It is currently Sun Dec 16, 2018 5:30 am

All times are UTC - 7 hours





Post new topic Reply to topic  [ 15 posts ] 
Author Message
PostPosted: Tue Nov 13, 2018 6:13 pm 
Offline
User avatar

Joined: Sat Mar 30, 2013 12:24 am
Posts: 371
Image
https://vgdensetsu.tumblr.com/post/1796 ... e-industry

_________________
www.mteegfx.com


Top
 Profile  
 
PostPosted: Wed Nov 14, 2018 5:41 am 
Offline
User avatar

Joined: Thu Aug 13, 2015 4:40 pm
Posts: 326
Location: Rio de Janeiro - Brazil
That article contains what I wanted to read: that at least at one point, the artists did make the pixels optimized for the CRT display. Some people consider this argument flawed and prefer LCD for retro gaming, but here we have proof that at least for some developers, the graphics are meant for CRT.

_________________
https://twitter.com/bitinkstudios <- Follow me on twitter! Thanks!


Top
 Profile  
 
PostPosted: Wed Nov 14, 2018 6:45 am 
Offline

Joined: Thu May 19, 2005 11:30 am
Posts: 709
The article is interesting, but the terminology used there in hair-raisingly inaccurate, to the point where it took me a moment to understand what it is talking about: "a computer screen and a CRT". What, the computer screen is not a cathode ray tube? What the article calls a CRT is properly called an NTSC television monitor with a composite connection.


Top
 Profile  
 
PostPosted: Wed Nov 14, 2018 7:49 am 
Offline
User avatar

Joined: Thu Aug 13, 2015 4:40 pm
Posts: 326
Location: Rio de Janeiro - Brazil
I would call that almost a typo and can be overlooked. I'm still reading the article and its amount of content is amazing.

_________________
https://twitter.com/bitinkstudios <- Follow me on twitter! Thanks!


Top
 Profile  
 
PostPosted: Wed Nov 14, 2018 9:28 am 
Offline

Joined: Sun Sep 19, 2004 11:12 pm
Posts: 20880
Location: NE Indiana, USA (NTSC)
So my native NES background graphics editor isn't that off the mark.


Top
 Profile  
 
PostPosted: Wed Nov 14, 2018 9:35 am 
Offline

Joined: Thu May 19, 2005 11:30 am
Posts: 709
nesrocks wrote:
I'm still reading the article and its amount of content is amazing.
That is true. I just do not want people perpetuating that mistake ("the graphics are meant for CRT"), because that would lead to confusion.


Top
 Profile  
 
PostPosted: Wed Nov 14, 2018 10:15 am 
Offline

Joined: Tue Nov 23, 2004 9:35 pm
Posts: 713
I think the article author was trying, imprecisely, to make a distinction between a computer monitor and a television. The computers on which the graphics were being digitized were often connected to CRTs with a sharper picture screen and a higher quality connection than the television next to them. But the player was going to use a television to view the graphics, so many programmers would have tried to optimize their graphics to look as good on the lower quality device as possible.

_________________
Nerdly Pleasures - My Vintage Video Game & Computing Blog


Top
 Profile  
 
PostPosted: Wed Nov 14, 2018 10:38 am 
Offline
User avatar

Joined: Thu Aug 13, 2015 4:40 pm
Posts: 326
Location: Rio de Janeiro - Brazil
Yeah, I think the article exactly demonstrates that the artists did indeed make the graphics to be viewed on regular tv sets, often making aesthetic choices that looked good there and not necessarily on the sharp image of a grid paper or high definition display.

Also interesting to see that some of them used a joystick to create the pixel art instead of a keyboard or mouse.

_________________
https://twitter.com/bitinkstudios <- Follow me on twitter! Thanks!


Top
 Profile  
 
PostPosted: Thu Nov 15, 2018 4:50 pm 
Offline

Joined: Wed May 19, 2010 6:12 pm
Posts: 2777
I knew SMB Mario originally had a blue shirt on, before Nelson Mandela's ghost changed it to brown.


Top
 Profile  
 
PostPosted: Mon Nov 19, 2018 2:30 am 
Offline
User avatar

Joined: Thu Sep 15, 2016 6:29 am
Posts: 833
Location: Denmark (PAL)
NewRisingSun wrote:
That is true. I just do not want people perpetuating that mistake ("the graphics are meant for CRT"), because that would lead to confusion.

I'd argue that the issue here is graphics meant specifically for 240p though, not composite video. I've started hearing a lot of people trying to argue for composite video over higher quality connections lately, and it's kind of infuriating.
"meant for CRT" is close enough IMO.


Top
 Profile  
 
PostPosted: Tue Nov 20, 2018 9:56 am 
Offline

Joined: Thu May 19, 2005 11:30 am
Posts: 709
No, the images in that article clearly show both chroma subsampling and cross-color artifacts, both signs of composite video, in addition to scanlines. 240p would only refer to the scanlines aspect. "meant for CRT" is just plain nonsense, because it does not specifically encompass any particular visual feature.


Top
 Profile  
 
PostPosted: Wed Nov 21, 2018 1:43 am 
Offline
User avatar

Joined: Thu Sep 15, 2016 6:29 am
Posts: 833
Location: Denmark (PAL)
240p is a pretty CRT-specific thing, and when people say graphics are designed with a CRT in mind, they are talking about scanlines. More specifically they are talking about the blank ones. This is not an NES-specific thing but also widespread in arcade games, where composite video isn't a thing, or other consoles where better options exist.

