Review Fuji Finepix S3 Pro UVIR
(English)
After having to wait for quite a long time UPS finally delivered a parcel from New York today (Sept 1st). As expected it contained the the much anticipated Fuji Finepix S3 Pro UVIR. After I had impatiently opened the parcel I put all the accessory parts on my table in a well-arranged layout in order to shoot the first photo. Then I had to charge the batteries and to set up the tripod. Because it tooks seemingly endless to charge the batteries, I took the net adapter and connected it with the camera.
I had enough time to think about different themes and to study several camera arrangements, but one question remained finally... what performance to expect from the camera when used within the visible spectrum in the range from 400nm to 700nm. This is important for a coherent documentation when comparing UV and IR - without evidence about the setup viewed under visible light a final assessment wont be possible.
Thus I took the camera, loaded a 8GB Sandisk extreme III CF card and then mounted it on my stable Gitzo. In the camera's menu I selected the manual mode, ISO 400 and activated automatic release. I used the standard Nikkor 45mm f/2.8 P lens. I went for ISO 400 as Fuji indicated on its website that using lower ISO might cause unsatisfying result. Of course I am aware to give away some quality in both the visible and the infra-red spectrum by using this setup. On the other hand I want to shot all pictures using the same parameters in order to obtain results which are easier to compare. The lens uses (45mm Pancake Nikkor) is a good benchmark for assesing the optical quality of a sensor. This is due to the Nikkor's high optical quality (almost no distortions or other aberrations). This lens manufactured by Nikon is a real highlight compared to the other standard lenses.
Maybe some of You are a little surprised by my test's style and thus I want to give you some explanations. I am familiar wth Bjorn Rorslet's comprehensive reviews and also with these beautifil tables and transcriptions done by all the other professionals trying to explain all sorts of things. It goes without saying that it is exciting to read about all these technical details, but it does not really help to assess if a certain lens is appropriate for myself. Therefore my review's style will be very idosyncratic. My review describes in plain English what I am doing with the camera and I post unprocessed pictures. Thus it is reproducible for everyone. I think the results speak for themselves and do not need any further comments.
As the old saying goes: One picture is worth a thousand words.
I will stick to this basic principle. I will just take photos and I will not write lengthy reports. Furthermore I think the motive is of significant importance for the evaluation of the review's pictures.
Not a patch on pretty flowers or interesting landscapes, but these are not adequate motives for the purpose of creating reproducible results. Hence I choose a maybe not too pretentious, but ever reproducible motive: The view from my attic window to my neighbour's house. The advantage of this arrangement is easy to understand. I can leave the tripod at the same location for weeks and can run tests with different lenses and whichever new parameters from the position. These circumstances make it much easier for people interested in my review to assess for themselves which lens is best suited for UV or IR. The motive's changes are limited to different wheather conditions or the trees losing their leaves. Also the focal length will usually change with the lens. As the motive does not change and as only the picture's angle might change sometimes, the different lenses' performance in regard to sharpeness and at different apperture settings are perfectly compareable even if weeks or even years have passed between different test-occasions.
Some of you might wonder about my motivation for choosing this test design. I think the rational is quite simple: I myself have never been able to entirely reproduce other persons' tests posted on the net. For example Bjorn Rorslet drew the conclusion that the legendary 50-300mm telephoto lens shows its best performance at aperture f/11. According to my own tests, however, it did much better at f/8. I wondered myself about the reason(s) for this discrepancy. Is it a matter of fluctuations in quality at the production stage? Or maybe the writer just made a mistake when counting the number of times he triggered the shutter and later on matched the shots taken with the wrong aperture? Whatever the circumstances were, I am great fan of Bjorn and I much appreciate his pioneering work in the field of IR- and UV-photography, but I do neither want to copy his test design nor do I aspire to do something similar to what he did.
