Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
03-08-2018, 07:15 PM - 21 Likes   #1
Otis Memorial Pentaxian
stevebrot's Avatar

Join Date: Mar 2007
Location: Vancouver (USA)
Photos: Gallery | Albums
Posts: 42,007
Stop-down metering, why it is a problem, and how to make it work

No, this is not another rant about the so-called "crippled mount". Instead, the intent is to provide a useful summary of why stop-down metering is often disappointing with modern dSLRs.

Stop-down metering

Stop-down TTL metering has a long history (see Pentax Spotmatic) and has long been considered the "Gold Standard" of TTL metering in that it measures the actual light at the taking aperture rather than an estimate based on a wide-open measurement. Even though all K-mount bodies support open-aperture metering, bodies having the so-called "crippled mount" (all Pentax dSLRs) limit this feature to lenses capable of automated aperture control by the body. In practice, this means lenses without an aperture ring or those with the "A" position on the aperture ring. Other K-mount lenses will still work using the aperture ring, but with in-camera metering limited to measurement with the lens stopped down.

K-mount auto-aperture actuation options include:
  • Maximum aperture only in modes other than M (default to Av)
  • Green button (instantaneous stop-down) in M mode to set shutter speed
  • EV scale stop-down using optical preview in M mode (menu option on some bodies) with shutter speed set using e-dial

Manual aperture (M42 and manual/preset aperture K-mount) options include:
  • Continuous metering in Av mode (all dial positions except M, X, and B)
  • Instantaneous metering using green button in M mode
  • EV scale using optical preview lever in M mode (menu option on some bodies) with shutter speed set using e-dial

Why it is a problem on many dSLRs
More properly, why is exposure with stop-down metering so often simply wrong?

My personal experience only goes back as far as the K10D and on that camera, stop-down metering simply sucked. Regardless of the method used, metered exposure was often 1 or more stops over or under-exposed. This was a puzzle since *ist D users at the time were happy with the feature on their cameras and stop-down metering was consistently good on Pentax film cameras. Much was made of the issue on this site and other places on the Web as users struggled to secure behavior similar to what they enjoyed shooting with K-mount film cameras when shooting with adapted M42 and non-A K-mount glass.

For example, the two graphs below show the meter variance (LV)* from expected** for two popular lenses using green button metering on the K10D equipped with the stock LF-80 focus screen. The X-axis is f-number.


The intended takeaway is:
  • Variance existed even at maximum aperture where the meter should have been seeing exactly the same light
  • Variance may be both positive and negative even with the same lens
  • The degree of error varies by both aperture and lens
Those two graphs are a mild example. I have lenses that metered as much as 2.5 stops from expected, but which provided properly exposed frames if set to the expected shutter speed aperture combination.
Question #1: "What does the meter see different than the sensor?"
Question #2: "What is different between a lens wide open with the aperture ring in the "A" position and the ring in the wide open position?"

Answer #1: The viewfinder optical system, specifically the focus screen
Answer #2: When the ring is off the "A" position, the body detects that it cannot control the taking aperture and has NO reference as to what it might be.
PDAF autofocus cameras all share a common fault in that light is diverted from the viewfinder to drive the AF sensor. To offset the resulting dim viewfinder, focus screens have been designed to optimize brightness. Two side-effects happen as a result:***
  • The light intensity viewed by the meter sensor is not representative of or even proportional to that which will strike the image sensor. How much it is off depends on the aperture size (rear pupil diameter to be precise) presented to the focus screen, not light intensity.
  • Ability to show the out-of-focus state is also compromised...apparent depth of field is increased down making manual focus difficult. This point is another discussion and will be ignored for the present.
The variance from expected metering is most likely due to the focus screen brightening. This hypothesis has been strengthened by user experience substituting screens from *ist D series bodies as well as results from non-brightened aftermarket screens. In short, variance changed depending on screen, with none of the aftermarket screens providing a clear improvement.
Conclusion? The camera applies a set of correction factors based on the known maximum aperture of the lens mounted when information is available from the contacts on the mount. Without correction factors, the meter is basically reading spurious light from an unknown fall-back position as far as exposure calculation. As a side note, this problem extends to other brands as well. Backward compatibility for many F-mount lenses on Nikon dSLRs depends on stop-down metering and their users were reporting similar issues.

