Category: production
Weta Digital – Manuka Raytracer and Gazebo GPU renderers – pipeline
https://jo.dreggn.org/home/2018_manuka.pdf
http://www.fxguide.com/featured/manuka-weta-digitals-new-renderer/
The Manuka rendering architecture has been designed in the spirit of the classic reyes rendering architecture. In its core, reyes is based on stochastic rasterisation of micropolygons, facilitating depth of field, motion blur, high geometric complexity,and programmable shading.
This is commonly achieved with Monte Carlo path tracing, using a paradigm often called shade-on-hit, in which the renderer alternates tracing rays with running shaders on the various ray hits. The shaders take the role of generating the inputs of the local material structure which is then used bypath sampling logic to evaluate contributions and to inform what further rays to cast through the scene.
Over the years, however, the expectations have risen substantially when it comes to image quality. Computing pictures which are indistinguishable from real footage requires accurate simulation of light transport, which is most often performed using some variant of Monte Carlo path tracing. Unfortunately this paradigm requires random memory accesses to the whole scene and does not lend itself well to a rasterisation approach at all.
Manuka is both a uni-directional and bidirectional path tracer and encompasses multiple importance sampling (MIS). Interestingly, and importantly for production character skin work, it is the first major production renderer to incorporate spectral MIS in the form of a new ‘Hero Spectral Sampling’ technique, which was recently published at Eurographics Symposium on Rendering 2014.
Manuka propose a shade-before-hit paradigm in-stead and minimise I/O strain (and some memory costs) on the system, leveraging locality of reference by running pattern generation shaders before we execute light transport simulation by path sampling, “compressing” any bvh structure as needed, and as such also limiting duplication of source data.
The difference with reyes is that instead of baking colors into the geometry like in Reyes, manuka bakes surface closures. This means that light transport is still calculated with path tracing, but all texture lookups etc. are done up-front and baked into the geometry.
The main drawback with this method is that geometry has to be tessellated to its highest, stable topology before shading can be evaluated properly. As such, the high cost to first pixel. Even a basic 4 vertices square becomes a much more complex model with this approach.
Manuka use the RenderMan Shading Language (rsl) for programmable shading [Pixar Animation Studios 2015], but we do not invoke rsl shaders when intersecting a ray with a surface (often called shade-on-hit). Instead, we pre-tessellate and pre-shade all the input geometry in the front end of the renderer.
This way, we can efficiently order shading computations to sup-port near-optimal texture locality, vectorisation, and parallelism. This system avoids repeated evaluation of shaders at the same surface point, and presents a minimal amount of memory to be accessed during light transport time. An added benefit is that the acceleration structure for ray tracing (abounding volume hierarchy, bvh) is built once on the final tessellated geometry, which allows us to ray trace more efficiently than multi-level bvhs and avoids costly caching of on-demand tessellated micropolygons and the associated scheduling issues.
For the shading reasons above, in terms of AOVs, the studio approach is to succeed at combining complex shading with ray paths in the render rather than pass a multi-pass render to compositing.
For the Spectral Rendering component. The light transport stage is fully spectral, using a continuously sampled wavelength which is traced with each path and used to apply the spectral camera sensitivity of the sensor. This allows for faithfully support any degree of observer metamerism as the camera footage they are intended to match as well as complex materials which require wavelength dependent phenomena such as diffraction, dispersion, interference, iridescence, or chromatic extinction and Rayleigh scattering in participating media.
As opposed to the original reyes paper, we use bilinear interpolation of these bsdf inputs later when evaluating bsdfs per pathv ertex during light transport4. This improves temporal stability of geometry which moves very slowly with respect to the pixel raster
In terms of the pipeline, everything rendered at Weta was already completely interwoven with their deep data pipeline. Manuka very much was written with deep data in mind. Here, Manuka not so much extends the deep capabilities, rather it fully matches the already extremely complex and powerful setup Weta Digital already enjoy with RenderMan. For example, an ape in a scene can be selected, its ID is available and a NUKE artist can then paint in 3D say a hand and part of the way up the neutral posed ape.
