Note: Andy discovered the issue behind some of these images’ longer render times. It’s a UI issue, with a somewhat hidden particles setting under the render settings as the culprit. Turn this setting off to drastically reduce render times! (See the weekly video at 1:14:42 for a demonstration.)

 

In Development

  • Lukas has been working on creating a(n almost) complete workflow test that covers rigging through coding to animation. See Lukas demo all the steps at 45:43 in the weekly video. Of perhaps the most interest is that the cache library is now linked, meaning you can now cache linked groups with proxy armatures and linked actions without needing to create any clumsy Python overrides.
  • Francesco spent more time working on the Cloud this week, building the foundations of a shot-management tool.
  • Gabriel spent the week fixing stuff, such as render farm file management (remember those missing frames?).
  • Antonis has implemented a new OpenGL algorithm that uses instancing and geometry shaders at the same time to drastically improve depth of field (see pics below). He expects to commit this addition to the Gooseberry branch this week.
Before

Before

After

After

  • In addition, Antonis fixed some sound synchronization issues caused by bugs in how the audio library calculated sound offset. This should be fixed now!
  • Finally, Antonis also worked on animation paths to allow animators to tweak motion paths while transforming things. See his video demo just below, or watch his demo with commentary at 55:18 in the weekly video. And expect to hear more about this at the next weekly!


24 Responses

  1. Joe Somerville says:

    Nice! I’m glad to get a first look at the tornado. I really like how it looks “watery,” it fits nice and well into the washing machine theme, although I would like to see some sharper, more frequent contrasts between the colors. I loved the first concept art, how it looked like a smoke simulation but with all the colors of the rainbow.

  2. Warren says:

    All I can say is … WOW!
    Things are really coming together now, It’s so amazing to see all the planning and hard work come together to make a real cohesive “world”.

    Finger’s crossed that some of those test tornadoes make it onto the Cloud, they’re some of the best examples of blender simulation I’ve seen yet, and I would Love to examine the blend files.

  3. Really liking the watery tornado. It shows off the colors well.

  4. Karlis says:

    By reading these posts and watching videos I am more and more convinced that it would actually have made more sense to make a good hair system first and then make a movie. In blender podcast also they said that hair and particles are basically used everywhere.. sheep and other characters as well as grass, moss and so on.. I mean.. Of course you can make a movie with tools Blender has now, but at least then don’t use features that everyone knows work bad. All I hear is how many problems there are in hair system every time and that sounds to me as time wasted by all of people, artists as well as developers. I know from my personal experience how time consuming and frustrating it is to work with that system and in the end you still can’t do what you want. That does not sound as a well planned project. Don’t get me wrong, I like all the great work which is done here, but don’t you agree at least a little bit?

    • Neil Burgin says:

      @KARLIS: Dude, they just overhauled the hair system in many ways. Haven’t you read the release notes for 2.74? Furthermore, these open movie projects, with the artists closesly collaborating with the development team, have always been seen as prime opportunities to FIX deficiencies in Blender. Before Tears of Steel, Blender didn’t do well at compositing with live action, that didn’t stop them, rather they improved the compositor WHILE IN THE EARLY STAGES OF MAKING Tears of Steel. And that’s what they have done with the hair system. They have been improving it as they have done this work, as preproduction on this comes to a close on this film, they have also more or less finished improvements to the hair system, these improvements have landed in the development version of 2.74 and will be in the stable release for users shortly.

      In short, they knew about this, and THEY FIXED IT. That’s not poor planning at all.

      • Neil Burgin says:

        Maybe I was a bit rude. To clarify: the people making the open movie projects are not using the stable release the other end users get. They use a development version with lots of fixes and other new goodies the rest of us don’t have yet. If you thought they were using the same version as the rest of us, then your confusion is understandable.

