Private video :-) It must be REALLY good!
On Tue, Aug 6, 2013 at 11:24 AM, Paul Doyle <[email protected]> wrote: > Some comments on Fabric that we recorded at Siggraph from a few familiar > VFX faces: https://vimeo.com/71818285 > > Eric is in this video, but don't hold that against us ;) > > > > On 6 August 2013 09:06, Eric Thivierge <[email protected]> wrote: > >> Raf basically said what I was trying to say in a shooting from the hip >> quick response that probably failed miserably. Either way I'm in agreement >> with Raf (look at that, you agree with an American, Raf!). The future >> seems to be platforms and frameworks to build out your pipeline tools as >> needed and in the way you want. You also have flexibility to change a deep >> level of structures your apps are running on. >> >> As Raf pointed out it's clear to me especially after Siggraph that there >> is not going to be a one app to rule them all. It's going to be a mixed bag >> of apps with standard formats supported across them to pass the data back >> and forth and use the app best for particular parts of the pipeline. >> >> So many companies branching out and incorporating open source solutions >> (again as Raf mentioned) and not being shy about it either. So many >> Siggraph talks this year talking about how they implemented an open source >> format and used it in their projects. >> >> Do I want an Uber Autodesk app? No. I've lost trust at this point in AD >> and it doesn't make sense. >> >> Eric Thivierge >> =============== >> Character TD / RnD >> Hybride Technologies >> >> >> >> On August-06-13 1:10:47 AM, Matt Lind wrote: >> >>> I think the ‘age of the platform’ assessment will be restricted to >>> film/video as I see a fork in the road developing between games and >>> film/video pipelines. Actually, it’s already been happening for many >>> years. >>> >>> Traditionally games have borrowed film/video tools for 3D work because >>> needs were simple and the film/video tools could be bent to service. >>> But now as graphics hardware improves, games requirements are much >>> more demanding and divergent from that which film/video caters. >>> >>> Film/video has always moved towards larger and larger datasets >>> requiring subdivision of labor to the N’th degree. Quality was the >>> overriding factor. If it takes N hours to render that one awe >>> inspiring frame, you do it. That growth requires asset management to >>> manage all the facilities and assets. The assets last only as long as >>> the production, unless there is a sequel. Each production typically >>> involves reworking and re-inventing the wheel unless you work at one >>> of the older mainstays that have significant R+D investment into their >>> pipelines. Basically assets are generated, a picture is taken of >>> them, then they are dumped into a box where they sit on a virtual >>> shelf until needed again. Kind of like the old gag on Popeye cartoons >>> where they chop down the redwoods, send them to the saw mill, then >>> whittle it down to a single toothpick where it’s shipped off in a box. >>> >>> In games, it’s a bit different. In the case of the MMO I’m working on >>> the assets must have a very long shelf life – measured in decades. >>> The assets contribute to live software environments, must be very >>> optimal, and are under constant iteration. While growth is also >>> occurring in the games pipeline, it’s moving in a different direction >>> than film/video. Games is moving fast towards ‘in context’ editing of >>> assets, as in, creating/editing the assets in the live game >>> environment. To accomplish the feat requires being very tightly bound >>> to the runtime environment of the game engine. Therefore a DCC >>> application which serves as a ‘platform’ will not serve any role where >>> the work is done in the game environment. I would venture to say that >>> many games developers are actively pursuing the route of removing DCC >>> applications from their pipelines completely. It will be many years >>> before it is actually accomplished, however. >>> >>> I remember a discussion with former Softimage PM Gareth Morgan back in >>> the late 1990s where he said they were actively working to make >>> ‘sumatra’ a game engine with DCC tools. That vision is not far off >>> from reality. The only part he got wrong is the DCC application isn’t >>> the host, it’s the guest. >>> >>> What you’ll see emerge in the games development arena