September 01, 2014, 10:59:26 PM

Author Topic: DxO & MTF Charts ... a little help please!  (Read 3997 times)

jrista

  • Canon EF 400mm f/2.8L IS II
  • *******
  • Posts: 4123
  • POTATO
    • View Profile
    • Nature Photography
Re: DxO ... a little help please!
« Reply #30 on: February 13, 2014, 05:26:00 PM »
that is a great explanation but i will never remember it all. is there a simple way to look at it to get a quick summary? such as the black line for a general indicator and how tight the lines are together? do these mtf charts generally translate into what is found in testing of real lenses?

thanks,

You can't really narrow it down. That's the fundamental fallacy everyone falls into, and why a lot of the stuff DXO, DPR, and the whole host of other "lens testers" produce is largely useless.

If you want to eliminate anything, eliminate the set of curves that don't fit with the apertures you will most use. If you will mostly shoot wide open, then you can ignore the blue curves. If you will mostly shoot stopped down, then you can ignore the black curves. You could ignore dashed curves, and just use the solid curves, but then your not really getting the whole picture.

There is no "one number tells all". That's just a fallacy. The attempt to utterly simplify everything is really what gets you into trouble.

When it comes to the difference between an MTF and a lens test, lens tests all ultimately run into the "sensor bound" problem. Sensors have a fixed resolving power...it's the same across the entire area of the sensor. Lenses, on the other hand, have a non-linear resolving power that falls off as you stop down. At apertures wider than f/8, the potential for a lens to resolve much finer detail than the sensor becomes very real. The problem is, final "output" resolution has an asymptotic relationship with the lowest common denominator. Since the sensor usually IS that lowest common denominator, at faster apertures, where lenses have the potential to resolve a LOT of detail, may all end up looking the same in the end. Why? Well, let's say your sensor can resolve 50lp/mm, and you have four lenses capable of resolving, at f/4, 100lp/mm, 130lp/mm, 150lp/mm, and 173lp/mm (the latter is the maximum diffraction-limited resolving power of an f/4 lens.) The problem is that all of these lenses will all appear to resolve somewhere between 45-49lp/mm with a "real world" lens test, like the kind that DXO does. They are all SENSOR BOUND! The SENSOR cannot resolve more than 50lp/mm, so that is your absolute limit on final output resolution (the resolution measured in the RAW images by computer algorithms.)

So, first off, your standard lens test that tests lenses attached to cameras are largely useless for any apertures above f/8, however from f/8 and narrower, the vast majority of lenses are diffraction limited, so they will all perform the same anyway.

Second, MTFs really don't have any relationship with artificial lens test results, because they are either performed algorithmically based on fairly accurate computer models that account for overall lens construction and design, as well as material traits; or they are performed with optical lens bench testing, which uses a special apparatus to test JUST the lens. Synthetic MTFs will usually indicate just a little bit better performance than Real MTFs generated with a optical test bench, however both will be largely similar, and neither will bear any resemblance to your "standard lens+camera" tests.

If you want to keep it simple: Pick one set of solid lines, for max aperture or f/8 (depending on whichever you use most), and go with that.
My Photography
Current Gear: Canon 5D III | Canon 7D | Canon EF 600mm f/4 L IS II | EF 100-400mm f/4.5-5.6 L IS | EF 16-35mm f/2.8 L | EF 100mm f/2.8 Macro | 50mm f/1.4
New Gear List: SBIG STT-8300M | Canon EF 300mm f/2.8 L II

canon rumors FORUM

Re: DxO ... a little help please!
« Reply #30 on: February 13, 2014, 05:26:00 PM »

golubiewac1

  • SX50 HS
  • **
  • Posts: 6
  • front ranger
    • View Profile
Re: DxO ... a little help please!
« Reply #31 on: February 13, 2014, 06:08:21 PM »
Thanks to jrista for the tutorial on MTF charts.  It will be very helpful to me.

