5 ms·

A big thing about pixel art and it's issues today is that what he mentions as pixel art, isn't so. I mean, he argues about the way the devices scale his art up

A big thing about pixel art and it's issues today is that what he mentions as pixel art, isn't so. I mean, he argues about the way the devices scale his art up and change aspect ratio etc. But on the other hand, he compares a picture of Yoshi at the end to an algorithmically smoothed version. However, Yoshi was never intended to look like the leftmost image. The thing is, pixels aren't square. In fact, when we played it on the SNES, it was rendered very differently, it was interlaced, scanlines were distinct, and each pixel was a bit of a blob. Because of this, an image like Yoshi looked just fine, because we weren't looking at a blocky grid of squares. The brighter colored glowed around the darker colors, the darkness between the scanlines caused your brain to interpolate the lines. We got a nicer looking image. When you blow it up to 500% on a crisp LCD panel it doesn't look the same. What used to be smooth lines are now stair steps. And just the same way that he complains about Auro being blurred or stretched, when you display pixel art designed for a CRT on an OLED display at 600% nearest neighbor scaling, it doesn't work. I like the symbolic style of pixel art. But we don't have the same limitations. When people made great art for the NES, it's because that's what they had to work with. We have more to work with. It's one thing to represent things in a symbolic, simplistic or minimalistic style. It's another thing to feign technological disadvantage.

