CHDK focal plane to suject distance, on G9 - General Help and Assistance on using CHDK stable releases - CHDK Forum

CHDK focal plane to suject distance, on G9

  • 8 Replies
  • 4346 Views
*

Offline hwntw

  • ***
  • 159
CHDK focal plane to suject distance, on G9
« on: 23 / May / 2010, 10:03:48 »
Advertisements
Hello,
For some time I have relied on the DOF misc setting SD to indicate the distance of the subject from my focal pane. I used SD to apply a focus pre-set and pre-set hyperfocal distance.
While experimenting with my G9 in P mode, I noticed that SD was changing as I zoomed in and out, while focus remained on the same, fixed, subject. The CHDK manual describes the SD setting rather vaguely (too vague for me to understand, anyway). Can others confirm what SD actually represents- is it a measured distance to the subject or is it a function of a number of variables?
Ixus 300HS S100


Windows 10

*

Offline hwntw

  • ***
  • 159
Re: CHDK focal plane to suject distance, on G9
« Reply #1 on: 24 / May / 2010, 04:48:11 »
More experimentation with my G9 shows that an accurate reading for SD will occur when focus is done in P mode, followed by confirmation of focus by Manual focus plus button (the one with a cross within a box). Tentatively, the Canon focus system is to blame for the different focal distances (SD) displayed when zooming in and out on a subject.
Simples
Ixus 300HS S100


Windows 10

*

Offline hwntw

  • ***
  • 159
Re: CHDK focal plane to suject distance, on G9
« Reply #2 on: 29 / May / 2010, 12:22:59 »
This discussion is highly relevant- http://www.cooptel.qc.ca/~rlemieu/rf2_a620_calibration.html.
Has anyone tried to use the method described with a G series model?
Ixus 300HS S100


Windows 10

Re: CHDK focal plane to suject distance, on G9
« Reply #3 on: 29 / May / 2010, 17:50:32 »
I noticed that SD was changing as I zoomed in and out, while focus remained on the same, fixed, subject.

That is correct, because the lens to focal plane distance varies as you zoom.

With SDM, that distance is displayed on one of the OSD's for many cameras (but not all).

I do not know if CHDK displays the lens to focal plane distance.


David


*

Offline hwntw

  • ***
  • 159
Re: CHDK focal plane to suject distance, on G9
« Reply #4 on: 30 / May / 2010, 07:21:07 »
Yes, the lens to subject distance changes, but not by much. The variations in SD I see are much larger than zooming the lens makes. The file rf_3.c in the CHDK source by Richard Lemieux and his discussion noted at the link above, provide plenty of insight into the way the Canon autofocus works/not works. His discussion is about the Powershot A620 only. Is there a way that his work can be applied to other Pwershots such as my G9? Can CHDK "cure" Canon's autofocus?
Ixus 300HS S100


Windows 10

Re: CHDK focal plane to suject distance, on G9
« Reply #5 on: 30 / May / 2010, 08:39:38 »
In SDM there is a rangefinder function.
It is only enabled at maximum zoom.
That is so the most accurate distance reading is obtained.

The reading is more accurate because field-of-view is smaller and so too is depth-of-field.

The position of maximum-contrast is presumably used to determine subject distance.

The wider the lens angle, the less accurate this will be.

*

Offline hwntw

  • ***
  • 159
Re: CHDK focal plane to suject distance, on G9
« Reply #6 on: 31 / May / 2010, 08:44:43 »
Thanks MFG, I will try SDM. How can I boot it, as I only use the Firmware Update method to start CHDK on my FAT32 sdcards?
Anyway, it is clear that the OSD displays showing DOF and especially SD, cannot be relied on, since the Canon autofocus itself is unreliable. Over to Richard Lemieux
Ixus 300HS S100


Windows 10

Re: CHDK focal plane to suject distance, on G9
« Reply #7 on: 01 / June / 2010, 05:17:21 »
SDM does not support the firmware update method.
sdminst.exe (included in Common_files.zip) can be used to create dual-partition cards.



*

Offline hwntw

  • ***
  • 159
Re: CHDK focal plane to suject distance, on G9
« Reply #8 on: 02 / June / 2010, 15:15:56 »
Postscript
DoF field values should be calculate drelative to the optical center

No .. to the anterior and posterior principal  points .. which you do not know.
You could determine them by experiment, they would change with every zoom position.
For some cameras that would be over 120 different positions.

Quote
this is can be easily correctable if the distance from the focal plane to lens front is known.
SDM displays that information on one of its OSD's.

Quote
  Finding this value or similar would help in correcting DoF calculations.
...
It has been found for some cameras, check the SDM source code.

The early versions of SDM had extensive DOF-based data on the OSD's.

I deleted the lot.

Why ?

Because the camera cannot measure distance accurately, so the whole excercise is a waste of time.


Ixus 300HS S100


Windows 10

 

Related Topics