candc

  • 6D
  • *****
  • Posts: 506
    • View Profile
Re: DxO ... a little help please!
« Reply #32 on: February 13, 2014, 06:51:15 PM »
ok, i am getting closer i think, bold lines on chart = bold lines in testing for contrast, fine lines = resolution. black is wide open, blue is f/8 solid lines are radians, dashed are perpendicular,   i am a bit fuzzy on what the radian and perpendicular lines tell you but i am getting the idea,

thanks

jrista

  • Canon EF 400mm f/2.8L IS II
  • *******
  • Posts: 4123
  • POTATO
    • View Profile
    • Nature Photography
Re: DxO ... a little help please!
« Reply #33 on: February 13, 2014, 06:52:43 PM »
ok, i am getting closer i think, bold lines on chart = bold lines in testing for contrast, fine lines = resolution. black is wide open, blue is f/8 solid lines are radians, dashed are perpendicular,   i am a bit fuzzy on what the radian and perpendicular lines tell you but i am getting the idea,

thanks

Sagittal (radian) and meridional (perpendicular) tell you about lens astigmatism (they will usually diverge when plotted across the chart).
My Photography
Current Gear: Canon 5D III | Canon 7D | Canon EF 600mm f/4 L IS II | EF 100-400mm f/4.5-5.6 L IS | EF 16-35mm f/2.8 L | EF 100mm f/2.8 Macro | 50mm f/1.4
New Gear List: SBIG STT-8300M | Canon EF 300mm f/2.8 L II

privatebydesign

  • Canon EF 300mm f/2.8L IS II
  • *******
  • Posts: 2268
  • Ermintrude says "moo"
    • View Profile
The best time to plant a tree is twenty-five years ago. The second best time is today.

candc

  • 6D
  • *****
  • Posts: 506
    • View Profile
Re: DxO ... a little help please!
« Reply #35 on: February 13, 2014, 07:04:52 PM »
i am going to have to read up on that, i hear about astigmatism in a lens all the time but don't fully understand what it is and how it translates to a lens performance. i know it is an unwanted characteristic in a lens and now i think i can see from the mtf chart that the closer together the dashed and solid lines of a particular pair are then the better the astigmatism is controlled?

candc

  • 6D
  • *****
  • Posts: 506
    • View Profile
Re: DxO ... a little help please!
« Reply #36 on: February 13, 2014, 07:24:51 PM »
http://www.luminous-landscape.com/tutorials/understanding-series/lens-contrast.shtml

http://www.luminous-landscape.com/tutorials/understanding-series/understanding-mtf.shtml

those are good articles thanks, i am going to have to read through them again a few times but what grabbed me is that you should look at the uppermost set of lines, the bold ones because they translate best to what you visually correlate with lens performance? that is to say that the contrast seems to be more important than absolute resolving power. i noticed that on the mtf chart for the 24-70ii, the bold lines were flat at the top and the fine lines dropped off significantly more.
« Last Edit: February 13, 2014, 07:43:42 PM by candc »

canon rumors FORUM

Re: DxO ... a little help please!
« Reply #36 on: February 13, 2014, 07:24:51 PM »

jrista

  • Canon EF 400mm f/2.8L IS II
  • *******
  • Posts: 4123
  • POTATO
    • View Profile
    • Nature Photography
Re: DxO ... a little help please!
« Reply #37 on: February 13, 2014, 08:12:38 PM »
http://www.luminous-landscape.com/tutorials/understanding-series/lens-contrast.shtml

http://www.luminous-landscape.com/tutorials/understanding-series/understanding-mtf.shtml

those are good articles thanks, i am going to have to read through them again a few times but what grabbed me is that you should look at the uppermost set of lines, the bold ones because they translate best to what you visually correlate with lens performance? that is to say that the contrast seems to be more important than absolute resolving power. i noticed that on the mtf chart for the 24-70ii, the bold lines were flat at the top and the fine lines dropped off significantly more.

It is easy to get absolute in an online article. I wouldn't say things are as simple as "contrast is more important than absolute resolving power". It depends. As a bird photographer, absolute resolving power is very important to me, so I pay close attention to the thin lines myself. Some lenses have great contrast with lower resolution detail, but they simply can't separate finer details. That would ultimately mean that I might be able to resolve the major details of a birds feathers, but not actually be able to cleanly separate the structure of the feathers themselves. I'd be ok with a slight loss of contrast for larger details, if it means I have the ability to separate finer detail.

