: What does the term raster and vector apply to? Are the term raster and vector only applicable to 2d graphics? The rendered image format might be one way to distinguish them but what about

@Si6392903

Are vectors, realy vectors?

First of all, where does the term vectors come from?

One definition is: a quantity possessing both magnitude and direction, represented by an arrow the direction of which indicates the direction of the quantity and the length of which is proportional to the magnitude.

In a 2D program you strech the nodes to form curves. Segments of a shape.

On 3D models you have more than just "vectors" You have diferent modeling techniques. The one similar to a 2D vector are bezier and other is nurbs. So you have 2 kind of modes to construct a model.

Another one is meshes. This is a coordinate system based. Just straight edges conecting verex. There is no vector there, becouse they have a magnitude of 0 with no direction.

You have thoose too in 2D programs, but the term vectors is more general. I am not sure, but I recall that some formats like EMF can record only straight segments of lines, not vectors properly speaking. If you are just drawing a rectangle on Ilustrator... probably that is not a vector.

Raster or bitmap?

In my opinion the correct term is bitmap. A map (2D coordinate system made of bits of information) Probably the correct term should be pixelmaps. Raster in my opinion is the proces of converting information into an image.

For example, one definition on CRT television is: A pattern of scanning lines covering the area upon which the image is projected in the cathode-ray tube of a television set.

The ray was "scraching" (raster) the tv to form an image.

On a 3D program rasterizing is the process now called Render. But inclusive just playing with the wireframe view needs some kind of rendering.

On a 2D program (vector) you also need rendering and rasterizing. After all, a 2D image format is just information you need to see as a human.

What is inside?


are 3d programs like blender raster or vector?


That is not the clasification used. They use vectors, coordinates and use inside bitmaps and can output "rasterize" bitmaps and videos. It becomes more complex than just thoose 2 terms.

The general clasification of 3D programs are:


Modeling (with a lot of sub categories here, like sculpting, nurbs, extruding, etc.) Ok... A "vector" like component.
Animation. You can use vectors or coordinate systems... Ok. Another "vector" like component.
Rendering (there are a lot of render engines for diferent modeling aplications) They take the "vector" component to export a raster one.


And some features that are more like accesories for the main initial categories, like phisics and particles simulators.

The same for a 2D program. One vector based software can also handle bitmaps and export them.

Edited For example, take a look at this category: en.wikipedia.org/wiki/List_of_3D_rendering_software and take a look at the bottom of the page, "see also".

Vote Up Vote Down

@Shelley591

For 3D apps, I think it's easiest to think of the 'building' aspect of the model as being based on vectors, and the 'rendering' aspect being based on raster.

To take a very simple example: Sketchup. You can build a 3-D box and rotate it all you want. It's easy to rotate this because it's a set of vector instructions that the computer can quickly re-calculate to make a temporary raster version for your pixel-based screen. You can then go a step further and add textures, lighting, shading, shadows, etc, and have Sketchup render a static, detailed raster version of a particular view.

Vote Up Vote Down

@Sherry646

A vector is a descriptive language that allows a person or computer to recreate the image by following a set of steps. It is metaphorically similar to a musical score. It has no resolution.

A raster is a string of numbers (brightness data) which have a one to one correspondence to a 2-d grid. The courseness of the grid (resolution) is set at the time of creation. This is metaphorically similar to a fixed grid of lights where each light is on a dimmer, and each number determines the dimmer setting for a single light. (originally, the raster data (a bitmap) was simply shoved as-is into the display memory)

All images must be rasterized for any form of output: display, print, etc. (plotters/laser cutters might be considered an exception to this). Raster is pre-baked, vector is made from scratch.

Raster images must be "resampled" if their pre-determined grid does not match the output device grid (their resolution does not match). You'd have to add or remove lights from the light bank and then {do magic} to recalculate the brightness to configure the light switches. This is usually very fast, and even faster when the grid is not altered.

Vector images have no grid (they are resolution independent), so the output device follows the instructions (the musical score) creating a new performance of it every time at the desired grid size (or the best possible for that device). This is computationally intensive and slower, but allows for the best possible resolution at any moment in time.

The main difference between a "vector image" and a 3d model is there is a 3rd dimension. They are both vector.

Back in the 20th century, 3d games would basically be wireframe or colored vectors (see Battlezone or Mechwarrior 2) and it was more obvious.

I haven't looked at Blender in a long time, but it was more a vector program and didn't really have a lot of raster image manipulation: you would do textures in Photoshop etc, and then attach them in blender, applying "shaders" etc to get lighting. What we perceive as an alien is usually what Illustrator etc would call a "raster effect."

The thing is that 3d models are almost always textured and those textures are almost always raster images and they have a resolution and the best way to get really good looking models is to have large-grid-size raster images. This is why, despite vectors being able to do more at a smaller file size than a raster image, games are upwards of 40GB anymore.

If the CPU power was there, they could probably go 100% vector and just render all surfaces on-the fly.

Vote Up Vote Down

@Murray976

Raster is bitmap image - Height and Width are defined with a certain number of pixels.

Vector - The image is formed by mathematical calculation and don't have a fix width and height. Dimensions can be changed whenever you want.

Vote Up Vote Down

@Yeniel278

Well, the terms do not really align in 3D and in 2D they are a bit elusive too. Most 3D applications are not something one could call raster, but it does not have to be vector either. It depends on exactly how you define the terms. Since the nuance of the definition is important, it is hard to say.

Generally I would say that 3D the images are most likely classifiable as vector*, if by vector one means the scale-ability of the end render. The description language of most 3D engines resemble more a page description language of a vector graphics engine.

However unlike 2D vector graphics, there is not necessarily a general re-rendering paradigm in 3D. So a 3D software has to output a raster image. This is true in most cases, but exceptions exist. So the end result is not a vector for your client.

Then there is the discrete versus continuous function discussion. Does raster imply discrete data sampling? Are raster data editors truly that, if you can change the text isn't that a vector description feature? So not even in 2D is the distinction so easy.

So in the end it depends on who's definition you use. The distinction is getting harder to make as time goes forward. I would say that in the end:


It depends on what you want to achieve and what you do. From a printers point of view, anything that is not vector content in a PDF is not vector. But since most data is mixed content, it's really hard to say.

Does something become vector even if there is just one vector element? Or is in fact all printer output vector on account that the page description languages image placement is a vector operation in its core?
The easiest way to deal with this is if your central work revolves around designing pixels and modifying individual samples - then it's raster. But since nothing is really that black and white it's up for grabs.


* Not all 3D apps are the same. There is more variation in 3D than in 2D. So no not all fall into this category. Some do lean much more on the raster side (voxels anyone).

Vote Up Vote Down