GFX100, a 15 fold increase of value compared to an XF IQ350 (per Bernard)

[mcbroomf]

No problem, but I understood what you meant and stand by my response .. yes a bit faster/easier but still much more time than you think.

[eronald]

No problem, but I understood what you meant and stand by my response .. yes a bit faster/easier but still much more time than you think.

I will try and get hold of someone who sells semi-custom sensors and ask them.
This seems the best way to get a handle on the realities of this business.

Edmund

[ErikKaffehr]

Hi Dan,

As far as I understand, sensor development has a three year cycle. So, design of the present design has probably started 3-4 years ago.

I would guess that most development is actually done small size imagers, like cell phone cameras, but the technology migrates to larger sensors and larger pixels.

Sensors are also a compromise. At least 24x36 sensor need to handle both stills and video. My understanding is that the around 60 MP is one that is optimized for video.

You are of course right that reusing the pixel design of the 102 MP 33x44 sensor would yield 61 MP on 24x36 mm.

Regarding the camera market, it may around like 8 million interchangeable lens cameras a year. Lens rentals reports around 4.2% Fujifilm market share for their customers. So, using those figures as an input I would guess that total Fujifilm camera sales are more like 350000. I guess 20000 out of 350000 may be a realistic figure.

Best regards
Erik

Hi Erik-
     Sounds about right, although Fuji probably needs to sell somewhat more than 20,000 cameras/year to support all that lens development (and three bodies). How much more? I don't know... 1/4 GFX 100 is kind of what I'm thinking, too...

     Does this count as an existing or a new pixel design? It is a new design, but shared among three sensors - the 33x44mm, the presumably very low-volume (unless there's some use that isn't photographic in our sense) IQ4 150 sensor and (importantly, because it could be a lot of volume over time), the X-T3 sensor, which will probably migrate into Sonys, Nikons and Pentaxes as well. There's a datasheet for a fourth sensor with this pixel, which we haven't seen in a camera yet - 24x36mm, ~65 MP. How much is medium format paying for the development of the pixel?

Dan

[BJL]

Not only is it not a new pixel design, but it's part of a whole family of chips with the same interfaces. In my view, it's a trivial scaling down of the 150MP chip. If I were feeling unpleasant, I would say that Sony were probably able to finalize the design for the 100MP in one day after doing the 150, by changing just a few parameters in a CAD macro file.

Edmund

I agree that the two new sensors, 100MP 44x33 and 150MP 54x40 might share a large proportion of R&D costs. But it is not as if Sony is going to recover them all with the mark-up on the (far lower volume?) 150MP sensor and then sell the smaller sensor at unit production cost plus a small margin. I would guess the opposite; that revenues on the 44x33 will have to cover the great majority of those costs.

Still, my bottom line is that, as usual, I trust Fujifilm’s business knowledge and decision-making over internet forum handwaving, guesswork, hearsay, 40-year old experience plus cynicism, but I also know that I will never persuade some self-styled teams of online experts on this!

[Dan Wells]

Isn't most of Sony's volume on that pixel design not going to be medium format at all? We haven't seen the 26 MP X-T3 sensor anywhere besides the X-T3 so far, but I suspect Sony will be using it themselves, probably starting with the new APS-C body that's rumored for announcement in the next few weeks. Fuji will almost certainly use it in everything for the next few years (they try to reuse APS-C sensors repeatedly to avoid making new X-Trans arrays). I'm sure Sony is hoping for large orders from Nikon and at least small orders from Pentax.

We haven't seen the ~61 MP sensor at all yet, but I'd strongly suspect it of showing up in the A7rIV first, then possibly another Sony (a fixed lens RX1, perhaps? Maybe a very high-end body as well?). Nikon will probably use it in D850 and Z7 replacements. If it's easy enough to replace the older 36 MP sensor with, Pentax may very well throw it in a K1 successor.

Each of these bodies individually (with the possible exception of a fixed-lens Sony or some of the Pentaxes) will probably outsell the combined sales of everything that uses the 100 and 150 MP sensors...

