Slashdot Mirror

To be fair, this was one of the criticisms Apple laid out against Samsung's original Galaxy S when it was released. That it used a pentile RGBG display instead of RGB like the iPhone, so it's 800x480 display resolution supposedly wasn't really an advantage over the iPhone 3G's 480x320 resolution. Evidently some iPhone owners still remember that, and Apple is now being hoisted by their own petard.

Your eyes are much better at resolving green than they are at red or especially blue. Nearly every method of storing video or photos has taken advantage of this - the old NTSC broadcast TV standard, color film composition, JPEG compression, digital camera sensors, even the latest h.265 video codec. All of them stored red and especially blue at a lower resolution than they do green. So you've been looking at the equivalent of pentile images all your life and never noticed it. Unless you peep at the pixels with a magnifying glass, there's no reduction in image quality from using a lower blue and red subpixel resolution than green. The only exception I've seen is due to a long-lived MPEG bug from the 1990s which still occasionally crops up as striations in blocks of solid color, especially red, which might not have been visible at a higher red resolution.

Unfortunately it was nearly impossible to convince iPhone owners and reviewers who'd drunk Apple's kool-aid of this fact, and Samsung eventually relented and used RGB versions of its OLED displays on their newer phones. So I'll shed no tears that Apple's chickens are now coming home to roost.

Oh right, huge pentile resolution controversy! We all remember that right? Right?

Well let me explain it to you all so I can set up how it came back to an even bigger nothing burger than before!

To be fair, this was one of the criticisms Apple laid out against Samsung's original Galaxy S when it was released. That it used a pentile RGBG display instead of RGB like the iPhone, so it's 800x480 display resolution supposedly wasn't really an advantage over the iPhone 3G's 480x320 resolution. Evidently some iPhone owners still remember that, and Apple is now being hoisted by their own petard.

Your eyes are much better at resolving green than they are at red or especially blue. Nearly every method of storing video or photos has taken advantage of this - the old NTSC broadcast TV standard, color film composition, JPEG compression, digital camera sensors, even the latest h.265 video codec. All of them stored red and especially blue at a lower resolution than they do green. So you've been looking at the equivalent of pentile images all your life and never noticed it. Unless you peep at the pixels with a magnifying glass, there's no reduction in image quality from using a lower blue and red subpixel resolution than green. The only exception I've seen is due to a long-lived MPEG bug from the 1990s which still occasionally crops up as striations in blocks of solid color, especially red, which might not have been visible at a higher red resolution.

Unfortunately it was nearly impossible to convince iPhone owners and reviewers who'd drunk Apple's kool-aid of this fact, and Samsung eventually relented and used RGB versions of its OLED displays on their newer phones. So I'll shed no tears that Apple's chickens are now coming home to roost.

• Some people interpret this to mean each pixel should be one arc-minute or smaller. If that's your standard, then the pixel size has to be 1 / tan (1 arc-minute) = 3438 smaller than the viewing distance. If you're holding your phone 12 inches from your eye, that corresponds to 286 PPI being all that's needed. That's what the Retina displays target - 326 PPI before the iPhone 6+.
• Other people interpret this to mean you need two pixels per arc-minute. After all, to form a line pair with a separation, you need a row of white pixels with a row of black pixels in between. So then the pixel size becomes 1 / (0.5 arc-minutes) = 6875.5 smaller than the viewing distance. If you're holding your phone 12 inches from your eye, that corresponds to 573 PPI.

It should be noted though that this is for luminosity resolution - your ability to resolve details of any color. This is why printers target 600 PPI - because they print in black and white. Within a specific color, your eye's resolution is substantially worse. Especially for blue, and somewhat for red (the density of your red and blue cones is lower than for green cones and rods). This is the basis behind Pentile displays, which cuts the blue and red resolution in half.* All the bad press coverage they've gotten is by ignorant reporters who compare magnified photos of it completely oblivious that magnifying it defeats the whole purpose. This strategy of reducing red and blue resolution has been used since NTSC video transmissions, and is still used today in JPEG and MPEG encoding. You've been seeing pictures on the web and digital videos all this time with reduced red and blue resolution. If you've never notice this before, then you've basically affirmed that Pentile works.

Anyway, it's a moot point on the Samsung displays because they design them to be used in the Gear VR headsets. Those provide a 96 degree wide angle of view, which to fool 20/20 vision would require 5760 pixels for each eye. Which which correspond to a 11520x6480 resolution display on a 5.7 screen, or 2319 PPI. Any lower than that and you can "see this pixels." This is why the 3D graphics and display screen industries still have a lot of room left to grow, even though CPUs have pretty much hit the point where a low-end CPU is "good enough" for most people's needs.

