CGTalk > Main Forums > General Discussion
To minimize the ads you see on this page create a CGTalk account and log in HERE
Thread Closed share thread « Previous Thread | Next Thread »
 
Thread Tools Search this Thread Display Modes
Old 06-04-2013, 05:16 AM   #16
earlyworm
car unenthusiast
 
earlyworm's Avatar
Will Earl
craftsperson
Grizzly Country, Canada
 
Join Date: Mar 2005
Posts: 1,679
I haven't used Houdini for lighting so I can't really compare it to Katana in that regard. I do imagine you could create a similar workflow to Katana with Houdini.

Katana is fairly limited and highly focused in it's functionality - it does lighting, lookdev and a little bit of compositing - that's it. It's very lacking compared to Maya, Houdini and whatever other all-purpose 3d application your using.

I honestly don't know anything about what Katana costs. But do consider that studios switching to it are most likely looking at it for the following reasons...

* So they can reduce or drop development of their in-house render translators, the idea being it's potentially cheaper and easier to let a third party handle the development and quality assurance.
* Better and more up-to-date than their current toolset. Most in-house render translators I've used are generally a lot more powerful and flexible than the off-the-shelf software available, especially when it comes to multi-pass workflows - but they're starting to show their age and Katana does have some useful features which are fairly unique to it. I think it has a great framework which has a better future than the in-house renderers I've used in the past (even if it's not quite as robust at the moment).
* Render agnostic. I think we're starting to see a switch to studios adopting multiple renderers - choosing the best tool for the job. Previously that was hard to do, each renderer having it's own particular workflow or doing something unique that made it hard to integrate with other elements. Katana comes included with PRMan and Arnold integration.
* Lighters generally work in Maya, make of this what you will, but I'm generally of the opinion that Katana has got to be competitively priced against a stack of Maya licenses plus current in-house development costs.
* There is also the potential for it to play nicely with Nuke and Mari.
 
Old 06-04-2013, 06:23 AM   #17
mr Bob
CG Sup
Mars
USA
 
Join Date: Dec 2003
Posts: 1,322
Quote:
Originally Posted by SebKaine
Hi guys !

I am desesperatly trying to find the Price of Katana , but can't find any info about this , how much does it cost ?

I would also like to know if some people have use it in production ? and what they think about it ?
Basically it can :
- read huge amount of assets
- setup light and shaders for those assets
- organise assets / lights / shader / pass the way you want
- export this to your render engine

But what does it have that houdini doesn't have ?
- you also have a tree to manage huge amount of assets
- you can manage light / shaders / assets in a similar way
- plus you have mantra and infinite render token for free ?

So what does Katana offer that houdini don't , and why big guys like SPI,ILM,D2 or MPC use it instead of Houdini ?

Many thanks for your lights guys !

Cheers

E


What are you talking about Katana has nothing to do with Houdini and just so you know Katana is not used over Houdini at Sony. MPC dont use Houdini and never have. Katana is not perfect and can be very slow and clunky.
 
Old 06-04-2013, 06:41 AM   #18
SebKaine
Expert
 
SebKaine's Avatar
portfolio
Emmanuel Mouillet
Fx Artist
Freelance
Paris, FR
 
Join Date: Dec 2005
Posts: 981
Quote:
Originally Posted by earlyworm
I haven't used Houdini for lighting so I can't really compare it to Katana in that regard. I do imagine you could create a similar workflow to Katana with Houdini.

Katana is fairly limited and highly focused in it's functionality - it does lighting, lookdev and a little bit of compositing - that's it. It's very lacking compared to Maya, Houdini and whatever other all-purpose 3d application your using.

I honestly don't know anything about what Katana costs. But do consider that studios switching to it are most likely looking at it for the following reasons...

* So they can reduce or drop development of their in-house render translators, the idea being it's potentially cheaper and easier to let a third party handle the development and quality assurance.
* Better and more up-to-date than their current toolset. Most in-house render translators I've used are generally a lot more powerful and flexible than the off-the-shelf software available, especially when it comes to multi-pass workflows - but they're starting to show their age and Katana does have some useful features which are fairly unique to it. I think it has a great framework which has a better future than the in-house renderers I've used in the past (even if it's not quite as robust at the moment).
* Render agnostic. I think we're starting to see a switch to studios adopting multiple renderers - choosing the best tool for the job. Previously that was hard to do, each renderer having it's own particular workflow or doing something unique that made it hard to integrate with other elements. Katana comes included with PRMan and Arnold integration.
* Lighters generally work in Maya, make of this what you will, but I'm generally of the opinion that Katana has got to be competitively priced against a stack of Maya licenses plus current in-house development costs.
* There is also the potential for it to play nicely with Nuke and Mari.


