[Mistral75]

I imagine the software has to be rather dramatically changed. Ripping out all things colour related is probably quite a bit of work. All to be recouped within a small sales volume.

I'm not so sure. In the case of Leica (sorry for speaking so much about them, but they are a good source of comparisons as far as monochrome cameras are concerned), the same firmware updates apply to the regular and the monochrome versions. As if the 'monochrome' firmware was identical to the 'Bayer' one, save for some disabled functionalities.

[house]

Wouldn't it be a simplified pipeline since there is no demosaicing involved anymore? You literally read the numeric value and that's correlated to a grayscale. No need for white balance, no need for pixel averaging to get colour data, no nothing.


If the camera can do a B/W conversion to convert the colour-corrected values to a luminance value, you can skip straight to the point with a monochrome sensor.

Yes it's technically simpler. Whether it's technically simpler or not only matters if you start two hypothetical cameras from scratch though. Pentax has 30 years of colour pipeline development, menu system design hardware adaptation based on colour cameras. This can also be used across many models. So even if the tech is simpler you'd have to redo a lot of the software and that can be more difficult than one imagines. They also have to keep this fork running with maintenance etc. Code tends to be a nasty interconnected web of hard to follow threads. It's rarely straightforward to just cut bits out.

So I imagine a monochrome pipeline and menu is a considerable cost for a low volume item.

---------- Post added 09-17-21 at 09:18 AM ----------

I'm not so sure. In the case of Leica (sorry for speaking so much about them, but they are a good source of comparisons as far as monochrome cameras are concerned), the same firmware updates apply to the regular and the monochrome versions. As if the 'monochrome' firmware was identical to the 'Bayer' one, save for some disabled functionalities.

Yes I'm sure you could develop the software like that but unless it was designed with such modularity in mind it's a considerable effort untangling it later. Rumours are that pentax firmware is new for GR and K-3 III so perhaps they put the architecture in place for a monochrome version.

[Mistral75]

(...) Rumours are that Pentax firmware is new for GR and K-3 III so perhaps they put the architecture in place for a monochrome version.

As far as the Ricoh GR III is concerned it's more than a mere rumour: its firmware is based on Android, whereas the firmware of the previous GRs was not.

[house]

As far as the Ricoh GR III is concerned it's more than a mere rumour: its firmware is based on Android, whereas the firmware of the previous GRs was not.

I remember looking through the files and seeing Linux stuff but it was busybox and such. Was there really any evidence of it being "based on Android"? Android being an os on the Linux kernel but I don't think it has a *nix userspace or does it? I really don't know.

---

Last edited by house; 09-17-2021 at 02:21 PM.

[reh321]

Android being an os on the Linux kernel but I don't think it has a unix userspace or does it? I really don't know.

I’m not sure what you mean - nor why it would matter.
I managed an “HP-UX” {a commercial variant of UNIX} system at one time, and I now have a “Linux” system.
I see no real difference between the two.

[house]

I’m not sure what you mean - nor why it would matter.
I managed an “HP-UX” {a commercial variant of UNIX} system at one time, and I now have a “Linux” system.
I see no real difference between the two.

Sorry I meant *nix not unix and i just found a link stating that busybox can run on Android. So that's neither here nor there regarding Android vs more traditional embedded linux on the K-3III and GR. I might misremember but I don't think we know what os the new cameras run.

[TwoUptons]

I’m not sure what you mean - nor why it would matter.
I managed an “HP-UX” {a commercial variant of UNIX} system at one time, and I now have a “Linux” system.
I see no real difference between the two.

I think the reference is to how easy it would be to get into the firmware and poke at it...

Conventional old-school firmware is notoriously hard to poke unless you're into that kind of thing (like the Canon Magic Lantern folks), but a more open Android/LINUX approach would make that easier for everyone involved.

Ricoh was even pushing the Theta API hard for a while to get people to develop things to work with it (because it uses an Android base).

As for LINUX vs UNIX, I'll bet whoever paid the bills noticed the difference between LINUX and UNIX

Where I work, we moved off HP-UX to some LINUX flavor for our non-Windows stuff a while back -- it's apparently worlds easier and cheaper to manage for the things we do with it...

-Eric

[reh321]

I think the reference is to how easy it would be to get into the firmware and poke at it...

Conventional old-school firmware is notoriously hard to poke unless you're into that kind of thing (like the Canon Magic Lantern folks), but a more open Android/LINUX approach would make that easier for everyone involved.

Ricoh was even pushing the Theta API hard for a while to get people to develop things to work with it (because it uses an Android base).

As for LINUX vs UNIX, I'll bet whoever paid the bills noticed the difference between LINUX and UNIX

Where I work, we moved off HP-UX to some LINUX flavor for our non-Windows stuff a while back -- it's apparently worlds easier and cheaper to manage for the things we do with it...

-Eric

At one time there was software called “GO-FER” - from the Univ of Minnesota.
Pre -“WWW”, it was a text-based system which gave text-based access to information.
This UNIX program had some “ifdef”s so it could be tailored for different ‘flavors’ of UNIX.
Unfortunately for me, HP-UX was a combination of them, so I had to go through them one-by-one to make it appropriate for our system.
Being a UNIX-like system does not make developing software for it easy - just slightly less hard.

