Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
02-27-2009, 12:00 AM   #16
Veteran Member




Join Date: Nov 2006
Location: Hong Kong
Posts: 1,934
More Technical Explanations

QuoteOriginally posted by HermanLee Quote
Rice High, your explanation is interesting and logical.
What does it mean with line pairs?
Line pairs mean a pair of white line and black line, which is equal to two "line widths".

QuoteQuote:
why the graph always stop at weighted MTF 40lp/mm: 0.5
Photodo only measured the resolution of the lenses up to 40lp/mm for the resolution test chart and pattern. With that spatial frequency, they measured the MTF values (and also for 10 and 20 lp/mm lower frequencies).

QuoteQuote:
Does it mean 0.5 is the limit for a lens to be considered acceptable?
Yes. But the measurement is in the other way. They (photodo) fixed the lp/mm and then measured the MTF values.

On the other hand, for Imatest or alike methodology, the MTF is selected and fixed (usually at MTF = 0.5 or 50%) and then the Line Width per "Picture height" (simply means vertical run of the picture frame) is measured by the software.

The LW/PH results for the K20D sensor were measured using the above second methodlogy by many reviewers previously, which were found to be around 1500 to 2400 LW/PH (at best), some examples here:

RiceHigh's Pentax Blog: GX-20 Vs K20D - Fully Compared

RiceHigh's Pentax Blog: First K20D Production Camera Formal Test

02-27-2009, 06:26 AM   #17
Pentaxian
Wheatfield's Avatar

Join Date: Apr 2008
Location: The wheatfields of Canada
Posts: 10,042
QuoteOriginally posted by jeffkrol Quote
Garbage......... completely ignores AA filter, color spectrum ect.
What does the AA filter have to do with the conversation? It's part of the optical system, and has to stay there.
Any discussion of sensor performance has to include the entire sensor system, including the Bayer filter, AA filter, etc.
RiceHigh's summary is actually pretty accurate if your idea of useful resolution is real world photographs and not bench tests of resolution charts.

Most of this stuff is little more than a distraction at best, as any DSLR camera that Pentax (or anyone else) has made in the past five or six years is quite capable of producing excellent results within the limitations of the responsiveness of the rest of the camera.
DXO is the worst of the bunch at the moment, their tests are meaningless outside of a testing lab.
02-27-2009, 07:57 AM   #18
Veteran Member
jeffkrol's Avatar

Join Date: Sep 2006
Location: Wisconsin USA
Photos: Gallery
Posts: 8,434
QuoteOriginally posted by Wheatfield Quote
What does the AA filter have to do with the conversation? It's part of the optical system, and has to stay there.
Any discussion of sensor performance has to include the entire sensor system, including the Bayer filter, AA filter, etc.
RiceHigh's summary is actually pretty accurate if your idea of useful resolution is real world photographs and not bench tests of resolution charts.

Most of this stuff is little more than a distraction at best, as any DSLR camera that Pentax (or anyone else) has made in the past five or six years is quite capable of producing excellent results within the limitations of the responsiveness of the rest of the camera.
DXO is the worst of the bunch at the moment, their tests are meaningless outside of a testing lab.
Trying to compare MTF of lenses to sensors based strictly on pixel math and stating "all lenses are outresolved" is baseless and wrong. The sensor will NEVER exhibit "theoretical" perfection. The TRUE MTF of the system, minus the lens can be as much as 50% lower then "calculated" MTF based on sensor geometry alone.... so no, he's wrong to assume that (quote) the sensor out resolves all Pentax lenses"(unquote). If it was sooo darn easy to determine which lenses are outresolved by the sensor, there would be little "haggling" about it.
The whole thing is a mess BUT the AA filter AND Bayer array CHANGE the MTF of the sensor IN REAL WORLD shooting. That is a fact.....
Where does this come from??? Does this include AA and Bayer array?
K20D has resolution from 1500 to 2400 linewidths/picture height at a MTF of 50%. So, just say 2000 lw/ph in average. As the sensor is 16mm height, which means the resolution figure in lp/mm is thus 2000/2/16 = 62.5 lp/mm.
this is the BASIS of all his conclusions. Do you agree with this? I don't.

