Andr
I was intrigued as well but didn't really get what he meant and the implications
Could anybody explain the final words of the presentation?

The implications are a GPU based version of Mantra comparable to what Octane and Redshift can currently do, along with a more visual and intuitive implementation of the MAT context. There was also an implication of the entire Houdini architecture leveraging GPU computing to speed simulations and other workflow elements even further.

BabaJ

So now that this is implemented even i could develop certain interactive tools.

Since I have very limited Python knowledge, I for one would watch one of your tuts on this topic - even if it's a paid one.

We intend to start building more of the viewport tools with python rather than c++, and the source will be available for them all in the Houdini install. That'll in turn push more HOM improvements. You'll be able to modify our tools, use them as a starting point for your own, and use others' tools as well. It's already looking to be a lot more powerful than some of the existing hooks, and much easier ro use

McNistor
So you didn't actually order it, you just happen to like it
DOn't get me wrong, I do not begrudge you or anyone that got their wishes fulfilled to one degree or another in this release, I am among them after all, but I am a bit disappointed in the apparent fact that money are not put where the mouth is when it comes to direct modeling.
It just might be a colossal misread of SESI's intentions starting with H14 or a failure of expectation management from my part regarding the time-frame.

Regarding the align feature itself, if it is as I understood it to be from the video presentation, seems like a more viewport interactive way of “transformation matching” which is currently achievable by clicking on the small buttons at each transform channel's side. The H17 is probably better, I don't deny that, it's just that there are more pressing features that need improvements or downright re/writing from ground up. And this alignment feature, it's very useful when you need it, but you don't need it that often compared to other incomparably more repetitive tools and features. You don't get a lot of mileage from improvements of not that often used tools. So why prioritize it? SESI not knowing what traditional modelers actually use most often could be an explanation.

Anyway, these threads with people going back'n'forth about what features should SESI implement next or whether the new ones have been the best choices, are a glaring consequence of a not so pretty side of our nature and I for one, will take a break from them. I'm limiting myself to filing RFEs, if that.

I says a lot that this one little tool seems to get more attention right now from the community than the whole project vellum.
( in truth I think the people excited about the vfx stuff just don't post it here )

OneBigTree
( in truth I think the people excited about the vfx stuff just don't post it here )

Well, I was in the middle of developing a glass breaking pattern HDA, can bin that one now.

Vellum is looking just amazing. And i am realy looking forward to easier setups of the constrains. One solver for cloth grains and wire is great to say the least. I just hope my GPU (1070) is good enough.

The new terrain tools look great. I just hope the errosion will be fast and nice. The other additions are also very welcome, masking for scattering was possible before but now its more streamlined. So setting up more complex biome rules should be much easier.

UVs look much better now, i just hope it can be done procedurally nicely and not only manually.

One of the things not mentioned (for good reasons) is the new renderman integration and i am looking forward to try this now. I had a quick look at renderman a year ago and the lacking interactive rendering just put me of. Now that should be working. (BTW for learning purposes renderman is free without watermark and other restrictions!)

Now i just need more TIME to learn all that great new stuff. Too bad H17 does not come with that.

Houdini is a monster in VFX industry.
The new algorithm Vellum is wonderful but I think for that reason a lot of user(me include) try to see new areas like modeling or render or animation.
No one have doubt about procedural tool that exist very better in Houdini but new user that came to Houdini is missing some improvements in our software thas is more clear in others aplication.
But SideFX is not a giant company like Autodesk or Adobe if we see Houdini 17 for this point of view Houdini is growing very well.
OK I will wait for use Houdini 17 and I will have a real opinion about it but I believe in SideFX and I respect all development in H-17.
Thank you SideFX for the new release.
Let's go procedural.

harryabreu
But SideFX is not a giant company like Autodesk or Adobe if we see Houdini 17 for this point of view Houdini is growing very well.

Honestly it seems to me like they're moving and developing a lot more quickly than someone like Adobe. I see an app like AfterEffects which seems to have only gained minor tweaks in the past 10 years vs. Houdini which seems to have come leaps and bounds forward. The only other people who seem to move as aggressively are the Blender folks.

