SX510 HS Dark Frame Subtraction Inverted options

  • 7 Replies
  • 1396 Views
SX510 HS Dark Frame Subtraction Inverted options
« on: 01 / July / 2015, 14:45:09 »
Advertisements
Hi everyone, new user here.

First of all I want to really thank all developers for such amazing work, it's absolutely fabulous!

I just got a refurbished SX510HS to practically burn it out with time-lapse and night star long exposure tests to start learning some of that, with CHDK it works amazingly well.

I came across the Dark Frame Subtraction option and started playing with the On/Off switches, I noticed that if I put it on Off, it takes twice the time to process the file, just as in Auto mode, If I put it in On the image is processed quicker. I did these tests with shutter speeds of about 2sec, also I have RAW (DNG) saving enabled, so can anyone confirm if these options are reversed or am I doing something wrong?

I could just leave it on Auto and let CHDK do it when exposure times are above the DFS threshold exposure time and let it kick in when necessary, is it OK to leave it in Auto mode for night photography? For both single star shots and for possibly doing multi-layered star trails?

The only two things that worry me about the Auto mode for a star trail project is the doubling of time where I'll might get gaps in my trails due to the extra processing time and my battery life. I'm sure I could squeeze more shots without DFS enabled.

Anyway, I just want to make sure if those options for DFS are reverted or is it me.

Thank you very much!
---------------------------------
Never tie a ship to a single anchor
nor life to a single hope

Re: SX510 HS Dark Frame Subtraction Inverted options
« Reply #1 on: 01 / July / 2015, 15:18:54 »
This might be of interest : Meteor Intervalometer with Dark Frame Management and answer some of your questions.
Ported :   A1200    SD940   G10    Powershot N    G16

Re: SX510 HS Dark Frame Subtraction Inverted options
« Reply #2 on: 01 / July / 2015, 15:59:39 »
Hey Thanks!

I will for sure give that script a try.

Although I would have to turn off the CHDK's DFS option, correct? Problem is, I'm still not sure if turning it off is really working on this camera or not since it does take twice the time to process the image when its disabled.

Thanks again!
---------------------------------
Never tie a ship to a single anchor
nor life to a single hope

Re: SX510 HS Dark Frame Subtraction Inverted options
« Reply #3 on: 01 / July / 2015, 16:30:48 »
I just did another few tests, and I can confirm that the "YES" and "NO" options are inverted for this option, "Auto" and "Off" take twice the time to process than "On", I also opened the tests in Darktable and saw hotspots on the ones that had the "On" option enabled. Doesn't really bug me but can confuse other people. How or where can I report this?

Thanks!
---------------------------------
Never tie a ship to a single anchor
nor life to a single hope


Re: SX510 HS Dark Frame Subtraction Inverted options
« Reply #4 on: 01 / July / 2015, 17:02:31 »
Just an update:

I just found out that if I leave it to On and turn off the camera and then turn it back on, then it does take twice the time to process (as it sould). I tried the same with Off and powering camera off and back on, it still takes twice the time, switching back to On in DFS then processes the image immediately.

Just weird....
---------------------------------
Never tie a ship to a single anchor
nor life to a single hope

*

Offline srsa_4c

  • ******
  • 3689
Re: SX510 HS Dark Frame Subtraction Inverted options
« Reply #5 on: 01 / July / 2015, 17:11:23 »
I just found out that if I leave it to On and turn off the camera and then turn it back on, then it does take twice the time to process (as it sould). I tried the same with Off and powering camera off and back on, it still takes twice the time, switching back to On in DFS then processes the image immediately.
Can you check whether changing the CHDK setting has immediate effect on the next photo or it only affects photos made after that? You may want to post a note in the porting thread, perhaps somebody else will notice.
If you're not afraid self-compiling CHDK, you can get hints for making little changes in the source to investigate...

Re: SX510 HS Dark Frame Subtraction Inverted options
« Reply #6 on: 01 / July / 2015, 18:18:12 »
@srsa_4c

Correct! The settings are ignored on the next shot right after changing CHDK settings, but the rest of the shots will obey the setting (hope it will with time-lapsing too), so I have to shoot one shot after changing so changes take effect! I can practically live with that! No big fuzz for me if only the first shot has to go.

It all resets or defaults to sameweird  behaviour after restarting camera.

Should I checkout SVN trunk 1.3 or 1.4 for further testing?

Maybe I could start helping out a bit if I start learning a bit of programming. For now I would probably like to get my feet wet, but for this particular "bug" (which isn't that buggy) which version to checkout?

Thanks!
---------------------------------
Never tie a ship to a single anchor
nor life to a single hope

*

Offline srsa_4c

  • ******
  • 3689
Re: SX510 HS Dark Frame Subtraction Inverted options
« Reply #7 on: 01 / July / 2015, 19:48:10 »
Correct! The settings are ignored on the next shot right after changing CHDK settings, but the rest of the shots will obey the setting (hope it will with time-lapsing too), so I have to shoot one shot after changing so changes take effect! I can practically live with that! No big fuzz for me if only the first shot has to go.

It all resets or defaults to sameweird  behaviour after restarting camera.
What I believe happens is that the CHDK setting overwrites a Canon firmware variable responsible for dark frame control, but when the overwrite happens, the firmware has already determined whether it needs to take a dark frame for the current photo - so the CHDK setting only becomes effective from the second shot on.
This could probably be solved by finding another dark frame related firmware variable, but this is not what I meant as "little change".
Your original report was about an "inverted" setting, my suggestion would have been changing the original definition of NR_OFF and NR_ON (defined in platform/generic/capt_seq.c) by defining them in platform/sx510hs/sub/capt_seq_hdr.c and then running make run-code-gen. Since this turned out not to be the problem, doing this is pointless.

Quote
Should I checkout SVN trunk 1.3 or 1.4 for further testing?

Maybe I could start helping out a bit if I start learning a bit of programming. For now I would probably like to get my feet wet, but for this particular "bug" (which isn't that buggy) which version to checkout?
This is in a "low level" part of the code, the port's capt_seq.c, it's similar or same in both trunk and 1.3. Finding the above mentioned firmware variable however requires some ARM assembly skills.


 

Related Topics