* The more clever displays take further advantage of this difference between color resolution and luminosity resolution, and use something called subpixel rendering. For an RGB stripe, this corresponds to shifting the "pixel" by 1/3 pixel increments. So if you're trying to display a white dot using two pixels, you actually have 4 possible locations. RGB rgb, rGB Rgb, rgB RGb, and rgb RGB. The problem (for displays you can rotate) is that this extra resolution is only along one axis - usually the horizontal. Windows subpixel rendering for fonts (ClearType) basically turns your 1920x1080 display into a 5760x1080 display for fonts. Pentile overcomes this by using a subpixel layout which is symmetric in both the horizontal and vertical axes. So you can use the same subpixel rendering algorithm regardless of whether the display is in landscape or portrait mode. It really is a superior subpixel layout, which has gotten a bad rap because early implementations had too low a PPI and thus the pixels were visible and lines and fonts were "fuzzy".

This is also why these super-high resolution screens aren't as important for Windows as they are for Macs. Subpixel rendering like ClearType shifts the location of letters by up to 1/3 pixel to make them line up with the subpixel grid. Apple knew their computers were used by most page layout professionals which would find this unacceptable. So their font rendering engine (based o

It may be my eyes, the angle at which I use my screen, the brightness and contrast I prefer, or something else, but the background color has always been almost undetectable to me.

Google's background is white - 255, 255, 255 (RGB). The old sponsored links background was 254, 247, 221 (RGB). Our eyes are pretty bad at detecting differences in blue, so it didn't take much of a bad monitor or settings to clip the colors to where you couldn't distinguish the two backgrounds. This mostly happens on poor quality low-gamut screens (laptops!) where the colors are so pale, the manufacturer or user pushes the saturation to try to make them stand out more. So 0 => 0, 100 => 125, 200 => 250, and anything 205-255 is clipped and the same shade since that's the most saturated color the screen is capable of displaying.

The new "Ad" icon is 234, 176, 53 (RGB). If you can't see it, you either need to replace your monitor (and marvel at all the wonderful pictures on the web you've been missing out), or get your eyes checked.

Here's another link with a more detailed handling of what the link I included shows. Including the same pic with red pixels enlarged.

Whether or not pentile sucks depends on the PPI. If the PPI is too low, then you can see the individual sub-pixels and pentile (RGBG) sucks relative to RGB. But if the PPI is high enough, then you cannot see the individual sub-pixels and RGBG is indistinguishable from RGB while using fewer sub-pixels. The reason is a quirk in the human visual system - our eyes' resolution in green is much better than in red and especially blue.

Pretty much every recorded image we see takes advantage of this. Nearly all digital cameras use a Bayer filter (RGBG overlay), so the images they capture have half the red and blue resolution as they do green. Unless you flip certain JPEG options, a JPEG image you create from a pure RGB scan will do the same thing - reduce the red and blue information that's stored relative to green. Same for MPEG and NTSC. Basically, nearly all the recorded images you've ever encountered in your life were brought to you in RGBG. That you never noticed is proof that it's indistinguishable.

It's only displays which were typically RGB, but that was because there were no "pixels" on CRTs, and LCDs typically had low PPI. Once the display's PPI becomes high enough, RGB becomes a waste. When the G sub-pixels in an RGB array are dense enough to surpass the the threshold of visual acuity, the R and B sub-pixels are far too dense and way past that point. That is, you have way more R and B sub-pixels than are actually needed. If you're at this point, then an RGBG display like pentile with the same pixel density (but lower sub-pixel density) will create an image that's indistinguishable from RGB but using fewer sub-pixels.

I picked-up a lot of knowledge from here: http://www.avsforum.com/

And here which describes NTSC in great detail: http://www.videointerchange.com/pal_secam_conversions.htm

And: http://nfggames.com/games/ntsc/colourresx.shtm

http://www.videouniversity.com/articles/video-recording-formats

I have an Atari 400 I still drag out from time to time when I get an itch to play the "definitive" (to me at least!) versions of Pac-Man, Donkey Kong and Defender.

Since when is the Atari 400/800 version of Donkey Kong the definitive version? In this page, compare the "Atari 800 Donkey Kong" sprite (first in the "Early Home Ports"; the second is from Mario Bros.) to the "ColecoVision" sprite (third), and then compare the "ColecoVision" sprite to the first sprite in "Mario 1.0". Nintendo thought the ColecoVision port of DK was so authentic that when it designed the Family Computer/Nintendo Entertainment System, it based the PPU's architecture directly on that of TI's TMS99xx VDP in the ColecoVision and Sega Mark I.