Last edited by jeffkrol; 02-27-2009 at 08:31 AM.
02-27-2009, 08:29 AM   #19
Veteran Member
jeffkrol's Avatar

Join Date: Sep 2006
Location: Wisconsin USA
Photos: Gallery
Posts: 8,434
QuoteOriginally posted by HermanLee Quote







Does it mean 0.5 is the limit for a lens to be considered acceptable?
Arbitrary cutoff. See this in part:

Re: What can I say?: Open Talk Forum: Digital Photography Review


and this:
Re: What can I say?: Open Talk Forum: Digital Photography Review
QuoteOriginally posted by HermanLee Quote
I'm waiting..

Yes, I'm reading the article from luminous-landscape. It's a good reading but still doesn't give simple explanation and not so clear conclusion.
Sorry, even the "experts" have little agreement on this subject. General consensus is w/ modern lenses,even in /aps-C, we have a way to go before ALL lenses are out resolved at all aperture.
I'm not perfectly comfortable w/ the m but here they are still pulling more resolution out of a 18-250 zoom regardless if it hit the wall or not which I doubt but is possible.
Pentax K20D review Cameralabs resolution results

PERFECT sensor and a nice simple description based on "perfect" lens and sensor

The issue is complex and often misunderstood.

First of all, lens resolution/sharpness does not sharply cut off. It drops off gradually as the detail it is trying to resolve gets finer. You can think of a lens (or any optical device I guess) as having various resolutions depending on the detail level in question.

So the issue isn't just one of resolving detail, but one of how well the detail is resolved. This is typically expressed in terms of MTF (google for more on that). A theoretically perfect lens (diffraction limited) at f/8 has an MTF 50 (relatively high contrast and good reproduction) of 95 lp/mm and an MTF 10 (very low contrast - barely showing detail) of 192 lp/mm.

Lets compare that to a fictional 24Mp DX sensor (6000x4000). If the sensor had no CFA (color filter array), it could resolve a maximum of 125 lp/mm. So it seems pretty clear that even a 24Mp DX sized sensor wouldn't completely outresolve a top-notch lens. But wait, there's more.

Most sensors do have a CFA. This means that the red and blue resolution is half that of a sensor with no CFA. So when shooting color, our 24Mp sensor can only resolve 62.5 lp/mm for red and blue. Even with a 24Mp sensor, this is far less than what an excellent lens can deliver. So it should be clear that even a 24Mp sensor should probably have a low pass or AA filter if we want to keep the lens from outresolving the red and blue resolution of the sensor - therefore reduce the possibility of color moire.

In short, modern DSLR sensors do not generally out resolve modern lenses. When they do it is because the lenses aren't very good, because they are used at very small apertures or perhaps both. Considering that modern lenses aren't "perfect," it seems likely that it will take about 100Mp on a DX sensor to get us to the point were we probably don't need an AA filter on the sensor. But even then, a nearly perfect lens will outresolve the sensor for reds and blues.

I ran the calculations using my Online Digiscoping Calculator. Just choose the "No Scope - Camera Only" option in the upper right to run calculations based on the camera information only. I do need to updated my camera/sensor database. It is getting quite a bit out of date.

http://forums.dpreview.com/forums/read.asp?forum=1018&message=29996262
http://www.jayandwanda.com/digiscope/digiscope_calc.html

Everything you need to know about MTF (but were afraid to ask)
http://www.imatest.com/docs/sharpness.html

