Second member of LHSA the R10 is confirmed

[Bob Ross]

Advertisement (gone after registration)

You cannot take two sensors and put them side to side to make a bigger sensor. The sensor in the M8 is a single chip with two output channels. Two channels are more or less standard, but there are also sensors with 4, 8, or 12 output channels – would you say that these are really 4, 8, or 12 sensors?

 

If you could just join two sensors to create a bigger sensor, this bigger sensor would only cost twice as much. But as this is impossible, you have to manufacture a chip of twice the size, at production costs much higher than twice as much .

Hi Michael,

I think that you took my explaination of joining two sensors a bit too literally. More accurately is would be printing two sensor units side by side with connecting circuitry. If you look at the Kodak sensor pages, at the production sensors and think in terms of 4/3rds (13 X17mm) as a unit, you'll see the multiple unit pattern. Each unit is an "output channel", as you put it. This is probably due to the stepping engines used to fabricate the sensors. Kodak has FF sensors, but only in their cataloge. BTW joining two separate sensors has been don on some of the early MF backs. Yes, bigger sensors cost more and so will the R9 if it has FF or larger. If a fabricator upgrade the stepping engines then you can add that to the cost, too. I suppose an anology would be those rubber floor matts that have dove-tail connectors to join each mat unit to others to cover a large area. To fabricate the mat for a whole large area, would require new fabrication lines & machines.

Bob

[lct]

Good sense would be FF + AF + IS with the current mount hence backward compatibility.

According to David Young (http://tinyurl.com/3ce4tb):

In the end, the "larger than FF" sensor makes sense to me, if in-camera IS is accomplished electronically (as has been done in various camcorders) as opposed to physically (as in the Pentax K10d). It need then not be a lot larger, and would allow the use of existing lenses. Remember, much of this may be simply good marketing. If we are all worried about compatibility for our existing glass, and they come out with (a) a ff camera with ( IS via electronic shifting on a 5% larger sensor and © AF that is implemented by electrical contacts in the existing mount (as per Pentax) we will all be so relieved, that we are likely to run, en masse, to the stores with our money in our hands!

[ptomsu]

Peter, please could you point me to my "bad and unqualified comments"? As I recall, my previous comments in this area were along the lines that I didn't see the need for a new lens mount, and I tried to explain why. My post above was also in the context of a configuration consistent with the existing R lens mount. So you are correct, no mind shift. I hope I try to comment constructively; I'm very pleased in return to be put right when I am factually in error.

 

Sorry, but I am simply too tired to point you anywhere. I only recall the "discussion" about a square shape mount ...... what type of humor was this then?

[andybarton]

Sorry, but I am simply too tired to point you anywhere. I only recall the "discussion" about a square shape mount ...... what type of humor was this then?

 

Lighten up a bit...

 

The "square mount rumour" was from me (so, take it out on me, not John), and if you took this at all seriously, then I apologise for wasting your time.

 

Given that everything written on these two threads (why do we need two?) is all speculation and rumour, hot air and postulation, it's still an interesting discussion.

 

It seems to me that the R10 is going to have to be twice the size and weight of the R9 in order to get all this stuff inside...

 

I can see that this is going to be a long year - let's hope we can all remain friends at the end of it...

[sdai]

I can see that this is going to be a long year - let's hope we can all remain friends at the end of it...

 

LOL

 

I have question, Andy ... have you or Andreas used the "red card" at all?

 

By the way, I want my pen name changed to "APO-Telyt", can you do that?

[stunsworth]

Perhaps it's time to consider a 'Digital R' sub-forum <grin>

[masjah]

Advertisement (gone after registration)

BARTON: I read somewhere that the new R mount will be square, not round.

HOWARTH: So will the glass.

 

Oh dear, I thought too that this was meant to be a joke - rather like the square wheeled bicycle. Sorry Peter, if this caused you offence. As Andy said, let's hope we can all remain friends.

[sdai]

BARTON: I read somewhere that the new R mount will be square, not round.

HOWARTH: So will the glass.

 

To add more confusion, I've said that R stands for rectangular. LOL

[mjh]

More accurately is would be printing two sensor units side by side with connecting circuitry. If you look at the Kodak sensor pages, at the production sensors and think in terms of 4/3rds (13 X17mm) as a unit, you'll see the multiple unit pattern.

Yes, of couse there are such patterns. If read-out woud proceed one complete row at a time, it would be quite slow, so the sensor is divided into two or more columns, each being read-out in parallel. For example, the Nikon D3 uses 12 read-out channels, so read-out procedes at the speed of a one-channel 1 MP sensor. Obviously, this implies that the sensor consists of 12 largely independent columns. Still, I wouldn’t describe this design as 12 sensors joined into one.

 

Yes, bigger sensors cost more and so will the R9 if it has FF or larger. If a fabricator upgrade the stepping engines then you can add that to the cost, too. I suppose an anology would be those rubber floor matts that have dove-tail connectors to join each mat unit to others to cover a large area. To fabricate the mat for a whole large area, would require new fabrication lines & machines.

My main point was that when you increase the chip size, the yield rate goes down. And the yield rate isn’t inversely proportional to the chip size, but drops much faster. Modularizing the sensor doesn’t help here (as it might with designing the sensor circuiry).

[mjh]

Good sense would be FF + AF + IS with the current mount hence backward compatibility.

