[keos]

I have a 77mm on a K3-II, and just took it out for a spin and almost wanted to chuck the lens. It tends to backfocus, but if I change the MFA setting, I can fix the center 3 points (the ones with the flux sensor supposedly), but the rest of the points start to front focus. This is the only lens I've tried that exhibits this problem (tried 55mm 1.4 and 16-50mm 2.8)
Since the sensor itself is sensitive at 2.8, and I tried lenses that were 2.8 or faster, can I eliminate the body as the cause of the error?

[stevebrot]

Since the sensor itself is sensitive at 2.8

Is it safe to assume we are talking about the FA 77mm f/1.8 Limited?

The center column focus points are sensitive at f/2.8. What that means is that it will detect OOF with the same precision at f/2.8 as wider maximum apertures. The "flux sensor" part is Ricoh talk and what it means may have been lost in the translation. The focus sensitivity has nothing to do with "flux", but has everything to do prism pitch.

That aside, make your adjustments using the center point and let the other points fall where they will. Be sure to evaluate the goodness of focus based on multiple attempts at magnification no greater that 100% (1:1). Also take care to use a flat target for the AF system to work against. A slanted or rounded target is too ambiguous for for focus evaluation.

Question...What distance is your camera from the target?

Remembering that the non-center points have focus sensitivity of f/5.6 (same precision to detect OOF at f/5.6 as at wider apertures), the matter of decreased precision should be taken seriously (may require many attempts to determine the whether apparent bias is actually present). I suspect your front focus when using AF points away from center is due to field curvature, a characteristic at some focus distances with that lens.* When focus is at center, the edges and corners would appear front-focused and when focus is further towards the edges, the center would appear back-focused. Edit: Bringing field curvature into the discussion was not appropriate.


Steve

* The FA 77/1.8 was not designed to have a flat field. This recent thread might be of interest...The FA Limited Development Story (also a little DA Limited) - PentaxForums.com

[keos]

Is it safe to assume we are talking about the FA 77mm f/1.8 Limited?

Yes, the limited

The center column focus points are sensitive at f/2.8. What that means is that it will detect OOF with the same precision at f/2.8 as wider maximum apertures. The "flux sensor" part is Ricoh talk and what it means may have been lost in the translation. The focus sensitivity has nothing to do with "flux", but has everything to do prism pitch.

That aside, make your adjustments using the center point and let the other points fall where they will. Be sure to evaluate the goodness of focus based on multiple attempts at magnification no greater that 100% (1:1). Also take care to use a flat target for the AF system to work against. A slanted or rounded target is too ambiguous for for focus evaluation.

Question...What distance is your camera from the target?

Remembering that the non-center points have focus sensitivity of f/5.6 (same precision to detect OOF at f/5.6 as at wider apertures), the matter of decreased precision should be taken seriously (may require many attempts to determine the whether apparent bias is actually present). I suspect your front focus when using AF points away from center is due to field curvature, a characteristic at some focus distances with that lens.* When focus is at center, the edges and corners would appear front-focused and when focus is further towards the edges, the center would appear back-focused.


Steve

* The FA 77/1.8 was not designed to have a flat field. This recent thread might be of interest...The FA Limited Development Story (also a little DA Limited) - PentaxForums.com

It's about 6 feet - infinity. Was testing outside. I didn't expect the field curvature to be so severe. I wasn't using an edge point, I was using any point not the center 3. It was consistently pushing the focal plane back by a just a bit.

[Digitalis]

The FA77 is a tricky lens to work with especially regarding AF,the DA70mm is a more precise lens but the DA lens rendering does not have the qualities of the FA77mm. I manually focus the FA limited lenses. I use the OME3 viewfinder magnifier and the standard focusing screen with no manual focus assist. The FA limited lenses have so many quirks I don't really trust any of the peripheral AF points at apertures below F/4*. I frequently use a focus-compose style of shooting with the FA limited lenses. AF inaccuracies is an unfortunate issue with SLR cameras.Sometimes the AF module itself can be misaligned with the focal plane, this was a huge issue with the Nikon DSLRs in the past.

* the same goes for all the AF cameras I work with from Leica,Nikon, Pentax, Canon, Phase one. Many of them have enhanced precision AF points that are clustered in the center of the frame, those are the points I work with most frequently. Phase one have developed an interesting approach: they test each lens they produce and program profiles in the lens ROM that get transmitted to the camera body that give the AF system a precise model of the optical characteristics of the lens. So the AF can correct for field curvature and other quirks as required, the drawback of such a uniquely advanced AF system and meticulous optical design is the spirit crushing price tag associated with high end medium format digital.

---

Last edited by Digitalis; 01-13-2020 at 07:00 PM.

