**M9 - 30+ mpix**

[lxlim]

Advertisement (gone after registration)

At 1.1x, you are talking about differences that fall within the margins of error anyway.

 

Leica "50mm" lenses are often close to 53mm in true focal length - and most others have some variation, also.

 

The extension factor you get focusing at 1m or .7m instead of infinity is about a 15% change (which is why the framelines are so rough at infinity - they have a built-in crop factor of 1.15x compared to the captured image at infinity).

 

Wide lenses tend to project the image slantwise "under" the edges of the film gate on film, making the final negative size larger than a frame shot with a 90 or 135 - the old 21mm Super-Angulons and 15 Hologons produced negs that sometimes impinged on the sprocket holes, they were so much larger. Which amounted to about a .95x crop factor compared to the longer lenses - from the same camera.

 

Slide mounts or negative carriers (unless filed out) impose their own crop factors of 1.05x or so.

 

I've shot Mamiya 6, Mamiya C330, Hassy, and Rollei TLR over the past decade, and every single one of them produced a slightly different quote-6x6-unquote negative size - none of which came close to 6cm x 6cm. Or even a perfect square. Varied from 53 x 55 to 54 x 54 to 55 x 56.

 

I imagine if you measured Nikon, Canon, Leica R, Leica M(21mm) and Leica M(90mm) negs/slides, you'd get a fair variation in actual image sizes as well.

 

Which raises the question "Whose full-frame? And what is it?" Exactly, precisely 24mm x 36mm? - or 23.5 x 35.5mm? - or 23mm x 35mm? - or any mixture therein?

 

I agree with those who think Leica won't bother with another "intermediate" format. When they can do "full-frame" - and I'll count 23 x 35 as full-frame - they'll do it. Until then, they'll stick with 1.33x.

 

Won't be anywhere near 30 Mpixels, either.

 

-------

 

(Edit) I just measured some old negs out of the same M body - "135" neg measured 23mm x 34.92mm, "21 S-A" neg measured 24mm x 35.7mm. From a Canon F-1: 23mm x 34.5mm. Tell me which of those is "full-frame" - and I'll tell you the "M9's" crop factor.

 

Thanks, Andy. You really know your film tech. When I worked with 4x5s I just made the shots and never worried about the edges nor the actual dimensions.

 

I know I can live with a 1.1 crop but it will take a little adjusting.

[jrc]

The same pixel size combined with newer technology could mean big improvements in the IQ department, which to me is far more important than a higher (30MP+) resolution. I hope for ISO 100-3200, with 1600 being clean-ish and 800 being perfect. I don't think this is unreasonable, given how good the M8 already was.

 

I'm with Carsten on this. I think it could be the M8 sensor expanded to FF, giving 18mp+, but with better tech around it to boost ISO performance. I would hope for 1/2 stop more than Carsten -- a perfect 1200, with 1600 being clean enough to be very usable with post-processing, and 3200 okay for emergencies.

 

I wouldn't be surprised if Leica engineers have spent the last four years or so working with the M8 sensor, doing various tweaks and trying out new tech on it -- remembering that Leica doesn't have either the financial or engineering resources to make the giant leaps of Canon or Nikon. Extending development of a known resource is much cheaper than leaping into the unknown.

 

I would also hope for other operational improvements, but I doubt that we'll get them. I would like focus confirm of some sort, to make the longer lenses more usable, and electronic framelines, for better framing, but I don't think we'll get either. If all we get is FF with minimal (one stop or less) ISO improvements, then I'll probably wait for the M10. Or not.

 

JC

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[pop]

Is this rumour correct?

 

That's mostly a question of the definition of the term "rumour".

 

Any statement repeated by "the public" which is not based on fact is a rumour. Hence, it is a "correct" rumour that Leica will anounce a new M model, that the new M model will have a FF sensor and that this sensor will have a gazillion of subpixels.

 

Wouldn't it be just too bad if the new model was a M8.n with a pixel geometry comparable to that of the Foveon, with true pixels at a density of roughly 10 "true" MP?

