Text and pictures © 2007-2024 Guillaume Dargaud
Last updated on 2021/11/05
"Photography is a reality so subtle that it becomes more real than reality." — Alfred Stieglitz.
I've never heard of the S1. The S2 and S3 were peculiar and clunky cameras producing good images and I though long and good about getting an S3. The S4 was never even produced for superstitious reasons... So here comes the S5 from my favorite film manufacturer. Its full name is a mouthful: the 'Fujifilm Finepix S5 Pro digital single lens reflex camera', although no film is involved anymore.
So why should you consider the S5 among the huge choice of other reflex cameras ? It's 40% more expensive than the Nikon D200 on which its hardware is based while having 40% less resolution... The whole difference can be summed up in the dynamic range of its imaging sensor which is 4 times wider. After using the Ricoh GRd for over a year, I came to the conclusion that the dynamic range of normal CCD sensors pretty much suck for outdoor photography: the contrast of the scene is simply often too high. One workaround is to use software HDR from multiple shots, a tedious technique. Or to purchase the Fuji ! Of course, after 2 decades of film photography some adaptation is required.
It's worth noting that with so many options you should sit down with the camera for a good hour and work through all the menu options at least once to figure out what the most arcane ones do. And repeat a month later.
The sensor is not even half the traditional film size: 23x15.5mm, aka APS size, leading to a diagonal focal ratio of 1.56 and a side ratio of 1.48:1, although the processing grinds this to a precise 3:2 ratio for non-interpolated images (3024x2016 or 2304x1536). Now it gets a bit complicated: there are 12.3M detectors on the sensor leading to a direct 6.1M pixels image or a 12M pixels interpolated image. It's just that half of those pixels are not used for resolution as you'd expect but to increase the dynamic range; in other words the maximum amount of light that the sensor (and hence the camera) can handle before reaching saturation. The large size of the base detectors ensure a good signal to noise ratio while the small size of the secondary sensors ensure an excellent ability to withstand overexposure or high contrast scenes. By excellent, I mean no less than 4 times better than the best the competition can offer. If you complain about burned out white clouds, snow or flash reflections, the S5 is for you.
|Focal equivalency table:
|A traditional focal of
|Becomes equivalent to this when mounted on the S5
|...or must be replaced by this for the same angle
|Available file modes (you can optionally save JPEG files together with RAW files). Note that there are other quality and resolution available if you use HS-V3 to convert from raw files:
|High quality version
|Raw D-Range 130~400%
|Normal quality version
|Raw D-Range 100%
|Since the S5 is based on the body of the D200, they invariably end up being compared:
|The D200 is cheaper by about 30%
|10 true megapixels for the D200, 12 interpolated from 12 'other' million for the S5
|The D200 has a bigger buffer and smaller RAW files leading to more images per seconds in most cases.
|More faithful colors for the S5
|You can have pre-programmed modes on the D200 and shift between them quickly
|Hands down for the high dynamic range of the S5
|Noise and low light
|Noise is better to much better on the S5 at all ISO sensitivities
|Better control of bright reflexion on the S5, yet again due to the high dynamic range
|Well, the S5 menus are not the most well organized...
|Nikon releases a new model about every year just to make you feel frustrated from your now outdated purchase ! C;-)
"If the D200 gave me anywhere near the color output of the s5, I would've kept the one I bought..." — Itinerant (on the S5 flickr discussion list).
Make sure to stay up to date with the latest Firmware upgrade from Fuji, hopefully it will correct some of the issues I raised above. The latest versions of the FinePix Viewer and the camera driver are available off Fuji's website as well. And there's a flickr discussion group dedicated to the S5 where you can find the latest gossip, gripes, flame wars and stupendous shots.
There are 4 possibilities of flash use: the built-in flash (guide number given as 13, probably somewhat less), one of the recommended external flashes (SB-800, SB-600 or SB-R200) mounted on the accessory shoe for full features, a standard flash on the accessory shoe with various levels of functionality or a studio flash via a PC-sync cable.
