Microcontrast

Started 4 months ago | Discussions
Great Bustard Forum Pro • Posts: 41,999
Microcontrast

First, let me define the term:  microcontrast is the contrast at high spatial frequencies.  With that definition in place, a few questions:

  1. If System A has a higher MTF-50 than System B, does that mean that it will have greater resolution at all other contrast levels?  For example, will the MTF-10 and MTF-90 of System A also be greater than System B?
  2. If (1) is true, does that necessarily imply the converse?  That is, the contrast of System A will be higher than the contrast of System B at any given resolution?
  3. Does halation affect the MTF-50 measurement?  If so, does the effect of halation result in greater degradation as the contrast level of the MTF measurement increases?
  4. Repeat (3) with veiling flare.

In the end, I wish to know if the term "microcontrast" as defined describes the photo in a photographically meaningful manner any differently than an MTF-50 measurement does.  That is, if System A has a higher MTF-50 than System B, does that automatically mean that the "microcontrast" of System A is greater than the "microcontrast" of System B?

bobn2
bobn2 Forum Pro • Posts: 58,507
Re: Microcontrast
2

Great Bustard wrote:

First, let me define the term: microcontrast is the contrast at high spatial frequencies. With that definition in place, a few questions:

  1. If System A has a higher MTF-50 than System B, does that mean that it will have greater resolution at all other contrast levels? For example, will the MTF-10 and MTF-90 of System A also be greater than System B?
  2. If (1) is true, does that necessarily imply the converse? That is, the contrast of System A will be higher than the contrast of System B at any given resolution?
  3. Does halation affect the MTF-50 measurement? If so, does the effect of halation result in greater degradation as the contrast level of the MTF measurement increases?
  4. Repeat (3) with veiling flare.

In the end, I wish to know if the term "microcontrast" as defined describes the photo in a photographically meaningful manner any differently than an MTF-50 measurement does. That is, if System A has a higher MTF-50 than System B, does that automatically mean that the "microcontrast" of System A is greater than the "microcontrast" of System B?

First stop is to read these two articles:

https://lenspire.zeiss.com/photo/app/uploads/2018/04/Article-MTF-2008-EN.pdf

https://lenspire.zeiss.com/photo/app/uploads/2018/04/CLN_MTF_Kurven_2_en.pdf

From the second, this figure which answers question 1 in the negative:

AiryDiscus Senior Member • Posts: 1,620
Re: Microcontrast
5

Great Bustard wrote:

First, let me define the term: microcontrast is the contrast at high spatial frequencies. With that definition in place, a few questions:

FYI, I believe it has more to do with MTF at low frequencies, but ok.

  1. If System A has a higher MTF-50 than System B, does that mean that it will have greater resolution at all other contrast levels? For example, will the MTF-10 and MTF-90 of System A also be greater than System B?

Certainly not. Here's a contrived example of a system with 1 wave zero to peak of low order spherical aberration, and one with 1 wave zero to peak of high order spherical aberration. The MTF at lower frequencies (and MTF50) is about the same. At high frequencies, it is very different.  MTF10 differs by more than a factor of 3, with the system that has the higher order aberration having much higher MTF10.

In general, the higher order aberrations (and scatter) reduce contrast at large scales (= low frequency) while the lower order aberrations reduce contrast at smaller scales (= high freq).

  1. If (1) is true, does that necessarily imply the converse? That is, the contrast of System A will be higher than the contrast of System B at any given resolution?

probably answered above?

  1. Does halation affect the MTF-50 measurement? If so, does the effect of halation result in greater degradation as the contrast level of the MTF measurement increases?

Halation in the sense of film?  CCD bloom?  Flare?  If you mean something along these lines, it has strongest effect at very low spatial frequencies which are typically rounded off into the "0" bin which everything is normalized by anyway.  So in that sense, no those things don't affect MTF50 unless the system is really tremendously bad.

  1. Repeat (3) with veiling flare.

See immediately above.