soup10 11y ago
I agree with your sentiment, I think the author makes some decent points but is pretty biased towards an aesthetic he prefers and the art tools he's most familiar with. The sf4 example is particularly demonstrating of that, people that do 3d modeling know that it's not as easy to get the arbitrary distortion effects you can do with frame by frame animation, that doesn't make it lower quality animation, it's a trade off of the medium. And art is about the overall impression it leaves on the viewer, not the technical specifics about how the pants look odd on 2 frames of a 100 frame motion
- the_af 11y ago
  How can he be biased, beyond his love of pixel art? He more or less concludes pixel art is not the best choice, and he sides with the public's opinion on this issue. Read the comments. He is not blaming anyone for this "pixel tax". To my eyes, the SF4 Chun Li is infinitely inferior to SF3, and I'm not even an animator. And haven't played either game, so I truly have no preconceptions at all in this case.
  - mrob 11y ago
    SF4 has bad animation because it mixes mostly realistic motions with cartoonish anatomy. SF3 consistently uses cartoonish style for both. The only obvious reason somebody might prefer SF4 is if they really hate pixel art. I've played far more SF4 than SF3 and I still prefer the SF3 look.
  - JohnBooty 11y ago
    I liked SF4's slightly goofy animation because to me, it looked like children's action figures - specifically, the wrestling figures I spent so much time with as a kid: http://www.thecanadianconnection.biz/images/figures/large/HulkHogan01.jpg http://www.thecanadianconnection.biz/images/figures/large/Hu... To me, SF4 looks like those. Except in motion, with lots of special effects. The way I imagined them as a kid. :)
  - saint_fiasco 11y ago
    The Chun Li comparison reminds me a lot of the messy transition between 2D and 3D of games like Final Fantasy and Mortal Kombat. In Mortal Kombat, the typical shaolin monk character has a typical "shoot fireball" special move. In the latest SNES game, a badass dragon made of fire shoots out of his hand and explodes on contact with the enemy. In the next game, for the PlayStation, it looks like he is flinging a flaming turd at his opponent. Final Fantasy had similar problems in the transition from awesome 2D (FFVI) to awful 3D (FFVII). Both Mortal Kombat and Final Fantasy eventually got 3D high definition graphics right, in my opinion. I hope eventually the artists will be as good and as experienced at making 3D art as they were at making 2D art in the SNES era.
    - bitwize 11y ago
      The thing that set Mortal Kombat apart back in the day -- besides blood and gore -- was its use of digitized video of live actors as its character sprites. This couldn't be done in the 3D era, and up Through Mortal Kombat (2011) the 3D character models looked hella goofy when compared with real people.
- egypturnash 11y ago
  > people that do 3d modeling know that it's not as easy to get the arbitrary distortion effects you can do with frame by frame animation So who wants to start writing a 3D package designed around making those kinds of things as easy as possible? With a way to easily drop them into a game and fire it off in realtime? Maybe not you, but maybe someone reading the comments here just found a side project. The big tactics 2D cartoonists use to stylize motion are squash and stretch, smears, and multiple images. How can these be represented efficiently in data structures? How can you make it easy for a character to leap into the air and turn into a giant eyeball with arms and legs sticking off of it[1]? And how can you make it easy for a non-techie artist to make it happen? This is probably not a quick weekend hack. Guilty Gear Xrd does a lot of this; can it be done in a game with a lot more characters running around at once? [1] https://www.google.com/search?q=clampett's+corneal+catastrophe https://www.google.com/search?q=clampett's+corneal+catastrop...
- jdbernard 11y ago
  Those arbitrary distortions are very important though: https://vimeo.com/93206523 https://vimeo.com/93206523 And it has been done with 3D animation, Pixar is an easy example: http://m.youtube.com/watch?v=GnfdLp0PcSg http://m.youtube.com/watch?v=GnfdLp0PcSg
emsy 11y ago
For anyone interested in the topic, here is an article written with dedication about it http://gamasutra.com/blogs/KylePittman/20150420/241442/CRT_Simulation_in_Super_Win_the_Game.php http://gamasutra.com/blogs/KylePittman/20150420/241442/CRT_S...
- qznc 11y ago
  Great article. This image from the article is probably the best summary: https://i.imgur.com/Pq2Yra4.png https://i.imgur.com/Pq2Yra4.png
the_af 11y ago
> It's one thing to represent things in a symbolic, simplistic or minimalistic style. It's another thing to feign technological disadvantage. Different in what ways? It's a perfectly valid and common artistic resource to "feign disadvantage", limiting the colors, the subject, the perspective or, yes, the resolution. If I understood the article, the author acknowledges the historical limitations of the "pixelated look", states he and an "insular" community of artists like it by choice, and at the same time he also acknowledges to the general public, the pixelated look is bizarre (what he calls the "pixel tax"), and therefore he understands he must change with the times. And in the end, he more or less concludes that good art is good art, regardless of the specific style.
Tyr42 11y ago
For anyone else interested in scaling up pixel art, there are alternative methods to just using nearest neighbour, such as EPX or something. I found this article on Datagenics to be pretty solid: http://www.datagenetics.com/blog/december32013/index.html http://www.datagenetics.com/blog/december32013/index.html
- greggman 11y ago
  I liked this one too http://www.piratehearts.com/blog/2014/03/28/crt-simulation/ http://www.piratehearts.com/blog/2014/03/28/crt-simulation/
- swimfar 11y ago
  It's linked to on that page, but here's a good paper on scaling/interpolating pixel art. They use a lot of SNES sprites as examples. Some of these are implemented in emulators so you can play around with them a little. http://research.microsoft.com/en-us/um/people/kopf/pixelart/paper/pixel.pdf http://research.microsoft.com/en-us/um/people/kopf/pixelart/...
mrob 11y ago
The SNES supported interlaced video modes, but Super Mario World did not use them. Early consoles achieved progressive scan by sending all the fields at the same polarity (eg. only odd field or only even fields), violating the NTSC standard. This is the reason why they had obvious scanlines on CRTs. Because this is a non-standard mode some modern TVs and capture cards will fail to support it and show interlacing artifacts, but this is not how it looked originally.
wodenokoto 11y ago
Maybe in the states. PAL and NTSC versions of games looks very different. Living in Europe, I could hardly recognize my favorite games when I saw them in American movies, or tv-show. i wonder how the Japanese PAL would render it, because despite being PAL it's still different from Europe.
- pandaman 11y ago
  Was there ever Japanese PAL? If so I cannot find any references to it. The PAL games indeed look different from the NTSC because they are different games. Other than the obvious difference in the video configuration there are usually content differences due to local laws, different marketing policies and technical limitations (e.g. an NTSC games in general did not have to support as many languages as PAL ones).
  - wodenokoto 11y ago
    I was wrong, Japan used a variation of NTSC http://en.wikipedia.org/wiki/NTSC-J http://en.wikipedia.org/wiki/NTSC-J
tlrobinson 11y ago
In fact, some emulators have modes to emulate interlacing/scanlines/etc.

