M9 high iso

[Jamie Roberts]

Advertisement (gone after registration)

{Snipped]

 

I have a feeling that the moment Leica uses a CMOS sensor in an M, everyone here who is dissatisfied with CMOS sensors today will suddenly sing their praises. It happened exactly that way in the Nikon world a few years back.

 

Jeff

 

If it has the same colour response as a DMR, then I will sing its praises; if it has the same colour response as my D3... feh... I'll stick with my M9 or use it for black and white only.

[ho_co]

... Nikon is officially on the record (with published statements) that Sony manufactures the D3x sensor for them. ...

Yes, precisely that: Sony makes the D3x sensor, but on a separate line according to what I've read.

 

Your post #70 seems to say that the Sony and D3x sensors are the same. But Rubén is right, I think: they're not.

Edited April 18, 2010 by ho_co

[Advertisement]

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

[Rick]

Rick,

 

Don't confuse what Leica does with color and what Kodak does. For the most part, all Kodak is doing is creating the color filter array (CFA). After that, it's all up to Leica. As I am sure you know, color is as much psychological in perception as it is purely technical. If you read Puts' review of the M9, you will see a good discussion of this fact. Based on a pure technical measure, Nikon and Canon actually do a better job in nailing color. Leica has shifted the colors in a way that many find very pleasing, but is not as technically accurate. You can see this in color reproduction testing. A pretty graph can be seen here: Head-2-Head Reviews. This actually matches my perception quite well, and explains why the M9 can frequently render skin tones a bit too pink/magenta--note how the reds are shifted pretty far too the magenta.

 

So, while you and I might really like the color rendering, it's not something you can ascribe to the sensor or Kodak--it's just Leica's way of interpreting the color. BTW, this is why some people like the D-LUX series and pay a premium over their identical Panasonic roots--Leica is applying their own color interpretation routines to give the image a look that many prefer. All the technology at the sensor level and IC level is the same.

 

Jeff

 

Jeff, First off I have to say that I'm just not buying into your technical presentation of the M9 color. Your technical explanations are not convincing, not at all what I see.

 

To say, don't confuse what Kodak does with color and what Leica does, makes no sense to me. Ok, what do they do? What ever it is, they both seem to be doing something right because the color looks better than Canon and I assume Nikon.

 

And, I don't get your point when you say, from a technical standpoint Canon and Nikon do a better job nailing color. Where did you get this? And, then you say that the technical measurements explain why the M9 renders skin tones too pink/magenta. Not for me. This problem we saw last fall has been solved now that we have better Adobe raw processor. How are you developing your RAW files?

 

And, of course you can ascribe the color to the sensor. The rendering has to start with something.

Edited April 18, 2010 by RickLeica

[adan]

As to whether the D3x and the Sony 24 Mpixel cameras have the same sensor - that depends mostly on how one defines sensor. Is it just the silicon wafer, or does one include the color filters, IR filters, AA filter(s), cover glass and mount with the electrical contacts?

 

That is pretty much a matter of opinion, without an authoritative answer.

 

Personally, I call the silicon the sensor - and the sensor plus all the other stuff the "sensor package." I suspect under that definition the D3x and A900/850 use identical bits of silicon from Sony's IC foundry (same sensor), but the D3X sensor packages have their own RGB arrays and AA filtering, to Nikon's own specs, added to the basic sensor on a separate assembly line.

[jonoslack]

As to whether the D3x and the Sony 24 Mpixel cameras have the same sensor - that depends mostly on how one defines sensor. Is it just the silicon wafer, or does one include the color filters, IR filters, AA filter(s), cover glass and mount with the electrical contacts?

 

Quite agree - certainly the results seem to me to be quite different, with the D3x offering much much better high ISO at the expense of colour subtlety at base ISO, where the A900 really does shine - certainly the best I've seen in a CMOS sensor.

[rosuna]

Ehh? And where is your source for this information? Nikon is officially on the record (with published statements) that Sony manufactures the D3x sensor for them. Yes, there are differences in the final implementation around processors etc, but the core sensor is the same. Much of Nikon's secret sauce is in the EXPEED processors.