In the end, I wish to know if the term "microcontrast" as defined describes the photo in a photographically meaningful manner any differently than an MTF-50 measurement does.

Microcontrast as either "MTF20" or "MTF80" (or pick your point at the tails) is poorly correlated to MTF50, except in the particular example of a diffraction limited lens in the f/5.6 to f/8 range, (for a 4-6 micron or so pixel) where the system MTF is similar to a triangle function and you would only need one point to know the slope of the line.

That is, if System A has a higher MTF-50 than System B, does that automatically mean that the "microcontrast" of System A is greater than the "microcontrast" of System B?

Nope

(unknown member) Regular Member • Posts: 484
Re: Microcontrast
3

Great Bustard wrote:

First, let me define the term: microcontrast is the contrast at high spatial frequencies. [...]

Modulo that no one knows what "microcontrast" means, I don't agree with your definition - but that's what's so great about microcontrast threads. We can each have our own definition, and all be right!

Back when dinosaurs roamed the Earth, a technician was measuring a 50 mm Summicron-R lens using a test chart and he said something along the lines of: "With most lenses, the contrast decreases steadily with the finer patterns, but with the Leica lens, it holds up really well, then suddenly winks out."

So I always held that as a kind of mental model of microcontrast - trading maximum possible resolution against higher contrast at low-to-mid spatial frequencies, which I suspect gives a visually pleasing image.

hjulenissen Senior Member • Posts: 2,170
Re: Microcontrast

It seems to me that "micro contrast" is a term like "pop" or "3d-ness" that can mean pretty much anything.

Did DXO release a paper at one time where they found that the subjective quantity "sharpness" was better correlated with the _derivative_ of MTF at zero frequency (tilting) rather than MTF50 or MTF10?

I googled it but could not find it in all of the noise.

Having a better understanding of "contrast" (0 frequency mtf?) and "clarity" (mid frequency mtf?) and "sharpening" (high frequency mtf?) affects subjective impression and putting them in a common framework rather than using arbitrary words seems like a worthwhile effort. As spatial scaling seems to be a significant factor, zooming in on 1:1 pixel with these new 40MP cameras and tuning until things "looks good" at that magnification might be counter productive.

Reminds me of the illusion where you can make an image of one person turn into another person by playing with spatial filtering and the human contrast sensitivity function:

einstein.monrow.gif

-h

fvdbergh2501 Regular Member • Posts: 480
Re: Microcontrast
5

Great Bustard wrote:

First, let me define the term: microcontrast is the contrast at high spatial frequencies. With that definition in place, a few questions:

  1. Does halation affect the MTF-50 measurement? If so, does the effect of halation result in greater degradation as the contrast level of the MTF measurement increases?

Others, like AiryDiscus, have already demonstrated some MTF curves that exhibit the properties you describe, i.e., two systems with very similar MTF50 values but very different contrast at higher frequencies. I can offer some simulated images generated with the MTF Mapper tools. Here is the comparison of the MTF curves:

Interestingly enough, my parameter choices appear to produce better contrast at both low- and high frequencies.

This is what the non-microcontrast lens (corresponding to the blue curve above) looks like:

Notice the total lack of detail between the "1" and the "0" marks on the trumpet

Next up, the "microcontrast" lens:

Notice the false detail (aliasing) between the "1" and the "0" marks (click to view at 100%)

For my "normal" lens, I simulated the lens stopped down to f/16. The simulated microcontrast lens was wide open at f/1.4, but with a large amount of w040 spherical aberration.

The "glow" in the image rendered with the lens with spherical aberration is not as pronounced as I had hoped, but at least these two examples demonstrate that a lens with more contrast at higher frequencies (here taken to mean beyond Nyquist) does appear to capture "more detail", but in reality that is just aliased false detail.

Just for fun, and to produce visible "glow", I ran another simulation of a system with an MTF50 of only 0.075 cycles/pixel:

Now that is a lot of spherical aberration!

Detail Man
Detail Man Forum Pro • Posts: 16,815
Re: Microconceptualizations
1

hjulenissen wrote:

It seems to me that "micro contrast" is a term like "pop" or "3d-ness" that can mean pretty much anything.

