Is Fujifilm testing the big Sony Sensor?

[hjulenissen]

Actually it's easier to interpolate the original color because each third line has as many greens, as blues and reds...

I don't think that you are making much sense here. We are talking about sparsely sampling a color scene using a CFA in front of the sensor that can only capture a single color channel at any one point. Bayer use a 2:1:1 distribution among green/red/blue while Fuji seems to use a 2.5:1:1 (slightly more green sensels). A simple (suboptimal) approach would be to use interpolation of the channels independently. In that case, Fuji has to cover larger "gaps" in the red and blue channels (2x2) than Bayer.

The CFA chosen by Fuji seems to be "quasi random". Or perhaps more generally, Bayer is "quasi-random" with a repeat pattern of 2x2. Fuji does "quasi-random" with a repeat pattern of 6x6. This makes it less likely that a particular color-contrast pattern will couple well with the CFA pattern, meaning that they can gamble on less AA filtering (for increased apparent sharpness).

I am guessing that _when_ one is unfortunate enough to shoot a scene that "resonate well" with the Fuji CFA, results can be quite bad?

If you wanted to guarantee that no aliasing would ever appear in any color channels (something that cannot practically be done), then the Fuji sensor would seem to demand an even lower (achromatic) spatial cutoff frequency than Bayer (1/4 the sensel pitch)

I wonder what is stopping them from making a truely randomized CFA pattern. Is it production cost or raw development processing cost? Once there is not pattern in the placement of red/green/blue, chances that some pattern in the scene would trigger a patterned artifact would be practically zero. You'd still see aliasing in the form of "unrealistically contrasty pixel values instead of a smoother edge-following function", but it seems that many photographers are perfectly happy with that.



-h

[ErikKaffehr]

Hi,

My understanding is that Fujinon makes most of the Hasselblad but assembly of cameras is now in Gothenburg, Sweden.

Regarding the lenses, all HC lenses are made by Fuji. The initial lens design is often made in Gothenburg but will be refined for manufacture by Fujinon.

Best regards
Erik

This is interesting as I had always thought that the x-trans had a different layout of pixels than the standard Bayer pattern.

If all it is instead is a color filter array over a Bayer pattern then I guess it would be quite possible.

I forgot they already have the glass out there as don't they make some of the Hasselblad glass? 

Paul C

[Theodoros]

But wouldn't this far mounting flange be a real advantage in digital?
Even Rodenstock hottest lenses have some retrofocus these days and some people are expecting for them to increase it even more.
Eduardo

Retrofocus is not relevant, all MF Dslr lenses are "retros". Longer mounting distance makes it harder to design decent WA lenses and causes shorter diameter image circles for the same focal length. It also causes more focal lengths to be designed like WAs as one approaches focal lengths near to "normal". Obviously, the problem is more evident if one uses a small image area. The P645 25mm lens is a good example, it required a complex design with huge (with respect to other MF) front element and they did it a 25mm (although they would like it to be a 24mm) ...eventually they stopped it. I doubt if the 28-45mm zoom will be able to project an image circle able to cover the full 6x4.5 image area, unless if it is zoomed to at least 35mm... (partly the reason why I believe that they won't be using the new FF Sony sensor on the P645).

All Hasselblad H, Mamyia 645 and Contax 645 are of 7 to 10mm shorter mounting flange, the P645 has a mounting flange distance that could serve a 6x6 image area, if one looks at the film days, the wider lenses of 6x6 cameras was of 40mm, while for 6x4.5 they where down to 35mm... The Bronicas are another example, they also had a much longer mounting distance which caused the ETRSi not to offer a wider lens than 40mm...

[BernardLanguillier]

The P645 25mm lens is a good example, it required a complex design with huge (with respect to other MF) front element and they did it a 25mm (although they would like it to be a 24mm) ...eventually they stopped it. I doubt if the 28-45mm zoom will be able to project an image circle able to cover the full 6x4.5 image area, unless if it is zoomed to at least 35mm... (partly the reason why I believe that they won't be using the new FF Sony sensor on the P645).

Is that is the case, why did they go through the trouble of re-designing very recently their 35mm and to make it compatible with FF sensors?

Odds are that their target photographers in Japan think that 35mm is wide enough. P1 is pretty much there as well, their 28mm is already sub-par on 36x48mm sensors.