 

Jeff

 

Companies do not publish statements on their private agreements regarding manufacturing or technology contracts. I would like to see those "published statements".

 

The differences between the two sensors' performance (very obvious) aren't due to the processor. Not at all. It is based on the sensor design. CMOS sensors have a lot of electronic components and processing built into the chip itself. Those two sensors are very different. What is "core sensor"? The number of channels for extracting the signal, the way this signal is processed, etc. All that is completely different. And you can see it in the RAW images.

 

This is a proof of how much two CMOS sensors may differ in terms of performance, depending on the design of the chip. CCDs are much simpler devices, and the system as a whole may determine some differences, but the sensors themselves don't (in a substantial way). The only differences may be in color filters, microlenses, etc.

[Alnitak]

Advertisement (gone after registration)

Jeff, First off I have to say that I'm just not buying into your technical presentation of the M9 color. Your technical explanations are not convincing, not at all what I see.

 

To say, don't confuse what Kodak does with color and what Leica does, makes no sense to me. Ok, what do they do? What ever it is, they both seem to be doing something right because the color looks better than Canon and I assume Nikon.

 

And, I don't get your point when you say, from a technical standpoint Canon and Nikon do a better job nailing color. Where did you get this? And, then you say that the technical measurements explain why the M9 renders skin tones too pink/magenta. Not for me. This problem we saw last fall has been solved now that we have better Adobe raw processor. How are you developing your RAW files?

 

And, of course you can ascribe the color to the sensor. The rendering has to start with something.

 

Rick, read carefully what I am writing in my posts. Color is as much psychological perception as it is pure spectral accuracy. Search the web. There are LOTS of sites that test cameras for color accuracy, and the general results are that Canon and Nikon do a good job of putting out good pure color from a purely technical reproduction standpoint.

 

That said, as I said earlier, many people like their pictures to have something more to them than pure technical color reproduction--which is why you and I both prefer what Leica does with the output of their cameras. Does the Kodak sensor have a role to play? Of course, and I didn't deny that--but much of what you see has to do with what Leica decides post sensor capture, and not what Kodak has done in sensor fab.

 

I am using the latest raw developer; the M9 still kicks out a red that is too shifted to magenta for skin tones, and I still shift it back a little. However, that's the only thing I don't like about the M9 color. Look, I thought the Canons produced a skin tone that was too yellow. It's personal preferences as much as it is pure color accuracy.

 

Jeff

[Alnitak]

Companies do not publish statements on their private agreements regarding manufacturing or technology contracts. I would like to see those "published statements".

 

The differences between the two sensors' performance (very obvious) aren't due to the processor. Not at all. It is based on the sensor design. CMOS sensors have a lot of electronic components and processing built into the chip itself. Those two sensors are very different. What is "core sensor"? The number of channels for extracting the signal, the way this signal is processed, etc. All that is completely different. And you can see it in the RAW images.

 

This is a proof of how much two CMOS sensors may differ in terms of performance, depending on the design of the chip. CCDs are much simpler devices, and the system as a whole may determine some differences, but the sensors themselves don't (in a substantial way). The only differences may be in color filters, microlenses, etc.

 

If you would like to see the statements, let me Google that for you:

 

Let me google that for you

 

I agree with you on the CMOS sensor side and how that relates to CCD sensors; that was the original point I was trying to get across with relation to how Leica is handling color output. That said, when you are having someone else fab a chip for you based on a standard core design, you aren't altering a lot of the per-pixel circuitry. If you were, there would be *zero* cost advantage to using the same core design. No, in the case of the same shared core, what's being altered are things like packaging and *possibly* the CFA, although I doubt even that is the case.

 

Look, if we can dramatic differences in color, sharpness, noise and overall image quality with different raw processors, why on earth do we feel the need to ascribe those differences to low-level sensor electronics, when they are much more easily implemented in the supporting ASICS and things like different AA filters?

