m11 camera shake VS m10-r

[harmen]

Advertisement (gone after registration)

8 hours ago, DadDadDaddyo said:

The motion smears the edge (or what have you) being captured across the boundaries between pixels. A black point becomes a grey streak. When the pixels are smaller, it takes less motion to smear an edge or point across the pixels capturing it. A shake resulting in movement of less than a pixel's shake should be invisible. But today's pixels are really really tiny. 

Not quite.

(1) Details that are smaller than a pixel still affect the overall amount of light being recorded for that pixel.  Think of edges of objects or shades along a surface rather than specs of tiny detail and you will realize that movement will always change the amount of light recorded by each pixel.

(2) Movement is the same regardless of the size of the pixels, so the impact on the overall image is the same.  It’s just that you can zoom in further when you have more pixels.

8 hours ago, DadDadDaddyo said:

Now remember, lenses are analog. They don't have megapixels. A lens that's capable of less resolution than the sensor being used will not be improved with a higher-res sensor. Rather, that higher-res sensor will simply give you a higher resolution rendition of the portion of the image circle produced by that lens that's captured by the sensor. (an aside: note I didn't use the term "rendering" - Gawd how tired I am the near infinite uses to which that incredibly sloppily-defined term has been put! If you're not sure how to describe what you see in a picture, be it an artifact of the camera, the sensor, the lens, or the technique, you can always say you like (or don't like) how it renders. (I reckon it means about the same as "comes out"....))

Not quite.  Lens and sensor are independent.  There is no specific value for lens resolution beyond which a sensor of a certain number of megapixels will suddenly have no impact on overall image resolution.  Sure, you many not want to invest too much in an amazing sensor when you’re better off buying a good lens, but there is no step-change; just a gradual change.

8 hours ago, DadDadDaddyo said:

Oh and by the way, the amount of motion (in terms of angle) that's imparted by the lens to the sensor is also a function of the focal length of the lens! Longer lenses are harder to hand hold. 

That depends on the kind of movement.  If you move the camera+lens without changing the angle, then it depends on the distance to the object you are taking a photo off.  If it’s the moon, you won’t notice any change.  If it’s your finger, it may no longer be in view.  For a rotating movement, yes, it’s purely a factor of focal length.

8 hours ago, DadDadDaddyo said:

So we practiced, back in those old dark days. We practiced hand holding exposures under available light. We developed all sorts of physical techniques to apply just before and during the exposure. We analyzed our results under a magnifying glass (a loupe, y'know). We practiced more. Over years, and thousands of exposures, our technique improved, and camera-motion-derived blur from any source became less and less of an issue among practiced practitioners. We got better at it. 

Yes, very valuable, certainly if you’re keen on using a magnifying glass or zoom in to pixel level.

8 hours ago, DadDadDaddyo said:

Sorry. End of rant. Guess I'd better go get something to eat. I've had too much coffee. My hands are trembling slightly. Doggone it, I'm never gonna to be able to get that shot this way...

Love it

[Advertisement]

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

[DadDadDaddyo]

Many thanks. Those are welcome clarifications!

[elmars]

The following was recently written by our forum friend 01af in the German forum  on the connection between camera shake and resolution: 

It convinced me. I had the German text translated by DeepL and inserted it in sections:

Der Autor begeht den in diesem Zusammenhang üblichen Denkfehler: Er geht irrigerweise davon aus, Digitalkameras könnten keine Strukturen abbilden, die kleiner sind als ein Pixel. Insbesondere könnten Unschärfestreukreise und Verwacklungsspuren die Bildschärfe nicht beeinträchtigen, solange ihr Durchmesser bzw. ihre Länge kleiner sei als der Pixelabstand des Sensors. Doch das ist falsch. Selbstverständlich kann die Verwacklungsspur eines Bildpunktes auch dann auf zwei oder drei Pixel fallen, wenn sie kürzer als der Pixelabstand wäre – wie man sich mühelos selber überlegen kann:

The author commits the usual error of reasoning in this context: he erroneously assumes that digital cameras cannot depict structures that are smaller than a pixel. In particular, blur circles and shake marks cannot affect the sharpness of the image as long as their diameter or length is smaller than the pixel pitch of the sensor. But this is wrong. Of course, the blur trace of a pixel can fall on two or three pixels even if it were shorter than the pixel pitch - as you can easily work out for yourself:

Die hier dargestellte Verwacklungsspur ist ein halbes Pixel lang und fällt trotzdem auf zwei Pixel. Sicher – für einen Bildpunkt etwas weiter links oder rechts wäre die Verwacklungsspur innerhalb des getroffenen Pixels geblieben. Aber vom Objektiv auf den Sensor projizierte Bildpunkte fallen nun einmal überall hin, nicht nur dorthin, wo es für die Verwacklung günstig wäre.

