www.astrocafe.hk

For mono CCD imagaing, people would use narrow band filters. I notice that the Astrodon has 2 standards, namely 3nm and 5nm. As a matter of fact, most astrophotographers use 5nm filters. Part of the reasons I believed is that the price. Both 3nm & 5nm filters are pricy. The prices of 3nm filters are really scary! I am sure that 3nm filters give higher contrast and detail images. Even under full moon situation, I guess 3nm filters are able to do the job! I am not sure about the 5nm filters are able to eliminate the sky gradient under full moon situation. Anyway, the point is that whether it is cost effective under certain situations. I found that for a reasonable 5nm narrow band Ha image, usually takes at least 10 mins. I guess for 3nm narrow band Ha may at take at least 15 mins! This is certainly a higher standard for auto guiding. So the question is that whether 3nm narrow band imaging is beneficially for Guerrilla style (打游擊) in Hong Kong. Any experience sharing?

willis 寫:I found that for a reasonable 5nm narrow band Ha image, usually takes at least 10 mins. I guess for 3nm narrow band Ha may at take at least 15 mins!

我對你的論據有點保留。

不管3或5nm，目標波段的通光率都接近100%，那同樣曝光例如10分鐘，哪款的訊噪比及訊息強度會高些? 當然最暗/深能拍到什麼，又是另一會事。

我在西藏的天文台正在用：Astrodon 5nm Ha、3nm OIII、3nm SII，因為太窄的Ha可能有反效果。

viewtopic.php?f=29&t=2884#p19196

目前其濾鏡盤還有一個閒置位置，究竟會放片3nm Ha或V測光濾鏡，隨時後者居多。

以西藏的天空計，在滿月時，SII及Ha仍要在40度以外拍攝才可避免Sky Gradient，OIII還要遠一點，其中一個原因是我的16803像場在目前組合下有3X3度以上，換作是8300會好些。在香港，這情況肯定要壞一些。

的確，窄頻濾鏡在香港很有發揮空間。我在香港拍得不多，3年前的作品現在重看也慘不忍暏，但希望給你一個概念以我當時的水平及環境可以拍到什麼，你一定可以拍得遠好得多：

viewtopic.php?f=17&t=621#p6177

viewtopic.php?f=17&t=302

回想過去三年，我個人的進步也算明顯，從不懂得分TIFF或FITS，不懂得用平場暗場，到從零開始在西藏覓地建台到現在已運作兩年，拍了兩千小時以上過百個目標，希望下月到德國學習PixInsight後可以正式處理手上的全部圖檔。

The crucial difference is not temporary setup or fixed observatory, but whether you suffer from severe light pullution.

For Tibert, Tai Chung Hill top, Yin Na Hill, East Dam, you would be better off with 5nm (even 7-8-13nm would do quite well) but in more urban areas like HK you should use 3-4.5nm.

Autoguiding is a must in either case but if the mount is a bit weak you may choose to use a higher QE (> 50%) sensor and shorter F-ratio (F2.8 to F5) to reduce single frame exposure time. 5 min is too short in Ha to gather sufficient S/N over read noise but 8-10 min would be minimum duration. A 15-20 min would be quite enough in Hong Kong even
in light pollution new towns but cloud coverage might spoil any sub-frame.

With proper polar alignment, a reasonable mount (eg EM11) and even commercially available mounts such EQ5 (Pro) would do good job of autoguiding with PHD Guiding to give
you round stars in 10-15 min sub-exposure.

I find 10 min sub exposure a good balance between S/N and risk of clouds in HK even for L frames.

15 min sub-frame for 4.5nm Ha would be more than enough in Hong Kong if the surrounding light pollution is not severe (5.0 - 5.5 visual magnitude in my observatory).

Example of
NGC281 http://www.raymondtse-lawyer.com/astro/ ... 14l-0C.htm

Bubble Nebula
http://www.raymondtse-lawyer.com/astro/ ... -0C-ha.htm

Witch Broom (Western Veil)
http://www.raymondtse-lawyer.com/astro/ ... -0C-ha.htm

Very good input from Raymond.

