Canon LF sensor

[jp]

http://www.canon.com/technology/future/cmos.html

This would rock in a RB SLR...

[Willie]

Maybe Hasselblad can get in line and try to restore some of its lost reputation as a Square format camera?

[Bob Salomon]

What do you mean an RB? This size could work for 5x7 or 4x5!

[Oren Grad]

RB Graflex, Bob. Think Digital Home Portrait, or Digital Super D.

[Bob Salomon]

RB Graflex, Bob. Think Digital Home Portrait, or Digital Super D.

Why would that be of interest? Limited lenses, lots of vibration, slow, there would be far better choices. Unless one is trying to do old time photography with modern tech. But the cost of that sensor would probably put it well out of the range of photographers!

[Oren Grad]

This particular sensor isn't going anywhere directly relevant to us. But since we're on a flight of fancy anyway, why not imagine a Graflex-like camera optimized for it? It doesn't have to be an actual 1923-vintage Home Portrait, with its assorted foibles.

On current trends, though, even high-end systems are moving in the direction of live view off the sensor, rendering the mirror box moot. So I guess we can lose the "flex" and have a Mirrorless Home Portrait Gra, or Mirrorless Gra Super D. They could still be "RB", though, if the mega-sensor isn't square.

[Leszek Vogt]

To be or not to be RB, eh ? I could be repeating, but the cost will likely be prohibitive and managing such size files
or stitched ones (gowd forbid) could be a sizable assignment....maybe a room of computers ? Yet, optimistically
we may see some competitive action....and that could effect the pedestrian sector.

Sure, my Yashi 124G could use one of those :>).

Les

[jp]

The sensor is fast enough for handheld shooting (designed for moonlight), and the Graflex RB is perhaps the easiest handheld LF camera series. It'd be kinda of a creative waste to let be strictly for tripod and tethered use. It would be disruptive, in a good way, to LF tradition if it is fast and good enough to use LF lenses without a tripod.

The 3x4 RB is a 4x4 camera with a 3x4 film holder on it.. 4x5 RB is a 5x5 film camera, etc... It could skip the mirror and have an electronics viewfinder (such as an ipad mini) of course surrounded by flocked chimney if we wanted. I do love square.

[Mark Sawyer]

Why would you use an RB (Rotating Back) with a square format?

[Havoc]

Sure, my Yashi 124G could use one of those :>).

Les

You mean such a sensor could use about 16 of those puny Yashicas. :D

[mmerig]

Maybe I missed it, but there is no information on pixel size for the big sensor. Given the trade-off between pixel size and light sensitivity, the pixels are probably quite large, so the resolution may not be that great. I think their goal was a very sensitive sensor, and to have reasonable resolution, they had to make it big.

[Havoc]

This is meant for scientific and industrial use. Frm the article it was used for astronomy at the KISO observatory in Japan. (check their site, thei have a great scanner for plates) So they probably also cooled it to 170K or something like that.

[Pere Casals]

Why would you use an RB (Rotating Back) with a square format?

PD: this information is wrong:

Mark, in this case R is for Reflex not for Rotating. Graflex RB vs Mamiya RB, they are using same letters from different words.

[Pere Casals]

http://www.canon.com/technology/future/cmos.html

This would rock in a RB SLR...

Big sensors are used in astronomical telescopes.

ESA's Gaia mission telescope had 106 MF sensors combined in a composite sensor covering 3800 cm2.

[Eric Woodbury]

When you order your copy of this sensor, don't forget to order a lens with 300 mm image circle and a usable aperture of f/1.6. Diffraction limited.

[Mark Sawyer]

Mark, in this case R is for Reflex not for Rotating. Graflex RB vs Mamiya RB, they are using same letters from different words.

Now you've got me curious, as I used to use an RB pretty regularly... None of Mamiya's other reflexes (Mamiya 645, C220, C-330, or other TLR or 35mm SLRs) used the letter R to distinguish that feature. And Camera-wiki.org states, "The naming of the camera RB67 stood for Rotating Back 6 × 7", (http://camera-wiki.org/wiki/Mamiya_RB67#Bibliography).

Do you have any source for RB standing for Reflex, not Rotating or Revolving? And what did the B stand for?

Either way, an 8x8-inch digital back would be pointless on a 6x7-cm camera...

[Pere Casals]

Now you've got me curious, as I used to use an RB pretty regularly... None of Mamiya's other reflexes (Mamiya 645, C220, C-330, or other TLR or 35mm SLRs) used the letter R to distinguish that feature. And Camera-wiki.org states, "The naming of the camera RB67 stood for Rotating Back 6 × 7", (http://camera-wiki.org/wiki/Mamiya_RB67#Bibliography).

Do you have any source for RB standing for Reflex, not Rotating or Revolving? And what did the B stand for?

Either way, an 8x8-inch digital back would be pointless on a 6x7-cm camera...

Mark, sorry, I was mistaken, the Graflex RB is for Revolving Back. https://graflex.org/articles/series-d/

I relied in a bad source.

[Havoc]

When you order your copy of this sensor, don't forget to order a lens with 300 mm image circle and a usable aperture of f/1.6. Diffraction limited.

I understand the need for the image circle, but why would it need to be f/1.6? Certainly if the sensor is touted to be usable in 0.3 lux scenes.

