Jump to content




The above adverts really do help Fotozones. Please click on them if they are relevant to you. Not seeing them? Just exclude Fotozones from your ad blocker. Thanks!


Photo
- - - - -

Reikan Focal Pro


  • Please log in to reply
17 replies to this topic

#1 Anker

Anker
  • Subscriber
  • 6 posts
  • LocationStow, Massachusetts, USA
  • Edit my pics?:Ask Me

Posted 12 December 2012 - 17:18

Found this software package just in time for my new D800.

Reikan Focal Pro automates sensor testing on Canon and Nikon high end cameras.

My primary need is testing for focus tuning. Using a target designed by Reikan the software controls the camera focus and shutter. It prompts you to change the focus microadjustment (can't be controlled by software on Nikon cameras) as it analyzes focus performance across the entire range of possible adjustments (-20 to 20). At the end it presents you with a recommended setting and allows you to save a pdf file with a full report of its analysis. Finding the recommended adjustment for a lens takes about 5 minutes.

Another extremely useful feature is analysis of different focus points. Some D800 and D800E bdies have problems with the leftmost focus points and I was anxious to test whether my new body might have the problem. To my relief my focus points tested fine. This process is 100% automated. You aim your camera at the target and click Start and 5 minutes later you have the report.

To get good result you need:
  • A stable tripod
  • Strong lighting of the target
Highly recommended.

Anker

#2 schwett

schwett

    Advanced Member

  • Life Member
  • 1,230 posts
  • LocationSan Francisco
  • Edit my pics?:Ask Me

Posted 31 December 2012 - 06:13

i've been trying out reikan focal pro as well, after realizing that a number of my lenses were pretty far off and my eyeball method of adjustment (shoot wide open a slightly tipped from horizontal gridded or ruled object) was yielding inconsistent results.

interestingly, the software is ALSO yielding inconsistent results on a number of lenses.

i ran through about 10 today and some worked well and consistently, some did not. i like the software a lot, but am curious what would cause some of the results i'm seeing.

this is the chart of (presumably) acuity at various fine tune settings for the 85/1.4D:
Posted Image

the recommended value was only +2
Posted Image

quite a few lenses, however, produced results which were unpredictable. this one was borderline, the 50/1.4G

Posted Image

the 50/1.8G returned a basically meaningless set of data: (you can see that the curve does not fit the orange points at all.
Posted Image

the 60/2.8G also returned an error - one wonders how the images at -10 could be sharper than both the images at -20 and -7, but less sharp than the fairly predictable looking cluster of images at +3 through +10.

Posted Image

did you have any lens/camera combos that caused the software to report that the results were not as predicted and that you should not continue?
http://photo.sfmthd.org/ [under construction]

#3 photoArne

photoArne
  • Life Member
  • 8 posts
  • LocationNorway

Posted 31 December 2012 - 07:31

Interesting. Does it take focus shift when stopping down into account?

#4 stenrasmussen

stenrasmussen

    Advanced Member

  • Life Member
  • 5,064 posts
  • LocationStavanger
  • Edit my pics?:Ask Me

Posted 31 December 2012 - 09:05

The software is very sensitive to light level (which has to be very bright and consistent), target and distance to said target.
Machina fotografica necesse est

#5 aerobat

aerobat

    Advanced Member

  • Life Member
  • 550 posts
  • LocationSwitzerland

Posted 31 December 2012 - 09:43

I guess rather than the software being sensitive to light levels it shows how sensitive our AF sensors are to mentioned conditions.

Regards, Daniel

#6 stenrasmussen

stenrasmussen

    Advanced Member

  • Life Member
  • 5,064 posts
  • LocationStavanger
  • Edit my pics?:Ask Me

Posted 31 December 2012 - 10:02

The AF system has to be sensitive in order to work at low light levels. Perhaps a not-through-the-lens AF system is required in order to nail focus with a f/1.2 lens.
Machina fotografica necesse est

#7 nfoto

nfoto

    Fierce Bear of the North

  • Administrators
  • 16,649 posts
  • LocationOslo, Norway
  • Edit my pics?:No

Posted 31 December 2012 - 10:21

Schwett: the 50/1.8G returned a basically meaningless set of data: (you can see that the curve does not fit the orange points at all.

Of course it does if fitted by least-squares regression analysis, which 99% likely is the tool of the trade. The reduced scatter towards both end points of the curve will strongly influence the run of the curve.

