36 + Mp SL?

[jrp]

Advertisement (gone after registration)

I have taken pictures at 16Mpx that have been blown up even bigger than Salgado's as posters on the walls of Pets at Home.

[CheshireCat]

I've always qualified these scores as : "scores better/worse at DXO"  I don't go with their self-proclaimed " the benchmark for image quality"

 

 

As far as I know, they at least publish their results. Unlike "scores better/worse at Jaapv"

[Advertisement]

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

[Paul J]

36MP will give you more pixels, for sure.  But let's look at that a little more carefully.  24MP gives you 6,000 pixels along the long edge and 4,000 pixels on the short edge.  As has been said on more than one occasion, in order to truly double the resolution you would need 12,000 x 8,000 - 96MP.  A move to 37.5MP (the S(007) sensor) would give 7,500 x 5,000, an increase of only 25% along each side.  So while an increase to 50MP seems colossal, it gives only a modest gain.

Well that's one way to word it, and it is a popular sales pitch and counter argument but in my experience you can't really think like this. The difference between 24 million pixels and 50 million pixels is 26 million pixels. That's twice the amount of pixels and 26 million pixels more of data in the same area, most of which translates into more detail, smoother tonal transitions and better colour. 50MP is not a "modest gain" from 24MP.

Edited March 16, 2016 by Paul J

[IkarusJohn]

I think you missed the first sentence. Let me rephrase, sure you get more pixels, but there's more to it.

 

I appreciate that you are very keen on more MP. There is,however, more to this than just numbers.

[CheshireCat]

Well that's one way to word it, and it is a popular sales pitch and counter argument but in my experience you can't really think like this. The difference between 24 million pixels and 50 million pixels is 26 million pixels. That's twice the amount of pixels and 26 million pixels more of data in the same area, most of which translates into more detail, smoother tonal transitions and better colour. 50MP is not a "modest gain" from 24MP.

 

 

Exactly, but these folks keep talking about linear resolution as if they would read images one line at a time.

 

Yes, there is only a 45% gain in linear resolution going from 24 MP to 50 MP.

Whoever is shooting 24 MP and thinks that 45% is just a modest gain, should go back to a 12 MP camera and enjoy much cleaner performance at the same ISO for a modest loss of just 30% in linear resolution.

 

I love numbers

[jaapv]

Not much wrong with the performance of the M8. In fact the per pixel acuity is probably better than that of newer cameras.

[CheshireCat]

Advertisement (gone after registration)

Not much wrong with the performance of the M8. In fact the per pixel acuity is probably better than that of newer cameras.

 

 

You mean: the perception of per pixel acuity.

I cannot disagree, as I can clearly see each and every pixel as a very well defined colored square in M8 images

[cpclee]

It's about the same resolution difference between APS-C and full frame. Not negligible but also not the only factor that matters.

 

Back in the day, numerous tests showed the 10mp DMR trouncing 16 to 20mp Canons by virtue of better optics and the absence of low pass filter. More pixels doesn't equal more information capture.

[talt03]

I think the OP already has something in mind based on the arguments and his purpose.

 

But we can buy a Sony A7RII (42MP), a Canon 5DSR (50MP) then forget about resolution, ISO, and everything else. Then hope we get the quality we wanted for our own purpose. If still not happy, we can go MF. The "want" will never end since these companies will keep producing increases in resolution as it seems to be the only camera characteristic that sell in the consumer market. But I still wonder why Sony is also pushing 12MP in one of their flagships. Same as Nikon which in fact has the D500 and D5 at 20MP, not following the trend set for their other models. Playing it safe? If hi-res does not sell, there is low-res. ✌

Edited March 16, 2016 by talt03

[CheshireCat]

But I still wonder why Sony is also pushing 12MP in one of their flagships.

 

 

Because videographers buy it.

[Paul J]

I think you missed the first sentence. Let me rephrase, sure you get more pixels, but there's more to it.

 

I appreciate that you are very keen on more MP. There is,however, more to this than just numbers.

 

 

I read it the first time. But 2 x 24 million does not equal 96 million what ever way you want to look at it, or whose ever sales pitch and justification you read. In this case there really isn't more to numbers. Twice the pixels is twice the data. The amount that translates will depend on many things, but most losses will be mitigated with a quality and innovative design and this is most likely what Leica is taking longer to get right.

 

I really don't understand pixel phobia. The large majority here act hostile (you too surprisingly, must have been having a bad day in this thread ) as soon as someone mentions it, yet I heard the same arguments when people didn't need more than 5-8mp, yet everyone is blissfully viewing their 24Mp cameras and mobile phones at small sizes, and at a time when the cost of an 8TB hard drive is less than what a 500GB hard drive cost back then, and we can view our work on a 5K monitor.

 

The only thing I can honestly see that is more than numbers is the quality and innovation of the design of the camera housing it (and the truthfulness of the marketing department). I've used the 5DS R quite a bit lately and I have to say that the designs are certainly coming along and mostly showing promise. I use that camera with out thinking of it and the results are outstanding.

 

Lets have more MP with little to no cost in DR and ISO, in a well designed and innovative body that makes us unaware of the upgrade when shooting. And Leica, sssshhhhh don't tell the forum when you do it.