soup10 11y ago

I agree with your sentiment, I think the author makes some decent points but is pretty biased towards an aesthetic he prefers and the art tools he's most familiar with. The sf4 example is particularly demonstrating of that, people that do 3d modeling know that it's not as easy to get the arbitrary distortion effects you can do with frame by frame animation, that doesn't make it lower quality animation, it's a trade off of the medium. And art is about the overall impression it leaves on the viewer, not the technical specifics about how the pants look odd on 2 frames of a 100 frame motion

the_af 11y ago
How can he be biased, beyond his love of pixel art? He more or less concludes pixel art is not the best choice, and he sides with the public's opinion on this issue. Read the comments. He is not blaming anyone for this "pixel tax". To my eyes, the SF4 Chun Li is infinitely inferior to SF3, and I'm not even an animator. And haven't played either game, so I truly have no preconceptions at all in this case.
- mrob 11y ago
  SF4 has bad animation because it mixes mostly realistic motions with cartoonish anatomy. SF3 consistently uses cartoonish style for both. The only obvious reason somebody might prefer SF4 is if they really hate pixel art. I've played far more SF4 than SF3 and I still prefer the SF3 look.
- JohnBooty 11y ago
  I liked SF4's slightly goofy animation because to me, it looked like children's action figures - specifically, the wrestling figures I spent so much time with as a kid: http://www.thecanadianconnection.biz/images/figures/large/HulkHogan01.jpg http://www.thecanadianconnection.biz/images/figures/large/Hu... To me, SF4 looks like those. Except in motion, with lots of special effects. The way I imagined them as a kid. :)
- saint_fiasco 11y ago
  The Chun Li comparison reminds me a lot of the messy transition between 2D and 3D of games like Final Fantasy and Mortal Kombat. In Mortal Kombat, the typical shaolin monk character has a typical "shoot fireball" special move. In the latest SNES game, a badass dragon made of fire shoots out of his hand and explodes on contact with the enemy. In the next game, for the PlayStation, it looks like he is flinging a flaming turd at his opponent. Final Fantasy had similar problems in the transition from awesome 2D (FFVI) to awful 3D (FFVII). Both Mortal Kombat and Final Fantasy eventually got 3D high definition graphics right, in my opinion. I hope eventually the artists will be as good and as experienced at making 3D art as they were at making 2D art in the SNES era.
  - bitwize 11y ago
    The thing that set Mortal Kombat apart back in the day -- besides blood and gore -- was its use of digitized video of live actors as its character sprites. This couldn't be done in the 3D era, and up Through Mortal Kombat (2011) the 3D character models looked hella goofy when compared with real people.
egypturnash 11y ago
> people that do 3d modeling know that it's not as easy to get the arbitrary distortion effects you can do with frame by frame animation So who wants to start writing a 3D package designed around making those kinds of things as easy as possible? With a way to easily drop them into a game and fire it off in realtime? Maybe not you, but maybe someone reading the comments here just found a side project. The big tactics 2D cartoonists use to stylize motion are squash and stretch, smears, and multiple images. How can these be represented efficiently in data structures? How can you make it easy for a character to leap into the air and turn into a giant eyeball with arms and legs sticking off of it[1]? And how can you make it easy for a non-techie artist to make it happen? This is probably not a quick weekend hack. Guilty Gear Xrd does a lot of this; can it be done in a game with a lot more characters running around at once? [1] https://www.google.com/search?q=clampett's+corneal+catastrophe https://www.google.com/search?q=clampett's+corneal+catastrop...
jdbernard 11y ago
Those arbitrary distortions are very important though: https://vimeo.com/93206523 https://vimeo.com/93206523 And it has been done with 3D animation, Pixar is an easy example: http://m.youtube.com/watch?v=GnfdLp0PcSg http://m.youtube.com/watch?v=GnfdLp0PcSg