for content >>> creation are application(s) which can attach live agents to the >>> content being created so it can be merged into the game environment. >>> In other words, something a game engine can host. The difficulty >>> comes in the area of viewing the work. Something like Fabric Engine >>> has its own language for compiling and preparing the assets for >>> display. This is the exact same responsibility of the game engine. >>> While the DCC application clearly isn’t a solution here, the Fabric >>> Engine model isn’t a hands-down winner either (but much closer to the >>> correct solution). It’ll be interesting to see how that problem is >>> addressed. >>> >>> Matt >>> >>> *From:*softimage-bounces@**listproc.autodesk.com<[email protected]> >>> [mailto:softimage-bounces@**listproc.autodesk.com<[email protected]>] >>> *On Behalf Of >>> *Raffaele Fragapane >>> *Sent:* Monday, August 05, 2013 9:23 PM >>> *To:* [email protected].**com<[email protected]> >>> *Subject:* Re: OT: Yost Group - related to the Naiad/SIGGRAPH discussion >>> >>> >>> Why Fanboi, and why conspiracy? >>> >>> I consider Paul and Co. to be smart enough to know that that is >>> EXACTLY what they should be shooting for. >>> >>> AD knows it themselves IMO, as does SideFX, and the Foundry, and many >>> others. >>> >>> The writing couldn't be plainer on all walls that the industry is >>> shifting again. >>> >>> >From blackboxed, fragmented specialistic apps in the end80s to mid >>> nineties, to the rise of the artist friendly monolith in the end 90s, >>> to the monolithic but moderately open app from end-90s until now, >>> we're now moving fast towards a common stream of OSS standards which >>> will be injected into by various small footprint, very specialized and >>> tailored apps (ZB, Mari, Katana etc.), and have a layer floating on >>> top to interface pipe and content/operation management on top of that >>> will be platform centric. >>> >>> You have pointed out bits of that youreself. >>> >>> Maya and Soft are more and more used as mere scene assembly and >>> animation platforms. That type of approach is becoming more widely >>> available by the minute to smaller and smaller entities, even to >>> individuals. It's only the middle end caught into hard software locks >>> at this point. >>> >>> The age of the platform is coming. >>> Everybody already manages shots with shotgun, assets with tank (or >>> perforce, or propietary, or what else you have it), models with ZB, >>> retopos with 3DC or Topogun, textures with mudbox or mari, does >>> effects in Houdini, or Realflow, hair is left to plugins (shave, >>> yeti), lights with katana, renders with PRMan, composites with Nuke, >>> finals with DaVinci... >>> >>> Who caches with something other Alembic (or propietary formats) or >>> writes images other than EXR? >>> >>> All UIs are Qt, threading is beind coalesced in fewer solutions by the >>> day, libraries emerge to abstract and generalise many things (OCL, >>> Thrust etc.). >>> >>> What little is left out has initiatives that might be caught up on >>> (OSL, partIO, openVDB), or will one day see an alternative that will >>> become the standard. >>> >>> What's left for Maya or Soft to do but assemblying assets and >>> rig/animation? Which are ultimately just scene Management tasks, a >>> specialized type of graph which, of the lot, is the most backwards and >>> dated of all sections of the pipe. >>> >>> There will be churn, as always for a few years one sub-field using CGI >>> is left better or worse serviced than others, one size more or less >>> competitive, but I don't think there will be a next-gen big app, not >>> one as big (proportionally) as Soft was, or Maya is. >>> >>> Fabric did the right thing, all they have to do is garner the >>> attention and sustenance to punch through the industry catching up to >>> the obvious through lean years. >>> >>> On Tue, Aug 6, 2013 at 12:57 PM, Matt Lind <[email protected] >>> <mailto:mlind@carbinestudios.**com <[email protected]>>> wrote: >>> >>> And to throw some fanboi conspiracy theory gas into the flames: >>> >>> If you integrate with all the DCC apps, you’ve essentially built up >>> the trust with all the user bases and have the ability to suck them >>> into your DCC of the future to reduce any and all risk of switching a >>> production pipeline to another base application. >>> >>> At least give us a ray of hope, Paul. ;-) >>> >>> Matt >>> >>> >> >