You can reduce the information in an MTF according to your needs. Don't let someone else tell you what your needs are, though. ;) If contrast is more important to you, well then sure, pay attention to the thicker solid lines.
My Photography
Current Gear: Canon 5D III | Canon 7D | Canon EF 600mm f/4 L IS II | EF 100-400mm f/4.5-5.6 L IS | EF 16-35mm f/2.8 L | EF 100mm f/2.8 Macro | 50mm f/1.4
New Gear List: SBIG STT-8300M | Canon EF 300mm f/2.8 L II

candc

  • 6D
  • *****
  • Posts: 506
    • View Profile
Re: DxO ... a little help please!
« Reply #38 on: February 13, 2014, 09:22:06 PM »
http://www.luminous-landscape.com/tutorials/understanding-series/lens-contrast.shtml

http://www.luminous-landscape.com/tutorials/understanding-series/understanding-mtf.shtml

those are good articles thanks, i am going to have to read through them again a few times but what grabbed me is that you should look at the uppermost set of lines, the bold ones because they translate best to what you visually correlate with lens performance? that is to say that the contrast seems to be more important than absolute resolving power. i noticed that on the mtf chart for the 24-70ii, the bold lines were flat at the top and the fine lines dropped off significantly more.

It is easy to get absolute in an online article. I wouldn't say things are as simple as "contrast is more important than absolute resolving power". It depends. As a bird photographer, absolute resolving power is very important to me, so I pay close attention to the thin lines myself. Some lenses have great contrast with lower resolution detail, but they simply can't separate finer details. That would ultimately mean that I might be able to resolve the major details of a birds feathers, but not actually be able to cleanly separate the structure of the feathers themselves. I'd be ok with a slight loss of contrast for larger details, if it means I have the ability to separate finer detail.

You can reduce the information in an MTF according to your needs. Don't let someone else tell you what your needs are, though. ;) If contrast is more important to you, well then sure, pay attention to the thicker solid lines.


That's the thing I am trying to get straight because its a bit confusing. What the mtf measures is micro contrast which we see as resolution at some point, not overall image contrast. The mtf chart for the 24-70ii  shows really good contrast center to edge but the resolution lines don't  look that impressive. Its known to be a super sharp lens by all accounts so that seems to reinforce the point that what we see as resolution is really closely related to contrast?

jrista

  • Canon EF 400mm f/2.8L IS II
  • *******
  • Posts: 4123
  • POTATO
    • View Profile
    • Nature Photography
Re: DxO ... a little help please!
« Reply #39 on: February 13, 2014, 09:35:55 PM »
http://www.luminous-landscape.com/tutorials/understanding-series/lens-contrast.shtml

http://www.luminous-landscape.com/tutorials/understanding-series/understanding-mtf.shtml

those are good articles thanks, i am going to have to read through them again a few times but what grabbed me is that you should look at the uppermost set of lines, the bold ones because they translate best to what you visually correlate with lens performance? that is to say that the contrast seems to be more important than absolute resolving power. i noticed that on the mtf chart for the 24-70ii, the bold lines were flat at the top and the fine lines dropped off significantly more.

It is easy to get absolute in an online article. I wouldn't say things are as simple as "contrast is more important than absolute resolving power". It depends. As a bird photographer, absolute resolving power is very important to me, so I pay close attention to the thin lines myself. Some lenses have great contrast with lower resolution detail, but they simply can't separate finer details. That would ultimately mean that I might be able to resolve the major details of a birds feathers, but not actually be able to cleanly separate the structure of the feathers themselves. I'd be ok with a slight loss of contrast for larger details, if it means I have the ability to separate finer detail.

You can reduce the information in an MTF according to your needs. Don't let someone else tell you what your needs are, though. ;) If contrast is more important to you, well then sure, pay attention to the thicker solid lines.


That's the thing I am trying to get straight because its a bit confusing. What the mtf measures is micro contrast which we see as resolution at some point, not overall image contrast. The mtf chart for the 24-70ii  shows really good contrast center to edge but the resolution lines don't  look that impressive. Its known to be a super sharp lens by all accounts so that seems to reinforce the point that what we see as resolution is really closely related to contrast?

An MTF measures both microcontrast as well as larger-scale contrast. The 30lp/mm is more of your "microcontrast" (sharpness and acutance), where as 10lp/mm is general contrast. That's why both sets of thin and thick line pairs are used.

The 24-70 is an excellent lens...but it IS both wider angle and it is a zoom. Both of those facts require that certain compromises be made. Compare the 24-70 II with the previous, or with third-party lenses in the same general range, and you'll see why it is considered a very high quality lens with high resolving power. It's got very good resolution for the focal length bracket it falls into.

