36 + Mp SL?

[ramarren]

Advertisement (gone after registration)

 

If I show a client a proof print which is 4x5 inches in size, and he/she asks for a final that is double that size, they mean they want an 8x10, which is 4x the area. 

 

Similarly if I show them an 16x20 and they say, "Oh, that's beautiful but it's too big. Can you make us one half that size?", they mean 8x10, which is 1/4x the area. 

It's not difficult to understand.

 

But neither is the correct usage.

 

So why not try to use the correct one?

 

I'm genuinely surprised to learn from this thread that you and many others are happy to preserve this confusing and unnecessary ambiguity. I accept that I'm the unusual one here, but It had never occurred to me that anyone seriously meant twice the length and twice the width when they said "twice the size". Particularly people in a business like photography where it makes a critical difference and patently leads to misunderstanding.

 

Are you seriously saying that if you show a client an A4 proof and they ask for it to be printed twice the size you'll assume they're asking for an A2 print?  Surely there's sufficient ambiguity to always want to clarify it with them isn't there? It certainly isn't standard usage, in my long experience.

 

In fifty years of doing photography, not a single person who said they wanted a print twice or half the size, according to the example I posed, misunderstood what was meant by that. Not a single person has ever been confused that what they meant was "twice the area" ... Not one. If I tried to use area measures, they'd all be confused. In my experience working with people whose countries work on the metric paper standard system, when they ask for an A2, A3, or A4 print, it's because they are familiar with what those sizes mean in terms of linear dimensions, NOT area. 

 

And, by the way, not a single person has ever asked what resolution the print might be. If they're looking at an image and want a large print, they ask me, "How large can you print that?" to which my answer is always length and width (or metric paper size), not area. 

 

I, personally, cannot change the conventions of the entire world by switching my terminology for something which is only arguably more understandable. That said, and being a precise kind of person, when I write up a print order for review, I list the print dimensions specifically ... the size of the paper (either US or Metric rules) and the size of the image area (either inches or cm) in some cases. 

 

Problem is, it's not difficult to understand, but no one guarantees that what you understand is correct.

 

Two rather simple observations: the term "size" appears not to be bound to any particular dimensionality, as opposed to "height", "length" or "width". Hence, double the size of a corn field and you double an area. Double the size of the corn or whiskey and you even double a volume. Or do you, now?

 

The term "resolution" can refer to different things. When used with reference to an optical device such as a lens, you'd use a term such as line pairs per mm. The same applies when talking about the "resolution" of a printing device, where dots per inch is the usual unit. When used with reference to a display device or a printed image, you'd often but not always mention the number of pixels or - perhaps - pixels per area. 

 

What's the dimensionality of the resolution taken at the eve of the New Year?

 

The word "resolution" is not only loaded, it's overloaded.

...

 

See above response to Peter H. 

Edited March 17, 2016 by ramarren

[IkarusJohn]

Interesting discussion; I apologise for heading us up a blind semantic alley. Perhaps print size and doubling the length and width was the wrong approach.

 

Of course, Paul J is right that doubling the pixels doubles the data and file size. But I think Wattsy raises the right question - is the image twice as "good", whatever that means?

[Advertisement]

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

[Peter H]

In fifty years of doing photography, not a single person who said they wanted a print twice or half the size, according to the example I posed, misunderstood what was meant by that. ................

Well, it sounds as though you and I are of a similar age and assuming a similar breadth of dealings with people, something must account for our very different experiences. Could it be cultural? I suspect it's something else, to do with our own way of expressing these things. Anyway, whatever the reason(s) I can assure you that there is some very considerable confusion when it comes to what is meant by doubling the size of a print, some of which is exhibited in this thread, let alone the outside world where many authors and writers on photography have gone to lengths to try to make it clear.

 

Be that as it may, it has been helpful to bring some clarity to some of the confusion evident in this thread, and resolve what appeared to be conflicting statements.

 

I agree though about resolution. No one has ever asked me about resolution, and almost everyone I've dealt with has assumed that photos can be enlarged to any size at all.

 

 

 

 

 

Interesting discussion; I apologise for heading us up a blind semantic alley. Perhaps print size and doubling the length and width was the wrong approach.

