Why not more pixels in the M camera?/ 36 MP {merged}

[01af]

Advertisement (gone after registration)

I find it very hard to understand how one pixel ...

That's because you're thinking about one pixel and one point. But a picture has many pixels, and a subject, even more points.

[jaapv]

Yes, but it is resolved in discrete points ( pixels) so it is legitimate to analyze what is happening at pixel level, as that is the only thing  that will end up in your file..

[pgk]

Like many things in life, motion blur is not as simple as we would like it to be. Its about linear motion and a number of angular deflections which can reinforce or cancel. Its also about lens resolving power and whether the sensor has already exceeded the ability of the lens to produce a discrete and distinct single pixel point. So trying to think of a complex situation in any simplistic terms is pointless. All in all this is a fruitless discussion.

 

We need to remember that photography is a practical application and that for the arguments being made here the final print is the only way to determine results - pixel peeping is all very well but fundamentally pointless except for very specific applications such as scientific ones. Anyone who really needs to print very large prints can look at available options - the current M rf cameras are one of these but there are and will be many others.

 

Personally if I had an absolute need to produce very large and extremely detailed prints I would probably not use an M rf camera - the emphasis being on absolute need. But for what I use my M rfs for they are great as they are. And FWIW, unless I had to I would not rely on any form of IS either - a good solid tripod is far better unless stability from a tripod is not possible.

[01af]

Yes, but it is resolved in discrete points (pixels) so it is legitimate to analyze what is happening at pixel level ...

Sure. But you have to do it the right way. If you only look at one case and keep overlooking the others then your analysis will be incomplete. The fact that the image is made up from discrete points doesn't mean the subject also was divided into the same discrete points which get 1:1 mapped onto the pixels.

 

For example, think of a motion that leads to motion streaks half a pixel wide. The streak of one point may stay within one pixel but the streak of the next point will cross a pixel border, causing visible blur. So yes, the blur for the picture as a whole will be less then with one-pixel streaks ... but still more than when there was no motion at all.

 

.

Edited February 6, 2018 by 01af

[Tailwagger]

I find it very hard to understand how one pixel could be able to differentiate between a pinpoint dot and a stripe that is smaller than itself.

 

Every camera on the planet experiences some amount of vibration and motion once the shutter button is pressed per Newton.  Yet somehow we don't (or do we?) see evidence of motion blur in every photograph we take.  IIRC wasn't the knock on the original A7R that it suffered from precisely this sort of problem as a result of shutter vibration? Yet, presumably the same shutter mechanism resulted in no artifacts or complaints  on the base A7.  If the difference between the two wasn't the size of the pixel site, then can someone please explain what was? 

[jaapv]

Sure. But you have to do it the right way. If you only look at one case and keep overlooking the others then your analysis will be incomplete. The fact that the image is made up from discrete points doesn't mean the subject also was divided into the same discrete points which get 1:1 mapped onto the pixels.

 

For example, think of a motion that leads to motion streaks half a pixel wide. The streak of one point may stay within one pixel but the streak of the next point will cross a pixel border, causing visible blur. So yes, the blur for the picture as a whole will be less then with one-pixel streaks ... but still more than when there was no motion at all.

 

.

Ok, that is correct. But halved motion blur is better than doubled by smaller pixels... Interestingly motion blur can be virtually eliminated by forensic pixel-shifting programs like Focus Magic.

[Martin B]

Advertisement (gone after registration)

Every camera on the planet experiences some amount of vibration and motion once the shutter button is pressed per Newton.  Yet somehow we don't (or do we?) see evidence of motion blur in every photograph we take.  IIRC wasn't the knock on the original A7R that it suffered from precisely this sort of problem as a result of shutter vibration? Yet, presumably the same shutter mechanism resulted in no artifacts or complaints  on the base A7.  If the difference between the two wasn't the size of the pixel site, then can someone please explain what was? 

 

A7R had a full mechanical shutter which caused at certain shutter speeds vibration issues. The A7 had an electronic first curtain shutter which avoided this vibration issue the A7R suffered from. 

Edited February 6, 2018 by Martin B

[tobey bilek]

I have two nearly new Nikon D3 @ 12 MP andI have made 12x18 at 1600 ISO.  Given away as gifts.    36 MP Nikon are better,  but I do not need bigger prints.  Any digital M produces files acceptable to me and I have used them all.  

 

For landscapes,  you can always stitch.   

[01af]

But halved motion blur is better than doubled by smaller pixels ...