[stevebrot]

I manually focus the FA limited lenses. I use the OME3 viewfinder magnifier and the standard focusing screen with no manual focus assist. The FA limited lenses have so many quirks I don't really trust any of the peripheral AF points at apertures below F/4*.

Strangely, I very seldom use AF with my FA 77/1.8, preferring to use manual focus (Katz Eye screen). Rare also is use of other than the center AF point. Selecting points is cumbersome and rare is the case where I allow the camera to choose my subject.


Steve

[BarryE]

My personal impression of the FA77 is that it is remarkable good using the 9 pt AF when focusing even in poor light for portraits. The poor light is because I keep the ambient light to zero (using flash).

It's not a lens I would use for landscapes when I looking for sharpness focus across the frame. I do like the FL for landscapes, but I discovered after getting my copy that it didn't work as I wanted. It is great for picking things out in the landscape, though.

The pixie dust argument did, I confess, persuade me to buy the lens. At first it was a disappointment, but slowly I came appreciate its strengths. It is a lens to love, but probably not after using it for only a few outings.

[keos]

wow, I'm surprised given the AF score this got that all of you commenting are manually focusing with this lens.
I love the results when it's in focus, the bokeh is... I can't describe the feeling when I see it on the screen.

But the AF using the non-central point is just so shifted, I wouldn't trust it to be able to take full body portrait shots (or upper body shots in landscape mode), at that distance I'd have to focus and recompose and not sure the field curvature isn't going to mess up the required focal plane either.

I do a lot of run and gun type shooting and move the focus points all over the place. This is the first lens I've used where that is seemingly impossible to do.


So conflicted

[mikesbike]

Like others here, I very seldom use the outer AF points anyway. In fact, usually I just use the central spot AF. I shoot mostly hand-held, and often opportunistic grab shots. With my years of having a camera wth a center AF point only to adapt to, I carry over the techniques I mastered into DSLR use. The FA 77mm LTD remains one of my favorite lenses, ever since I bought it back in the day for use with my film bodies.

[keos]

Like others here, I very seldom use the outer AF points anyway. In fact, usually I just use the central spot AF. I shoot mostly hand-held, and often opportunistic grab shots. With my years of having a camera wth a center AF point only to adapt to, I carry over the techniques I mastered into DSLR use. The FA 77mm LTD remains one of my favorite lenses, ever since I bought it back in the day for use with my film bodies.

My first AF camera was the E-20N, so I know all about only having 1 AF point. Except that camera+lens combo had a f10-11 equivalent depth of field with a relatively flat field curvature, which is must easier on focus and recomposing than the 77mm ltd lens apparently is. I'll take it out again next weekend and see how much trouble I get into with it.

[Digitalis]

Selecting points is cumbersome and rare is the case where I allow the camera to choose my subject.

Many manufacturers are adopting a joystick control interface for selecting AF points. If the recently unveiled Pentax APS-C flagship body is any indication - it would seem Pentax is doing the same. A move that is long overdue IMO.

It's not a lens I would use for landscapes when I looking for sharpness focus across the frame.

Actually around the f/5.6~f/11 aperture range the FA77 is a decent performer, on my optics bench my FA77@ f/8 has a very uniform resolution characteristic*. However In my testing I did encounter copies of the FA77 that were anomalies that performed a bit below/above average though.


Pentax K-1 - SMCP-FA77mm f/1.8 Limited @ f/4 [APS-C crop mode]

The pixie dust argument did, I confess, persuade me to buy the lens. At first it was a disappointment, but slowly I came appreciate its strengths. It is a lens to love, but probably not after using it for only a few outings.

This is pretty much the story of every user of the FA limited lenses in a nutshell, the pixie dust is very seductive:



* the extreme edges on full frame are quite good, especially for a portrait lens. My FA77 is one of the early black triple 0 copies made in Japan.

[stevebrot]

wow, I'm surprised given the AF score this got that all of you commenting are manually focusing with this lens.

AF on my K-3 works just fine with my copy of the FA 77/1.8, meaning that it is fast and accurate. My use of manual focus reflects my choice of subjects and work flow and is not an indication that there is a problem, only that manual focus is the better choice.

It would be helpful to know how are you determining the need for AF fine adjust and what is your process.

I must apologize for a bit of misdirection in my comment above.

The suggestion of field curvature being a possible explanation was off base and I will correct my comment above. Field curvature would show itself as peripheral aspects of a flat subject showing front focus (forward of the intended plane of focus) when the point of focus was placed at the center of the frame. This happens with both manual and auto focus. In the same situation, but with point of focus placed away from center, the center aspect of the frame will show back focus (behind the intended plane of focus).

The fine adjustment will affect all of the focus points the same and all should have the same relative accuracy as the center point. This will be the case regardless of any field curvature that may be present. In other words, field curvature should not effect the ability to attain accurate focus per se.