[like_no_other]

3x10 WOULD be a clever calculation indeed.

But the rumour is about common 30.000.000 bayer pixels.

[dkCambridgeshire]

Supposing they have a new Foveon sensor in the M9?

 

Cheers

 

dunk

[stevelap]

......it comes out at 24MP. If it used the 6.8 micron sensor, that would be 18.7MP.

 

Thanks for the clarification Mark.

 

Given the comments Leica has previously made about leveraging S2 technology, I would have thought that variations on the Maestro processor and new Kodak sensor theme would at least be likely for any new digital M. Especially given the concurrent development that must have taken place, if a FF M9 is indeed to be announced in ten days time and be available alongside the S2 shortly afterwards.

[Guest guy_mancuso]

Advertisement (gone after registration)

Personally I was hoping for the 6 micron and maybe not the case since I don't know how they would work it out to 18 mpx but given the firmware is already made for the S2 and that new technology would also have less moire. I have tested that difference on moire between 6 and 6.8 and 6 is less moire. Also the files from each system would match and they can use the same firmware basically as far as the profiles and such. I also believe it would be more apt for a higher ISO but maybe Leica found a way on the 6.8. Maybe a moot point.

[markgay]

Just dropped in and had to wipe the spittle off my face.

 

Hear people cussin' Leica, hollerin' they'd better provide this and that - or else!

 

There's nothing wrong with wish lists but from the tone it seems some people are squaring their shoulders and readying their fists.. against Santa Claus!

 

Puts me in mind of a

.

 

Just substitute full frame for hippopotmus.

 

I'm off to rob a bank.

[Guest Colt Seavers]

Not a rumour, my dealer says 39 Megaperls with an optional extra of 12 Megaperls for 1.500 bucks more

[markgay]

I'm ignorant of any significant advances in sensor technology. Perhaps someone can enlighten me.

 

As far as I understand, the only significant improvement with sensor performance is due to size. Which suggests we already should know the optimal number of m-pixels on a "full frame" sensor.

 

This guy (Stephen Scharf commenting in this thread) argues that the laws of physics mean you cannot alter the basics of photon size-pixel size.

 

Ie, the size of the pixel 'bucket' is fundamental and no amount of clever technology can get around the fact.

 

Most manufacturers are simply playing with the software to make basically the same sensors look as if they are producing better pictures.

 

Many new cameras - at least the premium ones - are going back to the basics: going back to the traditional camera size so they can accommodate a bigger sensor and a bigger, faster lens.

 

There are no significant sensor advances that affect basic image quality.

 

Here's the quote:

 

In low available light, the chances of a given pixel capturing photons become a Poisson statistics-driven event, and the shot noise from the gain on the sensor at higher ISO really degrades the image quality. There is no way around these physics-based limitations, ultimately. The only way around this is to go back to larger pixels in the same size sensor (so that the bucket capturing the photons is larger), or increase sensor size the way the Samsung DP-1 has. I really am sorry that they threw in the towel on this superb sensor and fell victim to the megapixel race.

Like many photographers who have owned and shot with these little gems, I'd rather have a clean beautiful 6 megapixel JPEG file than a noisy 12 megapixel RAW file with limited dynamic range any day of the week.

 

He's talking mainly about compact cameras - because that is where sensor size is more limited and the efforts of manufacturers to boost m-pixels most ludicrous - but the argument stands for larger sensors.

 

Mark

Edited August 30, 2009 by markgay

[like_no_other]

...

There are no significant sensor advances that affect basic image quality.

 

...

 

Mark

 

This is not quite true. According to your explanations the light receiving bucket size is important. This is increased by newer technologies that remove hiding wires from the front to the back site of the sensor where they don't cover light sensitive area.

That the manufacturer 'misuses' this to further increase pixel numbers or make the whole sensor smaller for the same pixel number is another story of course.