Summary points run something like this:
  • Both accuracy and linearity of stop-down metering are potentially compromised by the optical path present on modern dSLRs
  • The problem is not easily solved by simply dialing in a little exposure compensation
  • The degree of inaccuracy depends on the lens used and aperture set with some lenses being badly effected and others much less so

Is this still a concern? If so, what can we do?

The problem was widely known and had been acknowledged by Hoya/Pentax prior to the K20D release. There was great disappointment when stop-down metering was not improved on that model. When the K-7 was released in 2009, one of the announced improvements was better accuracy and linearity for stop-down metering. Happily, there was a significant improvement that joyously welcomed by many of us using vintage lenses. Since that has been almost a decade ago, one would think that incremental improvements should have resulted in near perfection.

Sadly, that is only partially true. Before sitting down to write this piece, I mounted up an XR Rikenon 50/2 to my K-3. In M mode at f/5.6, both green button and EV scale metering resulted in at least a 1.5 stop overexposure. Other users are encouraged to add their comments and experiences, but my present understanding of optical viewfinder metering runs like this:
  • The initial fix on the K-7 was limited to M-mode. Av metering in other modes continued according to the old rules from the K10D/K20D.
  • I believe that situation from the K-7 extends to the present. When I took delivery of my K-3, I checked both methods against the K10D and found that Av numbers were similar with both cameras. K-3 M-mode metering is better than I remember from the K-7, but still far from perfect.
  • Green button and EV scale metering in M-mode are fully equivalent. This applies to both K-mount and M42 lenses.
My personal recommendations run like this:
  • Use M-mode with either green button or EV scale for viewfinder metering with K-mount lenses
  • Similarly, use M-mode with green button for M42 lenses. I almost never use the Av option at present.
  • Seriously consider carrying a hand-held meter and gray card when shooting with non-A lenses. The hassle is often much less than chimping and figuring, and chimping some more.
  • I have been experimenting with center-weighted metering TTL in Live View using green button in M-mode. Once the exposure is determined, I change back to viewfinder and shoot at will until either the light or the subject changes. The results are very predictable and accurate. Paradoxically, I have been getting better results with M42 lenses using green button in live view than with Av mode. Go figure, they should be the same.
I anticipate some useful comments from our very competent members. Thanks in advance


Steve

* LV = Light Value, an expression of measured subject luminance equivalent to EV at ISO 100

** LV was about 7 on an evenly lit white wall using natural daylight. Expected values were metered with the lens aperture ring on the "A" position. A positive variance indicates an underexposure situation.

*** Both can be seen directly through the viewfinder when using the optical DOF feature. The screen does not darken appreciably until stopped down somewhere
around f/4.5. Similarly, there is little change in DOF as the lens is stopped down at wider apertures.


Last edited by stevebrot; 12-02-2018 at 01:51 PM. Reason: completeness
03-08-2018, 08:16 PM   #2
Senior Member




Join Date: Oct 2015
Posts: 142
I've found exposure results with a shifted full-manual lens are almost always significantly off (Laowa 15mm shift macro). I'd be curious what it does on a first-party shift lens.
03-08-2018, 08:33 PM - 2 Likes   #3
Pentaxian




Join Date: Sep 2006
Location: Mississippi, USA
Photos: Gallery
Posts: 854
The shift with exposure from the 'D' to the K10D was irritating. I adapted by using the Histogram and adjusting accordingly, leaning it to the right. Doing this gave enough wiggle room for exposure adjustment in post without adding noise. I use less K and M lens because of this and stick with Taks, PKA, and newer.
03-08-2018, 09:34 PM   #4
Otis Memorial Pentaxian
stevebrot's Avatar

Join Date: Mar 2007
Location: Vancouver (USA)
Photos: Gallery | Albums
Posts: 42,007
Original Poster
QuoteOriginally posted by fehknt Quote
I've found exposure results with a shifted full-manual lens are almost always significantly off (Laowa 15mm shift macro). I'd be curious what it does on a first-party shift lens.
I wonder if live view might work better?


