36 + Mp SL?

[adli]

Advertisement (gone after registration)

Or, if you are not satisfied with the SL, buy the new Nikon flagship camera for the pro market, the D5. Oh, only 21mp you said?   Or the new pro camera from Canon, the EOS 1D-X mark II. Oh, only 20mp?

 

Wonder why the top of the line pro cameras from Nikon and Canon has so "few" megapixels? Don't they know that more pixels always are better and their cameras will sell less because they have such an obsolete pixel count?

 

P.S. That was a rhetorical question....

[Advertisement]

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

[CheshireCat]

Or, if you are not satisfied with the SL, buy the new Nikon flagship camera for the pro market, the D5. Oh, only 21mp you said?   Or the new pro camera from Canon, the EOS 1D-X mark II. Oh, only 20mp?

 

Wonder why the top of the line pro cameras from Nikon and Canon has so "few" megapixels? Don't they know that more pixels always are better and their cameras will sell less because they have such an obsolete pixel count?

 

P.S. That was a rhetorical question....

 

 

... and here is a practical answer: most pros value more frame rate than resolution (sports, nature, events).

Other pros who need high resolution will get a Canon 5Ds with 50 MP.

[Peter H]

... and here is a practical answer: most pros value more frame rate than resolution (sports, nature, events).

Other pros who need high resolution will get a Canon 5Ds with 50 MP.

 

 

Or a Phase One probably, if the need is really there.

[AlanJW]

 

The world is ruled mostly by common sense,

B/K

Unfortunately, this is not true, whether for photography or more important things.

[CheshireCat]

No need to do so. The  iMac display does not have 45MP, no matter what terms you use. Try to understand the language you're using. i.o.w. watch it.

 

I am sorry - it is not your fault - but you and many others here have been fooled by the marketing monkeys.

 

The definition of pixel is the too generic: "picture element".

Some dictionaries define it as the smaller element that can be processed by a graphical system.

The marketing monkeys love generic definitions, so they can fool us.

And since we are talking Leica, here I am using Leica's terminology.

 

In Leica terms, if the SL EyeRes EVF (a display) has 4.4 MP, then the iMac has 45 MP.

 

https://us.leica-camera.com/content/download/129523/1620099/version/3/file/Datenblatt_SL-System_e.pdf

 

They didn't name the resolution correctly in the specs: "SXAG" instead of "SXGA+"; this prevents people from looking it up and understanding the trick.

Other companies use the term "dots" instead of pixels, to avoid the confusion with "full RGB pixels". A "full RGB pixel" is the picture element software deals with and is usually formed by 3 values, one for each color component.

 

About the 24 megadots (megasensels) in your bayer sensor, they are not "full RGB pixels".

The raw data from the sensors are converted to a full RGB 24 megapixels image by means of interpolation. This means that 2/3 of the information in your final output image is fake (was never captured). This is better understood in terms of planar (per-color) resolution, like I have already explained.

 

If you want to object to this, you may bring psychophysiological color resolution perception into the discussion. But just denying and telling me I am playing with words, is not interesting.

[IkarusJohn]

Or a Phase One probably, if the need is really there.

 

I'd love a PhaseOne XF - in medium format, to repeat adli's question, why does PhaseOne offer a back with "only" 40MP?  Actually, they offer digital backs from 40-100MP, at an astonishing price ...

[AlanJW]

Advertisement (gone after registration)

 

They didn't name the resolution correctly in the specs: "SXAG" instead of "SXGA+"; this prevents people from looking it up and understanding the trick.

You're not seriously saying that they did this deliberately as part of a "trick." are you?

[CheshireCat]

You're not seriously saying that they did this deliberately as part of a "trick." are you?

 

 

No, I don't think this is part of the marketing trick, otherwise they would have omitted the resolution name.

They are just incompetent.