 

Of course, Paul J is right that doubling the pixels doubles the data and file size. But I think Wattsy raises the right question - is the image twice as "good", whatever that means?

 

Agreed, 100%.

 

But there are so many other variables that I think it is sometimes helpful to reduce the argument to the unrealistic "all things being equal" level in order to isolate the variables one at a time, otherwise we end up with the sort of confusion that has bogged us down so far.

 

But in practice I agree, not only does double the pixels not produce double the quality, but there's so much that is more important in photography than the number of pixels that we'd probably be better occupied dropping the subject altogether!

[LocalHero1953]

Be that as it may, it has been helpful to bring some clarity to some of the confusion evident in this thread, and resolve what appeared to be conflicting statements.

 

That's great! We will never have another thread ever again that argues about doubling pixel numbers, resolution and print size! 

[ramarren]

Well, it sounds as though you and I are of a similar age and assuming a similar breadth of dealings with people, something must account for our very different experiences. Could it be cultural? I suspect it's something else, to do with our own way of expressing these things. Anyway, whatever the reason(s) I can assure you that there is some very considerable confusion when it comes to what is meant by doubling the size of a print, some of which is exhibited in this thread, let alone the outside world where many authors and writers on photography have gone to lengths to try to make it clear.

 

Be that as it may, it has been helpful to bring some clarity to some of the confusion evident in this thread, and resolve what appeared to be conflicting statements. ...

 

 

I think that what's going on is that people relate to photo sizing in linear terms because they're looking at a print and thinking whether it fits in the space they have over the couch, or in that little nook in the hall, or on the wall in the parlor between the windows. 

 

The fact that you know and are aware that the metric system of paper sizing is an area measure in its inception is, I think, a touch unusual, or maybe points to the fact that I live in the US where most things are sized in inches and feet, where we refer to prints as "4x5s", "8x10s", "16x20s" etc, not by a letter-number designated paper size code. Most people that I chat with abroad are not aware of the specific dimensions of an A4 or A2 piece of paper, but they 'know' what size each of them are by long familiarity, just like I do, not because they are conscious of what area they represent.

 

People would look at me funny indeed if I said, "Would you like that in 80 square inch or 154 square inch?" meaning 8x10 or 11x14 respectively.  Standard paper dimensions are always understood by the linear size, not by the area even if the system in which they are nominated is area-based. 

 

Oh yeah: I remember when a 4x5 camera was referred to as a "quarter plate camera", because the standard in view cameras was the 8x10 inch format, just to point out how much stronger conventions are than rational, reasonable metrics... 

[Peter H]

...............Oh yeah: I remember when a 4x5 camera was referred to as a "quarter plate camera", because the standard in view cameras was the 8x10 inch format, just to point out how much stronger conventions are than rational, reasonable metrics... 

 

 

The convention firmly conformed to the arithmetic in those days: half plate was half the area (size!) of full plate. Quarter plate was a quarter of the area of full plate and half of the area of half-plate. Just as it should be. Full plate was 8 1/2 x 6 1/2 inches, or 55 1/4 square inches, being exactly four times quarter plate which was 3 1/4 x 4 1/4, or 13 13/16  square inches. No one fell for the "double the length = double the size" trick back then! Like land, paper has always been measured by area, for highly logical and necessary reasons. Even though we in the UK also talk more about 8x10s than A4, quarto, octavo or whatever, I don't think many people think an 8x10 is half the size of a 16 x 20. They certainly don't when they see the prices!

 

But enough of this. We both know what we mean.

[IkarusJohn]

Advertisement (gone after registration)

That's all true, Peter; and I'm not sure there is any "trick" involved.

 

However, if you say to someone - this camera has a 24MP sensor and this other camera has a 50MP sensor, the assumption will be that the second camera is twice as good.  In real terms (if you managed to keep everything equal, which I think we all acknowledge isn't possible with sensors the same physical size), if you present prints to the same resolution, that same person will be surprised that the second print isn't bigger.  The maths may not back that up (I don't think anyone disagrees), but the simple point is that a print of double the area doesn't look (to the untrained eye) twice the size.