Steve

03-08-2018, 10:35 PM   #5
Loyal Site Supporter
Loyal Site Supporter
twilhelm's Avatar

Join Date: Feb 2009
Location: Florida
Photos: Gallery | Albums
Posts: 2,369
Great write up Steve.

When I put a “bright” split focus screen on my K20, I had all kinds of issues with metrering on old lenses. My very old light meter was put back in service then.

I haven’t done any real testing, but I’m noticing some of the same inaccuracies with green button metering in AV with my K-1. I do a lot of shooting in M mode. Hopefully I’ll get a bit of time to fully test out what it’s doing.
03-09-2018, 12:57 AM   #6
Veteran Member
LensBeginner's Avatar

Join Date: Sep 2014
Photos: Albums
Posts: 4,696
Interesting.
The next logical question is: have you/others tried it with the K-01?
03-09-2018, 01:51 AM   #7
Pentaxian
Dartmoor Dave's Avatar

Join Date: Aug 2012
Location: Dartmoor, UK
Photos: Gallery
Posts: 3,831
A superb summary of stop-down metering with legacy lenses, Steve. It would be great if the mods could make it a sticky, so that we can refer beginners to it whenever the question gets asked.

As for myself, I shoot mostly with Takumars but use an incident meter because it's what I've always been happiest with.

03-09-2018, 03:16 AM   #8
Site Supporter
Site Supporter




Join Date: Sep 2016
Location: Brisbane
Posts: 895
QuoteOriginally posted by thazooo Quote
The shift with exposure from the 'D' to the K10D was irritating. I adapted by using the Histogram and adjusting accordingly, leaning it to the right. Doing this gave enough wiggle room for exposure adjustment in post without adding noise. I use less K and M lens because of this and stick with Taks, PKA, and newer.
I just tried that with a landscape using an old ISCO preset this week. Haven't got to processing it yet but it looked more amenable to the task on the LCD at the time. I'm using a K5 with a focusing screen.com Nikon K3 split screen.
03-09-2018, 05:01 AM   #9
Pentaxian




Join Date: Dec 2011
Posts: 3,112
Given the fact that I never actually take pictures or perfect 18% grey flat subjects I can't even remotely see how the presented deviations of 0-1 EV are of any significance to anyone outside theory.

If I feel the exposure is off from my personal preferences by such a tiny bit, I adjust it in post in a second. And that will happen regardless of equipment.
03-09-2018, 06:10 AM   #10
Master of the obvious
Loyal Site Supporter
savoche's Avatar

Join Date: Jun 2011
Location: Lowlands of Norway
Posts: 18,307
QuoteOriginally posted by beholder3 Quote
Given the fact that I never actually take pictures or perfect 18% grey flat subjects I can't even remotely see how the presented deviations of 0-1 EV are of any significance to anyone outside theory.

If I feel the exposure is off from my personal preferences by such a tiny bit, I adjust it in post in a second. And that will happen regardless of equipment.
For 0-1 EV, sure. But as Steve notes:

QuoteOriginally posted by stevebrot Quote
Those two graphs are a mild example. I have lenses that metered as much as 2.5 stops from expected
...and if that is a 2.5 stop overexposure you might be out of luck even with a K-1.

Of course, when I shoot with manual lenses it usually means I'm not in a hurry and can chimp and retake if the highlights are blown. Still, it would be better if the metering was more consistent.
03-09-2018, 06:15 AM   #11
Moderator
Loyal Site Supporter
Loyal Site Supporter
pschlute's Avatar

Join Date: Mar 2007
Location: Surrey, UK
Photos: Gallery
Posts: 8,047
QuoteOriginally posted by twilhelm Quote
but I’m noticing some of the same inaccuracies with green button metering in AV with my K-1
Are you talking about modern lenses or the old K M series? If the latter, they will only work as wide open in Av mode.