You want to see some REALLY high resolution lenses? Here:

EF 300mm f/2.8 L II:


EF 600mm f/4 L II:


As you get to longer focal lengths, the falloff towards the corner drops. More light is collimated, you don't have to worry about bending highly oblique light. The more you bend light, the tougher it is to maintain resolution from center to corner.

Even the older EF 300 f/2.8 L Mark I was an unbelievably stellar lens:



But again, these are all primes, and they are telephotos. Mostly collimated light, no highly oblique incident light angles to deal with, and they use the highest quality optical glass around.

Also, keep in mind that you have blue and black lines, meaning f/8 and max aperture. Depending on what the maximum aperture is, the black lines may drop considerably due to optical aberrations. If you want to look at diffraction limited performance, use the blue lines. That is f/8, and it is easier to compare lenses at f/8, as more likely than not most lenses you compare will be diffraction limited, or close to it, at that aperture. You'll notice that f/8 resolution lines tend to edge higher up the chart.

Another example would be the 70-200 f/2.8 L II. At 70mm and 200mm:





Notice the differences here. At 70mm, a shorter focal length that has to deal with more oblique off-axis light, has greater falloff in the corners, and struggles more with meridional test lines (especially the high resolution ones). However at 200mm, you can see the whole set of lines has moved up the chart, and that faloff to corner is slower and less dramatic.

This is the power of an MTF chart. They tell you a LOT about optics, and generally at a glance (once you know how to read them.) Learning how to read them just takes some time and practice. Eventually, comparing lenses or focal lengths really boils down to some momentary glances at the charts, and you absorb a whole lot of information all at once. (At least...that's how it works with me...)
My Photography
Current Gear: Canon 5D III | Canon 7D | Canon EF 600mm f/4 L IS II | EF 100-400mm f/4.5-5.6 L IS | EF 16-35mm f/2.8 L | EF 100mm f/2.8 Macro | 50mm f/1.4
New Gear List: SBIG STT-8300M | Canon EF 300mm f/2.8 L II

mackguyver

  • Canon EF 300mm f/2.8L IS II
  • *******
  • Posts: 2760
  • Who Dares Wins
    • View Profile
    • My Personal Work
Re: DxO ... a little help please!
« Reply #40 on: February 13, 2014, 10:10:21 PM »
Another point on MTF charts - you don't want to compare a super telephoto to a wide angle lens - as you can see the IQ looks insanely higher on the former, but it's not really that way.  You should compare similar focal lengths to each other if using MTF charts, i.e 24mm prime to a 24-70 lens at the wide end.
EOS 1D X, 5DIII, M + EF 24 f/1.4II, 50 f/1.2, 85 f/1.2II, 300 f/2.8 IS II || 16-35 f/4 IS, 24-70 f/2.8II, 70-200 f/2.8II || TS-E 17 f/4, 24 f/3.5II || M 22 f/2, 18-55 f/3.5-5.6 IS || 1.4x III, 2x III

candc

  • 6D
  • *****
  • Posts: 506
    • View Profile
Re: DxO ... a little help please!
« Reply #41 on: February 13, 2014, 10:39:44 PM »
Thanks to you both for explaining all this, I undestand it a lot better than I did before which was:

Flat at the top=good
Ski slope to the right=bad

That's still true but I see why its relative and why

jrista

  • Canon EF 400mm f/2.8L IS II
  • *******
  • Posts: 4123
  • POTATO
    • View Profile
    • Nature Photography
Re: DxO ... a little help please!
« Reply #42 on: February 13, 2014, 10:48:45 PM »
Another point on MTF charts - you don't want to compare a super telephoto to a wide angle lens - as you can see the IQ looks insanely higher on the former, but it's not really that way.  You should compare similar focal lengths to each other if using MTF charts, i.e 24mm prime to a 24-70 lens at the wide end.

Oh, the telephoto focal lengths are indeed that good. Thats WHY I invested $12,000 in the 600mm f/4 L II! You have no idea how much better it is. And that conforms with the theory. At 600mm, the vast majority of incident light is collimated. There isn't much bending going on. Not nearly as much as at 70mm or 24mm. An MTF is an MTF. There are no special considerations for focal length. You can compare any MTF chart to any other. They are about as honest a review of lens quality at any given focal length as you can hope for.
My Photography
Current Gear: Canon 5D III | Canon 7D | Canon EF 600mm f/4 L IS II | EF 100-400mm f/4.5-5.6 L IS | EF 16-35mm f/2.8 L | EF 100mm f/2.8 Macro | 50mm f/1.4
New Gear List: SBIG STT-8300M | Canon EF 300mm f/2.8 L II

canon rumors FORUM

Re: DxO ... a little help please!
« Reply #42 on: February 13, 2014, 10:48:45 PM »