Cheers,
Bernard

[Theodoros]

Is that is the case, why did they go through the trouble of re-designing very recently their 35mm and to make it compatible with FF sensors?

Odds are that their target photographers in Japan think that 35mm is wide enough. P1 is pretty much there as well, their 28mm is already sub-par on 36x48mm sensors.

Cheers,
Bernard

Could be the reason... one can never be sure (as I said, it's my opinion that they won't, it doesn't mean that I'm sure...), another explanation may be simply to improve their 35mm to the (even) better...  If I was to ever decide for Pentax policies ( ), I would first redesign the body with a new mirrorbox/VF combination so that the mirror would be as short as possible, the shortest possible as to still project the full 6x4.5 film image area, that would allow some lenses (especially UWA/WA ones) to be designed with rear elements (and hence the whole lens) that would extent into the housing and thus solve the original design's handicap...

The other thing I would do, would be to change the balance of the body by re-positioning the grip further towards the mount as to balance the weight better when lens is attached. The third thing I would do (this is irrelevant to the original design's handicap) would be to design a new higher in the line body (yet retaining the "basic" ones) that would take interchangeable backs and would have interchangeable finders... Then I would use both Sony sensors as to create a MFDBs and would additionally offer the MFDBs for other MF camera platforms (or tech cameras) as to attract more customers.... 

Contax 645 is a very good example of how an MF camera should be designed (possibly the reason why it proves to be so popular 10 years after production stopped), not only because its balance feels about perfect or because of the interchangeable finders and backs, but it additionally has the shortest possible mirror so that it allows the lens rear element to be 8.5mm in the body before it messes with the mirror's travel... I believe that one could possibly design a 22mm/f4 lens for the Contax... I am now converting a Nikon 24mm f3.5 PCE lens that will be of Contax mount in two or three weeks time from now... it will even be of full electronic aperture control and have the lens data recorded in the files as well as metering information in the VF and all modes working as if it was a factory lens... Only AF won't work (due to the 24/PCE not being an AF lens), but with 24mm who cares? ....I never use AF with my WAs anyway.

[BJL]

Is that is the case, why did they go through the trouble of re-designing very recently their 35mm and to make it compatible with FF sensors?

Is that why Pentax redesigned the 35mm lens? I thought it was an update of a film-era lens design, so already compatible with the full "645" frame of about 56x42mm.

[BernardLanguillier]

Is that why Pentax redesigned the 35mm lens? I thought it was an update of a film-era lens design, so already compatible with the full "645" frame of about 56x42mm.

That's my guess, yes.

let's see what they come up with at the CP+ in Japan February end. If the camera isn't announced there they may not be one.

Cheers,
Bernard

[ErikKaffehr]

Hi,

I got the impression that Pentax dropped some WA designs that were cropped sensor only. Interesting to see what the future holds…

Best regards
Erik

That's my guess, yes.

let's see what they come up with at the CP+ in Japan February end. If the camera isn't announced there they may not be one.

Cheers,
Bernard

[Theodoros]

Hi,

I got the impression that Pentax dropped some WA designs that were cropped sensor only. Interesting to see what the future holds…

Best regards
Erik

There are two versions of the Pentax 25mm f4 lens, both versions are able to project an image circle that is big enough for the full 6x4.5 image area, but Pentax decided to extend the hood of the second version so that it shades more of the projected image circle... Of course one can always shave back a bit of the hood as to expand the image circle back...

[mi-fu]

i think now there is a definite answer that Fuji no longer produces sensors for their cameras.

From an interview with Fuji senior executive Toru Takahashi on DPReview:

What are the essential ingredients of the Fujifilm X-series?

TT: Product design is a key point of differentiation. We do this by ourselves. So sensor design, although we don’t make the sensors by ourselves. We design our processors, but of course we do not manufacturer them so we require other companies. But [whether we manufacturer a component or not] we stick to designs that we’ve come up with [in house]. So [the sensor in the X-Pro2] is a good example. This is a 24MP sensor that can produce something like 30-36MP equivalent resolution. Design is our strength I think. And lenses. We have very strong lens design capabilities. Lenses, we have our own technologies, we make lenses by ourselves.

http://www.dpreview.com/interviews/6258617860/fujifilm-interview-jan-2016