        • Karlis says:

          I am not offended in any way, don’t worry. I am not saying making movies is a bad thing, I think it is super great and that really tests Blender to the limits, but if you already know there is a major problem with something you will have to use at such a large scale then it would make more sense to solve it and then start making movie to test it in real working environment which would probably result in refining that feature even further. Or you could make it without hair at all, that is a style decision only. I really don’t think that movie would be better or worse because of that.. oh wait.. I actually think it would be better, because artists could spend more time on creative stuff and not tackle with technical issues, although they still will would have to do it because there are many other areas that they are improving and testing.

          • Charlie Ringström says:

            I don’t even think the simulations are baked here. The rig is made to simulate hair movements as good as it can, but later they bake it to make it look realistic. Am I wrong?

            By the way, watching Big Buck Bunny’s or Sintel again, the animation didn’t have “Pixar quality”. But watching this, I’m actually really “proud”! To be honest, isn’t there a Oscar for something like “best animated film”? Seriously, I would NOT be surprised to find this winning. Keep on the good work!

    • Ville says:

      Karlis has a good point: a good and supported development run before making the movie. But in intense project work, it may be the best accelerator to have artists and developers to be in the same situation: make it work as needed. As Rosendal has said: Blender is for artists. Open movies are done, so artists can break the limitations of Blender and say to developers what changes are needed to work better. That iteration; loop back from artists and enhancements from delvelopers makes the development very fast and accurate. Who ever invented this method, should be given a medal.

    • Elysia Brenner says:

      Hey Karlis, I’m afraid that whenever the hair (or any other) systems are made the process is going to be riddled with minor frustrations and initial bugs — just part of the development process in general! But by developing the systems side by side with the shots that actually use them, there’s unparalleled opportunity to test and get it right. ;)

    • Karlis: in contrary to what might appear – we planned it extremely well. Already last year in June the preparations started, including investigating a complete new particle/hair system. During September it was decided that a complete new system would be too risky. We just don’t have the development powers to design and implement a (node based) particle/physics/modifier system to satisfy production quality within a 3-6 month period.

      Instead the old system got fixed up (and quite well even!), and new modules (like Alembic caching) get added in a way that they are still usable in a new design. The experience Lukas will get with this production, to realize complex simulation shots, will be of invaluable importance for his insight in designing and implementing a new hair/particle system in the course of 2015.

      The fact people are cursing the particle or hair system (or actually cursing all of Blender) is just part of any production with committed team members. They’d be cursing the new particle system just as passionate :)

      -Ton-

      • Karlis says:

        Ok, just saying how it looks like from the outside. I no doubt agree with that cursing thing..

        About development powers.. I know you want people to subscribe to Blender development fund and have a stable income, but if you really have to make something you don’t have enough powers in, you could make a separate fundraiser to get more money, explain what will the benefits be for all. I don’t think that would interfere with current plan. For example, I am paying a little money every month and that is what I know I will be able to pay in future no problem, but once in a while I could invest a little bit more. Of course I can pay that amount any day if I would really want, but problem with people myself included is we have to be convinced that it’s worth it. It’s not like we don’t have other places to spend it. The best way to know that it’s worth it is to know where that money will go. I have heard you say many times that paying for features is not a good way of doing it, but is it really? If you are the one who initiates the fundraiser and you know what is needed, I don’t see a problem. There are many people who don’t like subscriptions at all, so you are not getting their money and you are not getting more money from others like me who are paying, but don’t want or can’t pay more monthly. Not asking people more money for particular projects just sounds a little arrogant to me or something like – we don’t want to make any promises, but we want them to pay us monthly anyway. Again, to not look as complete asshole, I must say that overall you are doing good job. ;)

        • Elysia Brenner says:

          Hi Karlis, that’s exactly what is going on now. The Blender Development Fund actually does not support this project or most of the development around it, as ongoing animation projects are not run by the Blender Foundation – the Blender Cloud is the fundraising arm of the Blender Institute (Blender’s business arm). Please see the Join page for more information about the benefits. It’s also possible to contribute one-time donations to the Blender Institute…see the donate button at the top-left of this page. :-)

          • Karlis says:

            Yes, I am member of Blender cloud. That’s not exactly what I meant.. but it is great, yes.