Rienzphotoz

  • Canon EF 300mm f/2.8L IS II
  • *******
  • Posts: 3322
  • Peace unto all ye Canon, Nikon & Sony shooters
    • View Profile
Re: DxO & MTF Charts ... a little help please!
« Reply #43 on: February 14, 2014, 01:26:00 AM »
Hi Jon,

Thank you very much for the AWESOME explanations .. I think I understood most of it and would need to re-read a few times to burn into my brain ... please continue the discusiion with as simple/stupid as you can make for people like me ... much appreciated and many thanks.

PS. As you may have noticed, I changed the title from "DxO ... a little help plesase!" to "DxO & MTF Charts ... a little help please!" ... with all the very useful info shared here, I think, that is a more appropriate title.

PPS. Also, thank you everyone for being very kind in obliging my request to not turn this into a war zone for/against DxO ... much appreciated, as this thread has provided some good education for me.
« Last Edit: February 14, 2014, 01:34:08 AM by Rienzphotoz »
Canon 5DMK3 70D | Nikon D610 | Sony a7 a6000 | RX100M3 | 16-35/2.8LII | 70-200/2.8LISII | 100/2.8LIS | 100-400LIS | 40/2.8 | 50/1.4 | 85/1.8 | 600EX-RTx2 | ST-E3-RT | 24/3.5 T-S | 10-18/4 OSS 16-50 | 24-70/4OSS | 55/1.8 | 55-210 OSS | 70-200/4 OSS | 28-300VR | HVL-F43M | GoPro Black 3+ & DJI Phantom

mackguyver

  • Canon EF 300mm f/2.8L IS II
  • *******
  • Posts: 2760
  • Who Dares Wins
    • View Profile
    • My Personal Work
Re: DxO ... a little help please!
« Reply #44 on: February 14, 2014, 09:11:04 AM »
Another point on MTF charts - you don't want to compare a super telephoto to a wide angle lens - as you can see the IQ looks insanely higher on the former, but it's not really that way.  You should compare similar focal lengths to each other if using MTF charts, i.e 24mm prime to a 24-70 lens at the wide end.

Oh, the telephoto focal lengths are indeed that good. Thats WHY I invested $12,000 in the 600mm f/4 L II! You have no idea how much better it is. And that conforms with the theory. At 600mm, the vast majority of incident light is collimated. There isn't much bending going on. Not nearly as much as at 70mm or 24mm. An MTF is an MTF. There are no special considerations for focal length. You can compare any MTF chart to any other. They are about as honest a review of lens quality at any given focal length as you can hope for.
jrista, this is something I had read from Canon (will have to look for it), and I guess I didn't phrase it correctly.  It's not to say that the MTF charts lie, just that wide angle lenses may look really soft in comparison to telephoto charts, but that doesn't mean that they are as soft as they appear, and the value of MTFs is better used to compare similar focal lengths to each other.

That makes me think of another great resource (even if it's not as current as the edition CPS members receive when they join) for learning about all of this stuff - Canon Lens Work.  You can download the 11 PDFs free from Canon's European CPS website:
http://www.canon-europe.com/Support/Documents/digital_slr_educational_tools/en/ef_lens_work_iii_en.asp

The 10th PDF, OPTICAL TERMINOLOGY & MTF CHARACTERISTICS, has tons of great information on optics and MTF charts.  It's not light reading, but it will tell you everything you wanted to know and probably a whole lot more!

One of the best explanations I've seen on how to read MTF charts is on page 14 of Sony's Alpha lens brochure which also contains great diagrams and explanations of all core lens concepts:
www.docs.sony.com/release/Alpha_Lens_Brochure_Fall_2012.pdf
EOS 1D X, 5DIII, M + EF 24 f/1.4II, 50 f/1.2, 85 f/1.2II, 300 f/2.8 IS II || 16-35 f/4 IS, 24-70 f/2.8II, 70-200 f/2.8II || TS-E 17 f/4, 24 f/3.5II || M 22 f/2, 18-55 f/3.5-5.6 IS || 1.4x III, 2x III

canon rumors FORUM

Re: DxO ... a little help please!
« Reply #44 on: February 14, 2014, 09:11:04 AM »