I didn't purchase the camera for its internal flash. I hardly ever noticed it was there in the first weeks of use, not firing it once, and knowing full well the usual problems associated with this kind of tiny flash: red-eyes, short range, hard light, shadows cast by large lenses and, I thought, limited capabilities. But then upon reading the camera manual in full, I discovered two important things: the internal flash can be used to control remote flashes through Nikon Advanced Wireless Lighting, and it offers complete manual control. Deep in the menu options you can set the flash in i-TTL, TTL, wireless, high-speed sync, manual, compensation, strobe... In other words it's good to know that you have a real and full-featured flash handy, even if its power is quite limited.
Right: Two example of flash photography with a SB-800 mounted on the S5. First image is in evening light, i-TTL with rear-curtain; nice mixup of ambient and flash light. 2nd image (MouseOver) is in complete darkness, front curtain and TTL to avoid pre-flash; it's a usual exposure trap but there's no burnt area, particularly where the flash reflects directly on the skin (nosetip and cheeks). Most settings set to default (Auto-WB, P mode, Auto-Iso, Auto-D-Range...). That's the Fulachier family.
The camera has an AF-assist illuminator which works independently from the flash; but only if your AF-mode is set to [S]ingle. I want it to work even in [C]ontinuous mode since I always use that focusing mode in association with the thumb AF-ON button with a disabled shutter focus. And why not also in [M]anual focus when pressing AF-ON so you can verify your focus ?
If you use a standard flash, that is to say anything except the very latest and most expensive Nikon SB-800/600, don't expect much. Even the recent SB-80DX which works on the first generation of Nikon DSLR won't offer many features: no TTL, no auto-zoom, hell, even no ISO transmission ! Back to the happy days of Auto-flash, like in the 70s. Well, maybe you can use it as a remote in SU-4 mode, provided you also have a SB-800 to act as master on the camera.
The high dynamic range of the S5 finds one more use here. Flashes are infamous for producing bright reflections, often saturating sensors into large all-white patches. The higher dynamic range ensures the overall contrast stays within usable range and it works very well indeed.
First example of remote flash photography on the S5: using the SU-4 remote mode. Camera setup: nothing particular. Mounted flash: SB-800, remote mode set to SU-4, exposure mode set to AA (with an 'S' looking arrow next to it). First remote: SB-80DX, Remote Sel ON, Mode A, remember to orient the side IR window towards the camera (or rather, towards the master flash); no other settings available nor necessary. Second remote: SB-24 mounted on SU-4; set the SB-24 to ON (not stdy), TTL mode, front curtain, no other settings (ISO, Aperture...) necessary; set the SU-4 itself to Auto. Main drawbacks of this method: yu cannot use TTL (unlike on a F100), not very reliable, you need a master on the camera (the internal flash cannot control a SU-4), the remotes must be in sight of the master which is hard to do if you don't want them in the frame, you cannot fine-tune the exposure by specific flash, you cannot use flash preview, the remotes will trigger with any visible flash. Main advantage: works with older flashes.
Second example of remote flash photography on the S5: using the Advanced Wireless Lighting (abbreviated as CLS, go figure): camera with the internal flash activated, [Setup][Flash/BKT][Builtin flash][Commander Mode] where you set TTL/TTL/TTL and a channel number. Remote flash: SB-800 set on remote, set the same channel number and choose a group. Pros: You can put many remote flashes, divide them in several groups with different compensation settings or exposure modes for each, you also have a choice of using TTL mode, it's more reliable than above (although the lone of sight problem still exist). Cons: you cannot use your old flashes, so you need to invest yet again.
There's more info about flash photography with the S5 pro in my much shorter review of the SB-800.
Let me state this clearly: I got the S5 so I wouldn't spend more time tweaking images on the computer than taking them. The idea is to have a camera that delivers the goods, and not a pile of expensive software to work around shortcomings. Understandably, some software is still necessary to go from the camera to the final print, but the workflow is somewhat lighter.
One thing comes out of all I've written above: the S5 produces excellent JPG files out of the box and gets even better tweaking the options. But then, what about going one step further an process the RAW files ? Let's see now what is possible.
Note: I've moved this review to a different page.
