Clarisse Overview

Non-linear-imaging, a new paradigm introduced by Clarisse

Clarisse's revolutionary new approach to 2D and 3D animation allows you to continually see the effect of your work on the final image. You can now work in context and, for the first time, finally have the big picture within arm's reach. You can easily make modifications, in any order, to anything in your project: from underlying geometry, to character animation, to lighting, to texturing, to image compositing... while seeing the result of your work on the final image!

With non-linear-imaging, you are free to work on your images without imposing any order of modification while working in context.

Cutting edge technologies supporting Clarisse

By making intensive use of the latest technologies in parallel programming, Clarisse delivers a fully multi-threaded evaluation pipeline from image filters, through geometry deformers to rendering.

It's just as simple as that: the more cores you have on your workstation, the faster Clarisse gets.

Loading speed in Clarisse is another major issue addressed by Isotropix developers. Unlike any other animation package, Clarisse's unique architecture provides on-demand data loading. Data such as images or geometries are dynamically loaded during evaluation.

As a result, typical production scenes that took several minutes to be loaded in other packages are now loaded in seconds. No more need for you to wait several minutes to modify a parameter in your project, such as antialiasing.

Running out of memory is a common problem for large film and game assets, and thus forces users to optimize manually their scenes. Most software will simply crash when this happens. However, Clarisse has been designed with a memory management system that, firstly, smartly detects redundant data and, then, automatically drops scene complexity when resources are low.

When Clarisse detects the same data has been loaded or evaluated twice, it automatically shares it between dependent objects. Not only this system saves a lot of memory but also, a lot of processing power as identical evaluations are also shared!

A new approach to image rendering

Until today, even when evaluation was computed in parallel, users had to wait for its completion or abort it manually before continuing their work. In other words, when dealing with final renders, applications are locked, even if a user wants to modify something that isn't related to the render.

This severe limitation led to the flourishing of progressive and interactive CPU or GPU based raytracers meant only to bypass this issue by displaying fast previews of user modifications on a 3D scene. Then again, users still have to make a final render, when pleased with the preview.

Clarisse redefines and simplifies the way rendering is done. Indeed, with Clarisse, you can do a lot more than working on 3D scenes, you can literally work on final images that potentially include a full compositing tree.

Thanks to Clarisse's unique evaluation system, you always see an automatic preview of your full work being gracefully refined into the final render. It's just as simple as that: you always work in context and have the ability to make any change at anytime, while seeing a result of your work progressively refined into the final image.

In practical terms, let's see what happens when you are tweaking a light, or even painting a texture applied to an object during rendering:

  • If your modification was relevant to the rendering image, Clarisse will transparently relaunch the render in the background and provide you the most faithful feedback.
  • On the other hand, if your modification wasn't relevant to the image, Clarisse will just continue it's rendering as if nothing happened.

This behavior is only possible because Clarisse beautifully keeps track of all information in parallel that are directly relevant to an object and thus, in this example, the rendering image.

A new workflow putting images at its heart

Traditional 3D animation software considers render images to be the result of a 3D scene, meant to be displayed in a window and saved to disk. These images have no real existence within the project and are merely the result of specific rendering settings applied to a 3D scene. On the other hand, traditional 2D software considers the image to be a central and reusable element resulting from an arranged network of nodes or a group of layers.

Clarisse is the world's first 3D software that puts images at the heart of its revolutionary workflow. Images become active elements of the project and can be manipulated and used as input.

In the same way 2D layers are applied to images inside 2D software, within Clarisse, 3D scenes simply become 3D layers applied to an image. Following this simple idea, if you modify one or more 3D layers or even 2D layers, an automatic update notification is sent to the image which if displayed renders only the needed layers to get fully updated.

Consequently, Clarisse's projects become much more than common scenes. They are a repository of shared objects (cameras, lights, geometries, layers, images...) plugged into each others which can eventually lead to images. Within a project, you can setup a single image, multiple shots or even multiple sequences : you are literally free to organize your work as you like to share objects as you want.

Another remarkable benefit that considerably improves the general workflow is image picking. With Clarisse, not only you can pick the color from an image, but you can also pick objects (layers, geometries, lights, materials etc.) directly in image space.

Finding and tweaking active elements that participate to an image has never been so easy: they are literally a mouse click away !