It's done on TV all the time and nobody complains (chrominance is separated from luminance and often transmitted at much lower resolution). As has been pointed out below, your eyes are made up of rods (which see black and white) and cones (which see color), and only a fraction of those cones are devoted to each individual red, green, or blue spectrum. So your color resolution is already significantly lower than your luminance resolution. You can even see photos demonstrating this with a 9x decrease in color resolution (3x in each linear direction). You're most sensitive to green, which is why the Bayer sensors commonly used in digital cameras divide each 4 pixels into GRGB.

It's done on TV all the time and nobody complains (chrominance is separated from luminance and often transmitted at much lower resolution). As has been pointed out below, your eyes are made up of rods (which see black and white) and cones (which see color), and only a fraction of those cones are devoted to each individual red, green, or blue spectrum. So your color resolution is already significantly lower than your luminance resolution. You can even see photos demonstrating this with a 9x decrease in color resolution (3x in each linear direction). You're most sensitive to green, which is why the Bayer sensors commonly used in digital cameras divide each 4 pixels into GRGB.

It's done on TV all the time and nobody complains (chrominance is separated from luminance and often transmitted at much lower resolution). As has been pointed out below, your eyes are made up of rods (which see black and white) and cones (which see color), and only a fraction of those cones are devoted to each individual red, green, or blue spectrum. So your color resolution is already significantly lower than your luminance resolution. You can even see photos demonstrating this with a 9x decrease in color resolution (3x in each linear direction). You're most sensitive to green, which is why the Bayer sensors commonly used in digital cameras divide each 4 pixels into GRGB.

I wrote this book, and I have to agree with most parts of this review. There's a good story behind it, but here's the short version:

Rotovision approached me about writing it based on some stuff they liked on my website. It was originally intended to be part of a 3-book series discussing game and character art. I don't know for sure what the other two were to be about, but they were canceled and only this one was given a green light.

Then the US publisher, Focal Press, wanted it to be about mobiles in what I can only assume was an attempt to be more trendy and hip. This compromised the book rather seriously, as from a design point of view there are no real graphics challenges when developing for a mobile over, say, a GameBoy or SNES.

The book was rushed. I was given four months to do it all, and for the first three I didn't hear a single word from the editor about the content. None of the promised guidance was forthcoming, and after a month of no contact I believed they had dropped the idea so I stopped working on it.

Suddenly, a month before the due date, they popped back up asking for the 50% complete text and 100% of the images. Say what? So I wrote 20,000 words in a week, enslaved some writer friends to help out, and started cropping out sprite images like a man posessed.

It worked too, I got it done and wrote some amazing stuff. Unfortunately the publisher decided time was too tight to actually use it, so several articles that were to be cut 'cause they sucked made it in, several good ones that were to replace them never did, images that were delivered for specific articles were ignored, random images were thrown in all over the place, and they didn't seem to bother labelling any images in the last quarter of the book.

It's not my fault! I've got a few copies of the book I can't even bring myself to read 'cause it was such a mess at the end.

I take full responsibility for many errors, and for the quality of the text. I totally screwed up by dropping the project for months at a time, but I really don't think I'd be the only person to make the assumption when the editor doesn't reply to your emails for six weeks.

I've re-written a few of the articles to have better image labels, added a few things I missed, and posted them online. For your reading enjoy, here they are:

Sonic Sprite History
Mario Sprite History (This one's been on slashdot before)
Castlevania Sprite History

And here's a rather lengthy errata.

I wrote this book, and I have to agree with most parts of this review. There's a good story behind it, but here's the short version:

Rotovision approached me about writing it based on some stuff they liked on my website. It was originally intended to be part of a 3-book series discussing game and character art. I don't know for sure what the other two were to be about, but they were canceled and only this one was given a green light.

Then the US publisher, Focal Press, wanted it to be about mobiles in what I can only assume was an attempt to be more trendy and hip. This compromised the book rather seriously, as from a design point of view there are no real graphics challenges when developing for a mobile over, say, a GameBoy or SNES.

The book was rushed. I was given four months to do it all, and for the first three I didn't hear a single word from the editor about the content. None of the promised guidance was forthcoming, and after a month of no contact I believed they had dropped the idea so I stopped working on it.

Suddenly, a month before the due date, they popped back up asking for the 50% complete text and 100% of the images. Say what? So I wrote 20,000 words in a week, enslaved some writer friends to help out, and started cropping out sprite images like a man posessed.

