# 登入區塊
帳號:

密碼:

自動登入 (請勿在公用電腦上登入時選擇「自動登入」選項,以防個人資料外洩。)

遺失密碼嗎?

何不馬上註冊!
# 主選單區塊
# 誰在線上區塊
 會員: 0
 訪客: 144

# 搜尋區塊
center   討論區主頁 » SLR綜合及副廠鏡版 » 您測了沒? CCD壞點研究無發表權

樹狀顯示 | 新的在前 前一個主題 | 下一個主題 | 頁尾
發表者 討論內容
jerloong
發表時間: 2004-12-02 13:19
版主
註冊日: 2004-10-18
來自:
發表數: 2768
[408] 您測了沒? CCD壞點研究
到底CCD壞點測試有否標準呢?
在網路上可以下載DeadPixelTest.exe
http://down.veryol.com/SoftView/SoftView_131.html
也可以參考以下網址解說測試
http://dc.itnow.com.cn/articles/skill/20040825/2004082511560433395-1.html
http://digi.blueidea.com/dc/experience/2004/882.asp

測試時最好注意以下幾點
1) 設定存取格式為TIFF最好
2) 鏡頭蓋蓋上
3) 關上觀景窗 不要讓光線由觀景窗進入
jerloong
發表時間: 2004-12-02 18:57
版主
註冊日: 2004-10-18
來自:
發表數: 2768
[409] Re: 您測了沒? CCD壞點研究
小弟測試E-20

一秒以內 沒壞點(Dead pixel); 沒熱雜訊點(Hot pixel)
二秒 沒壞點(Dead pixel); 有4點熱雜訊點(Hot pixel)
四秒 沒壞點(Dead pixel); 有24點熱雜訊點(Hot pixel)
八秒 沒壞點(Dead pixel); 有75點熱雜訊點(Hot pixel)
十五秒 沒壞點(Dead pixel); 有88點熱雜訊點(Hot pixel)
三十秒 沒壞點(Dead pixel); 有170點熱雜訊點(Hot pixel)
六十秒 沒壞點(Dead pixel); 有661點熱雜訊點(Hot pixel)
jerloong
發表時間: 2004-12-02 19:33
版主
註冊日: 2004-10-18
來自:
發表數: 2768
[411] Re: 您測了沒? CCD壞點研究
進一步分析
三次測試 兩秒鐘 熱雜訊出現座標是一致的
且重複出現在4" 8" 15" 30"
但並沒有在60"找到相同座標

推論較早出現熱雜訊座標 往後出現熱雜訊可能很高
相同曝光時間測試 如果出現較多熱雜訊 應該----
ITO
發表時間: 2004-12-02 21:16
網站管理員
註冊日: 2004-09-15
來自: 日本
發表數: 30762
[412] Re: 您測了沒? CCD壞點研究
引文:

熱雜訊出現座標是一致的


所以熱雜訊也屬CCD,CMOS的良率問題?
jerloong
發表時間: 2004-12-02 22:06
版主
註冊日: 2004-10-18
來自:
發表數: 2768
[413] Re: 您測了沒? CCD壞點研究
>熱雜訊也屬CCD,CMOS的良率問題?
有此可能吧!

許多人的認知:
熱雜訊產生 在不同照片會出現在不同位置?
若出現在同一位置應為壞點?
但小弟測試熱雜訊座標位置重複性高

根據小弟測試結果
1. E-20 手動曝光 快門設定最長為60秒(ISO80)
而60秒雜訊在二次測試中 座標一致性高
卻與2" 4" 8: 15" 30" 一致性較低
推論 60秒接近臨界值 溫度有大改變 故設計上將可設定最長時間 定為此 ( B快門不算)