Ein Foto besteht schließlich nicht aus nur einem Bildpunkt. Sondern aus unendlich vielen, die im Falle einer Verwacklung alle in gleicher Weise verschmiert sind. Und je nachdem, wo der ursprüngliche Bildpunkt hinfiel, kann er bei einer hinreichend kleinen Verwacklung innerhalb eines Pixels bleiben – muß aber nicht. Tatsächlich werden bei einer Verwacklung, deren Länge dem halben Pixelabstand entspricht, die Hälfte aller Verwacklungsspuren eine Pixelgrenze kreuzen – was selbstverständlich die Bildschärfe insgesamt beeinträchtigt.

Und sobald man sich das überlegt hat, wird auch sofort klar, daß höher auflösende Sensoren keineswegs verwacklungsempfindlicher sind als gleich große, aber niedriger auflösende. Ganz im Gegenteil! Höher auflösende Sensoren erzeugen immer das bessere Bild, weil sie alles höher auflösen – inklusive Verwacklungsspuren.

The shake trace shown here is half a pixel long and yet falls on two pixels. Sure - for a pixel a little further to the left or right, the shake trace would have remained within the pixel that was hit. But pixels projected onto the sensor by the lens fall everywhere, not only where it would be favourable for the blur.

After all, a photo does not consist of just one pixel. It consists of an infinite number of pixels, all of which are blurred in the same way in the event of camera shake. And depending on where the original pixel fell, it can remain within one pixel if the blur is sufficiently small - but it does not have to. In fact, for a blur whose length is half the pixel spacing, half of all blur traces will cross a pixel boundary - which of course affects overall image sharpness.

And as soon as you think about it, it also becomes immediately clear that higher resolution sensors are by no means more sensitive to blur than sensors of the same size but lower resolution. Quite the opposite! Higher resolution sensors always produce the better image because they resolve everything higher - including shake marks.

Dieselbe Verwacklungsspur wie oben auf einem linear doppelt so hoch auflösendem Sensor, also vierfache Pixelzahl. Was vorher auf zwei großen Pixeln abgebildet wurde, fällt nun auf zwei kleine. Ergo: höhere Bildschärfe. Zwar wird nun für nahezu jeden Bildpunkt eine Pixelgrenze überschritten statt wie zuvor nur für jeden zweiten. Dafür ist aber die Fläche zweier Pixel nur noch halb so groß wie die Fläche eines einzigen größeren Pixels. Trotz gleich starker Verwacklung ist das Bild also immer noch höher aufgelöst. Klar: Hätte man nur halb so lang belichtet, so wären die Verwacklungsspuren auch nur halb so lang und das Bild insgesamt schärfer ... aber das wäre es beim niedriger auflösenden Sensor auch.

Deswegen ist dieses Gerede von der angeblich höheren Verwacklungsempfindlichkeit hochauflösender Kameras so unsinnig. Ob eine Verwacklung unterschwellig bleibt oder störend ins Auge fällt, hat nichts mit der Pixelzahl oder -größe zu tun, sondern hängt allein von der Vergrößerung des Bildes ab. Die gleiche Verwacklung führt bei gleicher Vergrößerung zur gleichen Beeinträchtigung der Bildschärfe – einerlei, wie viele Pixel der Sensor hat. Je größer geprintet werden soll, desto höher der Anspruch an die Bildqualität und desto sauberer muß gearbeitet werden, basta. Das war auf Film schon so, und das ist mit digital ganz exakt genauso.

The same shake trace as above on a linear sensor with twice the resolution, i.e. four times the number of pixels. What was previously imaged on two large pixels now falls on two small ones. Ergo: higher image sharpness. It is true that a pixel limit is now exceeded for almost every pixel instead of only for every second pixel as before. But the area of two pixels is only half as large as the area of a single larger pixel. So despite the same amount of blurring, the image still has a higher resolution. Of course, if the exposure had been half as long, the blur would have been half as long and the image would have been sharper ... but that would be the case with a lower resolution image. but so would a lower resolution sensor.