However, one might wonder if the data set is sufficient for this, or any other kind of advanced statistical analysis.
Bjørn

#8 Dave Rosser

Dave Rosser

    Advanced Member

  • Life Member
  • 664 posts
  • LocationThornbury, near Bristol

Posted 31 December 2012 - 11:49

Very interesting results but what does the vertical axis represent? Further the third graph commits heinous crime number one, it has a false origin. Can you re-plot it with the same vertical scale as the other two? It will then be easier to compare the three graphs.
Putting on my (now 12 years retired) professional hat (CEng FRAeS) and eyeballing the 50mm F/1.8 data I see a straight line with a +- 450 unit scatter. If I knew what the vertical axis represented, (is bigger better?) I might then be able to interpret the results.

#9 nfoto

nfoto

    Fierce Bear of the North

  • Administrators
  • 16,649 posts
  • LocationOslo, Norway
  • Edit my pics?:No

Posted 31 December 2012 - 12:24

Vertical axis is some obscure "MTF50" statistic. Given as lines per picture height whatever that is. I infer "bigger is better" but if this is a parameter eligible for numerical analysis is unclear.
Bjørn

#10 Dave Rosser

Dave Rosser

    Advanced Member

  • Life Member
  • 664 posts
  • LocationThornbury, near Bristol

Posted 31 December 2012 - 13:21

Vertical axis is some obscure "MTF50" statistic. Given as lines per picture height whatever that is. I infer "bigger is better" but if this is a parameter eligible for numerical analysis is unclear.

Looking at the "results" on the other axis I see a scatter of between +-5 and +-10 AF fine tune units. I really don't know if the scatter is measurement error or error in the system being measured - a lot more data is needed to interpret these results - if a junior engineer (Reikan in this case) had presented me with these results I would have sent him away to try again. :angry:

Edited by Dave Rosser, 31 December 2012 - 13:23 .


#11 nfoto

nfoto

    Fierce Bear of the North

  • Administrators
  • 16,649 posts
  • LocationOslo, Norway
  • Edit my pics?:No

Posted 31 December 2012 - 13:51

I think the old computer adage 'garbage in, garbage out' strikes again. Save the money and go out shooting real pictures instead. Plus move AF to AF-ON button only and use AF-C mode. That'll fix most issues.

With a point scatter like shown on these charts, the better option obviously is setting AF Finetune = 0. Or use LV if feasible.
  • Dave Rosser likes this
Bjørn

#12 schwett

schwett

    Advanced Member

  • Life Member
  • 1,230 posts
  • LocationSan Francisco
  • Edit my pics?:Ask Me

Posted 31 December 2012 - 21:53

I think the old computer adage 'garbage in, garbage out' strikes again. Save the money and go out shooting real pictures instead. Plus move AF to AF-ON button only and use AF-C mode. That'll fix most issues.

With a point scatter like shown on these charts, the better option obviously is setting AF Finetune = 0. Or use LV if feasible.


some of us rely on accurate autofocus when we "go out shooting real pictures"... candids, portraits, sport, wildlife, etc are all difficult or impossible to deal with under LV and the pictorial desire for shallow depth of field or isolation plus moving subjects makes manual focus a real challenge.

the "usefulness" of this exercise aside, it's not affected by AF-ON or AF-C/S settings. the camera is triggered (with mirror lock up) from the computer, the finder window closed, and mounted on a sturdy tripod indoors. the irreproducibility of the results points only to the inability of the AF system to produce repeatable results.

more curious is that some lenses produce much more reliable and useful results. many of my lenses - particularly the newer ones - required an adjustment of as much as +18. the before and after results show a clear improvement and shooting with the camera at the default setting of 0 would result in a set of blurry photos, regardless of one's technique.

Posted Image

Posted Image
http://photo.sfmthd.org/ [under construction]

#13 schwett

schwett

    Advanced Member

  • Life Member
  • 1,230 posts
  • LocationSan Francisco
  • Edit my pics?:Ask Me

Posted 31 December 2012 - 21:56

Looking at the "results" on the other axis I see a scatter of between +-5 and +-10 AF fine tune units. I really don't know if the scatter is measurement error or error in the system being measured - a lot more data is needed to interpret these results - if a junior engineer (Reikan in this case) had presented me with these results I would have sent him away to try again. :angry:


i'm not sure how you can look at it from that standpoint - one would expect that there is some symmetry around the 'correct' value, so the spread is infinite, isn't it? i think the variable results within a given AF value (the vertical spread of the red dots) is likely due to the AF system struggling with relatively low EV.
http://photo.sfmthd.org/ [under construction]