Midphase

Andr
I was intrigued as well but didn't really get what he meant and the implications
Could anybody explain the final words of the presentation?

The implications are a GPU based version of Mantra comparable to what Octane and Redshift can currently do, along with a more visual and intuitive implementation of the MAT context. There was also an implication of the entire Houdini architecture leveraging GPU computing to speed simulations and other workflow elements even further.

If you're referring to LOPs mentioned at the end of the video, that's not what it is.

Enlighten me then.

hopefully all this vfx development hasn't taken away from the upcoming EOPs roadmap.. the blockchain FPGA neural-net powered Emoji context.

Midphase
Enlighten me then.

LOPs sounds more like Look Development Context.

I hope LOPs will be a new context that unifies COP and MAT Context.

Olaf Finkbeiner
Vellum is looking just amazing. And i am realy looking forward to easier setups of the constrains. One solver for cloth grains and wire is great to say the least. I just hope my GPU (1070) is good enough.

For Vellum your GPU probably won't be, Nvidia GPU's are notoriously bad at OpenCL. You will however be fine denoising, since that uses proprietary Nvidia tech.

I am more than a little perplexed why SideFX chose to suddenly go down the CUDA route for denoise, especially given how much OpenCL is already a part of Houdini. I always had the impression that SideFX preferred open standards to proprietary technology.

Lyr
I am more than a little perplexed why SideFX chose to suddenly go down the CUDA route for denoise, especially given how much OpenCL is already a part of Houdini. I always had the impression that SideFX preferred open standards to proprietary technology.

There is no prodution ready Ai denoiser made with opencl. So it makes sense to implement this one. To develop you own takes a lpt of time or when you want to use tensorflow you will end up the same it has only cuda support.

Doudini
LOPs sounds more like Look Development Context. I hope LOPs will be a new context that unifies COP and MAT Context.

That's the way I understand it too, hence my comment about a more visual and intuitive implementation of the MAT context. I've seen some screenshots of this new proposed LOP context which seem to confirm that.

Having said that, there was undoubtedly a heavy implication at the end of the SESI presentation that Mantra will be indeed moving toward a GPU model.

I do agree and hope that the implementation will not be necessarily tied to proprietary tech such as CUDA. I hope that SESI can figure out a way to implement GPU rendering in such a way that doesn't force the end user to only use Nvidia or AMD GPU's, but rather allows us to choose which GPU we want to use.

mandrake0
There is no prodution ready Ai denoiser made with opencl. So it makes sense to implement this one. To develop you own takes a lpt of time or when you want to use tensorflow you will end up the same it has only cuda support.

You are correct that there is no OpenCL denoiser for production yet, however Tensorflow is now available free of Cuda https://rocm.github.io/index.html [rocm.github.io]

Given the current state of things I guess it's better to have a Cuda denoiser than no denoiser.

Olaf Finkbeiner
One of the things not mentioned (for good reasons) is the new renderman integration and i am looking forward to try this now. I had a quick look at renderman a year ago and the lacking interactive rendering just put me of. Now that should be working. (BTW for learning purposes renderman is free without watermark and other restrictions!)

Same here. Renderman 22 looks great. Anyone know why the release of its Houdini bridge tool lags behind the Maya and Katana bridge tools? Given Houdini's implemenation of Disney's Principled Shader and close relationship to Pixar, I would have expected it to be as quick to develop.

For H17.0, the RenderMan bridge product needed a larger rewrite to support live rendering. It was demoed at SIGGRAPH. There is more info on the RenderMan forum.

https://renderman.pixar.com/answers/questions/90/is-houdini-supported-for-renderman-22.html [renderman.pixar.com]

Thanks, I didn't know they had previewed it at Siggraph this year. That's cool, althought the September RM forum post says “in the coming months…” which is a bummer.

why is it called Banshee?

bobc4d
why is it called Banshee?

No idea, but if sidefx sticks to these codenames, I wonder what the Dragon version will be like.

After seeing the promo for Houdini 3D 17, I'm now convinced to move over to it from Lightwave 3D. I really like what I see and I like seeing a company that is behind and supporting its software. So yes, I'll be a completely beginner in Houdini, but I'm going to start learning it. Looking forward to it!!