Well Will thanks for the very precise answer ! i think you 've highligted the main point, it's expensive to maintain a translator to an engine and now that raytracer are back in the field the ability to mix Prman / Arnold or a in House engine in the same UI is pretty poweful. Basically it's a toolbox where most of the job is done to communicate with a render engine , but you have all you need to customise it the way you want , on top of that you can have external R&D and support which reduce cost greatly ! From what you said it looks to be really for really big structure and not mr Anybody ...
__________________
emfx.fr

Last edited by SebKaine : 06-04-2013 at 06:47 AM.
 
Old 06-04-2013, 06:46 AM   #19
DagMX
Frequenter
Dhruv Aditya Govil
New Delhi, India
 
Join Date: Jun 2006
Posts: 266
Quote:
Originally Posted by SebKaine
Hi guys !

I am desesperatly trying to find the Price of Katana , but can't find any info about this , how much does it cost ?

I would also like to know if some people have use it in production ? and what they think about it ?
Basically it can :
- read huge amount of assets
- setup light and shaders for those assets
- organise assets / lights / shader / pass the way you want
- export this to your render engine

But what does it have that houdini doesn't have ?
- you also have a tree to manage huge amount of assets
- you can manage light / shaders / assets in a similar way
- plus you have mantra and infinite render token for free ?

So what does Katana offer that houdini don't , and why big guys like SPI,ILM,D2 or MPC use it instead of Houdini ?

Many thanks for your lights guys !

Cheers

E



Like it's been established, Katana's pricing is bespoke, but I'll try answering the other questions.

katana is really good for a certain set of tasks. These generally include lighting, lookdev and a small amount of comp. Granted, all of these can be done in other programs, but Katana's strength is handling large amounts of assets in a efficient and intuitive manner.

It's not a direct competitor to Houdini. Houdini is a general purpose application whereas Katana is a very focused one.

The advantages to Katana are:
*Very quick and intuitive for lighting and lookdev tasks. The UI and toolset have been streamlined to make this extremely quick.

*It's completely designed to fit and be molded into any sane pipeline. Houdini is great, but because of it's nature, you need quite a bit more RnD work to get it to be as pipeline streamlined as Katana

*Katana is render agnostic. Essentially, you can switch render engines on the fly, as long as your shaders can convert (OSL helps here) and your elements have similar analogies.
This means lighters don't need to know every render engine. That decision can be left to the show supervisors. The lighter can focus on the art of lighting


Now the reason Sony use Katana is because they made it and their pipeline is heavily based around it.
Other studios use it because it is a very good base that covers most things that studios used to build custom tools for.
Also as market adoption of Katana goes up, it means you need less training for new hires to jump right in.



And, having to have learned Katana at work, i found it very easy to get to grips with. I think it's more intuitive if you're coming from Nuke or a node based Houdini (rather than a pure UI based houdini).
 
Old 06-04-2013, 06:46 AM   #20
SebKaine
Expert
 
SebKaine's Avatar
portfolio
Emmanuel Mouillet
Fx Artist
Freelance
Paris, FR
 
Join Date: Dec 2005
Posts: 981
Quote:
Originally Posted by mr Bob
What are you talking about Katana has nothing to do with Houdini.


I do NOT share your POV on this one

Quote:
Originally Posted by mr Bob
MPC dont use Houdini and never have.


I know and sorry if my post was unclear on this ,anyway i've heard that they test it in 2012, but those info can be wrong. but yes indeed they are still Maya / Flowline / R&D base ...

Quote:
Originally Posted by mr Bob
Katana is not perfect and can be very slow and clunky.


Thanks for the info !
__________________
emfx.fr

Last edited by SebKaine : 06-04-2013 at 07:07 AM.
 
Old 06-04-2013, 06:59 AM   #21
SebKaine
Expert
 
SebKaine's Avatar
portfolio
Emmanuel Mouillet
Fx Artist
Freelance
Paris, FR
 
Join Date: Dec 2005
Posts: 981
Quote:
Originally Posted by DagMX
Like it's been established, Katana's pricing is bespoke, but I'll try answering the other questions.

katana is really good for a certain set of tasks. These generally include lighting, lookdev and a small amount of comp. Granted, all of these can be done in other programs, but Katana's strength is handling large amounts of assets in a efficient and intuitive manner.