Steve

[keos]

AF on my K-3 works just fine with my copy of the FA 77/1.8, meaning that it is fast and accurate. My use of manual focus reflects my choice of subjects and work flow and is not an indication that there is a problem, only that manual focus is the better choice.

It would be helpful to know how are you determining the need for AF fine adjust and what is your process.

I must apologize for a bit of misdirection in my comment above.

The suggestion of field curvature being a possible explanation was off base and I will correct my comment above. Field curvature would show itself as peripheral aspects of a flat subject showing front focus (forward of the intended plane of focus) when the point of focus was placed at the center of the frame. This happens with both manual and auto focus. In the same situation, but with point of focus placed away from center, the center aspect of the frame will show back focus (behind the intended plane of focus).

The fine adjustment will affect all of the focus points the same and all should have the same relative accuracy as the center point. This will be the case regardless of any field curvature that may be present. In other words, field curvature should not effect the ability to attain accurate focus per se.


Steve

shooting at a front lit flat target ~6 meters away, at f1.8, viewing the results on an iPad at 100%, around 10 shots to determine hit rate.

Then going outside and shooting a subject about the same distance and looking at the eyes and eyelashes. At that distance there's enough DoF that both should still be in focus if the focus is properly set.

edit: more tests (all lenses wide open)

K-1 II + 77mm: all focus points works
K-1 II + 50mm 1.4 SDM: all focus points works
K-1 II + Sigma 85mm 1.4: some focus points more accurate than others

K-3 II + 50mm 1.4 SDM: all focus points works
K-3 II + 77mm: center 3 focus points and rest of points differ
K-3 II + Sigma 85mm 1.4: all focus point works



And this is why I'm dipping my toes in mirrorless

btw, either I got the sharpest 50mm SDM ever, or the worst 3 16-50mms ever, but the 50mm at 1.4 is sharper or about equal to the 16-50mm at f4, and looks to be better than the 55mm SDM at f2. the lens is amazing. I'm wondering why the Tokina Opera version isn't as popular (as in, not popular at all based on the support raw processing software has given it)

[stevebrot]

shooting at a front lit flat target ~6 meters away, at f1.8, viewing the results on an iPad at 100%, around 10 shots to determine hit rate.

How are you determining front vs. back focus (accuracy, re original post) or are you just evaluating hit rate (precision)? One would expect significantly higher precision with center row (f/2.8) AF points than the remaining (f/5.6) points when shooting wide open. Out of curiosity, what are your hit rates using live view (CDAF).


Steve

[keos]

How are you determining front vs. back focus (accuracy, re original post) or are you just evaluating hit rate (precision)? One would expect significantly higher precision with center row (f/2.8) AF points than the remaining (f/5.6) points when shooting wide open. Out of curiosity, what are your hit rates using live view (CDAF).


Steve

indoors with a LensAlign mark 2.
outdoors by seeing if the fringing is green or pink.

I'm evaluating both hit rate and accuracy. I can get the center point to hit focus 8/10 times if I MFA for the center 3 points; I can get the outer points to hit focus 6.5/10 times if I MFA for the outer points. The MFA amount is different between the two sets. If I use the center MFA for the out points I hit focus 1/10 times; if I use the outer MFA for the center points I hit focus 0/10 times.


Live view seemingly no matter where I place the focus point I'm about 8-9/10 indoors but 3-4/10 outdoors (makes sense, target isn't exactly still)

edit: I may use a MFA value in between the two sets. I get less hit rate, but more than if I optimize for one and then use the other.

[stevebrot]

indoors with a LensAlign mark 2.

Are you using FocusTune along with the LensAlign? I use a different method,* but LensAlign + FocusTune would be my first choice of the commercial offerings, primarily because the software evaluates scatter before suggesting an adjustment.

FWIW, the manufacturer suggests a range of 1.9m - 3.8m target distance, using the distance tool at LensAlign.

As for your hit ratio, it will improve as you back away from evaluating at 100% and reading distance. After all, the classic DOF calculations are based on an 8"x10" print viewed at 12". As for the target not being "still", if you are shooting hand-held, neither is the photographer. The combination of distance shifts works against a quantitative approach to excellence.

BTW...MFA is Canon-speak. In Pentaxland, we talk about AF fine adjustment.


Steve

* Flat resolution target, natural light (north-facing window), 10X focal length, 20 repetitions, center AF point, recording the direction to correct in 100% live view. If about half correct near and about half correct far, no fine adjustment is needed. A bias one way or the other signals a try at improving by adjusting one step at a time. Once the distribution evens out, I field test using the center AF point, knowing that the other AF points will use the same correction factor, even though a calibration of the optical field using the f/5.6 sensor at that point may indicate a different correction. This possibility I choose to ignore.