Night timelapse (was Re: considerations for SX260 external power supply for time lapse situations) - page 2 - Script Writing - CHDK Forum  

Night timelapse (was Re: considerations for SX260 external power supply for time lapse situations)

  • 52 Replies
  • 23652 Views
*

Offline udo

  • ***
  • 123
Advertisements
OK...
So now I also have an S110 1.03 (!) to compare.
The S110 has larger aperture (2.0) and `bigger` (less tiny) sensor.

rawopint needs tweaking I guess?
I adjusted the max exposure and max ISO values.
What would be the reasonable settings for underexposure and overexposure control if we would do a timelapse run from dusk until dawn?
Priority would be the stars, milkyway andsuch, so a whiteout of the moon is not a big issue.

For the sx260 underexposure control is turned off, is that OK for such a dusk -> dawn run?
We'd like to have the sensitivity and exposure to go up as soon as possible when it gets dark....

*

Offline c_joerg

  • *****
  • 1248
Quote
So now I also have an S110 1.03 (!) to compare.
I have a S110 as well :)

I have not seen many differences between my SX230 and my S110 about noise. The jump between S110 and G1x is much greater. The G1x is much better on noise as my EOS 1200D. The picture show you comparable runs on ISO3200 (EOS 1200D, G1x, S110, IXUS500HS). I made the runs, before I got the Sx230.

I always want to made comparable timelapse run with EOS 1200D, G1x, S110, and Sx230 but I didn’t have the time, to do this.  Interesting as well would be, if the 2bits more resolution of G1x (G1x 14bits, S110 and Sx230 12bits) are really helpful for contrast.
 
Quote
for underexposure and overexposure control
Specially for night runs I would still switching underexposure off….

Quote
We'd like to have the sensitivity and exposure to go up as soon as possible when it gets dark....
We have full moon tonight but it looks like it get not very clear tonight…

M100 100a, M3 121a, G9x II (1.00c), 2*G1x (101a,100e), S110 (103a), SX50 (100c), SX230 (101a), S45,
Flickr https://www.flickr.com/photos/136329431@N06/albums
YouTube https://www.youtube.com/channel/UCrTH0tHy9OYTVDzWIvXEMlw/videos?shelf_id=0&view=0&sort=dd

*

Offline reyalp

  • ******
  • 14082
I can only playing a little bit with the weight. But without the meter window, the exposure will probably not increase...
Right, with only over exposure limit, exposure would only move in one direction, so you need some other kind of target.

For night I don't think having the meter on matters. With any reasonable value of Max Tv and Max ISO, the exposure will be on the limits after it gets dark. You can adjust the low limits, weight and/or prio so that it doesn't counter the over exposure limit much. The exposure calculations will only have an effect if there's enough over exposure to lift it off the limit.

You can set a large step size if the meter processing time is concern. You probably want a large meter area, so something like the moon passing through will not have a large overall effect.
Don't forget what the H stands for.

*

Offline reyalp

  • ******
  • 14082
rawopint needs tweaking I guess?
It shouldn't make much difference, you can probably use a higher ISO for the same quality.

Quote
What would be the reasonable settings for underexposure and overexposure control if we would do a timelapse run from dusk until dawn?
Priority would be the stars, milkyway andsuch, so a whiteout of the moon is not a big issue.
If you want to ignore the moon, and have underexposure protection disabled, then your Overexp thresh should be larger than the moon. To calculate this
  • Get your field of view. You can use this Angular Field of View Calculator with focal length in 35mm equivalent, multiplier 1 (or use actual FL and multiplier for your cam), 4:3 aspect
  • Moon is about 1/2 degree, so Moon width in pixels = 0.5*(sensor width in pixels)/(horizontal fov).
  • Moon pixel area roughly = width squared (yeah, I know it's round ;))
  • Moon pixel area in parts per 100k = (moon area)*100,000/(sensor area)
Your Overexp thresh should be larger than the value in #4. Note that overexposure protection has some effect before the threshhold.

The sun is about the same angular size, but it's so bright it will blow out a larger area and so will still trigger over-exposure protection to some extent. Both the sun and moon can overexpose a much larger area if there is some cloud.

The same calculation is useful if you want to correctly expose the moon. In this case, your Overexp thresh needs to be smaller than the value in #4, and your histogram step must be significantly smaller than the value in #2. You also need to make sure meter low limit doesn't counteract over exposure protection, by using prio or weight adjustments.