討論區的好處是匯聚不同角度更多意見。

補充數點：

- 無論是什麼濾鏡，我是用平宜的方法自製離軸導星器及在濾鏡後導星的(Guide Behind filter)的，就算是3nm濾鏡，不管什麼地平仰角，什麼天區，STi 3秒曝光下，除非天上有雲，2年下來在西藏的天空從未試過找不到導星，這個自製離軸導星器比市場上售賣的型號更不可能有剛性問題(flexure)，最近我其至簡化了調焦部份，乾脆用人手伸縮調焦，以卡尺定位：
viewtopic.php?f=20&t=1664

- 目前所見，我的組合出現導星誤差的只有兩種情況，那些長時間如單張30分鐘的，拍攝途中有雲掩蓋導星，或被導星是雙星或近距離內有另一顆相若亮度的星，否則我最長試過單張60分鐘曝光，100%出圖星點仍是Pinpoint的。前者是天氣問題無話可說，但後者出事機率甚低，至少不會出現不少愛好者打游擊時，拍半晚隨時因為機件設定或不知名問題，毀了百份之三五十的情況。另外我也會避開那些千多至兩千多海拔的地區，Willis那年和我到台灣領教過了，在我們這緯度，那正是"霧出沒地帶"，所以為了提升出動效益，貴精不貴多，若是打游擊，我會首選台灣高山而非廣東，當然在台灣我也不會留在清境。

- 若是軸外導星，又是單張曝光有二三十分鐘以上的，可能單靠極軸鏡校準赤道儀便會有些問題，那時最好做一下漂移法。

- 在香港拍Ha或SII窄頻有很大發揮空間，不少人初拍Ha時都有些喜悅，但OIII便吃力得多了，所以黑暗天空還是沒有替代品的。另外有些天體，因為本身實在太暗，以我的理解，就算是Ha，在香港仍是十分十分吃力，至少我見不到有很類近深度的香港土產照片：

Simeis 147 / Sharpless 2-240：
viewtopic.php?f=17&t=3319#p21609

Sharpless 303、308：
viewtopic.php?f=17&t=3460#p22268

因為欠缺後期處理技術，這些圖稍後會有更好版本出來。

以目前的情況說明，拍Ha在香港很有發揮空間，對於一些明亮天體如北美星雲、馬頭等，只要曝光略長，隨時拍出來的照片，明顯拉近跟西藏天空的差距，但一些真正暗，中光圈下仍要10小時曝光才是起步的天體，才是天空質量的更高指標。

再補充一點，若很仔細的對照比評，就算是一些明亮目標如馬頭等，合歡山及西藏拍出來的也有分別，剛台灣墾丁星空派對時WYK是這樣跟我說的，說那些星雲暗黑部份會呈現真正的黑云云.... 我沒深究他指馬頭天區的哪些部份，但他是高水平攝星者，會有一定道理：

viewtopic.php?f=13&t=3610

這帖移至相機及配件版，比較貼題。

社長 寫:

willis 寫:I found that for a reasonable 5nm narrow band Ha image, usually takes at least 10 mins. I guess for 3nm narrow band Ha may at take at least 15 mins!

我對你的論據有點保留。

不管3或5nm，目標波段的通光率都接近100%，那同樣曝光例如10分鐘，哪款的訊噪比及訊息強度會高些? 當然最暗/深能拍到什麼，又是另一會事。

What I mean here is that when we look at the histogram, usually we will judge the exposure is enough if the histogram is shifted to the right for about 25% in DSLR imaging. For example, ISO 1600, F4, without any light pollution filter, it may take 30s for histogram to shift about 25%. Perhaps with P2, it may take about 2 mins for the histogram to shift about 25%. So I guess the narrower the band width, the more time it takes to shift the histogram to shift the same amount. Of course, I have no CCD imaging experience, I don't know whether CCD imaging has DSLR like histogram after each shot.