[cpclee]

Sure, twice the pixels is twice the data. But it takes the quadrupling or four times the data to double the resolution, for example from a 150dpi print to a 300dpi print of the same dimensions. Which part of this math is confusing you?

[Peter H]

A few years back, about 6 or 7 perhaps, when cameras like the Nikon D700 were introduced and "full frame" sensors became popular and for a while sensor size overshadowed pixel count as the gearheads' principal obsession, we were frequently told that it's not just the number of pixels that matters but the quality of the pixels, and that larger pixels were better than small ones because they were more sensitive and therefore more accurate. The ideal therefore being large numbers and large pixels, but on a sensor of a given size, at a given state of sensor technology, there must be an optimum number of pixels: too few giving sub-optimum resolution, and too many giving sub-optimum sensitivity and accuracy.

 

The optimum was rarely stated explicitly, though the implication was that each manufacturer had found it, and offered it to us in their latest model.

 

Was it all nonsense? Or was it correct, but overtaken by improvements in sensor technology meaning that pixel count can be increased with no performance cost?

[thighslapper]

 

 

Was it all nonsense? Or was it correct, but overtaken by improvements in sensor technology meaning that pixel count can be increased with no performance cost?

 

correct ....... increased sensitivity .... and more importantly reducing the 'noise' that results from signal amplification 

[Paul J]

Sure, twice the pixels is twice the data. But it takes the quadrupling or four times the data to double the resolution, for example from a 150dpi print to a 300dpi print of the same dimensions. Which part of this math is confusing you?

You might find you would have a more mature and meaningful conversation if you left out the snide remarks at the end. Twice the pixels results in greater detail, better tonal graduations, and better colour. From camera to camera this amounts in a variable difference is that is not ultimately calculated with math, but perceived with the eye. Yes, that becomes confusing if you are looking solely at math, because the output varies from design to design and often the results are not in keeping with what should technically be reflected in the final result. Finally, relating metrics to something as unquantifiable as human vision and perception is also, arguably, somewhat hypothetical.

 

10% is usually enough to make a noticeable difference, yet the figure, 10%, is easily passed of as insignificant amount.

 

You are also missing the point and further confusing the matter relating ppi to dpi. dpi relates to the spacing of dots of ink, not pixels. Dots of ink are smaller than pixels and it takes several dots of different coloured ink to make one coloured pixel. Dots of ink are not layed perfectly like pixels, they are placed on top of and overlapping other dots of ink.

Edited March 16, 2016 by Paul J

[Paul J]

correct ....... increased sensitivity .... and more importantly reducing the 'noise' that results from signal amplification

A few years back, about 6 or 7 perhaps, when cameras like the Nikon D700 were introduced and "full frame" sensors became popular and for a while sensor size overshadowed pixel count as the gearheads' principal obsession, we were frequently told that it's not just the number of pixels that matters but the quality of the pixels, and that larger pixels were better than small ones because they were more sensitive and therefore more accurate. The ideal therefore being large numbers and large pixels, but on a sensor of a given size, at a given state of sensor technology, there must be an optimum number of pixels: too few giving sub-optimum resolution, and too many giving sub-optimum sensitivity and accuracy.

 

The optimum was rarely stated explicitly, though the implication was that each manufacturer had found it, and offered it to us in their latest model.

 

Was it all nonsense? Or was it correct, but overtaken by improvements in sensor technology meaning that pixel count can be increased with no performance cost?

The standard remains: bigger pixels do equal greater sensitivity but sensor technology is getting around, and mitigating this issue, reducing the gap and creating a new standard. The Sony BSI sensor, for example is far more efficient at collecting light with it's shallower light wells and has less signal noise with materials which are better suited to the task. So while the old truths remain, we can improve our technology and efficiency at collecting the light and also in converting it to usable data. At what point this stops, or even if it does, remains anyones guess with technological advances occurring faster with more ease.

Edited March 16, 2016 by Paul J

[cpclee]

Was not trying to be snide with that remark so I regret that it came out that way.

 

Having said that, resolution is resolution and there is a precise and technical (mathematical) definition.  And it's an intuitive definition: given a fixed screen (print) size, you need to double the pixels (dots) in both the vertical and horizontal directions to really double the resolution.  So it takes quadrupling the pixels.   Human perception doesn't change that fact. 

 

 

You might find you would have a more mature and meaningful conversation if you left out the snide remarks at the end. Twice the pixels results in greater detail, better tonal graduations, and better colour. From camera to camera this amounts in a variable difference is that is not ultimately calculated with math, but perceived with the eye. Yes, that becomes confusing if you are looking solely at math, because the output varies from design to design and often the results are not in keeping with what should technically be reflected in the final result. Finally, relating metrics to something as unquantifiable as human vision and perception is also, arguably, somewhat hypothetical.

10% is usually enough to make a noticeable difference, yet the figure, 10%, is easily passed of as insignificant amount.