Dan

[BernardLanguillier]

We haven't seen the ~61 MP sensor at all yet, ... Nikon will probably use it in D850 and Z7 replacements.

Out of topic in this thread obviously, but I think that Nikon will first release it in a full spec pro body. It wouldn't surprise me if they did to us a Z9x/Z9. And the Z9x would get the high res sensor while the Z9 would be a mirrorless D6.

If this is the case, the contribution of Nikon to the volume selling of the 61mp sensor may be limited at first. And this may be the deal btwn Sony and Nikon btw. Who knows?

Future will tell.

Cheers,
Bernard

[eronald]

Yeah, the MF is just presales for the coming 35mm market release - but it's also an absolute necessity to have a cutting edge remote sensing/mapping/military recon product out there. And in fact we see the 100MP has already been in use "officially" for something like a year, and we can assume that customers who were willing to pay had it some time before. My belief is that airborne sensors are used in clusters and spares are purchased; so the unit sales in airborne applications are an order of magnitude larger than one would assume.


Edmund

Isn't most of Sony's volume on that pixel design not going to be medium format at all? We haven't seen the 26 MP X-T3 sensor anywhere besides the X-T3 so far, but I suspect Sony will be using it themselves, probably starting with the new APS-C body that's rumored for announcement in the next few weeks. Fuji will almost certainly use it in everything for the next few years (they try to reuse APS-C sensors repeatedly to avoid making new X-Trans arrays). I'm sure Sony is hoping for large orders from Nikon and at least small orders from Pentax.

We haven't seen the ~61 MP sensor at all yet, but I'd strongly suspect it of showing up in the A7rIV first, then possibly another Sony (a fixed lens RX1, perhaps? Maybe a very high-end body as well?). Nikon will probably use it in D850 and Z7 replacements. If it's easy enough to replace the older 36 MP sensor with, Pentax may very well throw it in a K1 successor.

Each of these bodies individually (with the possible exception of a fixed-lens Sony or some of the Pentaxes) will probably outsell the combined sales of everything that uses the 100 and 150 MP sensors...

Dan

[kers]

Where are photos of the GFX100MP?
A lot of talk, but images should speak.

[32BT]

Where are photos of from the GFX100MP?
A lot of talk, but images should speak.

Otherwise they think you want to know what the camera looks like...

[faberryman]

Where are photos of the GFX100MP?
A lot of talk, but images should speak.

How are images from a 100MP camera posted as 800x600 72dpi jpgs on the web going to help?

[BJL]

Dan makes a better point that the pixel design costs will in part be spread over likely future sales of smaller sensors that also use it. However:
- there are other design issues for a larger sensor with longer signal paths, more stitching needed to fab, etc. (Edmund made a big-to-small comparison which could mostly avoid those extra costs)
- Sony is not going to set prices to defray all those costs on sensors in a competitive (vs Canon at least) and more cost-sensitive market, and then sell at “unit cost plus a small margin” in the medium format market where Sony has a monopoly and customers are a bit less price-sensitive.

It comes back to choosing the most profitable price, not any “charity for the privileged” idea that prices should be only modestly higher than unit costs.

[32BT]

Dpreview finished the reshoot. Much better resolving comparison, although the colorprofile in raw is very clearly less than optimal.

[SrMi]

Dpreview finished the reshoot. Much better resolving comparison, although the colorprofile in raw is very clearly less than optimal.

JPGs are still sharper than RAWs. Either Fuji sharpens JPGs more than others, or the RAWs require more sharpening. I do not see such a difference with, e.g., D850.

BTW, when loading  GFX100 raw files into Lightroom they are by default much sharper than what DPR displays in their studio scene.

[32BT]

JPGs are still sharper than RAWs. Either Fuji sharpens JPGs more than others, or the RAWs require more sharpening. I do not see such a difference with, e.g., D850.

