[candgpics]

Hi:

As some of you know I recently purchased the Sigma 18-35 f/1.8. I had some initial success, then had some of the often mentioned AF problems, and since have been able to reduce the AF problems with a global (-1) micro adjustment--a rather minor adjustment, but there still is some AF inaccuracy.

I decided to purchase the USB dock, read the Pentax Forums review, as well as the Lens Rentals review.

I think I know the answer to my "issue," but here goes:

The Sigma Optimization Pro software allows for micro adjustments at 4 focal lengths (35mm, 28mm, 24mm, and 18mm) and at 4 distances for each focal length, for a total of 16 possible adjustments.

The adjustments through the USB dock are from -20 to +20, in contrast to the lens adjustment in a Pentax camera of -10 to +10.

With the USB dock, for instance one starting point is at 35mm and .92 feet (approx 11 inches).

It looks as if I have to shoot one picture--let's say at zero, take a look at the image (I usually place a sheet of paper within the field of view noting the exact measurements/settings), then connect the dock, move the focus point, say, to -1, disconnect the dock, take a picture, go back and reconnect the dock and adjust to -2, disconnect the dock, take a picture, go back and reconnect the dock and adjust to -3, disconnect the dock and take a picture, rinse and keep repeating. This will take a LONG time, even for one who is relatively committed to this process. Of course, I do not have to go up and down through the whole range, but I would want to at least go up and down -5 through +5, unless it was obvious that I did not have to keep making adjustments.

Am I missing something. Is it this step-by-step and slow?

Thank you in advance!!!

[Adam]

There is no faster way to do it. However, if you have a focus test chart you could use that to approximate the severity of the FF/BF (if any) at the 16 presets, and then do trial-and-error from there.

[m1ss1ontomars2k4]

No, that's very wrong. Of course there is a faster way to do it.

Imagine we were playing a game of "guess the number", in the range of -20 to +20. You start guessing with 0. I say that you're too high. So, what's your next guess? It's not -1; that would be silly. Your next guess is -10. And so the game continues: I tell you -10 is too low. Your next guess isn't -9. Your next guess is -5. Then maybe I tell you that's too high now. So, your next guess isn't -6. It's -7 or -8 (technically, it should be -7.5, but let's pretend we're only allowing integers). Eventually you will just guess the right number, and that will be that.

The same procedure works for AFMA, using either the dock or the built-in AFMA in your camera. Higher and lower are nothing more than front and back focus. You can use longitudinal CA to detect front vs. back focus if your lens is susceptible to that sort of thing (a target with a lot of black and white patterns works best), or you can use a ruler or something.

I don't know where everyone is getting this ridiculous idea that you have to go one by one through all possible values until you find the correct one; you are not the first person on the internet I've seen with this misconception. Indeed, it seems that this is perceived as a major downside of the Sigma dock, that you need 41 tries per focal length range (or 164 total, if you have a zoom lens) to properly nail the focus. This is not correct. Simply going from -5 to +5 as you suggested is also not very effective. That's already 11 settings that you need to try, just to decide if you want to try more!

In computer science, the procedure you have described is formally known as a "linear search" and will get you laughed out of most job interviews. It can be proven that the method I describe, which is known as a "binary search", will take at most log(41)/log(2), or approximately 6, tries before finding the correct value. Since it can be a little harder to identify "focus adjustment is correct" vs. just knowing that the number you have chosen matches the number I'm thinking of, you might take 2 extra tests, +/-1 from the final number you settled on. This is still only 8 tries total and it is guaranteed to find you the correct number.

[Tom S.]

No, that's very wrong. Of course there is a faster way to do it.

Imagine we were playing a game of "guess the number", in the range of -20 to +20. You start guessing with 0. I say that you're too high. So, what's your next guess? It's not -1; that would be silly. Your next guess is -10. And so the game continues: I tell you -10 is too low. Your next guess isn't -9. Your next guess is -5. Then maybe I tell you that's too high now. So, your next guess isn't -6. It's -7 or -8 (technically, it should be -7.5, but let's pretend we're only allowing integers). Eventually you will just guess the right number, and that will be that.

The same procedure works for AFMA, using either the dock or the built-in AFMA in your camera. Higher and lower are nothing more than front and back focus. You can use longitudinal CA to detect front vs. back focus if your lens is susceptible to that sort of thing (a target with a lot of black and white patterns works best), or you can use a ruler or something.

I don't know where everyone is getting this ridiculous idea that you have to go one by one through all possible values until you find the correct one; you are not the first person on the internet I've seen with this misconception. Indeed, it seems that this is perceived as a major downside of the Sigma dock, that you need 41 tries per focal length range (or 164 total, if you have a zoom lens) to properly nail the focus. This is not correct. Simply going from -5 to +5 as you suggested is also not very effective. That's already 11 settings that you need to try, just to decide if you want to try more!