 

The more critical point I was trying to make is that if you want to really increase resolution in the linear fashion people assume with pixel count, you have to also increase the sensor size.  The S(006)'s 36MP sensor is properly twice the size of the M9's 18MP because everything else is equal.

 

Having used the D800E alongside the M9, I did not find the same gain I did when using the M9 alongside the Monochrom.

[CheshireCat]

CheshireCat (his name, I did not give it to him, no insult) came in a "logical" way to the absurd conclusion that a 5k Apple display has 45 MegaPixel resolution.

As a response I deducted in the "same logical way" that, if this is true, then a M246 has a resolution of 72 Megapixels, or even more absurd, that it can have up to 240 Megapixels.

 

Oh, I love absurd and surreal, but this time I was quite serious.

 

The M246 with three exposures [R,G,B] will still have 24 MP, because the R,G,B values are co-located.

Instead, in a sensor with bayer color matrix, and in a computer LCD display, the R,G,B are not co-located, and indeed this is used to render text at higher resolution using a technique that uses the same interpolation concepts as the bayer matrix on a color sensor (though R,G,B are arranged differently).

 

Check the following link. What people call "subpixels" on display, are what people call "pixels" in color cameras with a bayer matrix.

 

https://en.wikipedia.org/wiki/Subpixel_rendering

 

[Peter H]

That's all true, Peter; and I'm not sure there is any "trick" involved.

 

However, if you say to someone - this camera has a 24MP sensor and this other camera has a 50MP sensor, the assumption will be that the second camera is twice as good.  In real terms (if you managed to keep everything equal, which I think we all acknowledge isn't possible with sensors the same physical size), if you present prints to the same resolution, that same person will be surprised that the second print isn't bigger.  The maths may not back that up (I don't think anyone disagrees), but the simple point is that a print of double the area doesn't look (to the untrained eye) twice the size.

 

The more critical point I was trying to make is that if you want to really increase resolution in the linear fashion people assume with pixel count, you have to also increase the sensor size.  The S(006)'s 36MP sensor is properly twice the size of the M9's 18MP because everything else is equal.

 

Having used the D800E alongside the M9, I did not find the same gain I did when using the M9 alongside the Monochrom.

 

 

 

This is interesting me more than I think it should!

 

Just on "size" again for a moment: I think it's all a question of how it's presented. 

 

If you show someone an A4 sheet in the middle of an A3 sheet in the same orientation, they may well be surprised that the relatively narrow border, in other words the "extra" size of the A3 print doesn't look like a 100% increase or double the A4 sheet. So I agree with you there. But if you turn the A4 print to the opposite orientation to the A3 print and slide it across to one side it will become immediately apparent to everyone that the A4 is exactly half the size of the A3, and no one will have to measure anything or work out the area. If necessary you could put another A4 sheet on the A3 sheet next to the original to prove the point that the A3 is twice the size of the A4  beyond even theoretical or intuitive doubt. This much is obvious, and we shouldn't allow our perception, which we as photographers should understand better than most is hugely fallible and malleable, to cloud our judgement.

 

But, back to cameras: if we're talking about increasing the number of pixels on the SL sensor from 24 to 36, I agree that it won't make a radical difference, though it probably would be helpful to many photographers, particularly those who will crop anyway, and those who really want to print as large as they possibly can, because however strong their printing technique, the extra data can't be replicated any other way than by more pixels. And introducing those inconvenient variables again, I doubt that in reality we're talking about extra pixels at the cost of sensitivity (i.e. fidelity). So it would be a good thing at the extremes but not for most users most of the time. But Leica users in particular seem to me to be excessively concerned about the extreme. I shall return to this idea elsewhere at some point!

[Joshua Lowe]

Is pixel shifting what Hasselblad's doing with the H5D-200c "Multi-Shot"?

