Just as what we've been talking about, you can see the imaged "deformed" in the behind-focus images are almond shaped, exactly the same as your suggestion of the shape caused by vignetting... so i guess the "deform" is caused by blockage of light rather than "bending" of light.Wah!! 寫:I still doubt about it.anguslau 寫:....For a complex optical design (such as FSQ), it is normal to have unsymetrical in-out focus patterns. If you examine the in-out pattern of a very complex design such as a camera lens, this will be extremely obvious. All optical systems are designed to give optimal performance at focus only. To see the olived-shaped aperture shadow, the amount of in-out focus is very large already, obvious out of symetry is not surprising....
No matter how complex the optical system is, if it can converge light to a very small area (roughly a point), the shape should be symetric from in to out of focus.
If the light can converge to a point, but not symetric from in to out of focus, light MUST have been bent at the focus!!!
Takahashi FSQ 85 ED 光學討論
anguslau 寫:Thanks very much to 昇仔 for the analysis!To me, these are conclusive evidence of our analysis. Particularly the direct proof of olive-shaped out of focus aperture for off-center locations.
For a complex optical design (such as FSQ), it is normal to have unsymetrical in-out focus patterns. If you examine the in-out pattern of a very complex design such as a camera lens, this will be extremely obvious. All optical systems are designed to give optimal performance at focus only. To see the olived-shaped aperture shadow, the amount of in-out focus is very large already, obvious out of symetry is not surprising.
I believe we can conclude that this so-called dark band problem is inherent in most optical systems as long as the fully-illuminated circle does not cover the entire sensor in use. It can be seen as a limitation of a particular design. To minimise the occurrence of such phenomenon, one can choose an optical system which present minimal amount of vignetting covering the sensor desired. In general, such problem is typically more prominenent for faster systems. But the particular design of an optical system ultimately dictates its behaviour. For example, even though the FSQ106 is a f5 system, it has very little vignetting over a 44mm image circle and hence does not produce obvious dark band when imaging with a 35mm full-frame sensor. It is because the system was designed to give reasonable performance for a much larger image circle. On the other hand, FSQ85 was designed with a 44mm image circle in mind, hence the amount of vignetting is not insignificant within such circle. Therefore dark bands can be observed when imaging with a 35mm full-frame sensor.
I TOTALLY AGREE TO THIS SAYING .
Very good generallization indeed !
If vignetting is caused by blockage of light, then moving in front of and behind the focus may cause the light to be blocked by different extents, and hence results in the asymmetry in front and behind the focus.PaulNg 寫:Just as what we've been talking about, you can see the imaged "deformed" in the behind-focus images are almond shaped, exactly the same as your suggestion of the shape caused by vignetting... so i guess the "deform" is caused by blockage of light rather than "bending" of light.Wah!! 寫:I still doubt about it.anguslau 寫:....For a complex optical design (such as FSQ), it is normal to have unsymetrical in-out focus patterns. If you examine the in-out pattern of a very complex design such as a camera lens, this will be extremely obvious. All optical systems are designed to give optimal performance at focus only. To see the olived-shaped aperture shadow, the amount of in-out focus is very large already, obvious out of symetry is not surprising....
No matter how complex the optical system is, if it can converge light to a very small area (roughly a point), the shape should be symetric from in to out of focus.
If the light can converge to a point, but not symetric from in to out of focus, light MUST have been bent at the focus!!!
Moving behind focus --> more blockage --> more olive-shape
Why does it occur?
你的例子，也要看本身光学paramter 合不合当下拍出来的实验呀。PaulNg 寫:講下講下, 突然間覺得好似同個光學系統既設計有關
A: 有問題既設計--> Petzval 既折射鏡 (再加reducer一樣有問題), 相機鏡頭
B: 冇問題既設計--> 普通折射鏡+flattener
結論: 會唔會係同個"flattener"既設計有關?
A類設計flattener係內置, 即不能隨對焦既動作移動, 同sensor距離可以變.
B類設計flattener外加, 同sensor距離固定.