That's why all this talk about the allegedly higher camera shake sensitivity of high-resolution cameras is so nonsensical. Whether a blur remains subliminal or is disturbingly noticeable has nothing to do with the number or size of the pixels, but depends solely on the magnification of the image. The same blurring at the same magnification leads to the same impairment of image sharpness - regardless of how many pixels the sensor has. The larger the print, the higher the demand on the image quality and the cleaner the work must be, basta. That was already the case with film, and it is exactly the same with digital.

[Jeff S]

Magnification… page one… still not acknowledged or discussed by OP.

Jeff

[01af]

Yes—magnification is the crucial factor here. It's funny how many pixel peepers cannot understand that 100 % of 60 MP means a much higher magnification than 100 % of 24 MP. And on top of that there's the usual misconceptions about a pixel allegedly not being able to record anything that's smaller than itself.

Thank you, Elmar elmars, for transferring my older post from the German-language forum. Unfortunately there's an error in the automatic translation into English:

Zitat

After all, a photo does not consist of just one pixel. It consists of an infinite number of pixels ...

Here, the word pixel is not the proper translation of 'Bildpunkt'. It should read 'image point' instead. It's the projection of an infinitesimally small portion of the subject (a 'point') into image space—which eventually will fall on a pixel, or more precisely, a sensel (sensor element), along with many more image points.

So, to recapitulate: Loss of sharpness from camera shake DOES NOT depend on pixel count, pixel size, or pixel pitch. So the Leica M11 is not in any way more prone to shake than the M10-R, M10, M9, M6, M3, or IIIf. Or any other 35-mm-format camera.

[hdmesa]

3 hours ago, 01af said:

...So the Leica M11 is not in any way more prone to shake than the M10-R, M10, M9, M6, M3, or IIIf. Or any other 35-mm-format camera.

Excluding hardware differences that physically introduce motion/shutter shock.

[sometimesmaybe]

Advertisement (gone after registration)

6 hours ago, 01af said:

Yes—magnification is the crucial factor here

i assume the same logic applies with ISO noise of high MP cameras... but i guess thats something for a different topic

[01af]

vor 11 Stunden schrieb sometimesmaybe:

.. but I guess that's something for a different topic.

No. That's just another flavour of the very same topic.

And while we're at it—limited lens performance also is included in this topic. While hard-core pixel peepers believe they're forced to replace their whole lens collection with the latest and finest optics to keep up with their new super-high-MP behemoth camera (remember the foolish Internet discussions around the introduction of the first 36-MP 35-mm-format DSLR camera, the Nikon D800?), seasoned practitioners know that their lovely pre- or post-war lenses will perform even lovelier on the latest megapixel monster.

[Jeff S]

As Roger Cicala discusses in “Why Perceptual Megapixels are Stupid” ….

 

A better (higher MP) camera will reveal more of the character of the lens.

Jeff

 

 

Edited April 10, 2023 by Jeff S

[DadDadDaddyo]

Excellent material! Many thanks for the translations and the post!

[sometimesmaybe]

On 4/10/2023 at 1:25 AM, Jeff S said:

Magnification… page one…

i just went and looked at the images on my phone and they look acceptably sharp to my eyes... i guess it's this is the digital version of viewing distance for prints

i should add @ryanlowry there are AI tools that can help... i just ran your images (top row) through my google phone. the results (bottom row) doesn't look too bad 🤓

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[Heathom]

None of this is rocket science . I’m a photojournalist and was forced by the nature of my work into going digital in the late 90’s. By the early noughties when digital resolution began to outstrip 35mm film stock many of us began to notice camera shake at shutter speeds where it wouldn’t have been a problem with film . This has continued as resolution has increased , fortunately most modern high MP cameras have image stabilisation which helps to counteract this. The original 50mp Canon 5ds suffered terribly from camera shake and the new generation of mirrorless cameras now utilise stabilised sensors like the 48MP Canon R5 and the 100MP med format Fujis .

Of course technique helps to a certain extent but I know of experienced colleagues who have noticed the same issue since upgrading to the M11 . 60MP in a “full frame” sensor without stabilisation is pushing the boundaries, the Q2 47mp sensor camera has a stabilised lens which’ll help and it’ll be interesting to see where Leica go with the M series in the future . Perhaps those quirky little Leica camera shoulder stocks will gain popularity again, I’ll be sticking with my 24mp rangefinder for now  😂

[Jeff S]

11 hours ago, Heathom said:

None of this is rocket science . 

 

Indeed, yet you failed to address the key issue.  Read above.