emsy 11y ago

For anyone interested in the topic, here is an article written with dedication about it http://gamasutra.com/blogs/KylePittman/20150420/241442/CRT_Simulation_in_Super_Win_the_Game.php http://gamasutra.com/blogs/KylePittman/20150420/241442/CRT_S...

qznc 11y ago
Great article. This image from the article is probably the best summary: https://i.imgur.com/Pq2Yra4.png https://i.imgur.com/Pq2Yra4.png

the_af 11y ago

> It's one thing to represent things in a symbolic, simplistic or minimalistic style. It's another thing to feign technological disadvantage. Different in what ways? It's a perfectly valid and common artistic resource to "feign disadvantage", limiting the colors, the subject, the perspective or, yes, the resolution. If I understood the article, the author acknowledges the historical limitations of the "pixelated look", states he and an "insular" community of artists like it by choice, and at the same time he also acknowledges to the general public, the pixelated look is bizarre (what he calls the "pixel tax"), and therefore he understands he must change with the times. And in the end, he more or less concludes that good art is good art, regardless of the specific style.

Tyr42 11y ago

For anyone else interested in scaling up pixel art, there are alternative methods to just using nearest neighbour, such as EPX or something. I found this article on Datagenics to be pretty solid: http://www.datagenetics.com/blog/december32013/index.html http://www.datagenetics.com/blog/december32013/index.html

greggman 11y ago
I liked this one too http://www.piratehearts.com/blog/2014/03/28/crt-simulation/ http://www.piratehearts.com/blog/2014/03/28/crt-simulation/
swimfar 11y ago
It's linked to on that page, but here's a good paper on scaling/interpolating pixel art. They use a lot of SNES sprites as examples. Some of these are implemented in emulators so you can play around with them a little. http://research.microsoft.com/en-us/um/people/kopf/pixelart/paper/pixel.pdf http://research.microsoft.com/en-us/um/people/kopf/pixelart/...

mrob 11y ago

The SNES supported interlaced video modes, but Super Mario World did not use them. Early consoles achieved progressive scan by sending all the fields at the same polarity (eg. only odd field or only even fields), violating the NTSC standard. This is the reason why they had obvious scanlines on CRTs. Because this is a non-standard mode some modern TVs and capture cards will fail to support it and show interlacing artifacts, but this is not how it looked originally.

wodenokoto 11y ago

Maybe in the states. PAL and NTSC versions of games looks very different. Living in Europe, I could hardly recognize my favorite games when I saw them in American movies, or tv-show. i wonder how the Japanese PAL would render it, because despite being PAL it's still different from Europe.

pandaman 11y ago
Was there ever Japanese PAL? If so I cannot find any references to it. The PAL games indeed look different from the NTSC because they are different games. Other than the obvious difference in the video configuration there are usually content differences due to local laws, different marketing policies and technical limitations (e.g. an NTSC games in general did not have to support as many languages as PAL ones).
- wodenokoto 11y ago
  I was wrong, Japan used a variation of NTSC http://en.wikipedia.org/wiki/NTSC-J http://en.wikipedia.org/wiki/NTSC-J

tlrobinson 11y ago

In fact, some emulators have modes to emulate interlacing/scanlines/etc.