          • Karlis says:

            I meant that people would be more happy to pay one-time donations if they could choose the particular thing they are investing in. Like depthgraph development or particle system development or viewport and so on. On every project could be written what benefits would user get from that thing, for example, if depthgrapth is fixed, it opens up possibilities for better tools for rigging and animation, like editable motion curves and so on. Or you can choose some other ongoing project. Although of course everything is kind of connected and particles probably would benefit from depthgraph also, so the best part is, that Developers know all those things better and give us only a few choices.. maybe 3 max. That at least would be something.

          • Elysia Brenner says:

            It’s a nice idea, but I’m afraid it wouldn’t work in practice. There needs to be flexibility. And sometimes the most important thing for the project may not get the most funding, etc. This is also why almost no charities allow this specificity either. But thanks for the feedback, and keep the ideas coming! ;-)

  5. Neil Burgin says:

    Sorry for “yelling” (the all caps emphasis that comes across as angry). I saw you saying that they should make a better hair system before making this movie when they had already done so, and I posted a bit hastily and perhaps rather rudely. Normally if I do this I edit the post before anyone sees but this comment system doesn’t have an edit button. Sorry for any offense.

  6. JG Loquet says:

    Looking good ! Great progress. I’m in awe in front of the sense of self-sacrifice displayed by some animators, which seems to have no boundaries ;) going down on all 4 and all !

  7. BernAr says:

    Hi every one :-) You’re all doing a great job and I try to follow you live every Friday on the weekly when it’s possible for me. I wanted to point one thing that I realised last Friday watching the video of Franck’s hanging: it’s just a hint, I found it was not that clear for someone who doesn’t already know what’s going to happen ( after watching the different creative phases that Franck is hanging) and I would suggest the camera on the other side in relation to the rock, or Franck falling on the side of the camera if you prefer. I think the body and hair of Franck rocking before falling would be much easier to read… Bon c’est juste une remarque Mathieu, bon courage ça va être super :-)

  8. pawelstarz says:

    I have no idea why some people are so fascinated by Pixar’s movies. Do you really think that models or scenes they make are so awesome and much better than made by Gooseberry team? I think Pixar makes things just to sell them and most of that stuff is just boring, mostly it has no artistic spirit, because it’s made for ordinary people. Gooseberry team can achieve Pixar’s quality by deleting most of details, creating boring environment and expressionless characters.

  9. Kevin says:

    Many great things going on here. It seems improvements are being made in areas that really need them. I’m very excited about the new depth of field algorithm.

    As far as the film, I’m impressed with many things I see in the preliminary shots. However, what is impressing me the most, over the other open projects done in Blender, is the lighting/shading and surfacing/texturing. Very nice work. The “semi-fully” rendered images are very pleasing to look at and study. Franck’s Big Drag is great.

    I do notice that the mood in the video above is different than the still at the bottom of the post. Curious? Without knowing the story I’m not sure if the lighting is changing and this is just further along, is it just different or is it purposeful? The video has lighting that is darker and foreboding, Franck did just try to hang himself, that seems to fit the “feeling”. However, the still is brighter and seems less foreboding, in fact it feels promising. Is the lighting changing because he is about to overcome his feeling of doom or be shown a more positive prospect? Or did it change just to make the shot “prettier”? It is prettier, but from the little I see of the story, the darker seems more appropriate to the mood. Anyway, that’s my two cents. The work is beautiful either way. Keep it up Manu and Pablo. Everyone is doing great work. This is shaping up to be the best open movie project yet.

    Oh and kudos on this website too. This is the most professional layout and best looking website of any of the other projects. That goes for the content too – it looks better and is just more robust.

    • Elysia Brenner says:

      Thanks, Kevin! The shading and lighting decisions still aren’t final in these shots here, but you can bet Mathieu will make sure the right mood is nailed. ;-) (And, yes, things definitely get more hopeful for Franck!)

Leave a Reply

Your email address will not be published. Required fields are marked *