in another thread here...
haze on photos
it has become clear that the use (or even the need) for a lens hood has been raised. As there may be some confusion about some of the issues raised, it would be useful to try and clarify the issue for those forum users who may still be in doubt. I did not intend to discuss the deeper theories of optical laws here but do bear in mind that is why the lens hood exists in the first place.
The most important take home fact is that lens hoods are an essential part of the design of a lens and the lens will perform best if the lens hood is fitted to the lens at all times. There should never be a time when a lens hood is not fitted to the lens with which you are creating images. Where the on-camera flash unit is blocked by the lens hood and is casting ugly shadows as well as preventing the flash from reaching the subject, consider an off camera flash unit and bounced flash. In keeping with this simplified look at how both image-forming and non image-forming light has an impact on the final image, you should accept that this is a discussion for illustrative purposes only and it is not intended to be a definitively correct guide to the subject, in a purely scientific sense.
For the sake of this discussion, we are only going to be dealing with two types of light. Type 'A', which is the light that enter the lens of the camera, hits the sensor and will be used to form the image on the sensor. This light will eventually be processed into our captured image, first by the camera and then by image editing software. Type 'B' light is any other light that is not going to contribute to the formation of our image. As you may imagine, there is far more of this type of light (B) than there are rays that will ultimately be used to recreate our scene.
When we look at lenses, it is sometimes not quite clear that there may be many glass elements in a lens. Every separate piece of glass will impose another two glass-to-air surfaces on the lens design. For example, the Canon 180mm f/3.5L USM Macro lens has 14 separate glass elements arranged in 12 groups and this presents 24 glas-to-air surfaces instead of 28, which would be the case if all of the elements were separated from each other. If you follow the provided link and scroll down the page, you will see three cross sectional views that will show you how the lens elements sit in the lens barrel.
Google Image Result for http://www.mir.com.my/rb/photography/hardwares/classics/eos/EF-lenses/EF80200mmf28L/EF70210mmf3545USM_optic.jpg
All of these surfaces will help off-axis type B light to bounce around inside the lens, even though it wont be constributing anything towards the scene. What will happen is that the type B light will degrade the image by reducing the contrast because of its constant bouncing around inside the lens. The type A light, which we do want to reach the sensor, is on-axis and creates the image on the sensor having travelled through the most highly correct part of the lens. (described as on-axis but if you remember the central part of the glass, it will be good enough) As the type A light moves towards the edge of the field of view of the lens, it shows up any faults in the computation and configuration of the lens.
The general rule is that lenses designed for 35mm film format cameras are likely to perform better on a digital camera because the field of view crop (1.3x, 1.5x or 1.6x) forces the sensor to gather data from the central portion of the lens, so that the image-forming light is using the best part of the glass. If the lens hood (designed for a 35mm film camera) is fitted to lens on a 35mm film camera or a full frame sensor dSLR like the 1Ds or the 5D, it will be performing to a known standard. When it is fitted to the lens that is imaging a scene onto a sensor smaller than 35mm, then the lens hood is not stopping all non image-forming light (B) from hitting the sensor.
Some illustrations are in order. The small application (at the following link) will show you what the angular field of view will be for a given lens and body combination. The field of view can be calculated for horizontal, vertical or diagonal fields. The diagonal field is the largest and so it is the most convenient to use.
Nikon Lenses :: Angle of View // sweeting.org
The lens I have used to demonstrate the different lens performance on different camera bodies is the Nikon 180mm f/2.8. The lens hood prevents ALL light, that is not destined to be type A light, (determined by the angular field of view for the particular focal length you are using) from entering the field of view and degrading the image by reducing the contrast.
In the days of being able to use one lens construction for a single film format (swap film for sensor these days) the calculation was an absolute. If your lens could 'see' an angular field of view of say... 13°42' across the diagonal of the 35mm format, as in the case of the outstanding 180mm f/2.8 Nikkor, then the manufacturer could make a lens hood that would exclude the light from 13°43' onwards and know that enough had been done to exclude any type B light from the frame.
The identical lens and hood combination on various Nikon digital bodies will be different. The D70 body would permit the lens to 'see' 9° and it would be from the sharper central portion of the lens but the lens hood would be letting past itself what would, effectively, be type B light that extended from 9.1~13.43° While the lens can only 'see' what its calculated field of view for a given sensor size will permit it to, having a very bright light source get past the lens hood is guaranteed to be detrimental to the image.
Kind regards,
Jeff
Linear Mode