Now onto a different subject, it's been said that the S5 produces HDR images, but this is not quite true. Although the sensor has a high dynamic range, the image files it produces are classic. An HDR file should contain luminance information as expressed in intensity values and represented by open scale floating point number (12.6 is twice as bright as 6.3 and so on). Such images are usually produced by software, either in ray-tracing or by combining shots taken at various exposures, and as such cannot be displayed properly on current monitor or print technology (where 0 is black and 255 is white and those values are boundaries which don't exist in the real world). The S5 does indeed take the equivalent of two differently exposed shots at once thanks to its advanced sensor, but the data is combined internally to produce a standard JPG file, or a RAW file that can only be processed properly by Fuji software into a standard JPG or TIF file. So although the theoretical answer is yes, the practical one is no. But maybe a firmware upgrade or an update to FinePix Studio or Hyper Utility will one day allow direct HDR conversion, who knows. Or more likely there'll be a new version of Photomatix or another tone mapping program that can combine both sensor readings into a single HDR file; as for doing tone mapping from those two sets of sensors, forget it for now as there's no way to extract them separately that I know of. Please let me know if you know a way.
Click on the buttons below to see various test images taken with various options and lenses.
|The camera can be used in standard mode (100%), HDR modes (130 to 400%) or with simulated film types (not shown here). For instance F1 is a low contrast suitable for portrait (similar to Sensia is my guess) while F2 is high saturation (similar to Velvia). 100% correspond to the use of only the main photosites and thus leads to standard images. Any percentage above that (up to 400%) leads to HDR images created from a double input, so you get an extended dynamic range at the expense of a lower visual contrast (think about it). High contrast scenes can fit on the sensor instead of being clipped. This is where the S5 truly shines (see below the overexposure tests). Because of the drop in contrast I find that using 130% is a good choice (after all you didn't pay all those extra photosites for nothing), or more realistically the [Auto] option on the camera.
|Two series of two shots: with a dynamic range of 100% (Std) or 400% (W2), and overexposed by 3 stops or underexposed by one. Logically there is no difference when underexposing the image. But the overexposed image is 70% burnt at 100% while still fairly usable at 400% (with some additional post processing).
Those shots are taken with a 50mm/f1.8 lens in P mode.
|Two series of 5 images were taken: one set with 100% dynamic range, the other with 400%. In each set the images were underexposed to -5 and -2 EV, and overexposed to +2 and +5EV (that's a lot). Images were saved as raw (RAF) and then processed with Photoshop RAW module, either 'camera default' or 'auto'. There's not much difference between 100% and 400% for underexposed images, but the difference is huge for overexposed images: just compare the +5 shots after auto processing. It's easy to tell why the Fuji SuperCCD sensor wins hands down: even an image overexposed 32 times (that's 5EV) is still recoverable (even if the shade of the sky is way too purple); 100% is similar to what you get with other sensor brands. Note that RAF files for 130~400% files are twice as big since the data of the small photosites is saved as well.
In fact, I love the images rendered by the S5 so much that my GRd started looking bleak at once. So I looked at Fuji's compact camera lineup. They are all advertised as 'SuperCCD'. But beware: there's the SuperCCD SR, mounted on the S2, S3 and S5, and the HR, mounted on everything else. The SR technology with the double photosites implies that they be large, so if you have an itty bitty sensor (like in most compact where they are typically 5x9mm), the resolution would be insufficient. Hence the HR (High Resolution) denomination. But then you loose the great advantage of the Dynamic Range. In the days of film they could fit a 24x36mm 'sensor' in a camera half the size of a fist (including canister and winding area and motors), and now they claim they can't make one with even an APS-sized sensor. What a load of bullshit.
Here's a more useful example of image where the S5 pro really shines: astrophotography. Images taken with a Celestron Nexstar 5i (1250mm).
An image of a conjunction between Saturn and the moon taken with 100% dynamic range. In other words, this is the kind of image that could be taken from any digital SLR.
Now the same image with maximum dynamic range (400%). It may no seem obvious but there's a huge difference: on this image there's virtually no overexposed pixel.
To press the point, here's the first image with all the purely white or black pixels shown as red or blue. In this case we don't care about underexposure as the night sky is really supposed to be black. Notice that half the moon is 'burnt' ?
