Interesting post by Jim Kasson regarding X2d Autofocus

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Bob Foster

Link:https://blog.kasson.com/x2d/gfx-100-ii-x2d-auto-focusing-with-dark-evfs/

I've checked this duplicating Jim's test (but using the XCD 120) and with a "real world" setup.

As the X2D is normally used the XCD 120 simply fails to autofocus in very poor light conditions.

In manual mode with exposure simulation disabled the XCD 120 autofocus is successful. I'll note that the speed of the XCD 120 autofocus appears to be unchanged, it remains slow.

SrMi

Like many other mirrorless cameras, the subject's brightness under the focus point determines the accuracy and speed of AF. That is why turning off exposure simulation in M mode improves AF on X2D.
From my tests and what Jim wrote, it is not as much poor light that causes issues but backlighting that requires darkening the main subject to avoid highlight clipping.
One thing that mystifies me is whether X2D, like other cameras, always brightens EVF during AF if the subject is too dark. By temporary overriding exposure simulation during AF (brightness flash), the camera is more often successful in acquiring focus.

flash

Quote from: Bob Foster on January 13, 2024, 07:10:24 AM
Link:https://blog.kasson.com/x2d/gfx-100-ii-x2d-auto-focusing-with-dark-evfs/

I've checked this duplicating Jim's test (but using the XCD 120) and with a "real world" setup.

As the X2D is normally used the XCD 120 simply fails to autofocus in very poor light conditions.

In manual mode with exposure simulation disabled the XCD 120 autofocus is successful. I'll note that the speed of the XCD 120 autofocus appears to be unchanged, it remains slow.

The 120 macro is by some margin the worst focusing lens in the XCD line up. Try the XCD-V lenses or in the older mount the 21, 30, 90 or zoom and you'll have a very different experience.

Gordon

Bob Foster

Hello Gordon,

The XCD 120. being by a substantial margin the worst of the XCD series of lenses insofar as auto-focus is concerned, is precisely why I selected it for testing.

I created a low light and low contrast situation where the XCD 120 when mounted on an X2D with default settings failed to auto-focus. When exposure simulation is disabled in this sort of condition the lens does indeed auto=focus successfully at about the same rate as it does in far better lighting and/or contrast conditions.

Bob

JCM-Photos

That's an old well known story the XCD120 and autofocus

Very perfect autofocus but only when there is enough light
Sharpen your eyes not your files

asfeir

I was unable to replicate this with XCD 120 mm.

Incidentally, the manual states on page 87:
"• When autofocus is enabled and is analysing the subject, Exposure Simulation is deactivated to let the autofocus system operate in optimal conditions. When the autofocus process is completed, Exposure Simulation is automatically activated again."

SrMi

Quote from: asfeir on January 17, 2024, 03:20:27 AM
I was unable to replicate this with XCD 120 mm.

Incidentally, the manual states on page 87:
"• When autofocus is enabled and is analysing the subject, Exposure Simulation is deactivated to let the autofocus system operate in optimal conditions. When the autofocus process is completed, Exposure Simulation is automatically activated again."

Yes, that is the "flashing" that briefly appears as AF relies on proper subject brightness in the EVF/LCD.

JCM-Photos

Quote from: SrMi on January 17, 2024, 06:27:15 AM
Quote from: asfeir on January 17, 2024, 03:20:27 AM
I was unable to replicate this with XCD 120 mm.

Incidentally, the manual states on page 87:
"• When autofocus is enabled and is analysing the subject, Exposure Simulation is deactivated to let the autofocus system operate in optimal conditions. When the autofocus process is completed, Exposure Simulation is automatically activated again."

Yes, that is the "flashing" that briefly appears as AF relies on proper subject brightness in the EVF/LCD.
All X cameras work like this even the first X1D and all X cameras have some trouble with the XCD120 AF when light or subject is not best.
Sharpen your eyes not your files

Bob Foster

#8
"I was unable to replicate this with XCD 120 mm." -asfeir

I set X2D with XCD 120 as follows:

-Camera Firmware 3.1.0
-Lens Firmware 0.6.0
-Camera in MF mode.
-Exposure > Exposure Simulation in A/S/P/Auto not A/S/P/Auto/M
-I took a spot meter reading of the subject with my Sekonic 858D and set the X2D to about one stop less than what the meter indicated.
-Use the AF-D button to initiate an auto-focus cycle.

I repeated the test this A.M. On 9 of 10 attempts the camera did find focus successfully on the first attempt. One attempt did fail on the first try but succeeded after releasing the AF-D button and pressing it again for a second trial.

Bob




SrMi

Quote from: Bob Foster on January 18, 2024, 03:18:06 AM
"I was unable to replicate this with XCD 120 mm." -asfeir

I set X2D with XCD 120 as follows:

-Camera Firmware 3.1.0
-Lens Firmware 0.6.0
-Camera in MF mode.
-Exposure > Exposure Simulation in A/S/P/Auto not A/S/P/Auto/M
Were you shooting in M mode? Otherwise, that setting does not matter.
Quote
-I took a spot meter reading of the subject with my Sekonic 858D and set the X2D to about one stop less than what the meter indicated.
-Use the AF-D button to initiate an auto-focus cycle.

I repeated the test this A.M. On 9 of 10 attempts the camera did find focus successfully on the first attempt. One attempt did fail on the first try but succeeded after releasing the AF-D button and pressing it again for a second trial.

Bob

Bob Foster

Yes, I was shooting in Manual mode (as I usually do).

I used a copy of the once widely available test target from the days when DSLR auto-focus lenses had to be calibrated for acceptable focus. My copy of this target is simply labeled Cross Hatch Target (dot) png it is identical to what Jim Kasson identifies as a Horshack target. To duplicate the test I printed this target with greatly reduced contrast and dark, the second one even lower contrast and darker to produce good approximations of Kasson's targets. Light level about EV 3, exposure set for EV 2 for the first target, EV 1 for the second.

To insure that this was not an anomaly I was able to auto-focus on a cut "black" pansy against a dim background in similar lighting conditions. (I live in south-central Oklahoma: many pansy cultivars survive and flower in winter conditions as harsh as =18°C/0°F. One recent severe winter we saw the temperature drop to -26°C/-15°F; the plants did survive but flowered no more,,,)

Jgregg

Would folks who have experience of the 120mm XCD , in hindsight, avoid it like the plague? (I have the chance of a very nice condition used lens for a significant saving.)

mar-ko

Quote from: Jgregg on January 20, 2024, 12:56:37 AM
Would folks who have experience of the 120mm XCD , in hindsight, avoid it like the plague? (I have the chance of a very nice condition used lens for a significant saving.)

Depends on the use case: optically outstanding Macro lens, but only 1:2.
Very slow autofocus (not relevant for macro work, but may be an issue for portrait shooters).

Jgregg

Quote from: mar-ko on January 20, 2024, 01:34:26 AM
Quote from: Jgregg on January 20, 2024, 12:56:37 AM
Would folks who have experience of the 120mm XCD , in hindsight, avoid it like the plague? (I have the chance of a very nice condition used lens for a significant saving.)

Depends on the use case: optically outstanding Macro lens, but only 1:2.
Very slow autofocus (not relevant for macro work, but may be an issue for portrait shooters).

To be honest, it will mainly be studio based product and experimental photography. I would not expect it to perform well with any moving subjects

mar-ko