If you have under exposure protection enabled, it will limit how far the exposure can be shifted by over exposure protection, so for a night scene that is mostly black, the effect of over-exposure will be limited no matter what the Overexp thresh is.
Quote
For the sx260 underexposure control is turned off, is that OK for such a dusk -> dawn run?
Underexposure protection essentially varies how hard the script tries to resist exposure changes in the opposite direction. So if you aren't worried about correctly exposing the moon, it won't have much effect at night. It will change the behavior at sunset and sunrise, by allowing more over exposure to keep the darker parts of the scene more visible. Whether this is desirable is a matter of preference.
Quote
We'd like to have the sensitivity and exposure to go up as soon as possible when it gets dark....
With default settings, rawopint will try to keep the exposure at the target value until it hits the exposure limits. This actually looks quite strange in a sunset, because the color changes but the scene stays the same brightness and then drops very quickly when the limit is hit. I highly recommend using Bv Ev shift to get a more natural effect. This will lower the target exposure at night, but this doesn't really matter because the script will be on the limits once it gets really dark.

No matter what, I would recommend making some sunset or sunrise tests.
Don't forget what the H stands for.


*

Offline udo

  • ***
  • 123
  • Get your field of view. You can use this Angular Field of View Calculator with focal length in 35mm equivalent, multiplier 1 (or use actual FL and multiplier for your cam), 4:3 aspect
  • Moon is about 1/2 degree, so Moon width in pixels = 0.5*(sensor width in pixels)/(horizontal fov).
  • Moon pixel area roughly = width squared (yeah, I know it's round ;))
  • Moon pixel area in parts per 100k = (moon area)*100,000/(sensor area)
From this site I get a multiplier of 4.62; combining that with a 24 mm equivalent focal length it gives me:
FOV horizontal: 17.7
FOV vertical: 13.4
FOV diagonal: 22.1 (degrees)

Is this correct? (numbers appear small to me; is the equivalent focal length to be used?)

*

Offline c_joerg

  • *****
  • 1248
Quote
If you want to ignore the moon, and have underexposure protection disabled, then your Overexp thresh should be larger than the moon.

Would be a small script or CHDK function helpful, which gives you the percentage of pixel above a threshold?
Making everything black with is under the threshold and white which is above…
« Last Edit: 30 / August / 2015, 07:19:00 by c_joerg »
M100 100a, M3 121a, G9x II (1.00c), 2*G1x (101a,100e), S110 (103a), SX50 (100c), SX230 (101a), S45,
Flickr https://www.flickr.com/photos/136329431@N06/albums
YouTube https://www.youtube.com/channel/UCrTH0tHy9OYTVDzWIvXEMlw/videos?shelf_id=0&view=0&sort=dd

*

Offline reyalp

  • ******
  • 14082
From this site I get a multiplier of 4.62; combining that with a 24 mm equivalent focal length it gives me:
24 is the 35mm equivalent focal length, so if you use that the multiplier should be 1.

If you use the real focal length (5.2) then the multiplier should be 4.62 ( = 24/5.2 for this camera), and you get the correct result.
Don't forget what the H stands for.

*

Offline udo

  • ***
  • 123
24 is the 35mm equivalent focal length, so if you use that the multiplier should be 1.

If you use the real focal length (5.2) then the multiplier should be 4.62 ( = 24/5.2 for this camera), and you get the correct result.
Thanks, so the real and correct values for the S110 are:

FOV horizontal: 71.6
FOV vertical: 56.8
FOV diagonal: 84.1 (degrees)


*

Offline reyalp

  • ******
  • 14082
Thanks, so the real and correct values for the S110 are:
...
At the widest zoom setting, yes.

edit:
I should add that the number don't need be precise, the point is to have an idea whether the Moon covers 100 pixels or 10000 pixels when you are are setting the threshold values.
« Last Edit: 01 / September / 2015, 16:19:30 by reyalp »
Don't forget what the H stands for.

*

Offline udo

  • ***
  • 123
Thanks, so the real and correct values for the S110 are:
...
At the widest zoom setting, yes.
OK, so we use these for the calculation:

FOV horizontal: 71.6
FOV vertical: 56.8
FOV diagonal: 84.1 (degrees)


If you want to ignore the moon, and have underexposure protection disabled, then your Overexp thresh should be larger than the moon. To calculate this
  • Get your field of view. You can use this Angular Field of View Calculator with focal length in 35mm equivalent, multiplier 1 (or use actual FL and multiplier for your cam), 4:3 aspect
  • Moon is about 1/2 degree, so Moon width in pixels = 0.5*(sensor width in pixels)/(horizontal fov).
  • Moon pixel area roughly = width squared (yeah, I know it's round ;))
  • Moon pixel area in parts per 100k = (moon area)*100,000/(sensor area)
Your Overexp thresh should be larger than the value in #4. Note that overexposure protection has some effect before the threshhold.

Number 2 ((4000/71.6)/2) gives me ~27,932
Number 3 thus gives me ~780
Number 4 gets me 6,5

So we set the Overexp protection to 10. Or perhaps higher?


 

Related Topics