In computer science, the procedure you have described is formally known as a "linear search" and will get you laughed out of most job interviews. It can be proven that the method I describe, which is known as a "binary search", will take at most log(41)/log(2), or approximately 6, tries before finding the correct value. Since it can be a little harder to identify "focus adjustment is correct" vs. just knowing that the number you have chosen matches the number I'm thinking of, you might take 2 extra tests, +/-1 from the final number you settled on. This is still only 8 tries total and it is guaranteed to find you the correct number.

If you read Adam's post, that's basically what he is saying when he said 'trial and error' method.

[candgpics]

Hi:

Thank you all for your comments.

I won't go 1 by 1, but at least I know now that I do have to take individual photos, check, and then reset.

One issue that I did not mention, but am cognizant of, is the repeated mounting and dismounting of the lens to the dock. I am being very careful. As well, I am making sure I do not "nudge" the tripod or get dust on the sensor (I guess I could put another lens back on my camera--I am using one of my K5ii bodies--or the white lens mount cover, but the camera is on a 45 degree angle and it keeps falling off!!!).

Oh well, it's not that I do not enjoy this. Love all this testing stuff, but like others, family and work make putting aside the time for this kind of adjustment tricky.

Anyway, as always, thank you again!!!

[clackers]

Anyone attempting +1/-1 iterations should not apply to go on 'The Price is Right'. :-D

[candgpics]

Hi. Respectfully, I have tried larger AF increments in the past only to have to re-do the adjustments.

[baya]

Hi,

has anyone done customization over USB dock on this lens and used it on K-3? Is it possible than someone can share these setting? It would be a nice gesture and I would be very grateful for it .

Many tnx

baya

[candgpics]

Hi, Baya.

After several attempts with the dock I gave up. I almost returned the lens, but after speaking with Sigma tech support in New York (USA), I decided to send it to Sigma with my K3 (I paid shipping to Sigma, but the service is otherwise free and Sigma pays the return shipping). I was hesitant because I thought perhaps I received a defective lens and also because the lens can only be configured to match one camera at a time. Nonetheless, I sent in the camera and lens and received both back within 10 days (I live in Maine, USA). And, the settings are individual to each person's camera (my settings would likely not help you).

The lens now works well. Not perfect. I still get seemingly random, but recurring missed focus shots. So, I make sure to always take a couple of extra shots when shooting an important image. I know that is an unsatisfying resolution (at least to me), but when the lens is "on," which is most of the time, the images are excellent. I've learned to live with this issue since even at f/1.8 the lens is very sharp and contrasty.

I hope this is helpful.

[baya]

Hi, Baya.

After several attempts with the dock I gave up. I almost returned the lens, but after speaking with Sigma tech support in New York (USA), I decided to send it to Sigma with my K3 (I paid shipping to Sigma, but the service is otherwise free and Sigma pays the return shipping). I was hesitant because I thought perhaps I received a defective lens and also because the lens can only be configured to match one camera at a time. Nonetheless, I sent in the camera and lens and received both back within 10 days (I live in Maine, USA). And, the settings are individual to each person's camera (my settings would likely not help you).

The lens now works well. Not perfect. I still get seemingly random, but recurring missed focus shots. So, I make sure to always take a couple of extra shots when shooting an important image. I know that is an unsatisfying resolution (at least to me), but when the lens is "on," which is most of the time, the images are excellent. I've learned to live with this issue since even at f/1.8 the lens is very sharp and contrasty.

I hope this is helpful.

Thanks Candgpics,

here where I come from (Croatia, EU) things are not ruining so smooth. Just for the lens I was waiting 2-3 months. Pentax is exotic brand here. I can not even imagine how much it would take to send the equipment in Sigma.
This would mean than every camera and lens are unique Maybe that's why Sigma calls this series Art Idea for advertisement "Sigma Art lens - you get unique peace of Art" (I put a copyright on this slogan ).
It would be maybe good that everyone who has this customized settings post it here so maybe we can find some kind of "Golden middle". To set some start point.
What do you think about that?

BR

b@y@

[candgpics]

Baya, I will not have access to the lens and dock for about another two weeks. I am happy to hook up the dock and let you know the settings Sigma applied.

I can tell you that with both of my K3 bodies all of my lenses front focus and I have to use the micro adjustment to some extent. On my K5ii all of my Sigma lenses need a micro adjustment, too, but my Pentax lenses are fine.

I often shoot indoors at mma and martial arts events, with a wide aperture, so small focusing errors can stand out.

[baya]

Baya, I will not have access to the lens and dock for about another two weeks. I am happy to hook up the dock and let you know the settings Sigma applied.

I can tell you that with both of my K3 bodies all of my lenses front focus and I have to use the micro adjustment to some extent. On my K5ii all of my Sigma lenses need a micro adjustment, too, but my Pentax lenses are fine.

I often shoot indoors at mma and martial arts events, with a wide aperture, so small focusing errors can stand out.

No hurry. I find this as an challenge Same experience/opinion about Sigma/Pentax lenses.

BR

B@y@