Let us now take a look at the camera itself and share my first impression on my Fuji Finepix S3Pro UVIR. The camera by itself doesnt have really new features exept the live preview which is little bit tricky to activate. First you have to press the menu button, then go to menu 5 (which is the first one apears when you turn it on) and press live preview, then you are able to get an idea of the exposure you will get later. But its not perfect. First the old LCD screen is quite small compared to the new DSLR´s LCD size and second the preview doesn´t change with different exposure settings, just the Iso and the aperture settings make differences in the brigthness of the small prieview picture on the LCD screen. In true you need this nice feature only when using the camera handheld. If the cam is mounted on a study tripod you can make a single shot first to check out the true quality and correct the picture in the way you need.
Also the added manuel is from the original S3 Pro. The only consideration to this "special model" is a small sticker on the box and a yellow single paper size A4. Some importend notices are on this single sheet, but nothing really new, just the already posted informations from the fuji homepage. For the 700 bucks more then the Original Model you get only the Hyper-Utility Software HS-V2 Vers. 3.1 for "free".
The first shots I took without any filters in order to get an impression how this special camera performs when recording the whole spectrum from 350nm to 1000nm. For these pictures aperture was set to 8 or 11.
I uploaded these first test's shots in three different versions. The first link always leads to a unprocessed except for being reduced in size by "Save for Web (quality 70%)" to 1850x1238. The second link provides a crop (original size) of the unprocessed picture, again saved at 70% quality. Finally, the small thumbnail shows the third version which is processed. All the processed images have been edited in Adobe Photoshop according to the following sequence: Auto levels, contrast, saturation, unsharp mask, image size, unsharp mask.
The test's pictures' size is 1850x1238. Only registered users are permitted to view the pictures at this resolution. Non registered guests are restricted to view downsized images at a size of 800x600.
.
Image 01 (unprocessed original at 1800x1238 pixel)
Exif Data:
Aperture f/8.0
Exposure time: 1/500sec
Iso 100
Conclusion: All pictures show a very significant red and blue color cast which is hardly or not at all correctable using Photoshop.
Next I tried to shoot color neutral pictures by using a B+W UV/IR band-elimination filter 486. These images also show some light color cast. All in all the picture looks somewhat colorless. On the one hand this is a hint that there is some infrared contamination left which the B+W filter 486 was not able to block completely. On the other hand this is a good indication of the Fuji's infrared capabilities.
.
Image 02 (unprocessed Original at 1800x1238 pixel)
Exif Data:
Apreture f/8.0
Exposure Time: 1/250sec
Iso 100
The third test aims at completely blocking the visible and the IR spectrum in order to let pass only UV light to the camera's sensor. This is a somewhat difficult task to undertake, but achieveable by simple means when one has an IR filter at hand. For example one can use a B+W 093 filter which is supposed to block the complete visible spectrum below 800nm. I chose the extreme 093 filter deliberately. This filter is equivalent to the Kodak Wratten 87C or the Schott RG830 and is guaranteed to let no visible light at all pass through to the sensor. If I can succeed that easily we will see on the basis of the next test.
Image 03 (uncompressed Original at 1800x1238 pixel)
Exif Data:
Aperture f/11.0
Exposure Time: 1/60sec
Iso 400
The next step is to add the well-known B+W UV 010 filter to the B+W 093 filter. Der UV Filter soll sichtbares Licht passieren lassen, Ultraviolettes Licht aber gänzlich absobieren. Ganz gelingt diese Übung leider beiden Filtern nicht wie die folgende Graifk deutlich macht.
UV Filter transmissions graph
It has to be said that the diagrams in practice often are not applicable at all, since one does not know under which laboratory conditions and using which lightning source they were produced. Also one has consider that practically all filters can be bought in different thicknesses (1.2 or even 3mm) which then dramatically affects the attainable absorption.
In my case I have the filters of Hoya and B+W at hand for testing. With these two filters, however, we only block the ultraviolet part and not the visible part of the spectrum. Let's take a look at the results of the next test with an added B+W UV 010 filter.
