Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
05-19-2022, 04:13 PM - 6 Likes   #1
Loyal Site Supporter
Loyal Site Supporter




Join Date: May 2007
Location: Flagstaff, Arizona
Posts: 1,214
Precision Shutter Speed Measurements - Pentax K-3 III Third-step Error Fixed!!

I have just measured some accurate shutter speeds for my new Pentax K-3 III camera.

And, it looks like the shutter mavens at Pentax have finally learned to get it right on the 1/3 stop speed settings!

As I have measured and reported previously ( Shutter Speed Measurements: Pentax (K-1, K-3, Q7, *istD) and Canon (G15, Rebel XS) - PentaxForums.com ), all the cameras for which I have measured shutter speeds are very accurate when the shutter speed intervals are set for half stops. But, previously, my Pentax DSLRs have had a consistent error when it came to 1/3 stop shutter intervals. The intermediate speeds (i.e. those between the standard 1-stop values - 1, 2, 4, 8, 15, and 30 seconds (noting that “15" and “30" are really 16 and 32 seconds long!) and their reciprocals - 1/2, 1/4, 1/8 etc.) had very consistent 3% high / 3% low behavior.

This figure shows the problem for my K-3:




My K-1 and elderly *ist D show similar errors, although my Q7 did not - it was quite accurate.

Several Canon (Rebel XS and 5D) and Nikon (D-5200) DSLRs that I have tested are also quite accurate at both half- and third-speed settings - errors less than half a per cent or so.

So, when my new K-3 III appeared, I was eager to measure its shutter speeds, to see if Pentax had “seen the light.” Looks like they have.

I set out to use my 256 blinking LED array (see description in my reference above), setting it for an expected 240 flashes during an exposure. A setting of 250 flashes would be slightly better in terms of measurement precision, but if the shutter speeds were indeed off by 3%, the flash count could go “off scale” (3% of 250 is 7.5, so a 3% too long speed would give 257.5 counts - just beyond my 256 total number of dots. Some dots might be lit up twice. Turns out, that isn’t so bad! Read on.)

A quick run through of both half-stop and third-stop shutter speed steps indicated that the new K-3 III was MUCH better at the 1/3 step speeds. As close as I could measure (a fraction of a lit-up LED), in fact, all the results were good to better than half a percent.

In the process of making the measurements, at some point I had set my signal generator to the rate for the next shorter speed, but took the next shot with the previous (longer) shutter speed still set in the camera. That meant I should get a bunch of extra LEDs lit up - i.e. some number of LEDs would have flashed twice during the too-long exposure.

Since I knew both the “wrong” flash rate and the particular exposure speed, I could estimate how many LEDs should have been lit twice. Lo and behold, I could see that result quite obviously in the resultant exposure. I could easily count some 46 brighter dots!

That gave me a brain flash - why not deliberately set my speed too high for one round of LED flashes, and instead set it for an expected number of doubly-illuminated LED dots.

So, I took a second set of data for which I would expect to see 500 LEDs lit up, or, more easy to count, 12 LEDs that would be fainter than the rest (twice through the 256 array would have all 512 LEDs double exposed). This worked exceedingly well. The results reported here are based on such images. Since it is actually possible to detect a partially lit-up LED (i.e. one that is turning ON or OFF at the beginning or end of an exposure) by comparing its intensity to its fully lit neighbors, the precision I can obtain this way approaches half a dot out of 500, or 0.1 % (a part in a thousand)!

Here’s a cropped shot of an exposure with the 500 LED expected count (the darker dots are at the top of the second column on the left and the entire third column of the right-hand array pair -- 12 in total):




And, here are the results. As always, the shutter speeds for half-stop intervals are quite accurate. The total scatter for speeds from 1/15 second to 30 seconds is +/- 0.002 fractional error (in the sense of measured speed compared to expected speed).

Here’s the plot for th 1/3 shutter speed intervals:




I no longer see the zig-zag effect that is so obvious in my first plot at the beginning of this report. Again, as for the half-stop steps, all the speeds are within a one or two parts per thousand ( 0.1 - 0.2 per cent) of expected. Each point in the plot is an average of typically two measurements (there are three measurements for a couple of points). There is a slight tendency (also seen for the half-step results) for the exposures to be ever so slightly long (0.1 %) - you will never see any problems from that in your exposure settings, though!!!!!


Bottom line: it looks like Pentax has finally gotten their arithmetic correct when calculating exposure times for 1/3 stop interval shutter speeds. Shutter speeds for the K-3 III are all within one- or two-tenths of a per cent, at least for speeds down to 1/30 second.