According to David Young (FreeLists / leicareflex / [LRflex] R10 Speculation response. (also long)

 

In the end, the "larger than FF" sensor makes sense to me, if in-camera IS is accomplished electronically (as has been done in various camcorders) as opposed to physically (as in the Pentax K10d).

Nice idea, except that it won’t work. In still cameras, image stabilization is employed to avoid motion blur. In camcorders, it’s main purpose is to prevent the movie appearing jerky when the handheld camera shakes. An optical image stabilizer, be it lens-based or sensor-based, will achieve both goals at once and is thus preferred. An electronic image stabilizer only achieves the second end: adjusting the crop from a larger sensor image steadies the image, but it does nothing for reducing motion blur. This kind of image stabilization is thus entirely useless in a still camera (except in movie mode).

[wparsonsgisnet]

I was thinking about a sensor larger than 24x36. If my math is correct, a square sensor that would fit the existing diagonal is only 8 1/3 % larger.

 

That is, the diagonal of a 24x36 rectangle is 43.26662 mm.

 

If a circle is constructed with this diameter, the square that fits it comprises 2 isosceles triangles that are 30.59 on the 2 equal sides. That triangle is 43.26662 mm with 2 other sides of "a" and "a."

 

C-sq = 2 a-sq:

43.26662 squared = 1872

Half of that is 936

The square root of that is 30.59412.

 

This gives a square format that is 30.59412 by 30.59412 mm. The area of that is 936 sq mm. That area is 8.333 % larger than the 864 sq mm area of a 24x36 rectangle.

 

All lenses that cover a 24x36 rectangle will also cover a 30.6 by 30.6 mm square.

 

As it happens, the more I crop portraits of individual persons, the more I end up with square images.

 

Of course, Leica has to buy sensors that already exist. So, the sensor selected may dictate the image geometry.

[carstenw]

Leica does not have to buy sensors which already exist, or at least not completely. The M8 sensor was co-developed with Kodak.

[dfarkas]

Of course, Leica has to buy sensors that already exist. So, the sensor selected may dictate the image geometry.

 

Actually....

 

The VP of Kodak image sensor division also spoke at the LHSA meeting. He explained the hand-in-hand (Kodak+Leica) custom design process for the sensors in the DMR and later in M8. He stated that future camera projects would also be a collaboration.

 

Thus, Leica has never used an off-the-shelf sensor and probably never will, especially considering the custom microlens design and lack of AA filter.

 

The presentation was cool, BTW. He showed electron microscope imagery from the DMR and M8 sensors and showed the differences between the two.The M8 CCD is 1 micron thinner and has 40% more charge capacitance than the DMR CCD.Interesting stuff, really.

 

More folks from this foum should really join LHSA and start coming to meetings and spring shoots.The next one is in Asheville, NC April 30-May 4, 2008.

 

David

[andybarton]

 

More folks from this foum should really join LHSA and start coming to meetings and spring shoots.The next one is in Asheville, NC April 30-May 4, 2008.

 

David

 

Long way to come for most of us, though

 

Interesting stuff about the sensors. Thanks for posting.

[R10dreamer]

I live in Asheville. How would I go about getting tickets into this?

[dfarkas]

I live in Asheville. How would I go about getting tickets into this?

 

You join the LHSA. Membership is $40/year. Go to : :LHSA : : Welcome to LHSA! to learn more about the society.

 

David

[R10dreamer]

Thanks David. I just joined but it was 55 for the year. I think that it will be worth it and I look forward to the Asheville session.

[telyt]

How much color are you expecting when directly shooting into the sun, Doug?

 

Directy into the sun with the 560 isn't good for my eyes but with direct sunlight on the front element I expect this:

 

 

and to compare IS to non-IS designs you'd need to compare contemporaneous designs. I presume that the MTF charts were developed with the IS units locked in the center position. How are they when de-centered?

[sdai]

Doug, you need to "see" how IS works ... correcting the decentered light beam and bring it on axis is exactly its purpose.

 

Here is an interactive flash demo posted on Canon's web site, you can press the on/off button to compare different situations. If you believe the decentered light beam will result in decreased contrast, guess what, without IS when camera shakes ... everything will be off axis.

 

Canon in-lense image stabilizers — for easy photography without blur caused by camera shake

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[telyt]

Doug, you need to "see" how IS works ... correcting the decentered light beam and bring it on axis is exactly its purpose.

 

I have absolutely no desire to use a lens with 15 to 20 air/glass surfaces particularly when I have other means of keeping my lens stable and that lens has only 2 air/glass surfaces with minimal internal reflections. My experience with numerous lenses of several makes is very consistent: given contemporaneous designs, more air/glass surfaces results in degraded imaging performance with backlight. It's particularly seen in shadow and highlight detail and in color richness.

 

One really consistent thing I've noticed about Leica lenses vs. C and N (other than less sample variation) is that the Leica designs typically have fewer air/glass surfaces and also typically show better highlight & shadow detail and better color rendition (saturation and gradation). The correlation is much too consistent to ignore and it matches theory as well.

 

I can make non-IS lenses stable; I can't improve the color rendition of a lens with too many air/glass surfaces. Use Photoshop to increase saturation or to deepen blacks or brighten highlights and you lose gradation. I would much rather have an image stabilization technology in the body where it doesn't muck with optical design.

 

Optical stabilization made a lot more sense with film cameras because the film couldn't be moved to compensate. That's last century's technology and with in-body stabilization the full optical performance of the lens - center to edge, minimal flare or reflections - can be preserved along with allowing much slower hand-held photos.