Regarding the inaccuracies of stop-down with the K-10D camera + K and M lenses, I believe that was a result of the new design of focussing screen introduced with that model. I actually swappped my K10D screen for a split prism design. I got improved stop-down metering results with that than with the stock screen.
03-09-2018, 06:17 AM - 1 Like   #12
Pentaxian
photoptimist's Avatar

Join Date: Jul 2016
Photos: Albums
Posts: 5,104
Excellent analysis, write-up, and recommendations, Steve!

You are correct that much of the problem lies in the focusing screen, how it interacts with all the rays of light coming from the lens, and how the TTL meter sensor sees that light from the focusing screen.

Somewhere inside the camera's software is a numerical conversion factor that relates the light flux measured by the TTL sensor to the estimated light flux of the scene as modulated by the effects of the half-silvered reflex mirror and the efficiency of the focusing screen to capture the light from the lens.

Accurate stop-down metering would require a lens-specific calibration file that encodes how the focusing screen interacts with that specific lens at each aperture.
03-09-2018, 06:40 AM - 1 Like   #13
Veteran Member




Join Date: Feb 2010
Photos: Albums
Posts: 1,912
QuoteOriginally posted by photoptimist Quote
Accurate stop-down metering would require a lens-specific calibration file that encodes how the focusing screen interacts with that specific lens at each aperture.
So it's a big fix for a relatively minor problem
03-09-2018, 06:47 AM - 4 Likes   #14
Pentaxian
Dartmoor Dave's Avatar

Join Date: Aug 2012
Location: Dartmoor, UK
Photos: Gallery
Posts: 3,831
QuoteOriginally posted by mohb Quote
So it's a big fix for a relatively minor problem

An easier fix would be to stick a little paper label inside the lens cap with the corrections to use for that lens written on it.
03-09-2018, 06:50 AM   #15
Pentaxian




Join Date: Oct 2011
Location: Albuquerque, NM
Posts: 6,027
On the K-5 with my "S-type" screen I get fairly consistent results (which I got mainly for the ease of manual focusing fast manual lenses, but improved metering was a welcome bonus) -- comparing live-view metering to PDAF it stays within -+0.3 EV depending on aperture (so maybe 2/3rds of a stop total range) -- on some lenses I have to dial-in a consistent adjustment (usually -0.7 EV) but variance from that is pretty linear. I am using the S-type with no lines on it at all -- previously I had the same type but with a frame -- I think the plain one works best for metering. I'm just comparing CDAF vs PDAF, not to any grey card or ideal measurement. (With the way I shoot just being in the ballpark is fine, I am always adjusting to the scene anyway.) I know a lot of people got these screens for the K-5s, but it seemed like they were happier with the stock screens since the K-3 (at least for focusing). I'm planning on getting a KP sometime this year -- should I assume I'm gonna want a S-screen to go with it? (I use a ton of old manual lenses.)
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!
aperture, av, button, camera, dslr, k-mount, lens, lenses, light, mode, photography, stop-down, viewfinder
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
M42 on a K-Mount: Stop-down metering isn't the big problem (for me, anyway) taksharp Pentax SLR Lens Discussion 20 08-09-2017 04:01 PM
I don't understand why stop down metering doesn't work well at higher f stops. geekette Pentax DSLR Discussion 51 06-08-2011 10:39 PM
K10D Stop down metering doesn't work trixtroll Pentax DSLR Discussion 2 09-23-2009 06:05 AM
Down down down down... innershell Post Your Photos! 5 08-06-2009 05:45 PM
k100d stop-down metering problem stanjo Pentax DSLR Discussion 4 11-05-2007 12:37 PM



All times are GMT -7. The time now is 10:33 PM. | See also: NikonForums.com, CanonForums.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