Image 04 (uncompressed Original at 1800x1238 pixel)
Exif Data:
Aperture f/11.0
Exposure Time: 1/90sec
Iso 400
Its is easy to see that the image 04 shot with UV 010 filter practically does not differ at all from the preceding image 03 which was made without UV filter. This points quite clearly to the fact that pictures in the pure infrared spectrum are an easy task to undertake with this marvelous camera. Simply choose a suitable lens, mount the B+W of 093 filter and then let's go.
How about using the by far more common Hoya R72 filter? This filter lets pass through light in the spectral range not only starting from 800nm, but already starting from 700nm, which represents the visible light's boundary. At 750nm 93% are reached and maintained to about 2800nm. Not for nothing this is the most universal infrared filter around.
Image 05 (uncompressed Original at 1800x1238 pixel)
Exif data:
Aperture f/8.0
Exposure time: 1/750sec
Iso 400
Next for the most difficult test of my review: pure ultraviolet photography. Nothing is more complicated and more difficult to implement than this test's set-up. First one faces the problem to obtain a suitable lens. Unfortunately Nikon's range of products offered only one adequate lens. As only 3,030 units were manufactured it is today one of the rarest Nikkors. It is the UV-Nikkor 105mm f/4.0.
Unfortunately the actual focal length of this Nikkor mounted on a Nikon DSLR is 158mm, which is only appropriate for shooting florets in one's front garden. Taking pictures of complete buildings or landscapes, however, is hardly possible, except from quite a distance away. Moreover the quartz lenses used are a pure waste, as there is no sensor available that is capable of recording anything below the wavelength of the UV-A spectrum. The Fuji's technical data clearly evidence this fact. Even this special camera particularly developed for forensic purposes is incapable of recording light below a wavelength of 350nm. Thank God there are more affordable alternatives to the 105mm UV-Nikkor which are worth a closer look. One of these alternative lenses is the quite rare EL-Nikkor 63mm f/3.5. It is a magnification lens which forces photographers like us to temporarily work as handicraftsmen. This is due to its lack of both a Nikon bayonet and any means for focusing. Also these shortcomings make things not exactly easier it is nevertheless not impossible to succeed anyway as we will see later on.
Now, back again to the first lens reviewed, the Nikkor 45mm P f/2.8 which is also referred to as "pancake optics" because it as flat as a pancake. For this test two filters are needed. Instead of the best combination which I was not able to locate until now (Hoya U-360 plus the Schott BG-38/40) I will use the B+W 403 filter which blocks both the entire visible spectrum and small parts of the UV-A and the infrared spectrum (730nm to 770nm) and secondly, the B+W IR cut filter with the number 489 which blocks infrared light in the range from 780nm and above. As a matter of fact this combination of filters is not perfect and there is still much more contamination by infrared light left then Ultraviolet. Unfortunately however, the really perfect Baader U-filter is available only in the size of 31.8mm and therefore is difficult useable at all in combination with this lens. More on this topic later on.
Image 06 (uncompressed Original at 1800x1238 pixel)
Exif Data:
Aperture f/11.0
Exposure time: 1/1.5 sec
Iso 400
Unfortunately I suspect this last picture to be no real UV-photo, although it was made using two excellent filters produced by B+W. To be the real thing it was just too simple to produce. Furthermore I have got no information on the transmission capabilities of the Nikkor 45mm f/2.8 P. It is constructed using only 4 lenses in 3 groups. This does not, however, indicate anything without information on the sort of glass and the coating used. Assuming that the main part has still to be infrared light the question is raised why these extremely expensive filters are so poorly up to their taks? After all the manufacturer (B+W) claims that ultraviolet light is blocked completely by combining the filters with the numbers 403 and 489.
On the face of it there seems to be no way around for me to buy the filter manufactured by the company of Baader Planetarium. Then I will undertake some further testing by combining this filter with with the only lens really suitable for UV light, which is the EL-Nikkor 63mm f/3.5.
Until then I may thank you for your patience. I hope my review provided some useful information for you and say good-bye until next time, when I will upload a new test report.
Wolfgang Steiner