We called our system Manuka, as a respectful nod to reyes: we had heard a story froma former ILM employee about how reyes got its name from how fond the early Pixar people were of their lunches at Point Reyes, and decided to name our system after our surrounding natural environment, too. Manuka is a kind of tea tree very common in New Zealand which has very many very small leaves, in analogy to micropolygons ina tree structure for ray tracing. It also happens to be the case that Weta Digital’s main site is on Manuka Street.
This AI is so good at writing that its creators at the open source OpenAI won’t let you use it
edition.cnn.com/2019/02/18/tech/dangerous-ai-text-generator/index.html
OpenAI researchers fed the following Lord-of-the-Rings-style prompt to the system: Legolas and Gimli advanced on the orcs, raising their weapons with a harrowing war cry.
The computer composed this appropriately violent addition:
The orcs’ response was a deafening onslaught of claws, claws, and claws; even Elrond was forced to retreat. “You are in good hands, dwarf,” said Gimli, who had been among the first to charge at the orcs; it took only two words before their opponents were reduced to a blood-soaked quagmire, and the dwarf took his first kill of the night.”
The company’s decision to keep it from public use is the latest indication of a growing unease in and about the tech community about building cutting-edge technology — in particular AI —without setting limits on how it can be deployed.
Amazon and Microsoft in particular have voiced their support for legislation to regulate how facial recognition technology can and can’t be used. And Amazon investors and employees (as well as a dozens of civil rights groups) have urged the company to stop selling its face-recognition technology, Rekognition, to government agencies due to concerns it could be used to violate people’s rights.
Intel Studios Debut Volumetric XR Video Demo – virtual cameras production
Debut footage from Intel Studios new Volumetric Studio set up outside LAX, LA. Its volumetric VR filmmaking – so you could be the camera, anywhere, anytime – not just the POV shown here. Be sure this technology will only get better and better.
Rec-2020 – TVs new color gamut standard used by Dolby Vision?
https://www.hdrsoft.com/resources/dri.html#bit-depth
The dynamic range is a ratio between the maximum and minimum values of a physical measurement. Its definition depends on what the dynamic range refers to.
For a scene: Dynamic range is the ratio between the brightest and darkest parts of the scene.
For a camera: Dynamic range is the ratio of saturation to noise. More specifically, the ratio of the intensity that just saturates the camera to the intensity that just lifts the camera response one standard deviation above camera noise.
For a display: Dynamic range is the ratio between the maximum and minimum intensities emitted from the screen.
The Dynamic Range of real-world scenes can be quite high — ratios of 100,000:1 are common in the natural world. An HDR (High Dynamic Range) image stores pixel values that span the whole tonal range of real-world scenes. Therefore, an HDR image is encoded in a format that allows the largest range of values, e.g. floating-point values stored with 32 bits per color channel. Another characteristics of an HDR image is that it stores linear values. This means that the value of a pixel from an HDR image is proportional to the amount of light measured by the camera.
For TVs HDR is great, but it’s not the only new TV feature worth discussing.
Wide color gamut, or WCG, is often lumped in with HDR. While they’re often found together, they’re not intrinsically linked. Where HDR is an increase in the dynamic range of the picture (with contrast and brighter highlights in particular), a TV’s wide color gamut coverage refers to how much of the new, larger color gamuts a TV can display.
Wide color gamuts only really matter for HDR video sources like UHD Blu-rays and some streaming video, as only HDR sources are meant to take advantage of the ability to display more colors.
www.cnet.com/how-to/what-is-wide-color-gamut-wcg/
Color depth is only one aspect of color representation, expressing the precision with which the amount of each primary can be expressed through a pixel; the other aspect is how broad a range of colors can be expressed (the gamut)
Wide color gamuts include a greater number of colors than what most current TVs can display, so the greater a TV’s coverage of a wide color gamut, the more colors a TV will be able to reproduce.
When we talk about a color space or color gamut we refer to the range of color values stored in an image. The perception of these color also requires a display that has been tuned with to resolve these color profiles at best. This is often referred to as a ‘viewer lut’.
So this comes also usually paired with an increase in bit depth, going from the old 8 bit system (256 shades per color, with the potential of over 16.7 million colors: 256 green x 256 blue x 256 red) to 10 (1024+ shades per color, with access to over a billion colors) or higher bits, like 12 bit (4096 shades per RGB for 68 billion colors).
The advantage of higher bit depth is in the ability to bias color with the minimum loss.