You are also missing the point and further confusing the matter relating ppi to dpi. dpi relates to the spacing of dots of ink, not pixels. Dots of ink are smaller than pixels and it takes several dots of different coloured ink to make one coloured pixel. Dots of ink are not layed perfectly like pixels, they are placed on top of and overlapping other dots of ink.

[Paul J]

Was not trying to be snide with that remark so I regret that it came out that way.

 

Having said that, resolution is resolution and there is a precise and technical (mathematical) definition.  And it's an intuitive definition: given a fixed screen (print) size, you need to double the pixels (dots) in both the vertical and horizontal directions to really double the resolution.  So it takes quadrupling the pixels.   Human perception doesn't change that fact.

Yes, i understand the basics of that, but the reality of actual resolution increase is separate to what should technically be, with calculations. Also, while Camera data and print data may somewhat relate in conversation, in relative terms, they are two separate things entirely. A camera with greater resolution can often look better to a camera of lower resolution in the same print or jpg size because it is working with more rich level of detail and data to begin with.

Edited March 16, 2016 by Paul J

[LocalHero1953]

You might find you would have a more mature and meaningful conversation if you left out the snide remarks at the end. 

No more snide than your reference to pixel phobia in an earlier post. I don't think I've read a thread on megapixels that doesn't descend to similar irrational arguments on both sides. 

[Advertisement]

Hallo,

Look here: Leica L-Mount

[ramarren]

I'm not "pixel phobic" and I can do the math. 

 

I'm very happy with the doubling of resolution that going from the E-1 to the SL represents. I was quite happy with the doubling and tripling of data that using 10 and 16 Mpixel cameras represents, although it was only by going to 24 Pixels that the resolution from my E-1 baseline was doubled. More data is good, was good, and netted good returns. More resolution would be good too. 

 

However, Leica has chosen to stay with 24 Mpixels on this size sensor for the present rather than going to 31, 36, 40, or 50. Why? I don't know, I can only speculate. (I won't.) Moreover, there are such things as the "law of diminishing returns" ... How much more resolution and quality can you get out of moving to 50 Mpixels and what practical advantage does it offer? As as been said umpteen times in this thread, the gains are about 44% on linear resolution given the SL as the baseline; let's consider ISO and DR to be the same. Let's also presume you have lenses that can deliver the resolution required and not run into diffraction limitations in a useful aperture range. Yay! What are the costs and what are the gains? Costs are a doubling in the data handling requirements of the camera, a doubling in the data storage requirements of the camera and the off-line storage equipment, a factor of about four times the data processing requirements both in-camera and in-post with its concomitant increase in RAM and CPU power. Gains are more ability to capture fine detail and more processing overhead space. 

 

How much more detail? and for what kinds of outputs? Well, it's pretty obvious that you don't need much more detail for most computer displays; while there are 4K and 5K displays around, they're vastly outnumbered by displays that are far lower in pixel resolution, and 8700 pixels on the long edge are way more than you need for even a 4K display. Prints? My Epson P600 nets its best output at 360 pixels per inch and can handle 13x19" cut sheet. Printing borderless nets the need to deliver 6890 pixels (long edge) to achieve that resolution. So a camera optimized to deliver data for printing at that size should have 31 Mpixel resolution; 50 gives a bit more room for cropping. Obviously, for bigger prints more is better. Q: How many 13x19" or larger prints do you make? If the answer is "the majority of my prints", then you want more pixels as a baseline. (A 13x19 borderless print is pretty rare for me. I make mostly 6x9 image area on 8.5x11 paper and, less frequently, about 10x15 on 11x17 inch paper. But those are my needs, not yours.) 

 

There are gains to be had and costs as well. Maybe Leica could have made a 31, 36, 42, or 50 Mpixel SL at this time, charged a bit more for it, and they'd avoid all this nattering over pixel resolution. I would probably have bought it too if it weren't much more than another $1000 over the cost of the SL. Because $8500 or so would definitely have put me off enough that I would likely not go much higher ... the reason I haven't bought into the S system or other larger pixel/larger format is primarily the diminishing returns, the benefit/cost ratio, that doing so entails. 

 

BUT Leica chose to make the SL as it is and sell it at the price that it is. It is a good camera, and produces outstanding quality with the available lenses to use on it. It is priced at the same level as their other 24 Mpixel cameras and serves both the same audience as well as the expanded audience that wants to take advantage of the higher level of automation and features it delivers. It is what it is.

 

I see no point to bickering forever about the fact that it isn't 31, 36, 42, or 50 Mpixel resolution, or whining interminably that it must be, should be, etc. I am going to use the SL as much as I can and enjoy the heck out of it, just as it is. And when a 50 or 100 Mpixel version comes out, presuming I'm still breathing and still have enough money to consider the expenditure, I'll decide whether I want that then. I believe the current one satisfies my needs very well, and then some, but who knows? My needs and expectations might have changed by the time that new model comes about. I'm open to anything. 

 

Enough of this silliness. Let's get back to discussions of how to use the SL to best advantage, and how to make photographs with it that delight us and the people we share them with. 

 

This thread needs more photographs ... 

 

Polaroid SLR670a by MiNT

The Impossible Project 600 Color, revision 3, beta release

 

"... Equipment is transitory. Photographs endure. ..."