In contrast, the high dynamic version of the image fits within our histogram, save for a few pixels...
...meaning that it can be post-processed successfully into something even more interesting such as this image (local contrasts are enhanced as well as an unsharpen mask applied). With an average SLR you would need to take two images at different exposures and combine them.
The Fuji S5 pro seems to be a very good astrophotography camera due to a combination of factors:
The main drawback when mounted on a scope is that the camera is heavy, so you'd better tighten some screws and have a counterweight ready.
|Nikkor 50mm/f1.8 AF
|Comment: This is a reference lens, high definition all around.
|Sigma 24-70mm/f2.8 AFD EX DG
|Comment: at 24mm and also at full aperture the lens is pretty soft. When you combine both, you get abysmal performance... That's too bad as this is my only lens 'optimized' for digital cameras. Indeed that's so bad that I'm beginning to wonder if I got the focus right. Only 70mm stopped down performs well, but images are underexposed (why?)
|Sigma 28-70mm/f2.8 AFD EX
|Comment: Soft at full aperture, but not as bad as above.
|Nikkor 20mm/f2.8 AFD
|Comment: Excellent as soon as you leave 2.8, and not too bad there.
|Sigma 28-105mm/f2.8-4 AFD
|Comment: I wonder why the image is underexposed at 28mm...
|Sigma 70-200mm/f2.8 EX APO HSM AFD
|Comment: Only 70mm is a bit soft at full aperture, all other zoom settings being excellent. 400mm is obtained with the Sigma 2.0X APO Tele Converter which is soft in the angles (not a problem on a small sensor) but also suffers a hefty amount of vignetting (visible even here !).
|Sigma 35-70mm/f3.5-4.5 AF Anaglyphic
|Comment: Anaglyphic effect is maximum at full aperture, so a trade off with the weak sharpness must be accepted.
|Sigma 8mm/f4 Fisheye
|Comment: Unfortunately the sensor covers the whole 180° frame only horizontally, that is not to say that the lens becomes useless, for instance to build 180x360 spherical panoramas.
One of the first things I noticed upon using the S5 is that the exposure on some images is dead wrong. To be more specific, I can shoot a sequence of similar shots and there will be a huge ±2EV variation ! Operator inexperience ? Poor AE sensor ? Wrong setting somewhere ? Firmware bug ? Some other people have reported this issue while others have spot-on exposure. My luck if I got a lemon... Let's see if I can get to the depth of this.
|24-70/2.8 at 24mm
|24-70/2.8 at 70mm
|18-50/2.8 at 18mm
|18-50/2.8 at 50mm
|Comment: First on the Sigma 28-70mm/f2.8, the greyness of the top part of the image should be the same on all shots (since the exposure is done on it): it's not. Then notice that the 'gray' (from the 3 grey side) and 'grey' (from the white+black side) area should be absolutely identical and they are not. Also notice that while the exposure at 24mm is globally OK, at 70mm it is too dark on all but one shot.
Now let's look at the Sigma 18-50mm/f2.8: the exposure is perfectly stable almost everywhere (111 at 18mm and 123 at 70mm, a little lighter). At 18mm the blacks are a little overexposed and the whites a little underexposed.
So my guess is that there are problems with some lenses. Maybe the characteristics of some lenses, in particular 3rd party lenses like Sigma or older non DG lenses, have not been inserted inside the internal exposure database, who knows. Just to prove that if you want to be sure the S5 exposes correctly, you'd better check it out yourself. It also joins the fact that some people have been complaining about the exposure while others have reported no problem: it probably depends on what glass you have mounted. Note that since this test is only in spot, it shows nothing of the possible 'intelligence' of the other exposure modes.