[Peter H]

I'd love a PhaseOne XF - in medium format, to repeat adli's question, why does PhaseOne offer a back with "only" 40MP?  Actually, they offer digital backs from 40-100MP, at an astonishing price ...

 

 

I used to use Shneider Kreuznach lenses with my Rolleis.  What a fantastic outfit that was, which I sold when I committed to going fully digital.  I wish I still had it, although I couldn't bear setting up and operating a darkroom all over again to make the most of it.

 

So, a phase One XF with a 100mp back and a set of SK lenses is the obvious answer.

 

Would anyone like to buy a cottage in the Scottish Highlands?

[kaisernchen]

I use the Phase One since their P25, all the way to now I use P65+ and IQ180 on Hasselblad H system and the XF IQ380 system - it was designed primary for (studio & architectural & landscapes) production, with reasonable speed, but highest possible quality is their key efforts, of course they can be use for other thing.  This is not to say Nikon or Canon or SONY cannot do the quality of Phase One, it is simply a matter that such market is very small, few thousand new units in demand, and different set of know-how, and it is not what they are after.  

What we see in camera industries today is that each house offer something of their own, some very similar, but they are all basically a product with set of solutions that strong on some point, compromise on some point, while the compromises lessened when technology advances, but it is likely to be there still.

SL is a good system for a company like Leica, arguably their most advanced camera - except crippled by very limited native lens - which will take years and years to build, but it is also in a zone of strong competition.  There is little doubt Leica probably produce some best lenses currently available, most people recognize that, even those don't use Leica optics, but we can see all the international level of photo competitions and awards, Leica is not a popular camera there - it is a fact that Leica is a boutique brand, and naturally has a small market share, but it is also saying that there are great images can be made without using a Leica.  So with the SL, eventually the truth will be unfolded, and to meet mass market expectation will allure Leica go on that direction.  

Without question there are smart people in Leica and they are doing their job base on what their resources allow, M family is still likely to be their main line for some time to come.

 

B/K

[steppenw0lf]

Read again my previous post about per-channel resolution.

Then try to understand what bayer interpolation is, because clearly you don't know what you are talking about.

 

 

Thanks for your scientific answer (is it psychology or physics ?) That's really a big step for mankind.

 

And yes, here a small reminder:

Everyone is welcome, as long as they show respect for other members, their opinions and the forum rules.

[keithlaban.co.uk]

Gosh, it has never crossed my mind that I'm using a boutique brand.

[adli]

Reality bites

 

Gosh, it has never crossed my mind that I'm using a boutique brand.

[keithlaban.co.uk]

Reality bites

 

Well, how spiffingly modern am I?

[dgktkr]

I apologize in advance for feeding the troll.

 

And if your image is rescaled even a bit, say to fit your screen or the resolution of your printer, the resulting output image is fake! In general, the resampling interpolates the information from the original capture and models what the data "should" look like on a new grid. 

 

The horrors of digital imaging!

 

dgktkr

 

 

 

...

 

About the 24 megadots (megasensels) in your bayer sensor, they are not "full RGB pixels".

The raw data from the sensors are converted to a full RGB 24 megapixels image by means of interpolation. This means that 2/3 of the information in your final output image is fake (was never captured). This is better understood in terms of planar (per-color) resolution, like I have already explained.

 

...

[Jared]

 

About the 24 megadots (megasensels) in your bayer sensor, they are not "full RGB pixels".

The raw data from the sensors are converted to a full RGB 24 megapixels image by means of interpolation. This means that 2/3 of the information in your final output image is fake (was never captured). This is better understood in terms of planar (per-color) resolution, like I have already explained.