2. E-300 P & A 模式最長快門會因ISO值而變 因該與熱雜訊有關
ISO 100 ISO 200 ISO 400 ISO 800 ISO 1600
P 1 sec 1/2 sec 1/4 sec 1/8 sec 1/15 sec
A 50 sec 25 sec 13 sec 6 sec 3.2 sec
S 30 sec
M 30 sec + Bulb
jerloong
發表時間: 2004-12-11 10:43
版主
註冊日: 2004-10-18
來自:
發表數: 2768
[538] Re: 您測了沒? CCD壞點研究
小弟剛剛測試E-300
ISO 100 TIFF檔 手動曝光
1sec 4個熱雜訊
2sec 4個熱雜訊
4sec 8個熱雜訊
8sec 10個熱雜訊
15sec 14個熱雜訊
30sec 100個熱雜訊

與先前測試E-20一樣
一秒出現的4個熱雜訊在2" 4" 8" 15" 重複出現
但在30"測試中 找不到相同對應點
而30"熱雜訊暴增 這可能也是E-300訂定最長時間快門的依據吧
jerloong
發表時間: 2005-07-30 22:13
版主
註冊日: 2004-10-18
來自:
發表數: 2768
[3972] Re: 您測了沒? CCD壞點研究
最近針對Olympus pixel mapping 功能 再想了 一下
到底期Mapping標準為何?
用DeadPixelTest.exe測試
結果原本測得的熱雜訊Hot pixel 使用Olympus pixel mapping功能會把測得所謂的Hot pixel蓋掉

在軟體裡內定是閥值是60(但是拍全黑畫面在 LCD放大 還是肉眼可見熱雜訊) (且如同前幾次測試一樣 位置是固定)

使用Olympus pixel mapping後 約降低閥值至20以下 才會可能出現熱雜訊 故推估Olympus pixel mapping 是把20以上熱雜訊蓋掉

jerloong
發表時間: 2007-01-12 20:47
版主
註冊日: 2004-10-18
來自:
發表數: 2768
[21733] Re: 您測了沒? CCD壞點研究
J. Andrzej Wrotniak 對於雜訊的解說
Noise in Digital Cameras

令人深刻的註解
>Too much noise about noise
>Still, in my school any noise reduction, how smart it may be, is a necessary evil.
rorolala
發表時間: 2007-01-13 21:41
Quite a regular
註冊日: 2006-03-07
來自:
發表數: 58
[21767] Re: 您測了沒? CCD壞點研究
我有一個來自奧林巴斯的關於這個的說明,我覺得還不賴!給大家做個參考!

Olympus C-4040 Zoom, APM and an E-10 Upgrade.

With today’s higher end cameras, 3 megapixel and higher, there is some discussion about pixel issues. These include dead pixels, stuck pixels and hot pixels as well as noise. For most users this is not a problem and the condition will go unnoticed in their prints or on screen viewing. Digital cameras from all manufacturers can develop these CCD/CMOS errors over time, but very few will exhibit dead or stuck pixels when new. Also all CCD and CMOS sensors suffer from these conditions from the lowest to the highest pixel counts. To counter this Olympus puts all their digital cameras through several quality checks and maps the sensors before shipping and then re-checks for quality including pixel issues before shipping to their dealers. As a result we have a very low occurrence of the dead, or stuck pixel phenomena, well below .1%. What are “Stuck Pixels”, “Dead Pixels”, “Hot Pixels” and “Noise”? How do you deal with these issues, and what additional solutions does Olympus offer.

Dead Pixels: a pixel that reads zero or is always off on all exposures. This state produces a black pixel in the final image.
Stuck Pixels: a pixel that always reads high or is always on to maximum on all exposures this produces a white pixel in the final image.
Hot Pixel: a pixel that reads high on longer exposures, and can produce white, red (orange) or green (yellow green) pixels in longer exposures, the longer the exposure the more visible the hot pixels.
Noise: is the state with fixed pattern and random patterns created as the CCD heats up, in extended use or longer exposures. If you have a long enough exposure you can find hot pixels and noise in any digital photo. Along with heat higher equivalent ISO’s result in increased noise in the digital photographs.