Now onto a different problem: when taking shots in low-light conditions, the camera consistently under-exposes the image. This applies for instance when taking indoor portraits in ambient light at high sensitivity. I've found out that using a +1ev compensation brings out a lot more details and since the sensor handles perfectly even a solid overexposure, there's not much harm in doing this. You have to be a bit more careful when doing this when using the flash (you might want to then do -1ev on the flash itself). But this is more a matter of personal preference.
|Comment: Spot exposure is correct. Center-weighter seems to behave as if to expose on the brightest thing found near the center of the images (explaining why the wall image is so dark as there's a piece of outside in it). As for the Matrix, I don't know what to think: the outdoors shot is a little lighter that spot as expected; the wall shot is way off, meaning that if you have something bright covering 1/3 on the right of the frame, it will expose for it (don't expect good backlit images). Even the 'picture frame' shot is wrong, as when metering for it I had only a tiny bit of the window on the far right of the images. Conclusion: the camera is able to handle great overexposure thanks to its dynamic range, but they've set the meter to always expose on the highlights, which kinds of defeats the purpose... It is possible (in some sub-sub-menu) to customize the exposure to a certain amount of compensation for each specific mode (spot/center/matrix), but something more intelligent would be required.
|Camera on USB cable
|Memory card in CF card reader
|⇒ [FinePix Viewer] ⇒
|⇒ [ImageName.sh custom script]⇒
|⇒ [ImageName.sh Custom script] ⇒
|⇒ [Renaming of directory, deletion of obvious bad ones from thumbnails, Manual renaming in Explorer] ⇒
|⇒ [batch renaming in SilkyPix] ⇒
|YYYYMMDD_PlaceOrEventName/YYYYMMDD_HHMMSS_Something.raf, archival original.
|⇒ [Select thumbnails in FinePix Viewer] ⇒
|⇒ [Browse to wanted dir and select thumbnails in HSV3] ⇒
|⇒ [Open dir in SilkyPix] ⇒
|⇒ [Tools|Improve Image to launch FinePix Studio] ⇒
|⇒ [Click the 'Convert RAW-CCD RAW files using batch processing' button] ⇒
|⇒ [customize settings for each image, then convert all to TIF] ⇒
|YYYYMMDD_PlaceOrEventName/YYYYMMDD_HHMMSS_Something_1.RAG and YYYYMMDD_PlaceOrEventName/YYYYMMDD_HHMMSS_Something_1.TIF
|⇒ [Custom ForceProfile script] ⇒
|⇒ [optional post-processing in PSP or Photoshop, you may loose the Exif info] ⇒
|YYYYMMDD_PlaceOrEventName/YYYYMMDD_HHMMSS_Something.tif, this is your reference work image
|⇒ [JPG conversion script] ⇒ YYYYMMDD_PlaceOrEventName/YYYYMMDD_HHMMSS_Something.jpg, this is your distribution image
|⇒ [downsampling and sRGB conversion script] ⇒ wwwroot/.../YYYYMMDD_HHMMSS_Something.jpg, for web publishing
Here is a script I use as part of my workflow. It is an extension of the script I originally wrote for the Ricoh and works on any digital camera or memory card. It moves the files from the camera/card, renames them by date and moves them to dated directories. It works on RAF/RAW/RW2/DNG raw files, JPG/TIF images, sound files, videos and a few other specific files. You need Cygwin installed on Windows but it works directly on Linux. Read the comments in the file.
2019/07/03 — Published the source code on gitlab (find the ImageNames source code there).
I've been a loyal Fuji customer ever since I first tried Velvia. And Provia. And Sensia. And Reala. And I think I'll stick with them with their new Super CCD as it's the first time I see a sensor that gives me what I expect. And the camera around it is good old healthy Nikon hardware which I've been using since I first touched a camera. The few cons are fairly minor and most may even be solved by a future firmware upgrade, hopefully. Now if only I could afford to replace my whole range of lenses...
Fast forward a few months, August 2007, and Nikon coming out with the D3. I first feel a pang of regret: with its 24x36 sensor, this is the camera I've been waiting for for 10 years. Then when I look at the specs more closely one thing stands out: you cannot use your old 24x36 lenses without strong vignetting. And a few other details like a min ISO of 200 which is a drag if you want wide aperture portraits outdoors. So forget it, I'm happy with the S5 and will stop leering at the neighbors' teenage daughter. At least for a while.
After all this dry talk and ugly test images, what about some real images from the S5 ? Here are some S5 images:
Left: Upper part of the Glacier Noir. Notice the very high contrast of this image, but details are still visible all over.
Right: Jenny climbing at Les Ayes, above Briancon.