BTW, when loading  GFX100 raw files into Lightroom they are by default much sharper than what DPR displays in their studio scene.

What are you comparing exactly?

Note that the 850 and Z7 both show blooming on high contrast edges. Not sure if something was overexposed there, but shows a potential source for loss of microdetail, the kind of effect that would make one choose MF.

[SrMi]

What are you comparing exactly?

Note that the 850 and Z7 both show blooming on high contrast edges. Not sure if something was overexposed there, but shows a potential source for loss of microdetail, the kind of effect that would make one choose MF.

I am comparing results as seen in DPReview's studio scene (black on white text, banknote, ...).

[32BT]

Attached is what I see on my monitor. Not sure if it is just on my monitor.
In image 1 you can see magenta text for both bottom cameras.
In image 2 the camera engines have ameliorated the problem, but the lack of microdetail is very apparent. This has little to do with sharpness settings.

The comparative view is, if I remember correctly, 30mpx max. So all images are downsampled, but that would be favourable to all.

[SrMi]

Attached is what I see on my monitor. Not sure if it is just on my monitor.
In image 1 you can see magenta text for both bottom cameras.
In image 2 the camera engines have ameliorated the problem, but the lack of microdetail is very apparent. This has little to do with sharpness settings.

The comparative view is, if I remember correctly, 30mpx max. So all images are downsampled, but that would be favourable to all.

I was not clear, it seems. I did not want to compare Nikon's cameras with GFX100. Instead, I was noticing the much larger difference in sharpness between GFX100's RAW and JPG than I see between Nikon's RAWs and JPGs.

[32BT]

I was not clear, it seems. I did not want to compare Nikon's cameras with GFX100. Instead, I was noticing the much larger difference in sharpness between GFX100's RAW and JPG than I see between Nikon's RAWs and JPGs.

Ah, yes, like so.

I recall for the E3 I dialed down the default jpg sharpening for sooc files. Fuji seemed a bit heavy handed on the sharpening. I presume the results wouldn't be much different for the jpg engine in the GFX series.

[Dan Wells]

Of course, it's impossible to know how many sensors the aerial mapping/intelligence folks are buying... Some of that work is commercial, for everything from real estate development to agriculture and forestry to ecological science (even those numbers are hard to get, but there's no reason they should be impossible. A largeproportion, however, is classified spy secrets.

I once talked to an engineer from Cray, and I asked him where and how they tend to sell their machines. He said that a lot of them go to everything from universities to the National Weather Service to biotech companies (and at least one baseball team, which I found out years later). When someone like that buys a Cray supercomputer (for millions of dollars) it comes with some Cray engineers, who move into the customer site for weeks or even months, integrating the machine into the work environment and making sure everything is running well. On the other hand, they get orders from Langley, Virginia and Fort Meade, Maryland where they just drop the computer off on the loading dock and some folks in trenchcoats come out and move it into the building...

[eronald]

Of course, it's impossible to know how many sensors the aerial mapping/intelligence folks are buying... Some of that work is commercial, for everything from real estate development to agriculture and forestry to ecological science (even those numbers are hard to get, but there's no reason they should be impossible. A largeproportion, however, is classified spy secrets.

I once talked to an engineer from Cray, and I asked him where and how they tend to sell their machines. He said that a lot of them go to everything from universities to the National Weather Service to biotech companies (and at least one baseball team, which I found out years later). When someone like that buys a Cray supercomputer (for millions of dollars) it comes with some Cray engineers, who move into the customer site for weeks or even months, integrating the machine into the work environment and making sure everything is running well. On the other hand, they get orders from Langley, Virginia and Fort Meade, Maryland where they just drop the computer off on the loading dock and some folks in trenchcoats come out and move it into the building...

I once did an “on the record” interview of one of the creators of the commercial-use DES cipher. When asked about whether it was breakable, he averred that it would need acres of computers to break it.

Maybe the guys you were talking about had big buildings and large budgets, then. Of course they have smaller budgets these days, don’t they? 



Edmund