It's not a direct competitor to Houdini. Houdini is a general purpose application whereas Katana is a very focused one.

The advantages to Katana are:
*Very quick and intuitive for lighting and lookdev tasks. The UI and toolset have been streamlined to make this extremely quick.

*It's completely designed to fit and be molded into any sane pipeline. Houdini is great, but because of it's nature, you need quite a bit more RnD work to get it to be as pipeline streamlined as Katana

*Katana is render agnostic. Essentially, you can switch render engines on the fly, as long as your shaders can convert (OSL helps here) and your elements have similar analogies.
This means lighters don't need to know every render engine. That decision can be left to the show supervisors. The lighter can focus on the art of lighting


Now the reason Sony use Katana is because they made it and their pipeline is heavily based around it.
Other studios use it because it is a very good base that covers most things that studios used to build custom tools for.
Also as market adoption of Katana goes up, it means you need less training for new hires to jump right in.



And, having to have learned Katana at work, i found it very easy to get to grips with. I think it's more intuitive if you're coming from Nuke or a node based Houdini (rather than a pure UI based houdini).


Thanks for the very detail answer Dhruv ! The Katana profile is now pretty clear ...
- simple and optimise UI that allow to light in Arnold / Prman in the same way
- very strong assets managment capacity
- very easy to use and learn
- replace the cost of internal R&D and support
- ability to switch from an engine to an other on a shot or on a show ...
- strong toolbox to build your lighting / shading pipeline !
__________________
emfx.fr
 
Old 06-04-2013, 03:10 PM   #22
mr Bob
CG Sup
Mars
USA
 
Join Date: Dec 2003
Posts: 1,322
Quote:
Originally Posted by SebKaine
I do NOT share your POV on this one


I am interested then. Your an fx artist what is the connection ? . Are you considering buying Katana and using that instead of lighting your effects in Houdini ? . Your conclusions are very broad from the above post. Katana by no means is an application you can just open and hit render on your assets. It still requires some very good TD skills and ongoing support.

b

Last edited by mr Bob : 06-04-2013 at 03:18 PM.
 
Old 06-04-2013, 08:15 PM   #23
SebKaine
Expert
 
SebKaine's Avatar
portfolio
Emmanuel Mouillet
Fx Artist
Freelance
Paris, FR
 
Join Date: Dec 2005
Posts: 981
Quote:
Originally Posted by mr Bob
Are you considering buying Katana and using that instead of lighting your effects in Houdini ?

No

Quote:
Originally Posted by mr Bob
I am interested then. Your an fx artist what is the connection ?

I'm talking about a tool to manage the lighting workflow in a company.

Quote:
Originally Posted by mr Bob
Your conclusions are very broad from the above post. Katana by no means is an application you can just open and hit render on your assets. It still requires some very good TD skills and ongoing support.

This is enlightning ... Thank You
__________________
emfx.fr

Last edited by SebKaine : 06-04-2013 at 09:07 PM.
 
Old 06-05-2013, 12:41 AM   #24
rendermaniac
.
Simon Bunker
Shader Writer / TD
Australia
 
Join Date: Jan 2003
Posts: 1,040
Having used Houdini for lighting (but not Katana yet) I can definitely say Katana and Houdini work quite differently in several areas.

Katana is all about deferred loading. It only works with caches and only evaluates them at the last possible moment. This means massive scenes are easier to load / manipulate as they are a series of top level bounding boxes by default, but you can drill down to individual components if you need to. This does mean that you wouldn't want to do modelling or animation in Katana - because it is not designed to do that.

Katana only has one graph and each node can manipulate the whole scene. It is essentially as if each node is a take (render layer in maya terms). This means to remove an operation you just delete the node. This is very powerful as it gives you the ability to automatically build the camera / animation / surfacing etc and then layer lighting changes on top non destructively. I think this is the most powerful feature of Katana.

With Houdini you only have a single scene graph which you have to modify destructively before you export your render data. In Katana each render pass becomes a branch in your graph.

Supporting multiple renderers and linking nicely into asset management are also nice features to have. It is also designed to work nicely with interactive progressive rendering if the renderer supports it - it is nice not to have to write that yourself.

I'm not entirely sure how good it is for lookdev and if there are any tools to work together with Mari. That is my only concern. Building shader networks does not seem to be the easiest thing to do in Katana.

I'm not sure about pricing, but I would guess it's not cheap.