 

Jeff

[markowich]

very correct. obviously, M9 or any other camera's RAWs look very different in LR and in C1. so the statement 'camera X has better colors than Y' is rather meaningless. for example, D3x files look very much like M9 files in C1....probably phase tunes their profiling to their own color test. D3x files look much worse, or let me say, 'basic' in NC2, but is this bad? it only means that you have more parameters which you can tune.

those people who shoot test charts clearly conclude that D3x files are extremely accurate, much more than M9 files, or even 5DmarkII files. this however does not imply that you must like nikon's or phase's or adobe's color interpretation. and in printing, all this counts rather little, then color managment by CS or by the printer driver takes over anyway.

the most important issue with RAWs is not color interpretation but PP latitude. how much can you pull the shadows without creating ugly noise? how is the base iso noise level? and there D3x files shine, from base up to, say iso 1000.

 

peter

 

 

If you would like to see the statements, let me Google that for you:

 

Let me google that for you

 

I agree with you on the CMOS sensor side and how that relates to CCD sensors; that was the original point I was trying to get across with relation to how Leica is handling color output. That said, when you are having someone else fab a chip for you based on a standard core design, you aren't altering a lot of the per-pixel circuitry. If you were, there would be *zero* cost advantage to using the same core design. No, in the case of the same shared core, what's being altered are things like packaging and *possibly* the CFA, although I doubt even that is the case.

 

Look, if we can dramatic differences in color, sharpness, noise and overall image quality with different raw processors, why on earth do we feel the need to ascribe those differences to low-level sensor electronics, when they are much more easily implemented in the supporting ASICS and things like different AA filters?

 

Jeff

[Alnitak]

And while we are at it, like the D3x, I think shadow detail at lower ISOs is one of the areas the M9 really shines as well. I find that I have far more latitude with the shadows on the M9 than I ever did with any Canon body. It's changed how I shoot at lower ISOs, because at the same time, I also have less headroom (although the difference is not as pronounced) in the highlights. The adage "expose to the right" is far more applicable on the M9 than it ever was on my Canon and Nikon bodies.

 

It seems to me that like the D3x, the M9 is clearly optimized for high performance at lower ISOs. This is great unless you love to shoot in really dark conditions.

 

Jeff

[mjh]

Your post #70 seems to say that the Sony and D3x sensors are the same. But Rubén is right, I think: they're not.

Indeed the sensors are not the same, but it is the sensor in the Alpha900 that has on-board A/D converters (12 bit) whereas Nikon (as far as I know) uses external converters (14 bit).

[mjh]

No one, including Dalsa, says that "aggressive on-board noise reduction" had to be used with CMOS.

If nobody else says it, then I do … But as this is a well-known fact, I very much doubt I’m alone. CMOS sensors suffer more from non-uniformity than CCD-based sensors do. Because of amplifier circuits being integrated on the chip this is necessarily so; you cannot build millions of circuits with exactly the same amplification. Prior to the Canon EOS D30, CMOS sensors were only used for webcams and cheap toys because of the noise inherent in the CMOS design. That is also the reason why everyone reacted with disbelief when Canon announced they would switch to CMOS (I vividly remember the press conference back then). Until then it had been common wisdom that only CCDs could deliver the image quality expected from a professional DSLR. Canon was the first to overcome the fixed-pattern noise inherent in the CMOS design, and they implemented this (or some part of it) with on-board noise-suppression circuitry. So the noise typical for CMOS sensors came first, on-board noise-suppression came much later.

[Alnitak]

Michael,

 

I believe I brought that up earlier. As you point out, however, the issue was as much one of uniformity as it was actual noise; non-uniformity of response is a form of noise in and of itself. The fact is that the technology on CMOS side has pretty much moved past this issue, and the very design of the sensor has been turned to an advantage. Because a CCD is a serial read-out device, the developments in pixel-level noise reduction cannot be migrated across platforms.

 

Anyway, this is now way off topic....

 

How about that M9? I love mine! Unfortunately, I'm stuck in London while my M9 is back in Los Angeles. I thought I was here for just a week, but its now looking like it will be a lot longer.

 

Jeff

[Rick]

Rick, read carefully what I am writing in my posts. Color is as much psychological perception as it is pure spectral accuracy. Search the web. There are LOTS of sites that test cameras for color accuracy, and the general results are that Canon and Nikon do a good job of putting out good pure color from a purely technical reproduction standpoint.