[IkarusJohn]

 

... if we're talking about increasing the number of pixels on the SL sensor from 24 to 36, I agree that it won't make a radical difference, though it probably would be helpful to many photographers, particularly those who will crop anyway, and those who really want to print as large as they possibly can, because however strong their printing technique, the extra data can't be replicated any other way than by more pixels. And introducing those inconvenient variables again, I doubt that in reality we're talking about extra pixels at the cost of sensitivity (i.e. fidelity). So it would be a good thing at the extremes but not for most users most of the time. But Leica users in particular seem to me to be excessively concerned about the extreme. I shall return to this idea elsewhere at some point!

 

 

Everything else changes when you increase the number of pixels into the same sized sensor, and that's why it really doesn't interest me in the slightest.  It's a one dimensional demand.

 

To my mind, there is a good reason why Leica effectively produces one sensor for the M and for the S - the S just has more real estate.  The pixel size and everything else (I understand) is identical.  What would interest me is to understand what that actually means.

 

I appreciate that the S(007), SL and M(240) sensors are made by different manufacturers, but they are made to Leica designs, and presumably for good reason.  It will be interesting to see if the next digital M iteration departs from the fundamental sensor architecture, and moves to increase the pixel count.  I would be amazed if that was a core driver, but if the M goes to 37.5 or 50MP, presumably that change in pixel density will be for good reason and will be carried over to the S and SL?

 

I recall one of the manufacturers (Leica perhaps) talking about eliminating the gaps between the pixel sensors to get a more uniform image? or did I dream that.  There does seem to be a gulf between the perceptions we're given by the marketing departments (particularly in camera corrections and pixel count) and the underlying technology.  Whatever they're doing, Leica is doing a very good job in my view - particularly given their "less than state of the art" pixel boasting.

 

Cheers

John

[Peter H]

.......... There does seem to be a gulf between the perceptions we're given by the marketing departments (particularly in camera corrections and pixel count) and the underlying technology.  Whatever they're doing, Leica is doing a very good job in my view - particularly given their "less than state of the art" pixel boasting.

 

Cheers

John

 

 

 

Couldn't agree more.

 

Whatever the reality confronting certain specialist photographers who genuinely need every last pixel they can lay their hands on, for the majority of serious-minded photographers and less serious ones too, I firmly believe that 24mp never needs be a constraint on what can be achieved with a good idea, a good eye, a bit of imagination and a touch of technique. Fewer would hardly be a problem either, as many a fine photo attests.

[pop]

About the need for pixels.

 

My first somewhat usable digital camera was the Digital Ixus by Canon. It delivered 2MP. My first usable color printer printed 200Px/Inch at a color depth of not quite 16M colors. This combination yielded A4 prints with an ample border and a very nice color depth, even if they were a bit less than crisp. 

 

As I like shooting by framing in-camera (and not in PP), I reckoned that the sweet spot for my outfit would be reached with an 8MP camera and a printer capable of printing at 300 Px/Inch. This would allow for some conservative cropping and straightening and leave enough material for an A3 print.

 

Therefore, I really think that about 10MP would suffice for the kind of photography I prefer. I won't pretend that anyone else has to be so inclined. This also explains why I still think the Digilux 2 is a wonderful and perfectly usable camera if a bit slow.

 

In my mind, anyone  claiming that 50MP and more are desperately needed for photography to become useful either does not know what he's talking about or he does a very specific kind of photography and knows perfectly well what he's talking about. You can tell by their photographs, usually.

 

The rest is simple arithmetics. Some can do it, some can't.

[david strachan]

...

 

Therefore, I really think that about 10MP would suffice for the kind of photography I prefer.

 

...

 

Yep. Now that would be the size of the first digital M with nice BIG pixels.

And I mean BIG as...let's see, mmm...  ... a big bucket vs a small bucket.

Edited March 18, 2016 by david strachan

[steppenw0lf]

Check the following link. What people call "subpixels" on display, are what people call "pixels" in color cameras with a bayer matrix.

 

https://en.wikipedia.org/wiki/Subpixel_rendering

 

 

This is only an old and well-known method of improving the rendering of characters. Rendered images do not profit from it.

The screen of the Retina iMac can only display images of around 5120x2880 Pixels (14.7 M). That's it. No magic, no Harry Potter.

Edited March 18, 2016 by steppenw0lf

[CheshireCat]

This is only an old and well-known method of improving the rendering of characters. Rendered images do not profit from it.