Simon
__________________
http://www.rendermania.com/
 
Old 06-05-2013, 06:42 AM   #25
SebKaine
Expert
 
SebKaine's Avatar
portfolio
Emmanuel Mouillet
Fx Artist
Freelance
Paris, FR
 
Join Date: Dec 2005
Posts: 981
Quote:
Originally Posted by rendermaniac
Having used Houdini for lighting (but not Katana yet) I can definitely say Katana and Houdini work quite differently in several areas.

Katana is all about deferred loading. It only works with caches and only evaluates them at the last possible moment. This means massive scenes are easier to load / manipulate as they are a series of top level bounding boxes by default, but you can drill down to individual components if you need to. This does mean that you wouldn't want to do modelling or animation in Katana - because it is not designed to do that.

Katana only has one graph and each node can manipulate the whole scene. It is essentially as if each node is a take (render layer in maya terms). This means to remove an operation you just delete the node. This is very powerful as it gives you the ability to automatically build the camera / animation / surfacing etc and then layer lighting changes on top non destructively. I think this is the most powerful feature of Katana.

With Houdini you only have a single scene graph which you have to modify destructively before you export your render data. In Katana each render pass becomes a branch in your graph.

Supporting multiple renderers and linking nicely into asset management are also nice features to have. It is also designed to work nicely with interactive progressive rendering if the renderer supports it - it is nice not to have to write that yourself.

I'm not entirely sure how good it is for lookdev and if there are any tools to work together with Mari. That is my only concern. Building shader networks does not seem to be the easiest thing to do in Katana.

I'm not sure about pricing, but I would guess it's not cheap.

Simon


Many Thanks for your very detail and informative answer Simon ! From what you said it looks that there is definitly a reason why big boys use it ...

The ability to keep non destructively in your node tree all the lighting retakes in a readable way , with the possibility to go from version 28 to version 248 ... looks really gorgeous !

It also looks to really be a movie tool essentially where you will have the ability to load in a clever way 1.000.000.000.000 Polys and makes your 348 retakes before final ...

I would be curious to know if you have any idea of a tool or tricks to achieve excatly this non destructive lighting stuff in Houdini or Maya without heavy R&D ?

Thanks again for your light !

Cheers

E
__________________
emfx.fr
 
Old 06-05-2013, 08:23 AM   #26
Narann
Rendering, Pipeline, etc.
 
Narann's Avatar
portfolio
Dorian FEVRIER
Lighting/Pipeline TD
Atelier Animation
Montreal, Canada
 
Join Date: Mar 2005
Posts: 413
Quote:
Originally Posted by SebKaine
I would be curious to know if you have any idea of a tool or tricks to achieve excatly this non destructive lighting stuff in Houdini or Maya without heavy R&D ?

I will just answer for the "non destructive lighting" so don't compare my solution to "Katana".

If with "without heavy R&D" you mean: "possibly with good TDs" so my answer is: "Recursive text files".

Each text file is composed of simple commands to do something (assign shader, change position, attribute, etc...) and each file have a kind of "include" command to input another text file.

And you have something like this:

movie_cam.text
sequence_cam.text
shot_cam.text

movie_light_set.text
sequence_light_set.text
shot_light_set.text

set.text
seq_set_dress.text
shot_set_dress.text

You could have something like this:

movie_cam.text ->Include in-> sequence_cam.text ->Include in-> shot_cam.text

movie_light_set.text ->Include in-> sequence_light_set.text ->Include in-> shot_light_set.text

set.text ->Include in-> seq_set_dress.text ->Include in-> shot_set_dress.text

And finaly: shot_merge.text

shot_set_dress.text ->Include in-> shot_merge.text
shot_light_set.text ->Include in-> shot_merge.text
shot_cam.text ->Include in-> shot_merge.text

And you have a python file opening shot_merge.text and recompose you scene in rib/ass/vrscene/mi

This can be adapt but the idea is here. It's a kind of "overrider" system.

Each text file get a file content, tweak it and so forth.

PS: More more:

shot_merge.text ->Include in-> shot_layer_toto.text

I don't think you need R&D to do this but good TDs and some time.
__________________
My blog (fr)
 
Old 06-05-2013, 04:43 PM   #27
mr Bob
CG Sup
Mars
USA
 
Join Date: Dec 2003
Posts: 1,322
Narann : - Interesting approach I think you could execute all that through the mantra pre render script tab .

The closest thing you can use in Houdini is takes and ontop of that you can wedge parameters. No R&D needed . You just require a large renderfarm and masses of disk space

b
 
Old 06-05-2013, 05:16 PM   #28
Narann
Rendering, Pipeline, etc.
 