Common misrepresentations of dead and stuck pixels is to set the camera to a low ISO (100) put the lens cap on the camera or cover the lens, this results in a long exposure. When you look at the image you will see miss represented pixels some red or greenish in color some white. This is a normal state and should not be confused with dead or stuck pixels. It is noise and the red and greenish coloration is called Christmas tree artifacting in slang terms. The color is caused by the RGB and CMY colored filters of the CCD. “Noise Reduction” technology found in the C-4040 Zoom will reduce this effect. A truly dead pixel is dead at all times and would not show up in this scenarrio, since it is black. To test for a dead pixel or stuck pixel, simply shoot a white and black card in normal light. If the pixel is dead it will show as black against the white, and if it is stuck it will show as white against the black.

To fix dead or stuck pixels the standard procedure is to send the camera in and have the CCD re-mapped by the factory service center. Some manufacturers will do this during the warranty time as warranty service, others are charging for the service. Olympus America has chosen to re-map the CCDs with dead or stuck pixels as warranty service1. Over time all CCDs will produce stuck and dead pixels, meaning increased down time or retouching of the photographs. This is true of all CCD/CMOS cameras. To combat this the C-4040 Zoom includes a new algorithm to counter this phenomena, “Automatic Pixel Mapping” (APM). APM was developed to keep the camera operational over an extended period of time reducing the need to send the camera in for service by Olympus. APM is simple to use, all you have to do is go to the camera menu when the C-4040 Zoom is first turned on, and set the menu to pixel mapping, then go to start, and hit the OK button. The CCD re-mapping begins, taking about 10 seconds to complete. The APM only needs to be done about once a year. This functionality should bring years of easy extra use from the C-4040 Zoom. This feature cannot be added to previous C series cameras.

For the E-10, APM can be added as an upgrade when the camera is sent into the Olympus Service Center. The upgrade requires a physical re-mapping of the CCD and new firmware to be installed. After this upgrade, when the batteries are removed and replaced with new batteries the E-10 will automatically start the APM and re-map the CCD. This function takes 20 seconds or longer on the E-10. The E-10 will re-map the CCD each time the batteries are changed, when the pixel-mapping feature is turned on. For best results and longest life the B30-LBS grip system is recommended.

If a customer wants to have this function added and the camera has no problems (which will be the case with most E-10s), Olympus will do the re-mapping, checking the camera and add the new firmware for pixel mapping, charging only shipping and handling for this service. ($19.95 US2) Olympus will offer this additional function by the factory service center, but the E-10 must be sent in for this service. The addition of the APM function can only be done on [FCC labeled] Olympus America distributed E-10’s. Olympus does not repair or service gray market cameras.

1 Olympus Warranty is Limited to the area of intended distribution, which is for North and South America, and FCC approved units, and is one year in length form Olympus America Inc.
2 Shipping cost only, if additional repairs are required or requested the charge may vary, an estimate will be sent before any additional service is preformed.
jerloong
發表時間: 2007-01-14 01:26
版主
註冊日: 2004-10-18
來自:
發表數: 2768
[21768] Re: 您測了沒? CCD壞點研究
感謝rorolala提供的訊息
對於Olympus數位相機內建Pixel Mapping功能詳細解說
rorolala
發表時間: 2007-01-14 12:54
Quite a regular
註冊日: 2006-03-07
來自:
發表數: 58
[21776] Re: 您測了沒? CCD壞點研究
版主您言重了!長期潛水,也該回饋大家!

雖然我玩數位相機沒有很久,但是這個問題常常都被大家提起來討

論,至少比較有個準則供大家參考!
樹狀顯示 | 新的在前 前一個主題 | 下一個主題 | 頁首

無發表權
 

回首頁 | 新聞區 | 討論區 | 電子相簿 | 投票 | 網站連結 | 夥伴網站 | 文章分類