Alvy Ray Smith

1995 classic about this: http://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf https://entropymine.com/imageworsener/bicubic/

There is a bonus chapter on his website: http://alvyray.com/DigitalLight/default.htm I am still reading this but love it so far. I love seeing first-hand historical accounts for computing

Gamma should not have any effect on the alpha channel, which is linear by definition. The alpha channel represents the average coverage of (infinite) subpixels within the pixel. The color encoding of an object should not have any impact on the coverage within that pixel. I recommend reading the technical memo "Alpha and the History of Digital Compositing" by Alvy Ray Smith [1] to get a better intuition on the matter. [1] http://alvyray.com/Memos/CG/Microsoft/7_alpha.pdf

Yeah, it's funny how he actually ended up as a "ray smith" for a living. And even the "Alvy" part means something graphics related in some obscure language iirc. He wrote about that on his homepage, but I can't find it now - his homepage (at http://alvyray.com) is a bit... special.

This classic of his taught me much about rendering: http://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf However, if we're going to talk about actual genius, no one even comes close to Eric Veach; he's like the von Neumann of rendering. Fun bit of trivia: after he finished completely revolutionising all of rendering with his PhD thesis[0], he went on to develop a little thing called AdWords. [0] Robust Monte Carlo Methods for Light Transport Simulation - https://graphics.stanford.edu/papers/veach_thesis/

Probably not faster. Definitely better quality. As Alvy Ray Smith said, a pixel is not a little square, http://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf Edit: If you want to do even better than the tensor product Lanczos filtering, you can do a filter based on Euclidean distance. Make sure that you work in linear (non-gamma-adjusted) color space. A couple more resources to read: http://www.imagemagick.org/Usage/filter/#cylindrical http://www.imagemagick.org/Usage/filter/nicolas/ Edit 2: really weird that this is getting downvoted. There’s not really anything to dispute here. It is straight-forward to show that the Lanczos method has objectively better output. Moreover, Alvy Ray Smith’s paper is a classic which anyone interested in image processing should read.

They aren't wrong. Regarding the comments and images in the linked article, there's a key point made there that's fundamentally wrong. Let me pick out a quote: "We all get how pixels basically work. A computer divides a display into squares, and each square can be assigned one RGB value at a time. The total squares supported by the hardware is the device’s “resolution.” This square grid is the smallest possible subdivision of detail available to an artist." This is the bit that's fundamentally wrong. Pixels are not squares. And never have been. They are "point samples". The fact that "magnified" pixel art displayed here compounds the error by displaying big squares rather than using an appropriate reconstruction filter doesn't help correct the misconception. Here's a link to a classic paper on the topic: http://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf

Steve was someone I threw out as a person who had multiple acts. I threw out Elon as an example on this page too. We are not trying to solve who the best business man is? After all, what would any of that have to do with the Woz? If you want to compare a $2 billion business to one of the most valuable companies in the world, we could be here all day. By the way, I based my claim that Jobs was CEO of Pixar based on the co-founder of Pixar, and other facts I found on the Internet. Can you provide some facts to show that I'm wrong? http://en.m.wikipedia.org/wiki/Alvy_Ray_Smith http://alvyray.com/Pixar/PixarMyth1.htm

Why are you convinced of this, and can I help unconvince you? ;) What you describe is what’s called “Box Filtering” in the article. Box filtering is well studied, and it is known to not be the best possible output quality. The reason this is not the best approach is because a pixel is not a little square, a pixel is a sample of a signal, and it has to be approached with signal processing and human perception in mind. (See the famous paper linked in the article: A Pixel is Not a Little Square, A Pixel is Not a Little Square, A Pixel is Not a Little Square http://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf) It can be surprising at first, but when you analytically compute the area of non-overlapping parts of a pixel (i.e., use Box Filtering) you can introduce high frequencies that cause visible aliasing artifacts that will never go away. This is also true if you are using sub-sampling of a pixel, taking point samples and averaging them, no matter how many samples you take. You can see the aliasing I’m talking about in the example at the top of the article, the 3rd one is the Box Filter - equivalent to computing the area of the polygons within each pixel. Look closely near the center of the circle where all the lines converge, and you can see little artifacts above and below, and to the left and right of the center, artifacts that are not there in the “Bilinear Filter” example on the right.