Did DXO release a paper at one time where they found that the subjective quantity "sharpness" was better correlated with the _derivative_ of MTF at zero frequency (tilting) rather than MTF50 or MTF10?

I googled it but could not find it in all of the noise.

Here it is (co-authored by DxO Labs Chief Scientist Frederic Guichard - the basis of "BxU"):

https://pdfs.semanticscholar.org/d1b1/5bea10942277d7345d24bbf59418c95ba6ee.pdf

Another interesting article (see the software-simulated images near the end of the paper):

https://www.kth.se/social/files/542d1388f27654460dee76cf/Pinhole.pdf

Having a better understanding of "contrast" (0 frequency mtf?) and "clarity" (mid frequency mtf?) and "sharpening" (high frequency mtf?) affects subjective impression and putting them in a common framework rather than using arbitrary words seems like a worthwhile effort.

"Pick your poison(s)" of choice. "Bean counting" of sentient conscious experience abounds:

"Keep it simple, Einstein !"

From: "Analysis of Blur Measure Operators for Single Image Blur Segmentation" (2018)

As spatial scaling seems to be a significant factor, zooming in on 1:1 pixel with these new 40MP cameras and tuning until things "looks good" at that magnification might be counter productive.

Reminds me of the illusion where you can make an image of one person turn into another person by playing with spatial filtering and the human contrast sensitivity function:

einstein.monrow.gif

BobORama
BobORama Senior Member • Posts: 2,602
Re: Microcontrast
2

fvdbergh2501 wrote:

I think I have this lens! 

Just for fun, and to produce visible "glow", I ran another simulation of a system with an MTF50 of only 0.075 cycles/pixel:

Now that is a lot of spherical aberration!

 BobORama's gear list:BobORama's gear list
Pentax K-5 Pentax K-1 Sigma 10mm F2.8 EX DC HSM Diagonal Fisheye Pentax smc DA 18-55mm F3.5-5.6 AL WR Samyang 14mm F2.8 IF ED MC Aspherical +9 more
JimKasson
JimKasson Forum Pro • Posts: 21,561
Microcontrast definition
4

AiryDiscus wrote:

Great Bustard wrote:

First, let me define the term: microcontrast is the contrast at high spatial frequencies. With that definition in place, a few questions:

FYI, I believe it has more to do with MTF at low frequencies, but ok.

It's great that we all have our own definitions. Mine is a Goldilocks one compared to Brandon's and the OP's: contrast near 1/4 the Nyquist frequency.

With most of today's good cameras and excellent lenses, the highest spatial frequencies with significant contrast are above the Nyquist frequency, at least on-axis. With the OP's definition of microcontrast, the difference among excellent lenses would center on their ability to generate artifacts. I don't consider that a very useful metric.

-- hide signature --
 JimKasson's gear list:JimKasson's gear list
Nikon D5 Fujifilm GFX 50S Sony a9 Sony a7R III Sony a7 III +5 more
Jack Hogan Veteran Member • Posts: 6,679
Re: Microcontrast

fvdbergh2501 wrote:

Great Bustard wrote:

First, let me define the term: microcontrast is the contrast at high spatial frequencies. With that definition in place, a few questions:

  1. Does halation affect the MTF-50 measurement? If so, does the effect of halation result in greater degradation as the contrast level of the MTF measurement increases?

Others, like AiryDiscus, have already demonstrated some MTF curves that exhibit the properties you describe, i.e., two systems with very similar MTF50 values but very different contrast at higher frequencies. I can offer some simulated images generated with the MTF Mapper tools. Here is the comparison of the MTF curves:

Interestingly enough, my parameter choices appear to produce better contrast at both low- and high frequencies.

This is what the non-microcontrast lens (corresponding to the blue curve above) looks like:

Notice the total lack of detail between the "1" and the "0" marks on the trumpet

Next up, the "microcontrast" lens:

Notice the false detail (aliasing) between the "1" and the "0" marks (click to view at 100%)

For my "normal" lens, I simulated the lens stopped down to f/16. The simulated microcontrast lens was wide open at f/1.4, but with a large amount of w040 spherical aberration.