 

That said, as I said earlier, many people like their pictures to have something more to them than pure technical color reproduction--which is why you and I both prefer what Leica does with the output of their cameras. Does the Kodak sensor have a role to play? Of course, and I didn't deny that--but much of what you see has to do with what Leica decides post sensor capture, and not what Kodak has done in sensor fab.

 

I am using the latest raw developer; the M9 still kicks out a red that is too shifted to magenta for skin tones, and I still shift it back a little. However, that's the only thing I don't like about the M9 color. Look, I thought the Canons produced a skin tone that was too yellow. It's personal preferences as much as it is pure color accuracy.

 

Jeff

 

Read carefully what you are writing? Did you really want to insult me?

 

Ok, fine. First off, I'm not going to go search the web! I don't have to get my information on "psychological perception" (redundant term - yours). I can just turn around and pull my copy of Cornsweet's, "Visual Perception" off my shelf if I need to understand perception. It was a great graduate class taught by Rober Yolten, PhD, wish you could have been there.

 

Next, I wouldn't be surprised if Nikon or Canon technically measure more accurate color fidelity than the M9. I just asked you where you get this information. So far, you just state that it is this way or many camera sites have tested... Or, you gave a non-scietific web page earlier that Adan gave a excellent thrashing. I don't buy into pseudoscientific statements. And, this does not mean that I'm convinced that the M9 is more accurate. You just aren't backing up your statement.

 

And, thanks for back-peddling on the fact that the sensor does play a part.

 

You still haven't answered my question of what you are using to develop your RAW files. I can't be sure why you are perceiving too much magenta until I know how you are developing the RAW and for that matter if you even have a calibrated monitor.

[Alnitak]

Rick,

 

Clearly I have offended you, although that wasn't my intention. While you have now gone out of your way to disparage me unnecessarily, I will try to keep to factual responses, as I believe I have done all along. Additionally, I don't believe that the burden is on me to do all of the primary research when the information is readily and widely available on the Internet. Virtually every major camera review site includes color fidelity checking, and publishes their methodology. Such sites include dpreview.com, the site I linked to, the-digital-picture.com and many others. Go to any one of those and look at any review you please. If you don't choose to do so, that doesn't invalidate my comments.

 

As for the site I linked, you don't need to disparage them as well. Simply go back and read the link; they state that they use Imatest. Yes, you can influence the results of Imatest depending on what you do, as Adan points out, but that by no means nullifies the results. Used properly, Imatest will be very accurate--you can shoot a DNG and convert it to TIFF using Leica/JenOptik's specific profile for Imatest to analyze. In that way, you can remove the variable of color interpretation by the raw converter. If you would like to do a lot of technical reading on how Imatest does color matching, try these links:

 

Imatest - Colorcheck

Imatest - Using Colorcheck

Imatest - Colorcheck Appendix

 

Given that they are trying to present an accurate review online, I would err on the side of assuming they are trying to be as accurate and neutral as possible. However, if this does not satisfy you, perhaps you would be satisfied by Erwin Puts opinion, Puts is certainly well-regarded in the Leica circles.

 

M9, part 7

 

I don't think his conclusion is too different than what I have been saying: Canon and Nikon are bit more technically accurate, but that's about it. This doesn't diminish Leica or their color decisions any way; its a decision they have made in their color output. As Puts points out, color is a virtually impossible thing to be purely technically accurate with, so its not worth getting bent out of shape about it (his point, but I would concur).

 

Now to address your allegations. To be clear, I have not back-peddled. It was a similar negative remark by you earlier that led me to suggest that you read my posts thoroughly. This whole thing got started because I responded to your comments that implied that Kodak is behind the color; I responded that while they play a role, most of the color interpretation and presentation is in Leica's hands. I still put forth that if you gave Leica a CMOS sensor and told them they had to use it, you would get extremely similar color to what you get now, because most of what you are seeing is Leica's work and has very little to do with the sensor itself.