The screen of the Retina iMac can only display images of around 5120x2880 Pixels (14.7 M). That's it. No magic, no Harry Potter.

 

 

That is the proof that the iMac has 45 MP in terms of RGB spatial elements.

 

Rendered images from your 24 MP sensor do not profit from it because they have a lower resolution:

- Green channel resolution: 3000x4000 vs 5120x2880

- Blue channel resolution: 3000x2000 vs 5120x2880

- Red channel resolution: 3000x2000 vs 5120x2880

 

Interpolation of the Bayer matrix is creating fake pixels that were never recorded. That is your Harry Potter.

[Jared]

Good heavens, I can't believe how these discussions always devolve so badly.

 

How many megapixels are enough? Depends on how much cropping room you need, how large you print, whether you make prints at all, your personal requirements for rendering of fine detail, how good your technique is, what f-stop and focal length you tend to shoot at, how accurate your focus is, what subjects you are photographing, and probably several other factors that I can't even think of right now.

 

I think we can all agree that some subjects benefit from higher resolution cameras more than others. I think we can all agree that larger prints need more megapixels than smaller prints, at least until viewing distances start to really grow. I think we can also all agree that most Leica lenses at optimum focal ratio can out-resolve current sensors in Leica cameras so there is at least a potential benefit to more megapixels. Until that last is no longer true, there will be some photographers who crave higher resolution sensors.

 

Me? I ran tests of some of my own best (technically, not necessarily artistically) photos at various ppi on my Epson 3880 printer using a high quality glossy paper that holds detail really well to determine how many pixels I needed for my photography before more just wouldn't matter. We've already passed that number for me. I no longer desire more megapixels, but that doesn't mean others won't still crave 100mp and beyond. Clearly, their requirements don't match mine. Certainly not something that needs to be argued about, though.

 

I would encourage everyone who is passionate about this subject to make a stitched photograph with, say, 100 megapixels, print it as large as you would ever really want to make a print, and compare it to a down-sampled 50 megapixel version, 25 megapixel version, 12 megapixel version, and perhaps even an 8 megapixel version of the same photo to see, for real, what the differences are in IQ as it translates to real world prints. Then make your own determination as to what you need.

 

Like everyone else, I'm happy to take more megapixels as long as that feature isn't coming at the cost of poor dynamic range or poor high ISO performance. So far, each generation of CMOS imaging chips seems to be an improvement in all three areas so I won't mind getting more megapixels on my next camera. I just won't care much.

 

The arguments over whether a camera with 2x the megapixels have twice the resolution or just 1.4 times the resolution are just silly. I suspect everyone reading this forum understands the difference between linear measurements and area measurements.

 

- Jared

[Jeff S]

 

Me? I ran tests of some of my own best (technically, not necessarily artistically) photos at various ppi on my Epson 3880 printer using a high quality glossy paper that holds detail really well to determine how many pixels I needed for my photography before more just wouldn't matter. We've already passed that number for me. I no longer desire more megapixels, but that doesn't mean others won't still crave 100mp and beyond. Clearly, their requirements don't match mine. Certainly not something that needs to be argued about, though.

 

I would encourage everyone who is passionate about this subject to make a stitched photograph with, say, 100 megapixels, print it as large as you would ever really want to make a print, and compare it to a down-sampled 50 megapixel version, 25 megapixel version, 12 megapixel version, and perhaps even an 8 megapixel version of the same photo to see, for real, what the differences are in IQ as it translates to real world prints. Then make your own determination as to what you need.

 

 

You mean actually making real life test prints using one's own print workflow?  What would become of this forum?  

 

Jeff

[Jared]

I know, what was I thinking?

 

 

[Advertisement]

Hallo,

Look here: Leica L-Mount

[Peter H]

I know, what was I thinking?

 

 

This slightly sneering tone is a common response whenever a discussion turns a little theoretical.

 

In principle I'm completely in agreement with you but I do wish you didn't find it necessary to mock an attempt to understand and clarify some of the basic terminology that frequently crops up in conversations about digital cameras.

Edited March 19, 2016 by Peter H