[dosdan]

Looks more definitive than takings shots of a newspaper page.

AF microadjustment for the 1Ds mark III, 1D Mk3, 5D Mk2

Dan.

[konraDarnok]

Looks more definitive than takings shots of a newspaper page.

AF microadjustment for the 1Ds mark III, 1D Mk3, 5D Mk2

Dan.

Fascinating. Good Find.

[dosdan]

I just tested a DA 16-45/F4 on a K20D, set to a FL of 45mm, at 2.1m from a moiré pattern displayed on a PC LCD screen. (The distance is supposed to be at least 50x FL.) In LV (Custom Settings | option 1 of menu item 30) I used one right turn of the back e-dial to perform a quick zoom of x4 (Playback | Playback display | Quick zoom | x4) to fill the camera back LCD display with its own interpretation of the moiré pattern. What's captured in the shot is not the same as shown on the back LCD display. I then used the manual-focus override on the DA lens to see if moving from the AF-selected position in either direction increased the moiré as shown on the back LCD. It did not. It diminished it, indicating that the AF-selected position was the best.

When doing this test it's a case of "the moirér, the merrier".

Two shots. First the actual amount of the PC screen shot from a distance of 2.1m with a 45mm FL.

Now a crop of the same shot that was similar to what I saw with x4 quick zooming. Note that there was more moiré on the back LCD display than captured on the shot
(two extra columns of 4 smaller circles on the left & right of the centre pattern).

Dan.

[legacyb4]

Neat stuff.

Haven't tried using Liveview just yet so might give it a shot this weekend.

[Class A]

Whatever you shoot (newspaper, focus chart, laptop screen) make sure it has got the colour temperature that matches your real life situations.

For the moiré approach, this means that your laptop screen should be calibrated to daylight temperature, if you normally shoot outdoors.

[dosdan]

For the moiré approach, this means that your laptop screen should be calibrated to daylight temperature, if you normally shoot outdoors.

Colour temp has nothing to do with this particular test. It's looking for phase interference of the the pattern as shown on the PC's LCD display and interpreted by the the camera's back LCD display. So it's showing the accuracy of the AF (quality of the moiré pattern - as I understand it - if the camera is in focus it will interact/interfere more with the moiré pattern displayed on the the PC and thus show a a more distinct & complex moiré pattern on the camera's back LCD display) rather than the colour interpretation. So we not interested in any colour fringing on the moiré pattern, just it's intricacy.

Dan.

[blende8]

I have tried it, too, but in my view it is not very good.
The problem is to optimize for maximum Moiré.
It is a bit subjective, when this point is arrived.
You can turn the focus slightly and the pattern does not change significantly.

The idea is very good, though and I applaud to its invention!
Perhaps it can be improved in some way.

[Class A]

Colour temp has nothing to do with this particular test.

I'm pretty sure it does.

So we not interested in any colour fringing on the moiré pattern, just it's intricacy.

I am well aware of this.

However, in particular Pentax AF is sensitive to colour temperature. A focus chart shot under Tungsten light will show front focus for a lens that will be spot on under daylight.

If you make AF adjustments under Tungsten lighting so that the lens is spot on, it will BF in daylight.

Same thing with the moiré test; if you adjust the AF that it achieves optimum focus (maximum moiré) and the pattern you shoot has a colour temperature warmer than daylight, your lens/camera combination will BF under daylight.

It doesn't matter how you check optimum focus (focus bracketing, targeting the "zero" mark on a focus chart, maximise a moiré pattern), if you use adjust the AF system under lighting that is different from your normal shooting conditions, you will be off the perfect focus in your normal shooting conditions.

[dosdan]

However, in particular Pentax AF is sensitive to colour temperature. A focus chart shot under Tungsten light will show front focus for a lens that will be spot on under daylight.

If you make AF adjustments under Tungsten lighting so that the lens is spot on, it will BF in daylight.

In which case the predominant light source was the LCD back illumination. I estimate this is 5000K/Daylight White Fluorescent ("N").

[Class A]

In which case the predominant light source was the LCD back illumination.

It matters, though, what the panel lets through. And this depends on the colour settings/calibration.

I reckon its a good idea to double check one's result with realistic shooting conditions unless one has made sure that the colour temperature of the LCD screen is as desired.

[dosdan]

I based the WB on the camera response to a blank white screen page in Wordpad. It's 5000K. I don't adjust monitor contrast or brightness. It's always the same. I don't run a "dynamic" setting on the monitor. It stays in "photo" mode.

[Class A]

I based the WB on the camera response to a blank white screen page in Wordpad. It's 5000K.

Does your K20D tell you what colour temperature it sees? That'll be a cool feature.

Anyhow, good for you that things line up well in your case. My comments weren't not exclusively targeted at you but meant for anyone who may want to try this method.

Is 5000K the right temperature for viewing pictures? I know that it is 6500K for movies but that is probably a different case.

BTW, regarding the difficulty to find the maximum moiré: I asked the inventor of the moiré method about a variation I came up with and he responded that he thought it should be admissible: Slightly tilt the laptop screen and then the moiré pattern will move up and down the screen when you refocus. At its centre position, sharpness is optimised at the centre (camera needs to point at the centre). Not sure how much this helps. Never had the time to fully explore various focus adjustment techniques and validate how well they apply to real world shooting.

It would be interesting to learn from those who have applied the moiré method to see whether it leads to adjustments so that

• focus chart tests show no BF/FF (with tilted / non-tilted targets)
• real world results are as desired

For cameras with no live view, the moiré method isn't as convenient, but connecting the camera to the laptop and using Remote Assistant is perhaps an acceptable second best solution.

[royden]

Looks more definitive than takings shots of a newspaper page.

AF microadjustment for the 1Ds mark III, 1D Mk3, 5D Mk2

Dan.

Thanks Ken, I'll have to try this soon. My 20D need it.