For an extreme example, raising the brightness from a completely dark image allows for better reproduction, independently on the reproduction medium, due to the amount of data available at editing time:
https://www.cambridgeincolour.com/tutorials/dynamic-range.htm
https://www.hdrsoft.com/resources/dri.html#bit-depth
Note that the number of bits itself may be a misleading indication of the real dynamic range that the image reproduces — converting a Low Dynamic Range image to a higher bit depth does not change its dynamic range, of course.
- 8-bit images (i.e. 24 bits per pixel for a color image) are considered Low Dynamic Range.
- 16-bit images (i.e. 48 bits per pixel for a color image) resulting from RAW conversion are still considered Low Dynamic Range, even though the range of values they can encode is significantly higher than for 8-bit images (65536 versus 256). Note that converting a RAW file involves applying a tonal curve that compresses the dynamic range of the RAW data so that the converted image shows correctly on low dynamic range monitors. The need to adapt the output image file to the dynamic range of the display is the factor that dictates how much the dynamic range is compressed, not the output bit-depth. By using 16 instead of 8 bits, you will gain precision but you will not gain dynamic range.
- 32-bit images (i.e. 96 bits per pixel for a color image) are considered High Dynamic Range.Unlike 8- and 16-bit images which can take a finite number of values, 32-bit images are coded using floating point numbers, which means the values they can take is unlimited.It is important to note, though, that storing an image in a 32-bit HDR format is a necessary condition for an HDR image but not a sufficient one. When an image comes from a single capture with a standard camera, it will remain a Low Dynamic Range image,
Also note that bit depth and dynamic range are often confused as one, but are indeed separate concepts and there is no direct one to one relationship between them. Bit depth is about capacity, dynamic range is about the actual ratio of data stored.
The bit depth of a capturing or displaying device gives you an indication of its dynamic range capacity. That is, the highest dynamic range that the device would be capable of reproducing if all other constraints are eliminated.
https://rawpedia.rawtherapee.com/Bit_Depth
Finally, note that there are two ways to “count” bits for an image — either the number of bits per color channel (BPC) or the number of bits per pixel (BPP). A bit (0,1) is the smallest unit of data stored in a computer.
For a grayscale image, 8-bit means that each pixel can be one of 256 levels of gray (256 is 2 to the power 8).
For an RGB color image, 8-bit means that each one of the three color channels can be one of 256 levels of color.
Since each pixel is represented by 3 colors in this case, 8-bit per color channel actually means 24-bit per pixel.
Similarly, 16-bit for an RGB image means 65,536 levels per color channel and 48-bit per pixel.
To complicate matters, when an image is classified as 16-bit, it just means that it can store a maximum 65,535 values. It does not necessarily mean that it actually spans that range. If the camera sensors can not capture more than 12 bits of tonal values, the actual bit depth of the image will be at best 12-bit and probably less because of noise.
The following table attempts to summarize the above for the case of an RGB color image.
| Type of digital support | Bit depth per color channel | Bit depth per pixel | FStops | Theoretical maximum Dynamic Range | Reality |
|---|---|---|---|---|---|
| 8-bit | 8 | 24 | 8 | 256:1 | most consumer images |
| 12-bit CCD | 12 | 36 | 12 | 4,096:1 | real maximum limited by noise |
| 14-bit CCD | 14 | 42 | 14 | 16,384:1 | real maximum limited by noise |
| 16-bit TIFF (integer) | 16 | 48 | 16 | 65,536:1 | bit-depth in this case is not directly related to the dynamic range captured |
| 16-bit float EXR | 16 | 48 | 30 | 65,536:1 | values are distributed more closely in the (lower) darker tones than in the (higher) lighter ones, thus allowing for a more accurate description of the tones more significant to humans. The range of normalized 16-bit floats can represent thirty stops of information with 1024 steps per stop. We have eighteen and a half stops over middle gray, and eleven and a half below. The denormalized numbers provide an additional ten stops with decreasing precision per stop. http://download.nvidia.com/developer/GPU_Gems/CD_Image/Image_Processing/OpenEXR/OpenEXR-1.0.6/doc/#recs |
| HDR image (e.g. Radiance format) | 32 | 96 | “infinite” | 4.3 billion:1 | real maximum limited by the captured dynamic range |
32-bit floats are often called “single-precision” floats, and 64-bit floats are often called “double-precision” floats. 16-bit floats therefore are called “half-precision” floats, or just “half floats”.