社長 寫:- 在香港拍Ha或SII窄頻有很大發揮空間，不少人初拍Ha時都有些喜悅，但OIII便吃力得多了，所以黑暗天空還是沒有替代品的。另外有些天體，因為本身實在太暗，以我的理解，就算是Ha，在香港仍是十分十分吃力，至少我見不到有很類近深度的香港土產照片：

What do u mean by "但OIII便吃力得多了"?

社長 寫:這帖移至相機及配件版，比較貼題。

According to ur chinese description "Equipment (Eyepieces & Accessories) - 各類目鏡、濾鏡、平場鏡等光學配件 ". "3nm Vs 5nm ...." should be placed in ""Equipment (Eyepieces & Accessories)..."

Raymond TSE 寫:For Tibert, Tai Chung Hill top, Yin Na Hill, East Dam, you would be better off with 5nm (even 7-8-13nm would do quite well) but in more urban areas like HK you should use 3-4.5nm.

Thanks Raymond's input. So in my case, I should buy 3nm if I really do CCD imaging in HK. Burn $$ to solve the problem.

willis 寫:What do u mean by "但OIII便吃力得多了"?

OIII 的波段，正是光害的重災區，而對大多數深空天體來說，OIII本身的訊息強度又比Ha弱，所以你會看到不少愛好者在香港拍到很不錯的Ha照片，但OIII多數都強差人意，除非移到郊外天空質量較高的地方。

willis 寫:

社長 寫:這帖移至相機及配件版，比較貼題。

According to ur chinese description "Equipment (Eyepieces & Accessories) - 各類目鏡、濾鏡、平場鏡等光學配件 ". "3nm Vs 5nm ...." should be placed in ""Equipment (Eyepieces & Accessories)..."

無喎，你再睇真D?

willis 寫:

社長 寫:

willis 寫:I found that for a reasonable 5nm narrow band Ha image, usually takes at least 10 mins. I guess for 3nm narrow band Ha may at take at least 15 mins!

我對你的論據有點保留。

不管3或5nm，目標波段的通光率都接近100%，那同樣曝光例如10分鐘，哪款的訊噪比及訊息強度會高些? 當然最暗/深能拍到什麼，又是另一會事。

What I mean here is that when we look at the histogram, usually we will judge the exposure is enough if the histogram is shifted to the right for about 25% in DSLR imaging. For example, ISO 1600, F4, without any light pollution filter, it may take 30s for histogram to shift about 25%. Perhaps with P2, it may take about 2 mins for the histogram to shift about 25%. So I guess the narrower the band width, the more time it takes to shift the histogram to shift the same amount. Of course, I have no CCD imaging experience, I don't know whether CCD imaging has DSLR like histogram after each shot.

我在西藏拍窄頻，通常單張曝光30分鐘，就算在月夜下拍攝，減Dark後天空背景亮度只有100多而已，相對於65536峰值，還有很遠距離。我想以Ha本身的性質，是很難看到一個天體本身的Histogram推至右邊去(除了M8、M42這些吧)，那訊噪比可能更影響Stretch出來的結果。

而我的SII是3nm、OIII也是3nm，但Ha用5nm不用3nm，主怕有反效果，Cloudynights上有些相關討論，例如：

http://www.cloudynights.com/ubbthreads/ ... in/5640617

Quote :

As mentioned, 3nm H alpha cuts out NII which is common in many planetary and some other emission nebula. This changes the look from what you get with 5nm that picks up both. Nothing wrong with just the H alpha without NII just that it is "different". If LP is an issue then 3nm will help. Otherwise I'd go 5nm. The slight difference in star size can easily be controlled in processing with a bit of deconvolution or other means.

I have an issue you likely don't. I sometimes use Ha for very distant galaxies strong in H alpha. Even 5nm can be too narrow with redshift moving the H alpha line out of the passband! Happened to me even with my 6nm filter. Since my skies are very dark I'm thinking of replacing it with something wider for that reason. Obviously not an issue for typical narrow band imaging.

http://www.cloudynights.com/ubbthreads/ ... in/5760045

Quote :
Except for the HII regions, galaxies are very dim in HA light so I doubt going to 5nm will be of much help. What it does do is widen the bandwidth so you can pick up the HA in more highly redshifted galaxies. While nearby ones like M51 are fine in either 3 or 5nm as redshift increases the HA frequency moves out of the passband.