Did you mean [x,y,z,w]*M? For textbox use, my workaround is to write [x,y,z,w]^T when I want to be clear that I really mean a column vector. Incidentally, your comment inspired me to go spelunking through some old tech memos (Standard disclaimer: RenderMan dev, opinions my own, this is my speculation, yada yada). As best as I can tell, the row vector convention in RenderMan must have came from Newman and Sproull's "Principles of Interactive Computer Graphics" [1] by way of Alvy Ray Smith who referred to the convention as "computer graphics normal form" [2]. Interestingly, that book seems to have been inspired in turn by course material from Ivan Sutherland [3], who had previously built a head-mounted display using custom hardware with a row vector times matrix multiplier unit, among other things [4]. [1] https://archive.org/details/principlesofinter00newm/page/334/mode/2up [2] http://alvyray.com/Memos/CG/Pixar/Matrix64.pdf#page=6 [3] https://www.computer.org/csdl/magazine/an/2018/04/08506443/17D45Xh13tX [4] http://www.medien.ifi.lmu.de/lehre/ss09/ar/p757-sutherland.pdf#page=6

To expand on what the previous said, here's a link to a seminal paper on pixels - "A pixel is not a little square" by Alvy Ray Smith. https://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf

Obligatory "A Pixel Is Not A Little Square" https://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf

See also: A Pixel Is Not a Little Square (1995) [pdf] – http://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf

> A Pixel Is Not A Little Square! > This is an issue that strikes right at the root of correct image (sprite) computing and the ability to correctly integrate (converge) the discrete and the continuous. The little square model is simply incorrect. It harms. It gets in the way. If you find yourself thinking that a pixel is a little square, please read this paper. > A pixel is a point sample. It exists only at a point. For a color picture, a pixel might actually contain three samples, one for each primary color contributing to the picture at the sampling point. We can still think of this as a point sample of a color. But we cannot think of a pixel as a square—or anything other than a point. Alvy Ray Smith, 1995 http://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf

Pixels are point samples, they do not have corners, they are not little squares. [1] [1] http://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf

On the other hand, A Pixel Is Not A Little Square[0] would disagree, a raster is a grid sample of a continuous function. [0] http://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf

> Pixels are a unit of area like an acre or square meter. Some computer graphics experts don't agree with this: http://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf

There is a very good signal theoretic framework to discuss this. You can start reading from example Alvy Ray Smith's "A pixel is not a little square" http://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf But in the end, the sampling kernel you choose to sample your line, and, what sort of geometric primitive your line is in the first place, defaults back to matters of taste so there is a very good theoretical basis for discussing all of this, but, there is no ultimate absolutely correct answer that it will provide.

The first video clip on the page is about “what is a pixel”; here is an article Alvy Ray Smith (cofounder of Pixar) wrote specifically about that: http://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf And here he defines the HSV and HSL color coordinate spaces (transforms of RGB): http://alvyray.com/Papers/CG/color78.pdf Btw the book of his mentioned is “Biography of a Pixel” (2021). Cool guy.

Alvy Ray Smith wrote a good paper which might help to set things straight for people: http://alvyray.com/Memos/MemosCG.htm#ImageCompositing

Obligatory: http://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf

Many interesting learnings. One of them is how much the level of abstraction has been pushed up thru the years of technological improvements. A graphic designer in 2020 does not need to know how memory works in the CPU, not even most developers need to know that. The other learning, in the oposite direction, is about "A Pixel Is Not A Little Square" (http://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf). Abstractions are really useful, but they fail in some cases. Pixels are more complex that the abstractions we use. And, of course, the article has some beautiful graphics.

Related: "A pixel is not a little square". http://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf

"A Pixel Is Not A Little Square" http://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf

> Another way to look at is that digital audio is not like digital imaging. There aren't pixels. For the mathematically inclined, this would probably be a good time to repeat: pixels are not little squares[1]. [1] http://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf

I think he means that A pixel is not a little square: http://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf

Alvy Ray Smith may have the canonical paper on this, from 1995: http://alvyray.com/Memos/CG/Microsoft/4_comp.pdf

"A Pixel Is Not A Little Square, A Pixel Is Not A Little Square, A Pixel Is Not A Little Square! (And a Voxel is Not a Little Cube)" From: Technical Memo 6 Alvy Ray Smith, July 17, 1995 Link: http://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf

> Most anti-aliasing techniques do boundary smoothing by treating each pixel as a little square and visualizing the fraction of the pixel area that is contained inside the outline. Note that this is a kludge [1]. [1] A Pixel Is Not A Little Square: http://alvyray.com/Memos/CG/Microsoft/6_pixel.pdf

Elon maybe, but not Jobs. Jobs was Pixar's investor and an advisor, but he didn't run the company. http://alvyray.com/pixar/PixarMyth5.htm Although Jobs originally wanted to buy/run Pixar and essentially turn it into Apple's competitor (this was post-Apple and pre-NeXT). It wasn't until Jobs was occupied with NeXT that the Pixar guys were comfortable bringing him in.