In general, I am still a little confused on the differences between ray casting and rasterization. I understand that ray tracing works "backwards" from our own view, as opposed to rasterization, but in some aspects the end result from both processes seems to be the same?
arjunsrinivasan1997
I am still confused if there are types of situations in which rasterization would produce a better looking image than ray casting. What would these situations need to have as vital components in order to make ray casting less appealing?
Pinbat
@arjun, in general, it's a tradeoff between speed and quality. For video games and applications that require real-time graphics, rasterization is still prime (and it's gotten even better over the years). Pre-rendered things like movies will use ray tracing.
It's mentioned in lecture that both methods can be equivalent. I'm guessing the difference in quality is due to the fact that certain things may be more efficient using rasterization and certain other things may look better with ray tracing.
eliot1019
I don't think they would be equivalent for certain applications. Video games have been doing hacky things for years like cubemaps or screenspace reflections to generate pseudo realistic reflections or simulate natural lighting.
killawhale2
It is interesting to note that recently Nvidia has introduced real-time raytracing for games. Although this does come with a FPS drop, if ray tracing can be done in real time, shouldn't that indicate that the technology is moving away from rasterization?
cornrow-kenny
On the topic of real-time raytracing, thought I'd share a very cool real-time rendered demo that heavily features raytracing.
executable if you want to run it yourself: http://www.pouet.net/prod.php?which=61211
writeup from the coder: https://directtovideo.wordpress.com/2013/05/07/real-time-ray-tracing/
spopat
The introduction to this paper on “3D Rasterization" (http://cg.ivd.kit.edu/publications/p2012/3dr/gi2012.pdf), which is apparently an attempt to bridge some benefits of both ray tracing and rasterization, provides a summary of pros and cons of these two methods. For example they state that rasterization is used for interactive or real time graphics because of its low cost (“no need for floating point in 2D image space”), and that ray tracing requires “largely unpredictable memory accesses”. As mentioned in lecture they note that ray tracing operates directly on 3D space (and simple ray casting can be naturally extended to support different illumination effects).
In general, I am still a little confused on the differences between ray casting and rasterization. I understand that ray tracing works "backwards" from our own view, as opposed to rasterization, but in some aspects the end result from both processes seems to be the same?
I am still confused if there are types of situations in which rasterization would produce a better looking image than ray casting. What would these situations need to have as vital components in order to make ray casting less appealing?
@arjun, in general, it's a tradeoff between speed and quality. For video games and applications that require real-time graphics, rasterization is still prime (and it's gotten even better over the years). Pre-rendered things like movies will use ray tracing.
https://blogs.nvidia.com/blog/2018/03/19/whats-difference-between-ray-tracing-rasterization/
It's mentioned in lecture that both methods can be equivalent. I'm guessing the difference in quality is due to the fact that certain things may be more efficient using rasterization and certain other things may look better with ray tracing.
I don't think they would be equivalent for certain applications. Video games have been doing hacky things for years like cubemaps or screenspace reflections to generate pseudo realistic reflections or simulate natural lighting.
It is interesting to note that recently Nvidia has introduced real-time raytracing for games. Although this does come with a FPS drop, if ray tracing can be done in real time, shouldn't that indicate that the technology is moving away from rasterization?
On the topic of real-time raytracing, thought I'd share a very cool real-time rendered demo that heavily features raytracing.
video: https://www.youtube.com/watch?v=i8hSZGTXTx8
executable if you want to run it yourself: http://www.pouet.net/prod.php?which=61211
writeup from the coder: https://directtovideo.wordpress.com/2013/05/07/real-time-ray-tracing/
The introduction to this paper on “3D Rasterization" (http://cg.ivd.kit.edu/publications/p2012/3dr/gi2012.pdf), which is apparently an attempt to bridge some benefits of both ray tracing and rasterization, provides a summary of pros and cons of these two methods. For example they state that rasterization is used for interactive or real time graphics because of its low cost (“no need for floating point in 2D image space”), and that ray tracing requires “largely unpredictable memory accesses”. As mentioned in lecture they note that ray tracing operates directly on 3D space (and simple ray casting can be naturally extended to support different illumination effects).