If anyone thinks NES graphics look better via composite video than with an RGB mod, they are out of their mind. :P
I respect that people can have wildly different opinions on "clean" upscaled pictures vs. genuine CRT scanlines (and some people even preferring smoothing filters), but you can't really make an argument for this one.


Top
 Profile  
 
PostPosted: Wed Nov 21, 2018 2:01 am 
Offline

Joined: Sun Apr 13, 2008 11:12 am
Posts: 7832
Location: Seattle
The point isn't so much "better" in-as-much-as "graphics were clearly designed with the constraints of NTSC's limited bandwidth". I've seen Blaster Master cited as an example of intentionally using the reduced chrominance bandwidth to achieve a greater apparent gamut of colors without relying on dithering.

On the other hand, are the crosstalk artifacts ever desirable? Probably not: they seem to be too hard to control.


Top
 Profile  
 
PostPosted: Wed Nov 21, 2018 2:52 am 
Offline
User avatar

Joined: Sun Sep 19, 2004 9:28 pm
Posts: 3726
Location: Mountain View, CA
For me, the most notable image in that article is this actually from Wizardry circa 1987 -- that means original Famicom thus RF, not composite (AV Famicom didn't come out until December 1993): https://66.media.tumblr.com/8d2cf7adae9 ... t0_500.jpg

The difference here is pretty astounding. Pause and actually take it in. This is what caught my attention in the original:

The skeleton (in Wizardry 1: undead kobold) has a very long, drawn-out face with a distinct nose and open mouth and lower jaw. Likewise, the skull also appears more elongated as an effect of the shading. Also, the pelvic bone looks more like, well, a pelvic bone. The overall "grit" of the image (artefacting, etc.) causes it to have a substantially more eerie feel; you feel like you're looking at an old skeleton of sorts wielding a sword and buckler, while on the right, you know you're looking at what it supposed to be a skeleton with a sword/buckler but lacks fringe details that gives it a darker, menacing look.

Similar can be said for the orc in the lower shot: the face is substantially more defined (note the eyes, nose, and mouth are very distinguished, almost giving it a glow-like effect); the helmet the orc is wearing has very distinguished horns that look silver or white, and rounded (especially on the left), rather than just little ┕ and ┘ marks. The dagger (we're going off of original D&D here) looks more balanced/level. And he appears to be wearing gauntlets, based entirely on the sheen on his fist.

And if you want to see these in their pure/raw digital form, taken from an emulator and nearest-neighbour scaled up 4x for clarity, see bottom of this post. One thing I can't tell is whether or not the skeleton is wearing a metal shoulder guard, or if it's an intentional effect on the part of the artist (when intending to be viewed via RF) to define the skeleton's shoulder of his wielding arm. I can sit here pondering the intentional palette difference blah blah blah, but I could go either way.

Circling back to RF vs. composite and so on: I urge folks to take a look at this, also from Wizardry (the monster is Werdna, final boss of the game) (source):

Attachment:
DHmnmwxUQAE_UDI.jpg
DHmnmwxUQAE_UDI.jpg [ 439.84 KiB | Viewed 2602 times ]

The captions translated (ブラウン管 is "Braun tube", i.e. CRT):
Code:
CRT + RF connection            CRT + composite connection

Plasma TV + RF connection      Plasma TV + composite connection

Plasma TV + HDMI connection

I should note that the composite connections were done using an original Famicom that had been modded for composite/AV, while the HDMI output was accomplished using a RetroFreak unit (source).

So while in general I think people today like pixel art and pixel drawing on what we've come to use (LCDs etc.; something with high precision), please think about the amount of work good artists now have to put in to achieve what could previously be done through NTSC artefacting or reliance on noise.


Attachments:
003.png
003.png [ 1.76 KiB | Viewed 2602 times ]
001.png
001.png [ 1.64 KiB | Viewed 2602 times ]
Top
 Profile  
 
PostPosted: Wed Nov 21, 2018 4:42 am 
Offline

Joined: Thu May 19, 2005 11:30 am
Posts: 709
Scanlines, chroma subsampling and cross-color artifacts essentially all have the effect of modifying the visual impression of simple 8-bit low resolution graphics:
  • Scanlines have the effect of making the blockiness seem less jarring by adding a kind of texture in the vertical direction.
  • Chroma subsampling has the effect of creating transitory colors that are often welcome for consoles with limited native palettes.
  • Cross-color artifacts are the most difficult to use for effect, as their precise hues are usually semi-random for consoles like the NES. (Some home computers however with a constant phase-dot-clock relationship for every pixel use cross-color artifacts for additional colors, or in the case of the Apple II, exclusively for color generation.)
  • Cross-luma artifacts seem to be just an annoyance.
Emulating scanlines and chroma subsampling therefore should be sufficient to replicate the desirable aspects of contemporaneous displays.

A different aspect that usually gets overlooked, but that I have found to be vital to plausibly reproduce a 15 kHz monitor/TV on modern displays, is gamma-aware upscaling. Basically, to correctly upscale a 240p image to 1080 or whatever vertical resolution, one needs to first do a nearest-neighbor resize *2 to 480 lines, then convert monitor-gamma-relative RGB values to linear light RGB values (e.g. Rlinear =pow(R/255.0, 2.22222)), then upscale using Gaussian, Bilinear or whatever interpolation, then convert to monitor-gamma-relative RGB again(R =pow(Rlinear, 0.45)*255.0). This is usually not done by hardware acceleration, so one has to do it in software. Consider these videos for how this looks on a NES picture. They only use this upscaling procedure, but without chroma subsampling or scanlines.


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 15 posts ] 

All times are UTC - 7 hours


Who is online

Users browsing this forum: No registered users and 2 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Group