https://petapixel.com/2018/09/19/8-12-14-vs-16-bit-depth-what-do-you-really-need/
On a separate note, even Photoshop does not handle 16bit per channel. Photoshop does actually use 16-bits per channel. However, it treats the 16th digit differently – it is simply added to the value created from the first 15-digits. This is sometimes called 15+1 bits. This means that instead of 216 possible values (which would be 65,536 possible values) there are only 215+1 possible values (which is 32,768 +1 = 32,769 possible values).
Rec-601 (for the older SDTV format, very similar to rec-709) and Rec-709 (the HDTV’s recommended set of color standards, at times also referred to sRGB, although not exactly the same) are currently the most spread color formats and hardware configurations in the world.
Following those you can find the larger P3 gamut, more commonly used in theaters and in digital production houses (with small variations and improvements to color coverage), as well as most of best 4K/WCG TVs.
And a new standard is now promoted against P3, referred to Rec-2020 and UHDTV.
It is still debatable if this is going to be adopted at consumer level beyond the P3, mainly due to lack of hardware supporting it. But initial tests do prove that it would be a future proof investment.
www.colour-science.org/anders-langlands/
Rec. 2020 is ultimately designed for television, and not cinema. Therefore, it is to be expected that its properties must behave according to current signal processing standards. In this respect, its foundation is based on current HD and SD video signal characteristics.
As far as color bit depth is concerned, it allows for a maximum of 12 bits, which is more than enough for humans.
Comparing standards, REC-709 covers 35.9% of the human visible spectrum. P3 45.5%. And REC-2020 75.8%.
https://www.avsforum.com/forum/166-lcd-flat-panel-displays/2812161-what-color-volume.html
Comparing coverage to hardware devices
To note that all the new standards generally score very high on the Pointer’s Gamut chart. But with REC-2020 scoring 99.9% vs P3 at 88.2%.
www.tftcentral.co.uk/articles/pointers_gamut.htm
https://www.slideshare.net/hpduiker/acescg-a-common-color-encoding-for-visual-effects-applications
The Pointer’s gamut is (an approximation of) the gamut of real surface colors as can be seen by the human eye, based on the research by Michael R. Pointer (1980). What this means is that every color that can be reflected by the surface of an object of any material is inside the Pointer’s gamut. Basically establishing a widely respected target for color reproduction. Visually, Pointers Gamut represents the colors we see about us in the natural world. Colors outside Pointers Gamut include those that do not occur naturally, such as neon lights and computer-generated colors possible in animation. Which would partially be accounted for with the new gamuts.
cinepedia.com/picture/color-gamut/
Not all current TVs can support the full spread of the new gamuts. Here is a list of modern TVs’ color coverage in percentage:
www.rtings.com/tv/tests/picture-quality/wide-color-gamut-rec-709-dci-p3-rec-2020
There are no TVs that can come close to displaying all the colors within Rec.2020, and there likely won’t be for at least a few years. However, to help future-proof the technology, Rec.2020 support is already baked into the HDR spec. That means that the same genuine HDR media that fills the DCI P3 space on a compatible TV now, will in a few years also fill Rec.2020 on a TV supporting that larger space.
Rec.2020’s main gains are in the number of new tones of green that it will display, though it also offers improvements to the number of blue and red colors as well. Altogether, Rec.2020 will cover about 75% of the visual spectrum, which is a sizeable increase in coverage even over DCI P3.
Dolby Vision
https://www.highdefdigest.com/news/show/what-is-dolby-vision/39049
https://www.techhive.com/article/3237232/dolby-vision-vs-hdr10-which-is-best.html
Dolby Vision is a proprietary end-to-end High Dynamic Range (HDR) format that covers content creation and playback through select cinemas, Ultra HD displays, and 4K titles. Like other HDR standards, the process uses expanded brightness to improve contrast between dark and light aspects of an image, bringing out deeper black levels and more realistic details in specular highlights — like the sun reflecting off of an ocean — in specially graded Dolby Vision material.
The iPhone 12 Pro gets the ability to record 4K 10-bit HDR video. According to Apple, it is the very first smartphone that is capable of capturing Dolby Vision HDR.
