raspiraw "--expus" does not work for ov5647 (it does work for imx219)

[Hermann-SW]

Doing raspiraw captures with imx219 for "--expus 1000" and "--expus 2000" gives different bright frames after processing with dcraw.

Doing the same with ov5647 gives identical frames, no difference in brightness.

I tested that this bug exists on this repo as well.

This issues is different to the incorrect calculation from issue 20.

[6by9]

See #18.
Getting a calculation that can be parameterised and is correct for both sensors is trickier than it looks.

[Hermann-SW]

I used your latest raspiraw and setting exposure does not work at all.
I cannot find a repo of gustav-engdahl but only see his commits on your branch.
But the pull request is not integrated into your branch yet.
How can I test whether the three commits in that pull request make "-e" work on your repo?
(without copying over line by line by hand)

[GustavEngdahl]

Hi, You should be able to build my repo at: https://github.com/GustavEngdahl/raspiraw/commits/master It's master has the three patches needed to get the VTS/Exposure/gain values correctly written. The patches are not chip specific in any way, but I've only verified on OV5647. Regards, Gustav Den mån 4 mars 2019 kl 10:56 skrev Hermann Stamm-Wilbrandt < notifications@github.com>:

…

I used your latest raspiraw and setting shutter does not work at all. I cannot find a repo of gustav-engdahl but only see his commits on your branch. But the pull request is anot integrated into your branch yet. How can I test whether the three commits in that pull request make "-e" work on your repo? (without copying over line by line by hand) — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#25 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/Ap0oSRL7x-JmdJqKox_oplSyhG4jGEu8ks5vTO3UgaJpZM4bbK1J> .

[Hermann-SW]

Thank you, I did build your repo and it fixes the issue I reported here, closing this issue as duplicate now.
These are the (different now) frames with different --expus settings.

I did not look in detail into your changes, but "--expus" seems to have worked for imx219 because exposure bits are a multiple of 8 whereas for ov5647 they are not:

      .exposure_reg =         0x015A,
      .exposure_reg_num_bits = 16,

   .exposure_reg =         0x3500,
   .exposure_reg_num_bits = 20,

(scene is lit with 1000lm unfocused light)
--expus 4000
--expus 2000
--expus 1000
--expus 500

P.S:
Your version of raspiraw has 465 lines less than that of 6by9 ...

[GustavEngdahl]

Good to hear. Can you verify that my fixes doesn't break imx219? If nothing breaks, perhaps 6by9 could bring in #18 to master? Regards, Gustav Den mån 4 mars 2019 kl 12:52 skrev Hermann Stamm-Wilbrandt < notifications@github.com>:

…

Thank you, I did build your repo and it fixes the issue I reported here, closing this issue as duplicate now. These are the (different now) frames with different --expus settings. I did not look in detail into your changes, but "--expus" seems to have worked because for imx219 because exposure bits are a multiple of 8 whereas for ov5647 they are not: .exposure_reg = 0x015A, .exposure_reg_num_bits = 16, .exposure_reg = 0x3500, .exposure_reg_num_bits = 20, --expus 4000 <https://stamm-wilbrandt.de/en/forum/72kmh/out.0017.ppm.d.png> --expus 2000 <https://stamm-wilbrandt.de/en/forum/72kmh/out.0016.ppm.d.png> --expus 1000 <https://stamm-wilbrandt.de/en/forum/72kmh/out.0015.ppm.d.png> --expus 500 <https://stamm-wilbrandt.de/en/forum/72kmh/out.0014.ppm.d.png> — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#25 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/Ap0oSbZlt8z2L3ualXWEOGJ_0x1JD14Mks5vTQkJgaJpZM4bbK1J> .