[Eric Woodbury]

Lenses are limited in their ability to focus a point. It's called the diffraction limit. The faster the lens, the smaller the spot you can focus. The sensor in question here has 2 micron pixels. To get an infinitely small spot to focus into a pixel that size with green light (at least most of it, first null of the Airy disk), requires this insane f-number. There are some ways to get around this problem, but not practical for normal humans.

BTW, they still fly film in U-2s. Film is good stuff at a bargain price (relatively speaking).

[Pere Casals]

Lenses are limited in their ability to focus a point. It's called the diffraction limit. The faster the lens, the smaller the spot you can focus. The sensor in question here has 2 micron pixels. To get an infinitely small spot to focus into a pixel that size with green light (at least most of it, first null of the Airy disk), requires this insane f-number. There are some ways to get around this problem, but not practical for normal humans.

BTW, they still fly film in U-2s. Film is good stuff at a bargain price (relatively speaking).

f/1.6 is not necessary, 2 micron lines is 4 micron for a line pair, this is 1/0.004 = 250 lp/mm. Difraction limit at f/5.6 is 283 lp/mm, so in theory a f/5.6 lens can outresolve that giant sensor.

A suitable a lens for that sensor (if we want matching resolving power) may weight around 30kg, with best sharpness in the center at some f/4, and acceptable performance across all the field by some f/11, if one takes a look to some HiRes aerial designs that were popular.

But it would not be necessary to use a lens that outresolves the sensor. A Nikon D3400 also has usually much more pixels than the popular lenses are able to resolve...

[Havoc]

Ok, what are the f stops for a reflector telescope this would typically be fitted to? I'm serious, are these calculated similar to refractor lenses?

[Pere Casals]

Ok, what are the f stops for a reflector telescope this would typically be fitted to? I'm serious, are these calculated similar to refractor lenses?

I'd say that in a reflector the primary mirror diameter is taken to calculate the aperture, the rest should be the same.

Sky explorers like to speak in arc terms, but here you have it well explained: https://www.astronomy-electronics-centre.com.au/article_resolvingpower.htm

[jp]

But it would not be necessary to use a lens that outresolves the sensor. A Nikon D3400 also has usually much more pixels than the popular lenses are able to resolve...

For once I heartily agree with Pere Casals! A photographer or videographer would use a lens that allows them to makes the photos in the style they want.

[mmerig]

Lenses are limited in their ability to focus a point. It's called the diffraction limit. The faster the lens, the smaller the spot you can focus. The sensor in question here has 2 micron pixels. To get an infinitely small spot to focus into a pixel that size with green light (at least most of it, first null of the Airy disk), requires this insane f-number. There are some ways to get around this problem, but not practical for normal humans.

BTW, they still fly film in U-2s. Film is good stuff at a bargain price (relatively speaking).

The description for the large sensor does not say anything about pixel size. The 2.2 micron pixel size is mentioned in the 120 MB sensor, which is quite small (APS). If the large sensor (400 cm^2) had 2.2 micron pixels, there would be over 8 thousand megapixels. Even if it isn't color, the data coming out of it would take a very long time to get into a laptop or other storage devise

So the calculations for a suitable lens that follow are likely based on a too-small pixel size. Pixels could be > 10 microns, to be sensitive enough to the very low light levels.

Maybe someday, there will be 400 cm^2 digital sensor with 2-micron pixels that have low noise and are not prone to saturation, but save your egg money and pass on this one -- it is a very poor fit for normal large format photography.

[Pere Casals]

The description for the large sensor does not say anything about pixel size. The 2.2 micron pixel size is mentioned in the 120 MB sensor

... yes, but the article says that this sensor will be employed an astronomical Schmidt camera (https://en.wikipedia.org/wiki/Schmidt_camera).

In this kind of projects pixel size should be matched to the resolving power of the optical intrument.

[mmerig]

... yes, but the article says that this sensor will be employed an astronomical Schmidt camera (https://en.wikipedia.org/wiki/Schmidt_camera).

In this kind of projects pixel size should be matched to the resolving power of the optical intrument.

Maybe the pixel size is matched to the resolving power, maybe not. One thing the sensor may not be matched to is the curved focal plane that normal Schmidt cameras typically have, although the one they are using probably has a field flattener. But the flatteners can degrade resolution.

The Canon articles are apparently for marketing, and without much more information, how well the sensors would work is open to a lot of speculation; I am not going to speculate any further.

[Pere Casals]

Maybe the pixel size is matched to the resolving power, maybe not.

how well the sensors would work is open to a lot of speculation; I am not going to speculate any further.

You are right, we are speculating, we only know that the other CCDs used in the Kiso observatory are 24um and 17um, refrigerated and having low dark current levels. Other sensors in the VLT are 10um...

...Canon will make the pixel size that customers will find most useful...

I guess that Canon is presenting a new product line that is primarily targeted to astronomical applications, where the common choice is joining MF sensors in a mosaic when wanting to obtain really big sensors.

I guess that a composite sensor should need 2 exposures per image, being the second image a bit shifted, to fill the separations between sensors, perhaps they shift the sensor rather than moving the telescope. This way allows making monster sensors from industrially available components.

Of course that cannon development would simplify sensor developments and operation of telescopes not requiring monster sensors like the one in the Gaia.