Arrgh ... dammit, how hard can it be to get that foolish pixel-centered thinking out of your head!? No, smaller pixel WILL NOT (repeat: N-O-T) double motion blur! Return to post #118 and repeat reading from there.

[jaapv]

The blur itself will not change, the rendering will be halved  by larger pixels by your own post.

 

 

For example, think of a motion that leads to motion streaks half a pixel wide. The streak of one point may stay within one pixel but the streak of the next point will cross a pixel border, causing visible blur. So yes, the blur for the picture as a whole will be less then with one-pixel streaks ... but still more than when there was no motion at all.

[01af]

... the rendering will be halved by larger pixels by your own post.

Oh. My. God.

 

No, it won't 'be halved by larger pixels' by my own post. Instead, it will be doubled when the motion streaks are twice as long.

 

Really. Here's some good advice to you: Get those damned pixels out of your head. Think of pictures, not pixels. You'll come much closer to reality.

[jaapv]

Reality being the number of complaints about motion blur on high-MP cameras, the advice to use shorter shutter.speeds, etc.

 

Up until now you have only contributed shrill denials, no solid arguments. The geometrical explanation still stands.

[ramarren]

I sympathize with all those who believe they must have more pixels, but I don't agree.

 

I was printing 20x24 inch prints with a 5 MPixel DSLR with a quarter-frame sensor (Olympus E-1) a decade ago. Not one photo buyer, judge, or client ever complained that the photographs lacked sufficient detail. Seems to me that this approximate pixel density (FourThirds @ 5 MPixel, FF @ 24 MPixel) just works very very well. 

 

I don't need more. I have other cameras that do up to 80 MPixel ... I haven't seen one of my photos yet that actually benefitted from it. Since I'm standardized on the Leica M-D and SL now (and the Light L16 as well), and everything I've done for some years now is as good as it needs to be technically, I'll just stick here and enjoy what I have.

 

... And trust that whatever Leica does to the next round of SL and M products will be as good or better performers as what I have already—even if I don't need and probably will not buy anything for many years to come.

[pgk]

In a world where picture creation becomes technically ever better images seem to be viewed more and more poorly. On a very different forum last night some photos were posted and queries made as to how the raw files had been processed. They were 4k frames ..... . We really need to start thinking in terms of 'fit for purpose' rather than 'must have more' which ultimately will fail to deliver (and to an extent probably already is). Current photographic equipment is extraordinarily good and IMO and experience exceeds the need of many users already. Why not celebrate how fortunate we are instead of wanting 'better'? 

[Alberti]

I agree with the lower demanding folks.

I have perfect 40x60 M8 prints - where the pixels are seen and they are immensely sharp. They have an ultimate liquidity. 

Higher pixels of M240 do not give me more joy, in fact I have seen the black areas tend to be reduced a bit, leading to smurkiness in the background and (occasional) banding when they are pulled a bit.

It is not the 36MP but other trade-offs that will occur. Such as less DR = flatter images. I have seen it with my sisters Nikon, she went from the 200 (CCD) to 36 MP, and they just flattened, had less IQ - and she started to photograph less.

 

To me an 18MP CCD coupled with a Maestro processor (so the newest technology with the most silent image processing chips) to extract the ultimate detail is a sub goal on pixel count but might in every way (by looking at the output) beat the output of the higher mp chips.

 

That said, I do like the M240 but looking at the M8 files I have, those 10 MP were certainly not a sloth. 

[jaapv]

I’m not surprised; stills from 4K on a 20 MP sensor are 8 MP JPGs. Far in excess of Internet needs.

[hteasley]

National Geographic used photos taken with a Nikon D100 back in 2003. If NatGeo didn't harm its century-old reputation as the greatest photo periodical in history, publishing photos from a 6mpix APS-C camera with a max of 1600 ISO, then don't worry.

 

If you think your photography is so finely attuned that you can't manage your artistic vision with a paltry 24mpix, I don't know what to tell you except, "Up your game, son." 

[lct]

Hehe yes but great artists may feel the need to crop their masterworks. Composing with the camera is not up to date any more. The decisive moment happens in Photoshop nowadays .

[hteasley]

Hehe yes but great artists may feel the need to crop their masterworks. Composing with the camera is not up to date any more. The decisive moment happens in Photoshop nowadays .

 

"Faugh! I can only crop my M10 image by 75% before the quality might dip below National Geographic's standards! Blood and thunder!"

[Advertisement]

Hallo,

Look here: Leica M

[Lonescapes]

I LOLd.