Narann's Avatar
portfolio
Dorian FEVRIER
Lighting/Pipeline TD
Atelier Animation
Montreal, Canada
 
Join Date: Mar 2005
Posts: 413
Quote:
Originally Posted by mr Bob
Narann : - Interesting approach I think you could execute all that through the mantra pre render script tab .

For example.

In Vray you can execute python script directly in the standalone. All of this is a question of taste. I personally like the scene been reconstructed from scratch each time.

Quote:
Originally Posted by mr Bob
The closest thing you can use in Houdini is takes and ontop of that you can wedge parameters. No R&D needed . You just require a large renderfarm and masses of disk space

I didn't understand this. :(
__________________
My blog (fr)
 
Old 06-05-2013, 07:19 PM   #29
SYmek
PRO
 
Join Date: Mar 2004
Posts: 82
When talking about Katana and Houdini, keep in mind Katana was designed in a studio which uses Houdini extensively from version 1. For good and bad, that is, it clearly takes all good things from procedural, nodal, asset based workflows used in Houdini for 18 years, and extends it apparently a lot with very clever solutions. Also afaik Katana is really good deal for Maya based studios, because it solves so many Maya's shortcomings when it comes to rendering.

Houdini wise..., Katana seems to have less to offer (not saying nothing though). The thing is people usually don't realize how good Houdini is in lighting and rendering, although there are some major problems in that department too, unsolved mostly due to SESI focus on effects.

- fully deferred load pipeline is possible and natural way of working (with 12.5 alembic/vdb-auto-delayed-load-primitive even more).
- non destructive overrides with Takes, and objects variances.
- render agnostic with fully python scriptable output backend.
- fully featured shader authoring toolset and compositing context.
- hierarchical overrides* of render properties (most render settings could be per object/primitive assigned properties) (using camera to assign shader or opposite, objects with assigned custom shutter speed, or primitives with assigned different render qualities depending on camera distance etc)
- hierarchical and conditional render outputs (build networks of rule based render drivers).
- groups, bundles, hdas and procedural/script based scene creation - the most flexible thing out there. You can build scenes procedurally from saved assets with minimal-to-none TD effort. Making web page sending selected assets to render farm is a kid challenge (even more considering http:// is a proper path in Houdini)
- fully open deep raster system and built-in image processing*.
- flexible geometry container (note that "alembic hype" is completely artificial thing from Houdini perspective, as it allows to cache, exchange, save/load/delay-load geometry on any stage of the pipeline using single file format capable of storing anything from particles, volumes, nurbs, photons to animation caches, and it does it efficiently with minimal RAM footprint for years now. It wasn't an option for everyone, and surly Alembic has solved many of real life problems, but two studios working in Houdini on Antarctica could not even find a reason to develop Alembic.)

* - largely busted due to various reasons :(

Having task oriented and bullet proofed lighting program might be a real deal for studios like Sony/ILM. Houdini has many problems with final touching of many of its solutions (like takes, bundles, render properties hierarchy), specially that it's developed with very biased perspective. It's also a giant hard to customize fora big studio's scale needs, but it is also very efficient solution for small to semi size shops, simple because by nature its nodes dressed with properties freely added with GUI by anyone, anytime, are more capable when it comes to massive scenes assembling. Which doesn't mean you can't push this approach further, what Sony obviously did.
 
Old 06-06-2013, 05:00 AM   #30
earlyworm
car unenthusiast
 
earlyworm's Avatar
Will Earl
craftsperson
Grizzly Country, Canada
 
Join Date: Mar 2005
Posts: 1,679
Quote:
Originally Posted by SYmek
flexible geometry container (note that "alembic hype" is completely artificial thing from Houdini perspective, as it allows to cache, exchange, save/load/delay-load geometry on any stage of the pipeline using single file format capable of storing anything from particles, volumes, nurbs, photons to animation caches, and it does it efficiently with minimal RAM footprint for years now.


That's fine if all your working with is Houdini, but the thing is these days you've got to be able to play nice with other applications, departments and companies.

The thing (I'm not sure I'd call it hype) about Alembic is that it's an open format. Before that every large vfx studio around had (or still do have) their own unique file format for storing 3d data.
 
Thread Closed share thread


Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
CGSociety
Society of Digital Artists
www.cgsociety.org

Powered by vBulletin
Copyright 2000 - 2006,
Jelsoft Enterprises Ltd.
Minimize Ads
Forum Jump
Miscellaneous

All times are GMT. The time now is 06:28 PM.


Powered by vBulletin
Copyright ©2000 - 2016, Jelsoft Enterprises Ltd.