 

 

 

I can't agree that "interpolated" values are "fake" values.  Use of the word "fake" suggests trickery or something underhanded.  True, the demosaic process does result in lower spatial resolution than a sensor that took separate R, G, and B measurements for each pixel, but you are implying something dishonest in the approach and suggesting that quality differences are substantial.  Neither is true.  Based on my tests using both bayer and non-bayer cameras with the exact same underlying chip the differences in spatial resolution and even per-color accuracy are much smaller than your "2/3 of the information [...] is fake" comment suggests.  Since the vast majority of cameras use Bayer arrays to measure and interpolate color values and nearly all manufacturers report pixels in the same way, it is hardly misleading to state that the Leica SL is 24 megapixel camera or to state that an iMac 5K display is 14.7 megapixel display.

 

I do, however, agree with you that the viewfinder on the SL should be listed as 4.4 million "dots" rather than 4.4 MP.  What they are doing is misleading for exactly the reasons we have run into in this discussion.  Sony calls them "dots".  Fuji calls them "dots".  Even Epson, who presumably make the Leica EVF, calls them "dots".  Shame on Leica for getting that wrong.  I suspect it was a gaffr rather than an intentional marketing trick, but it's still something they should get right.  It's either 4.4 million dots or 1.4 megapixels.  4.4 megapixels is simply incorrect/misleading.  Just as it would be to label the large iMac as a 45 megapixel display.  

 

- Jared

[CheshireCat]

 

I apologize in advance for feeding the troll.

 

I see people who are in agreement with the admins, can happily call "troll" other forum users without no admin warning.

No worries, I am used to this treatment in this forum. Exemplary moderation

 

And if your image is rescaled even a bit, say to fit your screen or the resolution of your printer, the resulting output image is fake! In general, the resampling interpolates the information from the original capture and models what the data "should" look like on a new grid. 

 

I am not saying that the entire image is fake. I am saying that only those image values that have been "invented" by the camera (by means of interpolation) are fake.

 

Allow me to explain this in a super-simple way:

0) Let's call "dot" a sensor element (so we don't confuse sensels with pixels).

1) The M240 has 6 million dots that see red. The red channel image therefore contains only 6 million dots.

2) The M240 outputs a JPEG in which the red color resolution is 24 million dots.

3) Therefore the M240 has faked 18 million red dots. Faked. Never captured. Totally invented.

 

Do you agree with all of the above ? No ? Which one ?

I'll take a missing reply as a yes.

Now excuse me but I must run, or daylight will turn me into stone. 

[CheshireCat]

I can't agree that "interpolated" values are "fake" values.  Use of the word "fake" suggests trickery or something underhanded.  True, the demosaic process does result in lower spatial resolution than a sensor that took separate R, G, and B measurements for each pixel, but you are implying something dishonest in the approach and suggesting that quality differences are substantial.  Neither is true.  Based on my tests using both bayer and non-bayer cameras with the exact same underlying chip the differences in spatial resolution and even per-color accuracy are much smaller than your "2/3 of the information [...] is fake" comment suggests.

 

I am using the word fake with the meaning: Not true or real. Meant to look real or genuine but not real or genuine.

 

But you are right, smart people are working on interpolation algorithms to find the best way to literally trick our eyes. Nothing dishonest, think about it as a nice show by an illusionist: you see something there, but... there's nothing there

 

In a strict - information science - sense: I am totally right. Sorry, but this is just plain logic. There is no information captured if the sensor doesn't capture it.

 

I assume your tests on bayer vs non-bayer (full RGB) sensors compute some kind of MSE/PSNR value that your left-hemisphere can unequivocally read to produce your sharp reply to my original post. Correct me if I am wrong.

What your tests with these sensors prove, though, is that many photographic subjects are "flat", meaning that they lack the overall level of detail to require a higher sampling rate all over the frame (hence interpolation is a good approximation).

But detail is in the eye of the beholder ! Think about global detail vs local detail. Local detail is everywhere in the frame your fovea is looking at.

 

Finally, we can enter the fascinating world of psychophysiological vision.

 

First question: Assuming you are right, then how come a monochrome image from a M246 feels much better than a M240 image converted to b&w ?

[Advertisement]

Hallo,

Look here: Leica L-Mount

[Jared]

I am using the word fake with the meaning: Not true or real. Meant to look real or genuine but not real or genuine.