Just had to add this quote:
100MP/ CM squared......... :0 4.13cm squared = APS-C x 100 = 413MP sensors........ we got a ways to go
Joe, I suppose I am one of those somewhat knowledgeable people with my background in signal processing and as an electrical engineer. As to building a sensor "with pixels (photosites) crammed through its nose", we already have such in our compact camera sensors. I suspect there isn't any point at all in increasing photosite density much beyond about 100 MP per square cm, at which point the sensor will likely outresolve any glass that is put in front of it and still leave a little extra sampling to help remove moire artifacts from spatial frequencies above the Nyquist sampling limit. This is only a factor of three in packing density and less than a factor of two in linear dimensions, so we don't really have that far to go.


http://forums.dpreview.com/forums/read.asp?forum=1018&message=28506126


Last edited by jeffkrol; 02-27-2009 at 08:50 PM. Reason: OK I lied ;)
02-27-2009, 09:31 AM   #20
Veteran Member
pentaxmz's Avatar

Join Date: Dec 2008
Location: Seattle
Posts: 647
I concur with jeffkrol, that Ricehigh's explanation doesn't take into account all the factors leading to whether or not a sensor out-resolves a particular lens.

I did say this is a controversial topic and I'd rather not get into a spitting match with those with more knowledge or armchair optical engineers.

I was asked a similar question about two years ago and spent a way too much time coming up with a technical explanation and regretted the fall out as a result.

I am tempted to say, "who gives a flying flipper, just buy the best lenses you can afford".
02-27-2009, 09:55 AM   #21
julianactive
Guest




QuoteOriginally posted by pentaxmz Quote
I concur with jeffkrol, that Ricehigh's explanation doesn't take into account all the factors leading to whether or not a sensor out-resolves a particular lens.

I did say this is a controversial topic and I'd rather not get into a spitting match with those with more knowledge or armchair optical engineers.

I was asked a similar question about two years ago and spent a way too much time coming up with a technical explanation and regretted the fall out as a result.

I am tempted to say, "who gives a flying flipper, just buy the best lenses you can afford".
You are right about the flying flipper. Some people want to spend years doing technical tests and writing but they can't take a picture to save their hide.
My big passion is mountain biking. Same thing on Mountain bike forums. People are always talking tech trash and trading one bike for the other ad naseum, but when it come right down to it the person performing the best and having the best time is the one out there riding the most.
If you can't work around the nuanaces of your equipment be it AF, metering, or whatever, then get something else that works for you.
I know today I am going to be out in the lab riding my bike and taking award winning photos!
02-27-2009, 11:36 AM   #22
Veteran Member
pentaxmz's Avatar

Join Date: Dec 2008
Location: Seattle
Posts: 647
QuoteOriginally posted by julianactive Quote
You are right about the flying flipper. Some people want to spend years doing technical tests and writing but they can't take a picture to save their hide.
My big passion is mountain biking. Same thing on Mountain bike forums. People are always talking tech trash and trading one bike for the other ad naseum, but when it come right down to it the person performing the best and having the best time is the one out there riding the most.
If you can't work around the nuanaces of your equipment be it AF, metering, or whatever, then get something else that works for you.
I know today I am going to be out in the lab riding my bike and taking award winning photos!
Don't get me started on bikes!

I used to own a Gary Fisher Supercaliber... I loved that bike! Every once in a while I'd meet some [idiot] with a Ibex, Giant, Litespeed, etc... who made it clear that they had the best bike. I ALWAYS walked away from that nonsensical talk... with my head held high knowing I had the best bike. I kid, I kid!

Sadly, I had to sell it when I moved out of country...
02-27-2009, 03:31 PM   #23
Veteran Member
Ben_Edict's Avatar

Join Date: Jul 2007
Location: SouthWest "Regio"
Photos: Albums
Posts: 3,303
QuoteOriginally posted by krypticide Quote
Pretty sure it's out-resolved my 18-250mm, and probably my Sigma 70-200mm. With the sharper prime lenses, I doubt it.
I don't know which version of the Sigma 70-200 you have, but my old pre-macro 70-200/2.8 is as sharp as ever on the K20. One of my most beloved lenses.