Jeff

[Tailwagger]

One muses around what 100% magnification means in the context of film enlargement. 

[nwphil]

On 4/2/2023 at 5:08 PM, Harout62 said:

Can we go even lower???? how about 1/8th of a second 

All this was taken in my garage 50mm Summilux Asph Hand held 60mpx, The M11 has no issues Learn to use it or just sell it and move on 

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

I could not agree more with the statement - I am too guilty of poor handling technique, that not even faster shutter speeds cover up my mistake - maybe software can, but regarding the hardware part the fault it's on me. IS is a blessing and a wonderful tool. So is AF - Leica M does not have either, and that's the whole point....and a long line of learning how to use it.

OP, you need to improve your handling skills - that's all

Edited April 12, 2023 by nwphil

[01af]

vor 22 Stunden schrieb Heathom:

None of this is rocket science.

Indeed it is not. Yet ... uh, what Jeff Jeff S said.

.

vor 22 Stunden schrieb Heathom:

... many of us began to notice camera shake at shutter speeds where it wouldn’t have been a problem with film.

And why did you begin to notice? Because you were inspecting your exposures using a much stronger loupe than ever before. On a 24-inch 2K monitor, inspecting a 24-MP image file in 100 % view corresponds to beholding a 1 × 1.5 m print (42 × 64 inches) from reading distance. A 60-MP file, on the same monitor at 100 %, corresponds to a 1.7 × 2.5 m print (67 × 100 inches). Did you ever print a 35-mm negative this big in your life?

But when printing (as opposed to pixel-peeping), camera shake kept—and still keeps—being exactly the same problem as it ever was ... or even less of a problem when image stabilisation is available.

Edited April 12, 2023 by 01af

[jonoslack]

13 minutes ago, 01af said:

Indeed it is not. Yet ... uh, what Jeff Jeff S said.

.

And why did you begin to notice? Because you were inspecting your exposures using a much stronger loupe than ever before. On a 24-inch 2K monitor, inspecting a 24-MP image file in 100 % view corresponds to beholding a 1 × 1.5 m print (42 × 64 inches) from reading distance. A 60-MP file, on the same monitor at 100 %, corresponds to a 1.7 × 2.5 m print (67 × 100 inches). Did you ever print a 35-mm negative this big in your life?

But when printing (as opposed to pixel-peeping), camera shake kept—and still keeps—being exactly the same problem as it ever was ... or even less of a problem when image stabilisation is available.

Thank you for this O1af - so nicely put and thanks also to @elmars for. bringing it over here. . . . sadly explaining things properly doesn't often seem to persuade people!

[hdmesa]

6 hours ago, jonoslack said:

…sadly explaining things properly doesn't often seem to persuade people!

Some people post about their concerns in order to learn, but some are only seeking to have their opinions validated.

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[billm212]

On 4/5/2023 at 9:11 PM, SrMi said:

The tripod does not help with the shutter shock.

It certainly does, as it allows the use of slower shutter speeds without the interference, if any, from image stabilization. Before cameras came out with EFCS, that was the standard way of mitigating or eliminating shutter shock. In any case, the method used by the OP has too many variables to determine whether it’s his lens or something else that causes unsharpness, and  use of a tripod has always been standard in serious lens tests, for obvious reasons - it eliminates the effect of handholding technique, to state the obvious, in addition to mitigating or eliminating shutter shock or mirror slap.The OP is certainly entitled to shoot handheld, but if he really needs to find out the cause of unsharp pictures, he will not be able to with any degree of reliability.

[SrMi]

24 minutes ago, billm212 said:

It certainly does, as it allows the use of slower shutter speeds without the interference, if any, from image stabilization. Before cameras came out with EFCS, that was the standard way of mitigating or eliminating shutter shock. In any case, the method used by the OP has too many variables to determine whether it’s his lens or something else that causes unsharpness, and  use of a tripod has always been standard in serious lens tests, for obvious reasons - it eliminates the effect of handholding technique, to state the obvious, in addition to mitigating or eliminating shutter shock or mirror slap.The OP is certainly entitled to shoot handheld, but if he really needs to find out the cause of unsharp pictures, he will not be able to with any degree of reliability.

You need a very, very stable tripod to mitigate the shutter shock (caused by the first curtain). Typically, shutter shock is visible even on a tripod. Therefore, you need either EFCS or the electronic shutter to eliminate shutter shock completely.

Before mirrorless cameras, we had mirror-slap issues. Even on a tripod, we needed to raise the mirror before triggering the shutter.