 

As for my raw converter, I have used and do use many. My primary is ACR in both Photoshop and and LR, both in the latest versions of both of these products, and in the LR3 beta 2. However, I occasionally use a variety of other converters, including Capture 1 v4.x and Raw Developer (can't recall version). All have announced official support for the M9, and we all know that Adobe and Leica are working together on the raw converter support for the M9 in the Adobe products. It's only for skin colors that I find the M9's reds a bit off, and even then, its most noticable under red-shifted and or yellow-red-shifted light like tungsten or very intense sunlight. And yes, my screen is calibrated, and I use a high-quality Apple 30" ACD to start with. Finally, many others here have reached the same conclusion about the reds; Adan himself has provided very useful feedback on the adjustments he applies (thanks Adan) to dial back the reds.

 

So, back to my original point: While certainly Kodak plays some role as the manufacturer of the sensor, I hold that the real color output is nearly entirely determined by Leica as part of their process of developing the raw image data that comes from the sensor.

 

Jeff

[yanceman]

Ehh? And where is your source for this information? Nikon is officially on the record (with published statements) that Sony manufactures the D3x sensor for them. Yes, there are differences in the final implementation around processors etc, but the core sensor is the same. Much of Nikon's secret sauce is in the EXPEED processors.

 

Jeff

 

Yep. While Sony manufactures the sensors for Nikon, they are a Nikon design and built to Nikon's spec. It is not the same sensor as used in the Sony cameras.

[ho_co]

... perhaps you would be satisfied by Erwin Puts opinion, Puts is certainly well-regarded in the Leica circles.

 

M9, part 7

 

I don't think his conclusion is too different than what I have been saying: Canon and Nikon are bit more technically accurate....

 

Could you point out on the cited page exactly where Mr Puts says "Canon and Nikon are bit more technically accurate"? What I read on that page is that a) one shouldn't make such statements ("To claim that one camera is much better than the other is a dangerous game..."), and Canon's and Nikon's colors are tweaked more toward 'comfortable' rendering of flesh tones than those of the M9. He even begins the paragraph in question by saying "...colour psychology insists that accurate reproduction of colour is not always the best option...", which statement throws your argument to the winds.

 

IMO, you're becoming a nuisance. No matter how often someone points out that you're talking above your head, you come back with: 1) You're not reading what I wrote; and 2) What I wrote wasn't what I meant.

Edited April 18, 2010 by ho_co

[jaapv]

Yep. While Sony manufactures the sensors for Nikon, they are a Nikon design and built to Nikon's spec. It is not the same sensor as used in the Sony cameras.

Which begs the question - why? Surely Sony has an interest to build the best sensor technology they have into their own cameras - which brings us back to the conclusion that the difference is in the processing of the output of the sensor - at which Nikon appears to excel.

[Alnitak]

Could you point out on the cited page exactly where Mr Puts says "Canon and Nikon are bit more technically accurate"? What I read on that page is that a) one shouldn't make such statements ("To claim that one camera is much better than the other is a dangerous game..."), and Canon's and Nikon's colors are tweaked more toward 'comfortable' rendering of flesh tones than those of the M9. He even begins the paragraph in question by saying "...colour psychology insists that accurate reproduction of colour is not always the best option...", which statement throws your argument to the winds.

 

IMO, you're becoming a nuisance. No matter how often someone points out that you're talking above your head, you come back with: 1) You're not reading what I wrote; and 2) What I wrote wasn't what I meant.

 

Well Howard, you won't like this, but you are also ascribing things to me that I did not say. First, let me quote my post:

 

"I don't think his conclusion is too different than what I have been saying: Canon and Nikon are bit more technically accurate, but that's about it."

 

I should say again...you're not reading what I wrote if you misquote me. So be annoyed, but you just demonstrated exactly my point.

 

Puts says:

 

"In straight recording Nikon and Canon have an advantage as their colour profiles are optimized for skin reproduction and other often selected colours."

 

So...help me out here? Where does this differ from what I said above?

 

What is annoying the proclivity for many here to result to ad hominem attacks. Argue the facts, please.

 

Jeff

Edited April 18, 2010 by Alnitak
corrected sentence

[Advertisement]

Hallo,

Look here: Leica M

[Alnitak]

Which begs the question - why? Surely Sony has an interest to build the best sensor technology they have into their own cameras - which brings us back to the conclusion that the difference is in the processing of the output of the sensor - at which Nikon appears to excel.

 

Thank you Jaap--this is the point I have been trying to make all along.

 

Jeff