Over all, I am very impressed, practically in awe, at how accurate the shutter speeds actually are. Our shutters are amazing electro-mechanical devices that are expected to perform like this for several hundred thousand exposures!

Now, I would like to get my hands on a K-1 II, to see if the speed errors are fixed on it. I also wonder if a software upgrade would eliminate the +/- 3% effect on our older cameras.


As an aside: It turns out, one can be even a bit greedier, and repeat the overexposure trick at least up to 4 cycles through the array. i.e. if I set my signal generator to make an expected 750 counts during an exposure, I should see 3x256-750 = 18 LEDs that are darker than average (they been lit only twice rather than three times) and if I go all the way to an expected count of 1000 during an exposure, I should see 4x256-1000 = 24 darker LEDs. I’ve tested this and it works quite well. You do have to be a bit careful in counting the darker LEDs in these circumstances. A bit of playing with levels in PhotoShop can readily bring out the difference, but that gets tedious for some 50 frames!

05-19-2022, 04:25 PM   #2
Pentaxian




Join Date: Dec 2016
Location: Southeastern Michigan
Posts: 3,764
Pop Photo magazine, way back in the day, when reporting their test results of SLR film cameras, used to include "exposure accuracy" measurements, along with aperture and shutter speed accuracy. The Pentax Z/PZ 1p was among the most accurate from any brand. At one point, the Canon flagship was among the least accurate. Cameras with electronic aperture control tended to be more accurate than those using the aperture ring, although some using the ring were also very accurate- included among those was the Pentax MZ-S.

But such tests are no longer done when evaluating DSLR cameras, from what I have seen.

Last edited by mikesbike; 05-19-2022 at 04:54 PM.
05-19-2022, 05:13 PM   #3
Pentaxian
photoptimist's Avatar

Join Date: Jul 2016
Photos: Albums
Posts: 4,824
Great work and interesting results!

One interesting question is whether the shutter times are identical across the frame. Is there any way to set camera up so that it is looking at two flat mirrors (one visible at the top of the frame and the other visible at the bottom of the frame) and with both mirrors angled so the camera sees your clever LED array? That would measure the shutter time at the start and end of curtain travel.

Two views of the same timing array would tell you how well the two curtains match in travel time.

P.S. Is 30 seconds now 30 seconds on the K-3 III or is it still 32 seconds to make that shutter time exact 5 stops longer than 1 second?
05-19-2022, 06:03 PM   #4
Loyal Site Supporter
Loyal Site Supporter




Join Date: May 2007
Location: Flagstaff, Arizona
Posts: 1,214
Original Poster
QuoteOriginally posted by photoptimist Quote
One interesting question is whether the shutter times are identical across the frame. Is there any way to set camera up so that it is looking at two flat mirrors (one visible at the top of the frame and the other visible at the bottom of the frame) and with both mirrors angled so the camera sees your clever LED array? That would measure the shutter time at the start and end of curtain travel.

Two views of the same timing array would tell you how well the two curtains match in travel time.

P.S. Is 30 seconds now 30 seconds on the K-3 III or is it still 32 seconds to make that shutter time exact 5 stops longer than 1 second?
Interesting question on whether uniform across the frame. I'll see what I can do - either re-aim the camera between two measurements, or some kind of mirror. Generally, I have put the array image in the ~middle of the frame.

I have ALWAYS assumed 32 seconds, as well as all other exact values (square root or cube root of 2 steps). So far, seems to be the case - for any, and all, cameras.

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, count, dots, dslr, exposure, half, k-3, leds, lit, pentax, photography, shutter, speeds
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Are fixed aperture zoom really fixed aperture ? lotech Pentax SLR Lens Discussion 11 01-05-2022 01:20 PM
Shutter Speed Measurements: Pentax (K-1, K-3, Q7, *istD) and Canon (G15, Rebel XS) AstroDave General Photography 16 05-23-2021 10:33 PM
K-3 III - Photons to Photos sensor DR measurements Oxygenum Pentax DSLR Discussion 154 05-09-2021 04:49 AM
Optical Measurements of Pentax ME Super Shutter Speed AstroDave Film SLRs and Compact Film Cameras 15 07-01-2020 08:49 AM
True Shutter Speeds compared to Nominal Shutter Speeds - Actual Measurements AstroDave General Photography 14 02-08-2018 02:07 PM



All times are GMT -7. The time now is 01:17 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