Ben

02-27-2009, 03:36 PM   #24
Veteran Member
Ben_Edict's Avatar

Join Date: Jul 2007
Location: SouthWest "Regio"
Photos: Albums
Posts: 3,303
QuoteOriginally posted by HermanLee Quote
I've been reading some people having soft issue (even after adjustment) with their lens on K20D.

1. Does it mean that K20D's sensor has out resolved some of K-mount lenses?
2. If yes, then which lenses that have been out resolved?
3. How do you know when a sensor out resolve a lens? How do you measure it?

Thanks,
I notice a considerable contrast loss with the 500/4.5 on the K20. But this is really an old lens (though mine is a K-mount version) AND on top it was computed for the 6x7 camera, so a somewhat lower resolution was very acceptable. And as being a pre-ED-glass lens at that focal length, there is not much, one could have done better more than 30 years ago.

One lens, that was already out resolved with the K10 - but performed nicely on the istDS - was/is my Sigma 24-135mm, which I bought as a walk-around-lens, when I could not take more equipment. Even at 10 MPix it is so soft at f/8(!), that it is unacceptable to me.

Ben
02-28-2009, 12:53 AM   #25
Veteran Member




Join Date: Jan 2008
Location: Southern California, USA
Photos: Gallery
Posts: 726
Original Poster
Thank you for all the explanations guys.
I appreciate all of you taking your time to clear out this confusion which is a really confusing topic and may be debatable as you conclude.
But at least now I have some ground to start understanding how this sensor thingy work.

I know for some people, talking about this technical thing is a a waste of time.
But in order to use our camera as a tool, wouldn't it better if we know more what our tool can or cannot do?
That way we know what the limitations are so we can either perfect the tool or perfect ourselves.
02-28-2009, 09:01 AM   #26
Veteran Member




Join Date: Nov 2006
Location: Hong Kong
Posts: 1,934
QuoteOriginally posted by HermanLee Quote
Thank you for all the explanations guys.
I appreciate all of you taking your time to clear out this confusion which is a really confusing topic and may be debatable as you conclude.
But at least now I have some ground to start understanding how this sensor thingy work.

I know for some people, talking about this technical thing is a a waste of time.
But in order to use our camera as a tool, wouldn't it better if we know more what our tool can or cannot do?
That way we know what the limitations are so we can either perfect the tool or perfect ourselves.
Knowing more about the inside technics about cameras and systems not only will help you to take better pictures by making most out of your tools, it is also an interesting topic for us to know more about the technologies and more importantly to appreciate more the engineering arts and wisdoms as contained in each piece of camera gear.
02-28-2009, 01:10 PM   #27
Pentaxian




Join Date: Nov 2007
Location: Western Canada
Posts: 3,821
I'm not an optical or electronic engineer, math isn't my strong point but my common sense is generally not too bad.

I've been avid amateur photographer for 41 years, used mostly Pentax SLR's and Pentax lenses in that time. Also have medium format (Mamiya) and Leica products.

I have a K10D, not a K20D...but I'm not buying this about the K20D's sensor out resolving the Pentax and other lenses.

Could be cause I am not an engineer...but it doesn't make any sense to me. I f anything I would think that the camera needs some sort of adjustment. Making an expensive (K20D) camera that will not work well with lenses ...I dunno...it beggars belief of my poor old brain.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
camera, dslr, lens, lenses, photography, sensor
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Dust on K20D sensor ntrao Pentax DSLR Discussion 12 07-24-2010 08:09 PM
Pentax K20D freezing, locking up - RESOLVED JohnBee Pentax DSLR Discussion 0 01-28-2010 03:09 AM
K-7 sensor vs K20d sensor Mystic Pentax News and Rumors 33 06-21-2009 03:01 AM
FF sensor / lenses? Tom M Pentax DSLR Discussion 24 11-03-2007 07:59 AM



All times are GMT -7. The time now is 12:21 AM. | See also: NikonForums.com, part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top