Sensor vs Lens constraint

[ALD]

Advertisement (gone after registration)

"Do you have any idea on how large sensor resolution one can use with the current 50mm cron and lux before the lens becomes the constraint?"

 

 

The above sentence was a quote from another thread but it's interesting to me. Does anybody know what kind of sensor would be able to capture everything that the lens can possibly give and any sensor improvement beyond that is meaning less?

[jaapv]

Start reading.....

 

Understanding resolution and MTF

[01af]

"Do you have any idea on how large sensor resolution one can use with the current Summicron-M 50 mm and Summilux-M 50 mm Asph before the lens becomes the constraint?"

The lens never will be the constraint. Instead, it always is one constraint. Whenever a resolution-limited lens and a resolution-limited sensor collaborate to produce a picture, the image resolution will always be constrained by both the lens and the sensor.

 

 

Does anybody know what kind of sensor would be able to capture everything that the lens can possibly give and any sensor improvement beyond that is meaningless?

A sensor with an infinite number of pixels.

[adli]

Well, that is true, but at some point, higher sensor resolution will not give better resolution in the picture, because you have reached the limit of the lens. The better resolution of the lens, the higher pixel density sensor you can gain from.

 

 

The lens never will be the constraint. Instead, it always is one constraint. Whenever a resolution-limited lens and a resolution-limited sensor collaborate to produce a picture, the image resolution will always be constrained by both the lens and the sensor.

 

 

[01af]

... but at some point, higher sensor resolution will not give better resolution in the picture, because you have reached the limit of the lens.

Arrrgh!

 

Am I talking Greek or Chinese or something!? The sensor will reach 'the limit of the lens' at X megapixels where X is infinity. In other words—never. There is no such thing as a 'limit of the lens' actually.

 

You can introduce a kind of artificial limit when you define another parameter to the equation which would be something like 'return on investment.' The returns in terms of effective image quality will become smaller and smaller (but not zero) when the sensor resolution grows beyond any reasonable magnitude while the lens resolution does not. But that's an entirely subjective thing, not a physical or mathematical.

 

 

The better resolution of the lens, the higher pixel density sensor you can gain from.

This sentence is garbled and doesn't make any sense, grammatically. So I cannot comment on the logical sense because there is none.

 

However I do agree with the notion that higher-resolution lenses are better than lower-resolution lenses, if that is what you wanted to say. And that's true on any sensor.

[adli]

Try reading this:

Do Sensors “Outresolve” Lenses?

 

 

"So, do sensors outresolve lenses? It depends on the lens you use, the properties of the light, the aperture and the format. Small format sensors may have surpassed the limit, this is, in most cases they are lens-limited in terms of resolution."

 

Thank you and goodbye.

[01af]

Advertisement (gone after registration)

Try reading this: Do Sensors “Outresolve” Lenses?

Yes, I know this half-educated nonsense.

 

 

"So, do sensors outresolve lenses? It depends on the lens you use ..."

No, it doesn't.

[jdlaing]

71.994 Megapixels.

[haroldp]

I think what some folks here are trying to say, is that the resolving power of a lens is not a binary value but, represents reducing contrast at boundary conditions, which is what MTF tries to quantify ( I know this is a simplification ). It is not a step function even though stats like lp/m try to represent it as such.

 

The greater the resolution of the sensor, the better it will represent these contrast changes around boundary conditions.

 

Therefore, a higher resolution sensor will always better represent what any lens is projecting.

 

Just like higher resolution (finer grain) film will always better represent what any lens is projecting.

 

Lenses also evolve and improve. Today's are generally better than they were 25 years ago because of :

 

- CAD

- Better glass availability,

- CNC machines and modern casting make it practical to produce on an industrial scale, shapes that were prohibitive in the past.

- Digital sensors and 'pixel peeping' (and finer grained films ) displayed the flaws in lenses that were once considered perfect, and created a market for better lenses..

 

There is no reason to believe that as higher resolution sensors enable greater enlargement, that lens performance will not be driven to improve by the market this creates.

 

Performance of the best military and industrial optics is at a much higher level than available in consumer (even professional) products.

 

Regards ... Harold

[jank]

Can we put it this way:

 

The best design is best compromise ( for the money spent).

The best design would mean that either slight deterioration of the lens resolution or sensor resolution will became noticeable.Giving best match for the money spent.

I do not think any of us will will accept it, for we all hope for either will became better tomorrow.

Jan

[Michael Geschlecht]

Hello ALD,

 

Welcome to the Forum.

 

Hello 01af,

 

The answers you gave to ALD's & adli's questions/comments appears to be a combination of an explanation of the general parameters of lens resolution & sensor ability combined w/ a semantical analysis of what was added by adli.

 

Ironically in your response to adli in Post #5 of this Thread you answered the question which you said was effectively unanswerable.

 

When I began teaching I realized my students were quite bright but sometimes they did not ask the questions they wanted answered.

 

Sometimes part of teaching is to teach people what they need to know.

 

For them to know what question to ask.

 

To get the answer to the question that is in their mind.

 

When you proceed w/ a semantical critique: "There is no such thing as a 'limit of the lens' actually." Or when you comment on the written English of someone who's first language is not English. Then you are taking away from yourself. You are a better person than that.

 

It is clear you have a good understanding of this topic & it would be good if you would share your knowledge w/ us & answer the questions meant. As you began to explain in Post #5 of this Thread:

 

"When the sensor resolution grows beyond any REASONABLE magnitude while the lens resoloution does not".

