by Jin Kim




Nexus One “Resolution” Debate Continues


In Nexus One PenTile Matrix OLED Display (January, 20, 2010), I orig­i­nally argued that because of the dif­fer­ent sub-pixel struc­ture used in the Nexus One the claimed 800×480 pixel for­mat was not accu­rate. A sim­ple cal­cu­la­tion yielded more along the lines of 533×480 because of the 2×2 sub-pixel struc­ture instead of the typ­i­cal 3×1 found in most LCDs. Soon after the blog was pub­lished I was con­tacted by Joel Pollack, Sr. VP of Strategic Sales and Marketing at Nouvoyance and sub­se­quently had a meet­ing involv­ing Joel, Candice Brown Elliott (CEO) and Tony Botzas (Director of System Apps and Engineering) on February 3.

After some con­sid­er­a­tion, I updated the post by stating:

I came out of the 2 1/2-hour meet­ing impressed and con­vinced that the 3.7-inch PenTile Matrix OLED dis­play used in the Nexus One does indeed have a res­o­lu­tion that is equiv­a­lent to a 3.7-inch TFT LCD with a 800×480 pixel for­mat using red, green, blue 3×1 striped sub-pixels.

I argued in Display Showdown Part 1a: Nexus One that mere hard­ware specs can­not deter­mine res­o­lu­tion. The Video Electronics Standards Association (VESA) Display Metrology Committee has a def­i­n­i­tion for res­o­lu­tion in Section 303-7 of Flat Panel Display Measurements Standard Version 2.0: “… the num­ber of alter­nate black and white lines that can be dis­played with a stated min­i­mum con­trast mod­u­la­tion.” I also added that the require­ment that these black and white lines not only be dis­played but viewed by the human visual sys­tem and concluded:

So, let’s move from vision back to hard­ware spec­i­fi­ca­tion: does the Nexus One’s PenTile Matrix OLED dis­play have a res­o­lu­tion equiv­a­lent to a 800xRGBx480? The answer is: yes, it does.

I thought the debate ended there, but Luke Hutchison in Secrets of the Nexus One’s screen: sci­ence, color and hacks on Ars Technica showed the res­o­lu­tion of the Nexus One screen is not as high as claimed. Luke claims that VESA’s res­o­lu­tion mea­sure­ment method­ol­ogy in Section 303-7 in the FPDM Standard Version 2.0 is “insuf­fi­cient to test the display’s true res­o­lu­tion, at least for the way the PenTile dis­play is con­fig­ured on the Nexus One.” Luke then sup­ported his argu­ment by look­ing at diag­o­nal white and black lines, which “exhibit sig­nif­i­cant color band­ing…” He then concluded:

My main point about screen res­o­lu­tion in this arti­cle is that both the Nexus One and the Droid screen were specced at almost the same res­o­lu­tion so they should look the same (I think this is a fair assumption)—and they sim­ply don’t.

In the most recent High Resolution newslet­ter (March 2010) pub­lished by Veritas et Visus, Luke con­tin­ues his asser­tion that the Nexus One’s PenTile Matrix OLED dis­play is not as high as claimed in Resolution and false color images on the Nexus One:

PenTile’s cre­ater, Nouvoyance (for­merly Clairvoyante), claims that the PenTile’s res­o­lu­tion is exactly com­pa­ra­ble to an LCD dis­play of the same num­ber of phys­i­cal pix­els. It is com­pletely impos­si­ble for this claim to be cor­rect, given that there are fewer total sub­pixel ele­ments on the PenTile screen than on an RGB-striped LCD screen of the same num­ber of phys­i­cal pixels.

Luke pointed to effec­tive pixel size, sub­pixel posi­tion­ing, and offered links to full-resolution sam­ple black and white stip­pled images that induce color arti­facts on the Nexus One dis­play. In the same High Resolution newslet­ter Candice H. Brown Elliott and Joel Pollack responded in PenTile OLED res­o­lu­tion to Luke’s orig­i­nal post on Ars Technica:

In Mr. Hutchison’s arti­cle he tries to exam­ine the func­tion­al­ity of the lay­out inde­pen­dently from the algo­rithms. That is not pos­si­ble, as the algo­rithms are an inte­gral part of how PenTile tech­nol­ogy works.

I think the con­fu­sion lies in not accu­rately cre­at­ing the bound­aries of dis­cus­sion. As I have men­tioned already in Display Showdown Part 1a: Nexus One there is a dif­fer­ence between hard­ware spec­i­fi­ca­tions and vision. The PenTile Matrix OLED dis­play was devel­oped and tuned for the way the human visual sys­tem works. Our vision sys­tem is not sim­ply the sum of the bio­log­i­cal capa­bil­i­ties within the eye but works in con­junc­tion with our brain. That sounds sim­ple but it is really quite com­plex. I don’t pre­tend to fully know how the brain works when it comes to vision, but I was intrigued to find out that most of what we see are things that we expect to see. Only in a few instances does our brain react to what we see; most of the time our visual sys­tem is con­firm­ing what the brain is already expect­ing, which points to a strong cor­re­la­tion of vision to our expec­ta­tions, a func­tion of the brain. So in agree­ment with Candice and Joel, I must say the PenTile Matrix OLED dis­play can­not be exam­ined with­out the under­ly­ing algo­rithms that make it work.

Candice and Joel goes to great lengths to pro­vide ample analy­ses, some of which I will hon­estly admit the author does not fully under­stand. In regards to color band­ing on diag­o­nal black and white lines:

… the PenTile OLED is capa­ble of recon­struct­ing black and white line pairs, with at least 50% Michelson Contrast, out to the full Fourier space of the input con­ven­tional image.

Close-up text images are then shown as exam­ples for the reader to exam­ine in response to Luke’s claim that text were fuzzy. From my point of view, read­ing the High Resolution newslet­ter on a 17-inch MacBook Pro at 1920×1200, the close-up text images look quite clear with­out fuzzi­ness. And in response to color arti­facts on stip­pled images Candice and Joel point to the cause:

… this arti­fact is caused by a very slight color error that occurs at the bound­aries between two dif­fer­ent fil­ters in our locally adap­tive SPR [Subpixel Rendering] algo­rithm. Mr. Hutchinson has lit­er­ally found a “hack” in which he forces the SPR engine to switch between fil­ter sets in a highly struc­tured man­ner, forc­ing a neg­li­gi­ble, but real color error, to be repeated over and over again until that error is visible.

It wasn’t stated in the arti­cle but I am sure Nouvoyance, Samsung, Google, and HTC will make sure the next ver­sion of their smart­phone (Nexus Two?) will sport a much smarter dis­play allay­ing fears or doubts that the PenTile Matrix OLED dis­play tech­nol­ogy is up to the task of deliv­er­ing a view­ing expe­ri­ence that meets con­sumer expec­ta­tions for a “800×480″ dis­play. Even then, I’m still a sucker for high-end LCDs… iPhone HD anyone?








Shop at Amazon.com and support DisplayBlog