While I have responded to several blogs that appeared in English it would appear as though there has been misinformation also published in Mandarin. Today’s blog would like to respond to one of the more detailed blogs that appeared a last year (10/26/10) by Shan-Shan as Shan Shan’s Diary – 分类： 科技时代 |字号 订阅 (http://bit.ly/heD5HL)
之前有回應過一些部落格文章, 但他們皆是英文的, 也許在有些中文部落文中會有些許誤導的資訊. 今日文章即將針對過去一年中, 其中一篇較詳盡的部落文作為解說. (10/26/10)
Shan Shan was certainly correct that PenTile technology can have some pattern visibility that can appear grainy when this technology is applied at too low of a resolution. This is only evident when one displays solid fully saturated red color. In general photos, videos and all unsaturated colors are not troubled by such issues with grainy appearance.
在”山山的日志”中, 山山在對某些 PenTile Technology的解說是正確的. 比如說, 在低解析度下, 有些圖像會呈現顆粒感, 但是, 這其實只會發生在全飽和紅色的區域; 在一般的相片, 影片和非飽和的影像是不會有這種顆粒感的狀況存在的.
Most of this blog deals with PenTile OLED that has an RGBG pattern. The most significant error in this blog was the author’s attempt to guess how PenTile OLED would render an image of finely spaced white dots. When he examined such patterns hew saw that we had rendered this with R, G, and B subpixels, but thought that this was some sort of error. He defined what he thought should have been done and highlighted the problem that still existed with doing so. His error was in understanding well enough how subpixel rendering is accomplished.
多數是針對 PenTile OLED 的RGBG 子像素排列. 該文最明顯的錯誤是, 作者嘗試猜測分析 PenTile OLED 如何渲染一個精細的點陣圖. 很清楚地, 作者並未真正的去做, 否則, 他會了解錯誤發生在哪裡. 作者預測一個黑白點陣圖像如下:
intended to look like:
Desired Image of White Dots
經過PenTile 顯示”應該”會如下, 並且要開啟藍色子像素, 他認為這是失敗的呈現.
Shan-Shan say should look like:
Shan-Shan Suggests this Rendering
但實際上, 原圖經過PenTile 顯示會如下圖:
Whereas, in reality it would appear as below.
Note that the luminance of green is 100%, while the luminance of both the red and the blue would be 50%. The human vision system would see these as a well-balanced white chrominance and would see the luminance center just where you would have expected this to be.
在此請注意, 綠色子像素的亮度是100%, 紅和藍則是50%. 經過人眼系統後, 會巧妙平衡成白色的彩度(Chrominance), 並且亮度中心點會被調整至原本期望的位置.
Correct PenTile OLED Rendering for White Dots
The key to understanding PenTile technology is to understand that we do not confine subpixel addressing to just 2 subpixels, but can use up to 10 subpixels to render any given pixel. PenTile algorithms take great care in placing the luminance center where it is intended to be located, but also uses adaptive filters to sharpen edges. There is a bit more flexibility in where the center of chrominance is located, since the human vision system is far less capable of localizing the position of the center of chrominance.
了解PenTile Technology的主要關鍵是, 我們並不侷限只用兩個子像素來顯現原本的一個像素, 相對的, 我們使用, 甚至高達至10個子像素, 來渲染原本的一個像素. PenTile 演算法絕佳巧妙地將亮度中心放在它原本應該被置放的位置, 並且還使用了自適應濾波器來銳化邊緣. 也正因為人眼系統並無法準確地決定彩度的中心位置, 更為彩度中心位置添加了些靈活性.
Later in this article it speculates that PenTile OLED suffers from chromatic aliasing, where there is color error at edges. This is not the case. Surely if one zooms in to see individual pixels there will be one color at any edge, but this is no different than would be the case for legacy RGB stripe displays.
該文之後又推測PenTile OLED 在邊緣會有色差, 然而並不是這樣的. 當然吾人若將圖像放大去看每個像素, 邊緣會看似錯誤, 但是傳統RGB 顯示器也會有同樣的情形.
The last part of this blog quotes blogs by Ray Soniera of DisplayMate who compared the PenTile OLED in
the Nexus One to an iPhone display. Quantization error seen in the Mars sunset and the color wedges was long ago realized by Mr. Soniera to be caused by the used of 5-6-5 color in this phones SW rather than 24-bit color that is built into all PenTile OLED and PenTile LCD displays. Some users finally realized that this quantization only existed in gallery applications, but not elsewhere, and even disappeared when using the touchscreen, clearly removing the PenTile algorithms from any blame.
最後一個部分是, 作者引用Dr.Soniera 比較Nexus One 的PenTile OLED 和iPhone 顯示器. Dr.Soniera 其實早就明白, “火星日出和色彩漸進圖的量化誤差” 是手機的軟體本身使用5-6-5 color, 而非24-bit color 導致的. PenTile OLED 和PenTile LCD 顯示器本身是使用24 bit color. 有些使用者後來甚至發現, 此量化誤差只發生於”影像應用程式”, 並不是每個地方都會發生. 這並不是PenTile 演算法的錯誤.
Mars Sunset from DisplayMateQuantized Color Bands
Quantized Color Wedges Due to 5-6-5 Color
DisplayMate also criticized the Nexus One display as being too saturated since it has a bigger gamut than sRGB. In fact, this is a property of OLED, not PenTile. OLED displays have a gamut that is close to NTSC gamut, which may well be more appropriate than sRGB for displaying video.
DisplayMate 也批評Nexus One呈現太飽和的顏色, 但這是因為相較於sRGB, 有較大的色域. 事實上, 這是OLED的特色, 並不是PenTile. OLED 顯示器有比較大並接近於NTSC的色域. 因此也比sRGB更適合撥放影片.
DisplayMate also concluded that the Nexus One phone and this display used dynamic color and contrast. This again was an incorrect conclusion. The reason he felt that this was being done was that R+G+B luminance did not equal white luminance. The real reason for this was due to loading of the display power supply and had nothing at all to do with dynamic color and contrast.
DisplayMate 並下結論, Nexus One顯示器是使用動態色彩和對比. 再一次聲明這不是一個正確的結論. 他下此結論的原因是, R+G+B的亮度不等於W. 但主要原因其實是顯示器本身電源的負載, 和動態色彩和對比一點關係也沒有.
Please come to this blog to find the real information about PenTile technology. This is considerable misinformation about PenTile technology being blogged today. We are willing to admit our deficiencies where they may exist, but we strive to correct those inaccuracies, no matter what language in which they are being written.
請參見此部落格, 並找出真正有關PenTile Technology的資訊. 因為誤解導致此文如此陳述PenTile Technology. 如果有不足, 我們非常樂意接受指教批評, 但同時我們也會努力指正解釋被詮釋錯誤和曲解的地方, 無論該文是使用何種語言.