[Smolk]

I would hope it would be a bit more expensive than the X model, not something like 1500€.

Here's a man hoping...
We don't know, meanwhile, whether the lens will be 18/28 or 26/40mm.

[ZombieArmy]

I agree with a hypothetical GR IIIm being priced a bit higher than the regular GR III, because of (i) higher sensor costs due to lower volumes (still under $100 though), (ii) the incremental costs associated with a specific SKU and (iii) Ricoh Imaging seeking higher unit margins from this niche, low-volume product.

However, all the above doesn't justify a doubled price. $100 or $200 more, why not. Twice the price, definitely not.

Try thinking of this problem from the supplier side. Sony gets in an order for what is essentially a new custom sensor. Just like any other chip they would need to setup the manufacturing for it, which costs them time and resources. If Ricoh is only buying a small amount of chips (expecting lower demand) then the price per chip increases dramatically.

These aren't generic parts, they're not expected to have the security that someone else would use or buy them. The only customer buying these exact sensors would be ricoh. So no it won't be 100-200 increase, it'd likely be much more unless Ricoh at some point needs to put in a larger order for these chips to meet demand (or to include in another product) reducing the price per chip.


There's probably other issues that increase the cost as well (firmware, advertising, shipment to stores, etc...) but I'm fairly certain a large reason for the price increase is just because it straight up costs more to make.

[RobA_Oz]

Try thinking of this problem from the supplier side. Sony gets in an order for what is essentially a new custom sensor. Just like any other chip they would need to setup the manufacturing for it, which costs them time and resources. If Ricoh is only buying a small amount of chips (expecting lower demand) then the price per chip increases dramatically.

In principle, though, doesn’t this involve substituting one (near) end-of-line component for another (a plain transparent layer for the Bayer filter), which I wouldn’t have thought was as intensive a change as setting up an entirely new production line? Granted, there is the matter of disruption to the regular process including the development of a modified QM procedure, but these things are presumably made in batches, rather than in a continuous process.

It depends, of course, on what you call “dramatic”.

[ZombieArmy]

In principle, though, doesn’t this involve substituting one (near) end-of-line component for another (a plain transparent layer for the Bayer filter), which I wouldn’t have thought was as intensive a change as setting up an entirely new production line? Granted, there is the matter of disruption to the regular process including the development of a modified QM procedure, but these things are presumably made in batches, rather than in a continuous process.

It depends, of course, on what you call “dramatic”.

The only people that could answer that would be Sony. But I'd assume that any changes to the chip requires unique manufacturing just like with most chips.

[TwoUptons]

In principle, though, doesn’t this involve substituting one (near) end-of-line component for another (a plain transparent layer for the Bayer filter), which I wouldn’t have thought was as intensive a change as setting up an entirely new production line? Granted, there is the matter of disruption to the regular process including the development of a modified QM procedure, but these things are presumably made in batches, rather than in a continuous process.

It depends, of course, on what you call “dramatic”.

I remember there was a Canon EF-mount camera that was manual focus back in the 90s.
A small group of Canon employees would go down the line, grab a couple of cameras, go back to their desks, remove the autofocus parts, fit what were essentially blanks, and package them up.
Canon was really surprised when people wanted the cameras to be cheaper... "but they cost more to make..."
Turns out it was cheaper and a lot easier to just put a MF switch on the regular ones...

But hasn't Nikon traditionally used "mildly tweaked" Sony sensors?
Or at least said they did, so Sony couldn't use exactly the same ones in their own cameras.
So presumably the Sony chip folks are used to small tweaks to a basic architecture.
And I'd expect removing a Bayer array would be easier than most of those changes...

So I'd expect the cost on the chip side for a monochrome sensor to be much less than the cost on the Ricoh end of things.
New packaging, changes to the firmware, testing, a different color plastic ring for the body, etc.
Say it costs $500,000 extra to make the camera and you sell 10,000 of them.
That's $50 of extra cost to Ricoh on each one, so retail would need to be $100 more to make that worth it for all involved...
And who thinks they'll sell 10,000 of them?

-Eric

[Serkevan]

Try thinking of this problem from the supplier side. Sony gets in an order for what is essentially a new custom sensor. Just like any other chip they would need to setup the manufacturing for it, which costs them time and resources.

Nope.
The CFA is basically bolted on top of the sensor. Sony simply skips one step during production.
We can extrapolate that this shouldn't impact cost too much because the presence/absence of an OSPDAF array (similar level of complexity) or the microlens array (different depending on mount) doesn't change the camera costs dramatically.

[ZombieArmy]

Nope.
The CFA is basically bolted on top of the sensor. Sony simply skips one step during production.
We can extrapolate that this shouldn't impact cost too much because the presence/absence of an OSPDAF array (similar level of complexity) or the microlens array (different depending on mount) doesn't change the camera costs dramatically.

Even if that's true it's still a separate production line AND it's low volume.