 

Definition of REASONABLE needed. An interesting discussion.

 

Change a word here & there & you could explain the parameters for lenses.

 

Sometimes things are very simple retrospectively once you ALSO know something which was apparently unimportant before the fact.

 

Wheels which are retrospectively very simple & reasonable were only invented relatively recently. For most of the entire existence of modern people just like us on this Planet there were none.

 

Best Regards All,

 

Michael

[leicafan!]

When you proceed w/ a semantical critique: "There is no such thing as a 'limit of the lens' actually." Or when you comment on the written English of someone who's first language is not English. Then you are taking away from yourself. You are a better person than that.

 

I've noticed an increase of 'jerk-ness' on the forum of late! What gives guys?

[rramesh]

"Do you have any idea on how large sensor resolution one can use with the current 50mm cron and lux before the lens becomes the constraint?"

 

 

The above sentence was a quote from another thread but it's interesting to me. Does anybody know what kind of sensor would be able to capture everything that the lens can possibly give and any sensor improvement beyond that is meaning less?

Wouldn't the constraint be the human eye?

[lars_bergquist]

The (original) question is extremely pedantic. The main restrictive component in the equation is the Mk1 eyeball, which has been with us longer than we have been with us – it is the basic vertebrate eye. When our ancestors not too long ago lived in wooded savannas where it was a vital matter to spot the leopard in the grass, there was an intense evolutionary pressure to increase visual acuity (and to improve the image-processing wetware which is what we actually do see with). So the equipment we have is likely the best that we can have. And it has limitations, beginning with the famous retinal limit of c. 0.7 minutes of arc, and continuing with wetware artefacts like visual illusions.

 

This constraint, I know, is not valid for technical and scientific applications, where the captured image can be used in entirely different ways, circumventing these constraints, but it is eminently valid for purposes of 'general photography'. Vide e.g. the ongoing discussions of valid depth-of-field criteria.

 

So I am just opening the door, sticking my whitened head in to make this comment, and then I leave discreetly, closing the door behind me without banging it.

 

The old man from the Kodachrome Age

[sdk]

The human eye has a pretty lousy lens but the retina is an excellent sensor. Post processing removes the lens’ chromatic aberration and other aberrations, corrects distortion and enhances edges. Even more complex processing occurs next: flip the image so up appears up, to detect vertical horizontal and diagonal lines and movement of objects against backgrounds. Meanwhile the eyes are controlled to autofocus, scan the scene with saccadic movements and the visual cortex tracks moving objects, while persistence of vision maintains the illusion that the person looking at a scene is seeing an entire scene at once, but in reality the eye only sees a tiny part of it at any given set of milliseconds. While RAW processing software is able to do some post processing, such as removing chromatic aberrations and distortion, cameras still don’t have the pixel density to remove other image flaws in post processing. Some day we will get there.

 

I think the example of the vertebrate eye and the complexity of processing in the visual cortex of the brain points to the sensor and image post-processing being somewhat more important than the lens. However, a better lens will certainly help. People who have cataract surgery that replaces their damaged lens with an artificial lens often report being able see better than before they developed cataracts (at the focus distance of the new lens and at the expense of being able to focus to many distances because the artificial lenses are inflexible compared to the native lens). So all parts of an imaging system are important to minimizing image flaws and maximizing image quality.

[giordano]

When our ancestors not too long ago lived in wooded savannas where it was a vital matter to spot the leopard in the grass, there was an intense evolutionary pressure to increase visual acuity (and to improve the image-processing wetware which is what we actually do see with). So the equipment we have is likely the best that we can have.

 

"The best" only in terms of the trade-off between our repertoire of genes, time available for evolution to operate, available space, energy consumption and so on. Some other vertebrate species - e.g. birds of prey - have evolved much greater visual acuity.

[lars_bergquist]

The human eye has a pretty lousy lens but the retina is an excellent sensor.

 

I think the retina is a pretty lousy sensor too. Look at the wiring – all the connections in front of the light-sensitive cells, not behind them, where a sensible (1) designer would put them. Our visual experience is actually a piece of virtual reality, built up from cues arriving from the eyes, plus lots of innate preconceived notions about what reality is like. Sometimes these notions are wrong, giving rise to the well-known illusions and 'impossible figures' (Necker's cube, Escher's prints etc.)

 

The old man who can't shut up

[adli]

To conclude, a lousy sensor and a lousy lens gives the best result

 

I think the retina is a pretty lousy sensor too. Look at the wiring – all the connections in front of the light-sensitive cells, not behind them, where a sensible (1) designer would put them. Our visual experience is actually a piece of virtual reality, built up from cues arriving from the eyes, plus lots of innate preconceived notions about what reality is like. Sometimes these notions are wrong, giving rise to the well-known illusions and 'impossible figures' (Necker's cube, Escher's prints etc.)

 

The old man who can't shut up

[pgk]

Does anybody know what kind of sensor would be able to capture everything that the lens can possibly give and any sensor improvement beyond that is meaning less?

Yes. A simple enough answer is that it would be a sensor which could record each and every individual photon arriving from the lens;). What this would achieve I have no idea....

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[01af]

Yes. A simple enough answer is that it would be a sensor which could record each and every individual photon arriving from the lens;). What this would achieve I have no idea ...

A light meter.