當然，進入數碼時代，十多年下來，我發現不少愛好者初踏足窄頻攝影，發現Ha的效果跟以前截然不同，北美星雲、馬頭、M42、玫瑰幾乎是必到之處，看到那種對比度層次感的提升，誤以為自己的攝影水平大躍進。

正如拍行星月面，從菲林到webcam，然後到天文攝像頭，又或者從菲林到數碼單反，然後到冷凍CCD，每一次儀器更新換代，都會湧現一批新的"大師"，但下一輪的新技術新器材推出市場，原有一批"大師"又被淹沒了。

若你滿足於那種層面，那就不用太深究了。

社長 寫:

willis 寫:

社長 寫:這帖移至相機及配件版，比較貼題。

According to ur chinese description "Equipment (Eyepieces & Accessories) - 各類目鏡、濾鏡、平場鏡等光學配件 ". "3nm Vs 5nm ...." should be placed in ""Equipment (Eyepieces & Accessories)..."

無喎，你再睇真D?

彈弓手！

社長 寫:

willis 寫:

社長 寫:

willis 寫:I found that for a reasonable 5nm narrow band Ha image, usually takes at least 10 mins. I guess for 3nm narrow band Ha may at take at least 15 mins!

我對你的論據有點保留。

不管3或5nm，目標波段的通光率都接近100%，那同樣曝光例如10分鐘，哪款的訊噪比及訊息強度會高些? 當然最暗/深能拍到什麼，又是另一會事。

What I mean here is that when we look at the histogram, usually we will judge the exposure is enough if the histogram is shifted to the right for about 25% in DSLR imaging. For example, ISO 1600, F4, without any light pollution filter, it may take 30s for histogram to shift about 25%. Perhaps with P2, it may take about 2 mins for the histogram to shift about 25%. So I guess the narrower the band width, the more time it takes to shift the histogram to shift the same amount. Of course, I have no CCD imaging experience, I don't know whether CCD imaging has DSLR like histogram after each shot.

我在西藏拍窄頻，通常單張曝光30分鐘，就算在月夜下拍攝，減Dark後天空背景亮度只有100多而已，相對於65536峰值，還有很遠距離。我想以Ha本身的性質，是很難看到一個天體本身的Histogram推至右邊去(除了M8、M42這些吧)，那訊噪比可能更影響Stretch出來的結果。

而我的SII是3nm、OIII也是3nm，但Ha用5nm不用3nm，主怕有反效果，Cloudynights上有些相關討論，例如：

http://www.cloudynights.com/ubbthreads/ ... in/5640617

Quote :

As mentioned, 3nm H alpha cuts out NII which is common in many planetary and some other emission nebula. This changes the look from what you get with 5nm that picks up both. Nothing wrong with just the H alpha without NII just that it is "different". If LP is an issue then 3nm will help. Otherwise I'd go 5nm. The slight difference in star size can easily be controlled in processing with a bit of deconvolution or other means.

I have an issue you likely don't. I sometimes use Ha for very distant galaxies strong in H alpha. Even 5nm can be too narrow with redshift moving the H alpha line out of the passband! Happened to me even with my 6nm filter. Since my skies are very dark I'm thinking of replacing it with something wider for that reason. Obviously not an issue for typical narrow band imaging.

http://www.cloudynights.com/ubbthreads/ ... in/5760045

Quote :
Except for the HII regions, galaxies are very dim in HA light so I doubt going to 5nm will be of much help. What it does do is widen the bandwidth so you can pick up the HA in more highly redshifted galaxies. While nearby ones like M51 are fine in either 3 or 5nm as redshift increases the HA frequency moves out of the passband.

Very useful discussions indeed!