The "glow" in the image rendered with the lens with spherical aberration is not as pronounced as I had hoped, but at least these two examples demonstrate that a lens with more contrast at higher frequencies (here taken to mean beyond Nyquist) does appear to capture "more detail", but in reality that is just aliased false detail.

Just for fun, and to produce visible "glow", I ran another simulation of a system with an MTF50 of only 0.075 cycles/pixel:

Now that is a lot of spherical aberration!

Good show Frans, so do I take it that the latest version of MTF-Generate-Rectangle can now simulate SA?  What else is new?

Jack

NateW Junior Member • Posts: 26
Re: Microcontrast
1

I don’t have a great definition, but I am positive that the perception of micro-contrast has to do with a lot more than MTF 50 resolution. If I had to, I would describe it as the ability of a lens to render object surfaces with a high degree of tactile presence/corporeality across a broad range of image magnifications ie, no matter what size the image, it gives the impression of real light falling on a real object. My hunch is that it has a lot to do with the clarity and purity of color and specular highlights and little to do with the kind of resolution measured by MTF charts. When a lens has exceptional micro-contrast the tactile impression of depth seems to extend even into parts of the image that are out of focus, suffering from astigmatism, etc.

Hard for me to define, but you know it when you see it! While it’s far from the sharpest or best corrected lens, my 50mm Summilux Pre-ASPH has micro-contrast in spades and renders with a certain textural clarity that my Sony/Zeiss lenses lack even though they deliver higher resolution. (A7III, f/1.4, ISO 640)

Full res: https://www.dropbox.com/s/nqjl1fda2p21tip/_A731782.jpg?raw=1

fvdbergh2501 Regular Member • Posts: 480
Re: Microcontrast

BobORama wrote:

fvdbergh2501 wrote:

I think I have this lens!

Well, I like it! It has tactile qualities almost entirely unlike dragging your fingernails across a chalkboard, with just a hint of plum and tobacco

fvdbergh2501 Regular Member • Posts: 480
Re: Microcontrast
1

Jack Hogan wrote:

fvdbergh2501 wrote:

Now that is a lot of spherical aberration!

Good show Frans, so do I take it that the latest version of MTF-Generate-Rectangle can now simulate SA? What else is new?

I have not yet released this version, but rendering of systems with SA and defocus will definitely be included in version 0.7.4.

Another new-ish feature in the 0.7.x series is that the GUI now uses OpenGL for displaying images (and MTF Mapper results), so zooming now works much  better, and I have tweaked the "lens profile" and "grid" output styles to work with Imatest-style test charts (although those are quite sparse).

-F

57even Forum Pro • Posts: 12,757
Re: Microcontrast

There is no direct correlation between micro-contrast, which is a perceptual phenomenon, and MTF, which is a measured one.

Our peak spatial sensitivity when looking at an image peaks at about 5-8 cycles/degree. So, MTF at the corresponding image frequency has most influence on micro-contrast.

Depending on the size, resolution and the viewing distance, this may correspond to just about any spatial frequency in the image. There is no way to tell unless it is specified.

Most of the time it corresponds to higher MTF values, such as MTF 80, but only because images are normally printed so that the angular resolution is conducive to viewing, which is around 70 pixels/degree or 35 cycles.

-- hide signature --

Reporter: "Mr Gandhi, what do you think of Western Civilisation?"
Mahatma Gandhi: "I think it would be a very good idea!"

 57even's gear list:57even's gear list
Fujifilm X-Pro2
OP Great Bustard Forum Pro • Posts: 41,999
Thank you all!

