Fuji S5 Pro - is good for used 'till today times (like hobby)?

fPrime wrote:

Jared Willson wrote:

fPrime wrote:

In my humble opinion both the D200 and S5 Pro are CCD color kings in their own right. Because of its skin tone color science I consider the S5 Pro to be slightly better for photographing people while, thanks to its higher resolution, the D200 is slightly better for photographing landscapes.

Is it still worth it to shoot the S5 Pro in 2021? Absolutely! There’s a certain image magic to the output of this camera that is both filmic and almost surreal. I kid you not when I say that I would not trade my S5 Pro for a GFX 100. The latter is a digital device that records digital images at high resolution, the former is an analog device that draws images with passion, fire, and life.

fPrime

I’m glad you like your S5, and I’m sure it was and remains a wonderful camera capable of producing great images in the right hands.

However, the claims that CCD’s are somehow better in terms of color or are somehow more “film like” are nonsense. CCD’s have discrete pixels, just likeCMOS. CCD’s convert voltages into a digital count, just like CMOS chips. Manufacturers spec the Bayer filter, including the colors of the dyes and the amount of overlap between colors. They can and must do this for each new chip design, CCD or CMOS, as different chips have different sensitivities. They also pick the strength and cutoff of the IR filter (if present) as well as the strength of the low pass filter (if present). Then the software takes over in the RAW conversion (whether in camera or in post) and re-maps all of the colors based on algorithms developed by the manufacturer and/or software vendor.

So, there can be and are differences in color response, color accuracy, micro contrast, and aliasing from one sensor to the next based on the choices of the manufacturer and the software vendor, but the process works the exact same way for CCD’s and CMOS. Neither is any more or less “digital”—they are both the exact same, converting photons to analog voltages, then reading out those stored voltages as digital counts. Neither has necessarily better or worse colors. And certainly, there is nothing about CCD’s that creates passion, fire, or a “filmic” look. Passion and fire exist in the photographer alone, not in the camera.

Hi Jared,

The reason I said the S5 Pro renders images that look more filmic are all directly related to the sensor. I don't know about you, but I recall 35mm film as being somewhat low res and grainy, but with great dynamic range particularly on the high end where highlights rolled off softly instead of clipping. This is exactly what we have with the S5 Pro:

• At 12MP it's low res and slightly soft due to interpolation from 6MP.
• It's a bit grainy after applying strong sharpening to offset the softness.
• The unique S+R pixel configuration allows it to render a soft highlight roll off just like film.

Back when we were all shooting 35mm film, we generally got lower resolution shots. While photographic film is capable of holding nearly as much detail as the current crop of sensors (depending on chemistry, of course), it rarely did so in practice. Aspheres didn't exist, so corner performance was generally not as good. Multi-coatings were not as good, so flare, ghosting, etc. were worse. Focus was rarely all that accurate, and there was a loss in acutance compared to the best modern sensors due to the simple fact that the film "plane" actually had a bit of depth to the various dyes. So, yes, I agree that 12MP--or even 6 MP--often seems to capture about the same level of detail as I routinely achieved with 35mm photographic film. Of course, anyone interested in achieving as much detail as possible probably wasn't shooting 35mm in those days.

So, if what you meant by "filmic" is lower resolution and increased grain/noise compared to a more current digital camera, I would agree that the S5 was more filmic. As to the soft highlight rolloff, I'm not really in a position to say. I had a D200 back in the day, but never owned an S5 so can't speak from experience there.

I’m not telling you a more modern, higher resolution camera is necessarily better. If you like the output of your S5 and you know how to get the results you want out of it, more power to you. Keep it, and don’t trade till you have to. But to attribute the colors to it being CCD based? Or to suggest it is somehow less digital than a newer camera? Or to ascribe emotions to it or it’s output? Sorry, there is just no basis for that. It’s poetic, but it’s not real. Your S5 is EXACTLY as digital as a GFX. It does not create magic, passion, or fire. If you are a good photographer, YOU create magic, passion, life, and fire. It isn’t the chip.

I disagree... the chip matters. As stated above, the SuperCCD chip has a unique ability to render highlights. But the colors are also quite unique as well thanks to Fuji having used a special CFA for the S5 Pro. This CFA seems to have been specifically tuned for skin tone rendering over other photographic considerations like landscape. And color nuances are intrinsic to the sensor. Here are five custom calibrated cameras shooting the same skin tones. Are any alike? Are some better than others?

Image credit to Saltydogstudios

So to summarize, I'm not saying that CCD chips have any intrinsic color advantages over CMOS chips based simply on their silicon. I am saying instead that the strong CFA's used in some older CCD cameras like the S5 Pro were optimized for color fidelity. Today's weaker CFA's seem to be more optimized to support high pixel densities and high-ISO performance.

fPrime

I didn't mean to suggest the chip has no effect on the image. Of course it does. Just that the technology of CCD vs. CMOS doesn't. With every new chip family, whether CMOS or CCD, the chip manufacturer must figure out the appropriate Bayer dyes. The width of the bandpass varies. The sensor sensitivity by wavelength varies. These are not interference based color arrays, so there is no hard "cut off" at a given frequency. So, definitely, the hardware matters, it's just that the claims of CCD's giving better colors than CMOS are not accurate as they face the exact same challenges. So, I think we are on the same page there.

As to the rolloff as you approach full well capacity (white point), there is also some variation. The anti blooming gate function built into both CCD's and CMOS result in a loss of linearity near the white point, just as with film, though it isn't the same shape curve. In addition, no RAW file is actually RAW, so additional curves are applied between what the camera captures and what you see. Generally, that's a good thing. Makes for much less work in post. Ultimately, though, the dynamic range in the S5 even with its dual pixels is significantly lower than a modern sensors--both in terms of noise robbing the chip of sensitivity in the shadows, and in terms of full well capacity in the highlights.  It's not really even close. At its best, the S5 has eight stops of dynamic range.  An X-T4 has a little over ten.

Honestly, I don't recall older CCD's fondly with regard to color fidelity, though I admit to not having owned an S5. As to the samples you provided, obviously they are different in terms of skin tones. I find two of them have significantly more green than I would like, but that's me. There are also differences in contrast and tone/brightness, not just in color. If you like what the S5 does in a portrait shoot, by all means sing its praises. Certainly we all get far too obsessed over minute differences in detail and contrast that have exactly nothing to do with what constitutes a good image. I'm sure the S5 was as and is a great camera.

To summarize, yes different sensors and different software implementations (in camera JPG conversion or RAW conversion in a PC) have a profound effect on colors and tones. The choices the chip manufacturer or camera manufacturer make in terms of CFA have a significant effect as well. So there will be differences from one chip family to the next and from one camera manufacturer to the next. It's just that one shouldn't attribute those differences to CCD vs CMOS or to the year the chip was made or anything other than the decisions of the engineers in how to implement the CFA, adjustment curves, debayering algorithms, and color adjustments. Those determine the end result.