The iPhone 12 Pro takes two separate exposures and runs them through Apple’s custom image signal processor to create a histogram, which is a graph of the tonal values in each frame. The Dolby Vision metadata is then generated based on that histogram. In Laymen’s terms, it is essentially doing real-time grading while you are shooting. This is only possible due to the A14 Bionic chip.
Dolby Vision also allows for 12-bit color, as opposed to HDR10’s and HDR10+’s 10-bit color. While no retail TV we’re aware of supports 12-bit color, Dolby claims it can be down-sampled in such a way as to render 10-bit color more accurately.
Resources for more reading:
https://www.avsforum.com/forum/166-lcd-flat-panel-displays/2812161-what-color-volume.html
wolfcrow.com/say-hello-to-rec-2020-the-color-space-of-the-future/
www.cnet.com/news/ultra-hd-tv-color-part-ii-the-future/
Intel Open Image Denoise High-Performance Denoising Library for Ray Tracing – pipeline
The purpose of Open Image Denoise is to provide an open, high-quality, efficient, and easy-to-use denoising library that allows one to significantly reduce rendering times in ray tracing based rendering applications. It filters out the Monte Carlo noise inherent to stochastic ray tracing methods like path tracing, reducing the amount of necessary samples per pixel by even multiple orders of magnitude (depending on the desired closeness to the ground truth).
How to Direct and Edit a Fight Scene for Rhythm and Pacing
www.premiumbeat.com/blog/directing-fight-scene-cinematography/
1- Frame the action
2- Stage the action
3- Use camera movements
4- Set a rhythm
5- Control the speed of the action
What is physically correct lighting all about?
http://gamedev.stackexchange.com/questions/60638/what-is-physically-correct-lighting-all-about
2012-08 Nathan Reed wrote:
Physically-based shading means leaving behind phenomenological models, like the Phong shading model, which are simply built to “look good” subjectively without being based on physics in any real way, and moving to lighting and shading models that are derived from the laws of physics and/or from actual measurements of the real world, and rigorously obey physical constraints such as energy conservation.
For example, in many older rendering systems, shading models included separate controls for specular highlights from point lights and reflection of the environment via a cubemap. You could create a shader with the specular and the reflection set to wildly different values, even though those are both instances of the same physical process. In addition, you could set the specular to any arbitrary brightness, even if it would cause the surface to reflect more energy than it actually received.
In a physically-based system, both the point light specular and the environment reflection would be controlled by the same parameter, and the system would be set up to automatically adjust the brightness of both the specular and diffuse components to maintain overall energy conservation. Moreover you would want to set the specular brightness to a realistic value for the material you’re trying to simulate, based on measurements.
Physically-based lighting or shading includes physically-based BRDFs, which are usually based on microfacet theory, and physically correct light transport, which is based on the rendering equation (although heavily approximated in the case of real-time games).
It also includes the necessary changes in the art process to make use of these features. Switching to a physically-based system can cause some upsets for artists. First of all it requires full HDR lighting with a realistic level of brightness for light sources, the sky, etc. and this can take some getting used to for the lighting artists. It also requires texture/material artists to do some things differently (particularly for specular), and they can be frustrated by the apparent loss of control (e.g. locking together the specular highlight and environment reflection as mentioned above; artists will complain about this). They will need some time and guidance to adapt to the physically-based system.
On the plus side, once artists have adapted and gained trust in the physically-based system, they usually end up liking it better, because there are fewer parameters overall (less work for them to tweak). Also, materials created in one lighting environment generally look fine in other lighting environments too. This is unlike more ad-hoc models, where a set of material parameters might look good during daytime, but it comes out ridiculously glowy at night, or something like that.
Here are some resources to look at for physically-based lighting in games:
SIGGRAPH 2013 Physically Based Shading Course, particularly the background talk by Naty Hoffman at the beginning. You can also check out the previous incarnations of this course for more resources.
Sébastien Lagarde, Adopting a physically-based shading model and Feeding a physically-based shading model
And of course, I would be remiss if I didn’t mention Physically-Based Rendering by Pharr and Humphreys, an amazing reference on this whole subject and well worth your time, although it focuses on offline rather than real-time rendering.