Great Bustard wrote:

First, let me define the term: microcontrast is the contrast at high spatial frequencies. With that definition in place, a few questions:

  1. If System A has a higher MTF-50 than System B, does that mean that it will have greater resolution at all other contrast levels? For example, will the MTF-10 and MTF-90 of System A also be greater than System B?
  2. If (1) is true, does that necessarily imply the converse? That is, the contrast of System A will be higher than the contrast of System B at any given resolution?
  3. Does halation affect the MTF-50 measurement? If so, does the effect of halation result in greater degradation as the contrast level of the MTF measurement increases?
  4. Repeat (3) with veiling flare.

In the end, I wish to know if the term "microcontrast" as defined describes the photo in a photographically meaningful manner any differently than an MTF-50 measurement does. That is, if System A has a higher MTF-50 than System B, does that automatically mean that the "microcontrast" of System A is greater than the "microcontrast" of System B?

Apologies for not chiming back in earlier.  I want to say I very much appreciate all the great responses!  It's going to take me a bit to think of intelligent questions to ask.

My goal is to give/create a more precise interpretation of "microcontrast", even if it cannot be quantified as a scalar, as implied in the OP.  One might question the utility of such a goal, but given that the term is often used in connection to the visual properties of a photo, I thought it worthwhile to define the term more precisely in a manner that is consistent with what people are talking about when using the term.

D Cox Forum Pro • Posts: 21,524
Re: Microcontrast definition

JimKasson wrote:

AiryDiscus wrote:

Great Bustard wrote:

First, let me define the term: microcontrast is the contrast at high spatial frequencies. With that definition in place, a few questions:

FYI, I believe it has more to do with MTF at low frequencies, but ok.

It's great that we all have our own definitions. Mine is a Goldilocks one compared to Brandon's and the OP's: contrast near 1/4 the Nyquist frequency.

With most of today's good cameras and excellent lenses, the highest spatial frequencies with significant contrast are above the Nyquist frequency, at least on-axis. With the OP's definition of microcontrast, the difference among excellent lenses would center on their ability to generate artifacts. I don't consider that a very useful metric.

It could be a useful warning.   

BobORama
BobORama Senior Member • Posts: 2,602
Re: Microcontrast
3

NateW wrote:

I don’t have a great definition, but I am positive that the perception of micro-contrast has to do with a lot more than MTF 50 resolution. If I had to, I would describe it as the ability of a lens to render object surfaces with a high degree of tactile presence/corporeality across a broad range of image magnifications ie, no matter what size the image, it gives the impression of real light falling on a real object. My hunch is that it has a lot to do with the clarity and purity of color and specular highlights and little to do with the kind of resolution measured by MTF charts. When a lens has exceptional micro-contrast the tactile impression of depth seems to extend even into parts of the image that are out of focus, suffering from astigmatism, etc.

In lenses with pattern bokeh the slightly OOF areas appear sharper than normal because the edges are multiplied via the zoning in the bokeh.

Like the tube amplifiers that some audiophiles and and musicians seek out.   Its not faithful reproduction, but rather distortion that they crave,

 BobORama's gear list:BobORama's gear list
Pentax K-5 Pentax K-1 Sigma 10mm F2.8 EX DC HSM Diagonal Fisheye Pentax smc DA 18-55mm F3.5-5.6 AL WR Samyang 14mm F2.8 IF ED MC Aspherical +9 more
NateW Junior Member • Posts: 26
Re: Microcontrast
1

BobORama wrote:

NateW wrote:

I don’t have a great definition, but I am positive that the perception of micro-contrast has to do with a lot more than MTF 50 resolution. If I had to, I would describe it as the ability of a lens to render object surfaces with a high degree of tactile presence/corporeality across a broad range of image magnifications ie, no matter what size the image, it gives the impression of real light falling on a real object. My hunch is that it has a lot to do with the clarity and purity of color and specular highlights and little to do with the kind of resolution measured by MTF charts. When a lens has exceptional micro-contrast the tactile impression of depth seems to extend even into parts of the image that are out of focus, suffering from astigmatism, etc.

In lenses with pattern bokeh the slightly OOF areas appear sharper than normal because the edges are multiplied via the zoning in the bokeh.