 

You addressed dynamic range as a function of single pixel size. The differences between 6, 7 or 8 micron sized pixels are small compared to what you can gain if you have an in-camera HDR feature that makes two shots very quick in a row and automatically combines them into one picture while shooting hand-hold without tripod and being able to adjust the EV difference up to 3 EV between the two pictures. This is a function implemented in a new DSLR and seems to work good. It's a workaround of course, but it works.

[SJP]

This is increased by newer technologies that remove hiding wires from the front to the back site of the sensor where they don't cover light sensitive area.

This is only relevant for small pixels like in P&S, for the M8 pixel size this technology will not contribute anything significant (less than 5% increase in light gathering capacity).

[gwelland]

You addressed dynamic range as a function of single pixel size. The differences between 6, 7 or 8 micron sized pixels are small compared to what you can gain if you have an in-camera HDR feature that makes two shots very quick in a row and automatically combines them into one picture while shooting hand-hold without tripod and being able to adjust the EV difference up to 3 EV between the two pictures. This is a function implemented in a new DSLR and seems to work good. It's a workaround of course, but it works.

 

I personally think that Fuji had the right idea with SuperCCD combination of sensor elements to extend dynamic range. It's just a pity that it's still not quite realized in practice beyond their cameras, JPG processing and software. I suspect that someone like Phase One could have probably extracted much better images from their RAW data.

[like_no_other]

This is only relevant for small pixels like in P&S, for the M8 pixel size this technology will not contribute anything significant (less than 5% increase in light gathering capacity).

 

That's my opinion as well. But it was meant as an answer to a general remark.

This 5% value is new for me. It's the first time that I read a concrete value for larger sensors at a forum.

I wasn't sure about how insignificant the value would be because I thought that we find a complete layer with integrated wires on the top that makes all indiviual pixels lying deeper in their 'caves'. If you should know more about this technically it would be welcome if you could share a little bit more of your knowledge.

[SJP]

This is based on a guestimate I posted a few days ago. For smaller sensors the difference is significant. However the pixels are very small in the sensors involved. Using this you can calculate the surface area occupied by the electronics and rescale to the M8 pixel size. If I recall correctly the electronics occupies 1.2 x 1.2 micron (squared), so with 6.9 micron (squared) the M8 pixels are a lot bigger. Some simple arithmetic then yields 1.2 ^2 = 1.44 and 6.9^2 = 47.6, and 1.44/47.6 = 3% or so.

[4season]

Just curious, but how badly do you guys really want a FF sensor? Badly enough to replace your 24mm and wider lenses if need be?

 

There's still other areas for Leica to explore within the confines of the 1.3x crop format. Improved low light/high ISO performance for one. 16-bit output for another. And who wouldn't mind a slimmer profile and maybe an OLED screen? Better battery life, a counter that goes beyond 999? Better moisture resistance?

[gwelland]

Remove the IR filter dependency, extend ISO & DR support and I'd be pretty happy. FF would be nice but not really a deal breaker.

 

Ergonomically the size is no problem IMHO. In fact I'd rather it stay the same size so that my existing grips & cases would still fit. A detent to avoid hitting S mode by mistake would be useful though.

[ennjott]

He's talking mainly about compact cameras - because that is where sensor size is more limited and the efforts of manufacturers to boost m-pixels most ludicrous - but the argument stands for larger sensors.

It doesn't. There is some point at which the signal really falls apart, but before that, the size of the pixels don't matter.

 

Noise depends on sensor size, not on pixel size. A 30 MP sensor has the same amount of noise as a 12 MP sensor at the same print size.

[t024484]

It doesn't. There is some point at which the signal really falls apart, but before that, the size of the pixels don't matter.

 

Noise depends on sensor size, not on pixel size. A 30 MP sensor has the same amount of noise as a 12 MP sensor at the same print size.

A 30 Mp Sensor has 4dB more noise compared to a sensor with the same size having 12 Mp.

 

Hans

[Advertisement]

Hallo,

Look here: Leica M