It worked too, I got it done and wrote some amazing stuff. Unfortunately the publisher decided time was too tight to actually use it, so several articles that were to be cut 'cause they sucked made it in, several good ones that were to replace them never did, images that were delivered for specific articles were ignored, random images were thrown in all over the place, and they didn't seem to bother labelling any images in the last quarter of the book.

It's not my fault! I've got a few copies of the book I can't even bring myself to read 'cause it was such a mess at the end.

I take full responsibility for many errors, and for the quality of the text. I totally screwed up by dropping the project for months at a time, but I really don't think I'd be the only person to make the assumption when the editor doesn't reply to your emails for six weeks.

I've re-written a few of the articles to have better image labels, added a few things I missed, and posted them online. For your reading enjoy, here they are:

Sonic Sprite History
Mario Sprite History (This one's been on slashdot before)
Castlevania Sprite History

And here's a rather lengthy errata.

I wrote this book, and I have to agree with most parts of this review. There's a good story behind it, but here's the short version:

Rotovision approached me about writing it based on some stuff they liked on my website. It was originally intended to be part of a 3-book series discussing game and character art. I don't know for sure what the other two were to be about, but they were canceled and only this one was given a green light.

Then the US publisher, Focal Press, wanted it to be about mobiles in what I can only assume was an attempt to be more trendy and hip. This compromised the book rather seriously, as from a design point of view there are no real graphics challenges when developing for a mobile over, say, a GameBoy or SNES.

The book was rushed. I was given four months to do it all, and for the first three I didn't hear a single word from the editor about the content. None of the promised guidance was forthcoming, and after a month of no contact I believed they had dropped the idea so I stopped working on it.

Suddenly, a month before the due date, they popped back up asking for the 50% complete text and 100% of the images. Say what? So I wrote 20,000 words in a week, enslaved some writer friends to help out, and started cropping out sprite images like a man posessed.

It worked too, I got it done and wrote some amazing stuff. Unfortunately the publisher decided time was too tight to actually use it, so several articles that were to be cut 'cause they sucked made it in, several good ones that were to replace them never did, images that were delivered for specific articles were ignored, random images were thrown in all over the place, and they didn't seem to bother labelling any images in the last quarter of the book.

It's not my fault! I've got a few copies of the book I can't even bring myself to read 'cause it was such a mess at the end.

I take full responsibility for many errors, and for the quality of the text. I totally screwed up by dropping the project for months at a time, but I really don't think I'd be the only person to make the assumption when the editor doesn't reply to your emails for six weeks.

I've re-written a few of the articles to have better image labels, added a few things I missed, and posted them online. For your reading enjoy, here they are:

Sonic Sprite History
Mario Sprite History (This one's been on slashdot before)
Castlevania Sprite History

And here's a rather lengthy errata.

I wrote this book, and I have to agree with most parts of this review. There's a good story behind it, but here's the short version:

Rotovision approached me about writing it based on some stuff they liked on my website. It was originally intended to be part of a 3-book series discussing game and character art. I don't know for sure what the other two were to be about, but they were canceled and only this one was given a green light.

Then the US publisher, Focal Press, wanted it to be about mobiles in what I can only assume was an attempt to be more trendy and hip. This compromised the book rather seriously, as from a design point of view there are no real graphics challenges when developing for a mobile over, say, a GameBoy or SNES.

The book was rushed. I was given four months to do it all, and for the first three I didn't hear a single word from the editor about the content. None of the promised guidance was forthcoming, and after a month of no contact I believed they had dropped the idea so I stopped working on it.

Suddenly, a month before the due date, they popped back up asking for the 50% complete text and 100% of the images. Say what? So I wrote 20,000 words in a week, enslaved some writer friends to help out, and started cropping out sprite images like a man posessed.

It worked too, I got it done and wrote some amazing stuff. Unfortunately the publisher decided time was too tight to actually use it, so several articles that were to be cut 'cause they sucked made it in, several good ones that were to replace them never did, images that were delivered for specific articles were ignored, random images were thrown in all over the place, and they didn't seem to bother labelling any images in the last quarter of the book.

It's not my fault! I've got a few copies of the book I can't even bring myself to read 'cause it was such a mess at the end.

I take full responsibility for many errors, and for the quality of the text. I totally screwed up by dropping the project for months at a time, but I really don't think I'd be the only person to make the assumption when the editor doesn't reply to your emails for six weeks.

I've re-written a few of the articles to have better image labels, added a few things I missed, and posted them online. For your reading enjoy, here they are:

Sonic Sprite History
Mario Sprite History (This one's been on slashdot before)
Castlevania Sprite History

And here's a rather lengthy errata.