当然，结构会是一个方面。
其实结构和光学paramter应当都要考虑在里，而并不是分开来看的。
都是同一合成的一个整体问题。
最後由 MANDII 於 週二 15 6月, 2010 11:44 編輯，總共編輯了 1 次。
The olived-shaped aperture shadow is a direct picture of the blockage obstructing part of the aperture. This caused vignetting (lost of light) and at the same time the diffraction pattern we've observed (ie dark bands). This blockage existed somewhere along the optical path of the system. Moving the observing point in-out of focus shifts the distance between the observing point and the blockage. When such such is large, you should expect to see a slightly different shadow. I think the out of focus shadow probably give you a better view of the "actual" shadow.PaulNg 寫:Just as what we've been talking about, you can see the imaged "deformed" in the behind-focus images are almond shaped, exactly the same as your suggestion of the shape caused by vignetting... so i guess the "deform" is caused by blockage of light rather than "bending" of light.Wah!! 寫:I still doubt about it.anguslau 寫:....For a complex optical design (such as FSQ), it is normal to have unsymetrical in-out focus patterns. If you examine the in-out pattern of a very complex design such as a camera lens, this will be extremely obvious. All optical systems are designed to give optimal performance at focus only. To see the olived-shaped aperture shadow, the amount of in-out focus is very large already, obvious out of symetry is not surprising....
No matter how complex the optical system is, if it can converge light to a very small area (roughly a point), the shape should be symetric from in to out of focus.
If the light can converge to a point, but not symetric from in to out of focus, light MUST have been bent at the focus!!!
Wah!, light does not bend in our case. Remember we are defocusing to try to observe the aperture shadow at focus point. The shadow picture is not the image at focus point. The shadow picture at in-out focus positions are approximate pictures of the shadows. Each light ray converges to a point at focus. But the picture at in/out focus positions can be different.
The problem should not be directly related with a specific design, such as Petzval. It should be directly related to the amount of vignetting inherited in the design.PaulNg 寫:講下講下, 突然間覺得好似同個光學系統既設計有關
A: 有問題既設計--> Petzval 既折射鏡 (再加reducer一樣有問題), 相機鏡頭
B: 冇問題既設計--> 普通折射鏡+flattener
結論: 會唔會係同個"flattener"既設計有關?
A類設計flattener係內置, 即不能隨對焦既動作移動, 同sensor距離可以變.
B類設計flattener外加, 同sensor距離固定.
1) we know that an olived-shaped aperture will produce dark bands
2) we also actually see the olived-shaped aperture shadow in Sing Chai's experiment, directly correlating with the dark bands
3) for any system with vignetting, part of its aperture is blocked; the question is what is the shape of the partially blocked aperture? I believe it is generally olive-shaped for most optical designs; if so, this will produce dark bands similar to the FSQ85
4) if you agree with 3) above, it means most optical systems which exhibits vignetting should have dark band problems; in fact, unless you believe the partially blocked aperture shape is not olive, the characteristic dark band should result
Light goes straight line, then when all light pass through a single point, both in/out focus light cone should be similar.anguslau 寫:Wah!, light does not bend in our case. Remember we are defocusing to try to observe the aperture shadow at focus point. The shadow picture is not the image at focus point. The shadow picture at in-out focus positions are approximate pictures of the shadows. Each light ray converges to a point at focus. But the picture at in/out focus positions can be different.
Can you agree with this?
Defocusing is just putting the sensor to different parts of the light cone.
Moreover, the sensor plan doesn't change direction, they're parallel to each other when we put the sensor to different location. As a result, the shapes showing on the sensor should be similar, the only difference is 180-degree rotation after passing the focus plan.
(Assume that the elements of optical system don't move when we move the sensor.)
I agree that the proof is pretty convincing... but I've been looking at some images, like this... taken by TMB130 with flattener, 6X7 E200. It shows no dark banding like that...anguslau 寫:The problem should not be directly related with a specific design, such as Petzval. It should be directly related to the amount of vignetting inherited in the design.PaulNg 寫:講下講下, 突然間覺得好似同個光學系統既設計有關
A: 有問題既設計--> Petzval 既折射鏡 (再加reducer一樣有問題), 相機鏡頭
B: 冇問題既設計--> 普通折射鏡+flattener
結論: 會唔會係同個"flattener"既設計有關?
A類設計flattener係內置, 即不能隨對焦既動作移動, 同sensor距離可以變.
B類設計flattener外加, 同sensor距離固定.
1) we know that an olived-shaped aperture will produce dark bands
2) we also actually see the olived-shaped aperture shadow in Sing Chai's experiment, directly correlating with the dark bands
3) for any system with vignetting, part of its aperture is blocked; the question is what is the shape of the partially blocked aperture? I believe it is generally olive-shaped for most optical designs; if so, this will produce dark bands similar to the FSQ85
4) if you agree with 3) above, it means most optical systems which exhibits vignetting should have dark band problems; in fact, unless you believe the partially blocked aperture shape is not olive, the characteristic dark band should result
誰在線上
正在瀏覽這個版面的使用者：沒有註冊會員 和 5 位訪客