#14 schwett

schwett

    Advanced Member

  • Life Member
  • 1,230 posts
  • LocationSan Francisco
  • Edit my pics?:Ask Me

Posted 31 December 2012 - 21:59

Very interesting results but what does the vertical axis represent? Further the third graph commits heinous crime number one, it has a false origin. Can you re-plot it with the same vertical scale as the other two? It will then be easier to compare the three graphs.
Putting on my (now 12 years retired) professional hat (CEng FRAeS) and eyeballing the 50mm F/1.8 data I see a straight line with a +- 450 unit scatter. If I knew what the vertical axis represented, (is bigger better?) I might then be able to interpret the results.


it doesn't really matter what the vertical axis is since the purpose isn't to compare lenses or cameras. it's certainly some measure of 'sharpness' or 'contrast.' i'm not sure why that one chart doesn't have the vertical axis starting at 0. you can't adjust it and there's no obvious reason that one would be different...
http://photo.sfmthd.org/ [under construction]

#15 nfoto

nfoto

    Fierce Bear of the North

  • Administrators
  • 16,649 posts
  • LocationOslo, Norway
  • Edit my pics?:No

Posted 31 December 2012 - 22:24

A lens with axial colour (longitudinal CA) will throw off any AF system. Visual focusing is then preferable. My old 105 DC belonged in this category and the AF 85/1.4 is a candidate as well.

Strangely enough, I never had AF issues with any other lens and any body before the D800, the latter obviously showed the infamous left hand AF issue. After it was fixed by Nikon Nordic, now again the usual situation: no AF problem. And it goes without saying, no AF finetune is conducted on any lens (of which I have "some").


It is definitely wrong to ignore what scale or implication the vertical axis has when one tries to interpret the analysis.
Bjørn

#16 schwett

schwett

    Advanced Member

  • Life Member
  • 1,230 posts
  • LocationSan Francisco
  • Edit my pics?:Ask Me

Posted 31 December 2012 - 22:33

...
Strangely enough, I never had AF issues with any other lens and any body before the D800, the latter obviously showed the infamous left hand AF issue. After it was fixed by Nikon Nordic, now again the usual situation: no AF problem. And it goes without saying, no AF finetune is conducted on any lens (of which I have "some").


would your recommendation then be to return a lens which requires a large af fine tune adjustment to perform correctly under optimum and controlled AF conditions? e.g., the 28 f/1.8g shown above. the image of the target at AF Fine Tune 0 is clearly out of focus. not motion blurred, but clearly out of focus. this using the center AF point under conditions in which any AF system should perform easily and accurately. the image at -16 is clearly not out of focus, or at least not by very much.
http://photo.sfmthd.org/ [under construction]

#17 Tejpor

Tejpor

    Advanced Member

  • Life Member
  • 303 posts

Posted 31 December 2012 - 22:44

Yes, I think that is what he means. If you have both the lens AND your camera calibrated by Nikon, it should be OK without any fine tune adjustment. Both the lens and camera will be adjusted to a standard separately - a standard Nikon body with your lenses and an 50mm lens with your camera.

The AF fine tune should only be used as intermediate solution or for really special lenses, especially as it is not adjustable over different focus distances.

#18 stenrasmussen

stenrasmussen

    Advanced Member

  • Life Member
  • 5,064 posts
  • LocationStavanger
  • Edit my pics?:Ask Me

Posted 01 January 2013 - 01:22

Don't be fooled into thinking that Reikan focal is accurate. I've run several test rounds and every time the results come out differently. I had my 50/1.4G at +4, -8 and 0....so I ended up at...0.
My 84/1.8G is best left at...0.
And then there is close range focus and far distance focus. Both require slightly different AF-tunes.
Nah, I will revert and trust manual or >f/2.8 focus method.
Machina fotografica necesse est




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users





The above adverts really do help Fotozones. Please click on them if they are relevant to you. Not seeing them? Just exclude Fotozones from your ad blocker. Thanks!


An appeal to all Fotozones visitors: please help me to keep this site going by starting your gear purchases using any one of the affiliate links shown below:

Amazon.com | Amazon.co.uk | Amazon.de | Adorama.com | thinkTank Photo | DigitalRev.com | OWC | B&H or Donate via PayPal

Starting your shopping here doesn't cost you anything more, but by using the links above (or any others found on the site) you are advising the affiliate that you support this website. This results in a small commission that helps with the running costs. If your preferred outlet isn't among those listed above you can also support the site by making a donation of any amount via PayPal (no PayPal account required). Any donation will be most appreciated.