 

But you are right, smart people are working on interpolation algorithms to find the best way to literally trick our eyes. Nothing dishonest, think about it as a nice show by an illusionist: you see something there, but... there's nothing there

 

In a strict - information science - sense: I am totally right. Sorry, but this is just plain logic. There is no information captured if the sensor doesn't capture it.

 

I assume your tests on bayer vs non-bayer (full RGB) sensors compute some kind of MSE/PSNR value that your left-hemisphere can unequivocally read to produce your sharp reply to my original post. Correct me if I am wrong.

What your tests with these sensors prove, though, is that many photographic subjects are "flat", meaning that they lack the overall level of detail to require a higher sampling rate all over the frame (hence interpolation is a good approximation).

But detail is in the eye of the beholder ! Think about global detail vs local detail. Local detail is everywhere in the frame your fovea is looking at.

 

Finally, we can enter the fascinating world of psychophysiological vision.

 

First question: Assuming you are right, then how come a monochrome image from a M246 feels much better than a M240 image converted to b&w ?

 

 

I agree with you that values are interpolated.  However, I still do not agree that this is trickery or, as you are characterizing it, magic or a "nice show".  

 

As far as my tests... They were performed on astronomical subjects using short focal length optical tubes with significant under sampling to ensure that resolution was limited by the sensor, not by the telescope, the tracking, or the local seeing conditions.  In other words, there is no question that higher resolution out of the sensor would result in measurably higher resolution in the final image.  I compared full width/half max values on point sources--non saturated/non blooming stars--both with a Bayer matrix in place and with an otherwise identical camera that used a monochrome chip, separate exposures taken through individual RG&B filters.  I consistently found that monochrome chips that individually measured RG&B values at each pixel had roughly 1.3 times the linear resolution of the demosaic'd Bayer chip when looking at the final RGB image.  That's a significant advantage for measuring each color at each pixel, but hardly the dramatic difference that you are implying would result.  

 

Does psychophysiological vision play a part in this?  Of course.  That's why, for example, the green pixels outnumber the red and blue.  We perceive detail better in the middle of the visual light spectrum.  It's almost the reverse process of JPG compression or audio compression where psychoacoustics need to be taken into account--how we hear details in a stream of sound.  

 

Even when you measure the loss of detail, though, using hard numbers rather than perceived values, the differences in resolution are significant, but not as profound as you seem to be implying.

 

As far as why the monochrome produces what feels like sharper results... That is completely consistent with my argument above.  It is capturing more resolution with its 24 megapixels than an M(240) would.  I just don't think it's the profound factor of 3 that you are implying.

 

- Jared

[adli]

You are twisting the reality a bit. The M240 has 24 million picture elements capturing light. All of them register the luminance. So you have 24 million samples of luminance values. However, the picture elements can only capture one color each, so you have 12 million capturing green values, and 6 million capturing red and blue respectively.

 

So, for half of the 24 million picture elements, you don't know if they were green, but you do know what luminance value the green should be, so you actually have a lot of information about that cell. In addition, you know the pattern of green around the cell based on all the green cells in the area around the cell. This means your software is in a very good position to estimate if the cell should be green or not. Based on this, you will get 24 million elements for each color which all will have the correct luminance value and a very high probability of having the correct color.

 

 

 

 

Allow me to explain this in a super-simple way:

0) Let's call "dot" a sensor element (so we don't confuse sensels with pixels).

1) The M240 has 6 million dots that see red. The red channel image therefore contains only 6 million dots.

2) The M240 outputs a JPEG in which the red color resolution is 24 million dots.

3) Therefore the M240 has faked 18 million red dots. Faked. Never captured. Totally invented.

 

Do you agree with all of the above ? No ? Which one ?

I'll take a missing reply as a yes.

Now excuse me but I must run, or daylight will turn me into stone.