Like the tube amplifiers that some audiophiles and and musicians seek out. Its not faithful reproduction, but rather distortion that they crave,

True enough, I’m not typically a fan of lenses with “swirly” bokeh, but it isn’t usually this apparent with the Summilux and the lens has a lot of other great qualities. That said, you could the same thing (regarding that “craving” for distortion) about the desire for lenses that use extensive corrections to flatten the field and produce “perfect” uniformly smooth bokeh that renders the background into a 2 dimensional plane of color. That isnt faithful reproduction either. Personally, I find that older, simpler, high quality lenses draw in a way that feels much more similar to the way things actually look through human eyes.

57even Forum Pro • Posts: 12,757
Re: Thank you all!

Great Bustard wrote:

Great Bustard wrote:

First, let me define the term: microcontrast is the contrast at high spatial frequencies. With that definition in place, a few questions:

  1. If System A has a higher MTF-50 than System B, does that mean that it will have greater resolution at all other contrast levels? For example, will the MTF-10 and MTF-90 of System A also be greater than System B?
  2. If (1) is true, does that necessarily imply the converse? That is, the contrast of System A will be higher than the contrast of System B at any given resolution?
  3. Does halation affect the MTF-50 measurement? If so, does the effect of halation result in greater degradation as the contrast level of the MTF measurement increases?
  4. Repeat (3) with veiling flare.

In the end, I wish to know if the term "microcontrast" as defined describes the photo in a photographically meaningful manner any differently than an MTF-50 measurement does. That is, if System A has a higher MTF-50 than System B, does that automatically mean that the "microcontrast" of System A is greater than the "microcontrast" of System B?

Apologies for not chiming back in earlier. I want to say I very much appreciate all the great responses! It's going to take me a bit to think of intelligent questions to ask.

My goal is to give/create a more precise interpretation of "microcontrast", even if it cannot be quantified as a scalar, as implied in the OP. One might question the utility of such a goal, but given that the term is often used in connection to the visual properties of a photo, I thought it worthwhile to define the term more precisely in a manner that is consistent with what people are talking about when using the term.

Contrast level at 5 cycles per degree in a standard sized image would be a good start

-- hide signature --

Reporter: "Mr Gandhi, what do you think of Western Civilisation?"
Mahatma Gandhi: "I think it would be a very good idea!"

 57even's gear list:57even's gear list
Fujifilm X-Pro2
AiryDiscus Senior Member • Posts: 1,620
Re: Thank you all!
1

57even wrote:

Great Bustard wrote:

Great Bustard wrote:

First, let me define the term: microcontrast is the contrast at high spatial frequencies. With that definition in place, a few questions:

  1. If System A has a higher MTF-50 than System B, does that mean that it will have greater resolution at all other contrast levels? For example, will the MTF-10 and MTF-90 of System A also be greater than System B?
  2. If (1) is true, does that necessarily imply the converse? That is, the contrast of System A will be higher than the contrast of System B at any given resolution?
  3. Does halation affect the MTF-50 measurement? If so, does the effect of halation result in greater degradation as the contrast level of the MTF measurement increases?
  4. Repeat (3) with veiling flare.

In the end, I wish to know if the term "microcontrast" as defined describes the photo in a photographically meaningful manner any differently than an MTF-50 measurement does. That is, if System A has a higher MTF-50 than System B, does that automatically mean that the "microcontrast" of System A is greater than the "microcontrast" of System B?

Apologies for not chiming back in earlier. I want to say I very much appreciate all the great responses! It's going to take me a bit to think of intelligent questions to ask.

My goal is to give/create a more precise interpretation of "microcontrast", even if it cannot be quantified as a scalar, as implied in the OP. One might question the utility of such a goal, but given that the term is often used in connection to the visual properties of a photo, I thought it worthwhile to define the term more precisely in a manner that is consistent with what people are talking about when using the term.

Contrast level at 5 cycles per degree in a standard sized image would be a good start

We have reached a simplistic SQF and caught up to the 1980s...

Keyboard shortcuts:
FForum MMy threads