社長 寫:當然，進入數碼時代，十多年下來，我發現不少愛好者初踏足窄頻攝影，發現Ha的效果跟以前截然不同，北美星雲、馬頭、M42、玫瑰幾乎是必到之處，看到那種對比度層次感的提升，誤以為自己的攝影水平大躍進。

正如拍行星月面，從菲林到webcam，然後到天文攝像頭，又或者從菲林到數碼單反，然後到冷凍CCD，每一次儀器更新換代，都會湧現一批新的"大師"，但下一輪的新技術新器材推出市場，原有一批"大師"又被淹沒了。

若你滿足於那種層面，那就不用太深究了。

This is exactly what I always remind myself. The leap in my images is due to my equipment or my skill.

Dear Willis,

If you wish to image in urban areas, like school's roof-top with neigbouring street lights, then 3nm will always help you.

If you image in PTC or sub-urban areas, then 5nm will still be fine.

Try to buy par-focal narrowband filters with the other LRGB, since you would not have to adjust focus much for each filter.
If you use off-axis guider, par-focal will be more important, as you don't have to adjust the focusing of the guider when you change from LRGB to
narrowband.

I find the same problem during remote imaging with my 7x2 inch FLI filter wheel where I use 1 brand of LRGB and another brand
CS 4.5nm narrowband, since usually there is no way to adjust the focus of the guider remotely.

Raymond

Raymond TSE 寫:Try to buy par-focal narrowband filters with the other LRGB, since you would not have to adjust focus much for each filter.
If you use off-axis guider, par-focal will be more important, as you don't have to adjust the focusing of the guider when you change from LRGB to
narrowband.

I find the same problem during remote imaging with my 7x2 inch FLI filter wheel where I use 1 brand of LRGB and another brand
CS 4.5nm narrowband, since usually there is no way to adjust the focus of the guider remotely.

"Parfocal" 這字可圈可點，只說出了故事的一部份。因為用在折射攝星鏡上，LRGBOIII還可以，但Ha尤其SII在波段的遠端，就算頂級APO也要調焦了。

不過我用的自製離軸導星器，是Guide behind filter的，郤沒有什麼問題：

viewtopic.php?f=20&t=1664

我的離軸導星器面對所有最壞的情況：

- 在90mm像場邊緣導星，以APM的TMB 3.5"平場鏡來說，也是良像圈的邊緣了，當然其他折星鏡只會更壞。

- 我的採光棱鏡是固定的，不能旋轉或伸縮選取導星，因為我要確保導星器的剛性/穏定度，但兩年來拍攝任何天區包括暗星雲都沒有找不到導星問題。

- 因為導星器在濾鏡後的距離已固定，從B到SII兩個波段極端，當然星點會有焦準變化，但以我不變的3秒曝光，總有三數顆，通常二三十顆可被導星去選。

- 我的導星器是人手調焦的，最被時用過Borg調焦座，後來經實操驗證以上情況後，我換16803相機時索性改用一個1.25"套筒，以人手拉推導星器調焦及卡尺量度，去年6月調好至今都沒動過，一直用到現在。

所以STi是SBIG真正做得好的導星器：輕巧、有機械快門、一條標準USB連線供電同時操控、高靈敏度，其細小輕功體積，可以隨時用在任何離軸導星器上，就算最薄的TS9也沒問題。SG4就妄想了，可謂高不成低不就，所以我寧可用LVI及STi各司其職，分別配合單反及冷凍CCD相機拍攝時發揮自己優勢，加起的價格仍比SG4平宜。

Lucky if you could adjust guider's manually.

Last week I could remotely image L frames on
M1 for hrs, since the guider's focus has been set for L frames.

When Ha filter was remotely focused by Robofous,
the guider Lodestar could always show guide stars but the
guide stars were all out of focus.

不，我的導星器也要自己在天文台內人手調，我在香港便動不了，也避免請站上人員去做。

因為我是用L通道去調，那是整個光譜的加權平均最佳焦準位置，誤差為+/-0.1mm，所以在光譜遠端的SII通道，便會失焦，但不影響導星，實際上還可某程度上避免Guide on Seeing。所以按SBIG的說法，導星的焦準並不太嚴謹。