Posts: 4
Threads: 1
Joined: Oct 2017
Ultimately this would be implemented in DSLRDashboard and not LRTimelapse but I do think it would be covered in the LRTimelapse section of that program. So I am submitting this here for review and discussion before making a feature request to DSLRDashboard.
As a general statement the night images in my auto timelapses are too bright and there is no way to correct this during the timelapse other than manually. If shooting raw then this brightness can be of course reduced/recovered. However, I may not want that brightness because it could be that the slower shutter speed introduces more blur to moving objects thank I want or a higher ISO causes more overall noise in the image.
Additionally, although not directly related to the brightness, I may want to ensure I get more frames after dark without having to have the timelapse run for longer than necessary. Due to the longer shutter speeds after dark my guess is that a 2 hour timelapse before and after sunset results in a video which is 80% daylight and 20% night (of course with a transition). Yes setting the maximum shutterspeed can reduce the object blur but overall not having the night image brighter than necessary is a good thing.
So my suggestion how to handle the night brightness is to have a percentage value factored into the algorithm calculating exposure. I can set it as a percentage of the initial captured image. What percentage to set may not initially be intuitive but it is something that would come with experience.
There are also some constants that can help. Sunset time is known, golden hour is known and blue hour is known. This information is available for any location online and therefore these times could be used to calculate when to introduce the percentage value in the exposure calculation. Additionally if I am shooting into the sun or away from the sun will have an impact on the sky brightness for a period of time after sunset. These timings are generally fairly constant and would be known prior to timelapse running.
So in summary the factors would be:
Golden Hour Start Time
Sunset Time
Blue Hour End Time
Percentage of Brightness of Reference Image.
Thoughts and comments welcome.
As a general statement the night images in my auto timelapses are too bright and there is no way to correct this during the timelapse other than manually. If shooting raw then this brightness can be of course reduced/recovered. However, I may not want that brightness because it could be that the slower shutter speed introduces more blur to moving objects thank I want or a higher ISO causes more overall noise in the image.
Additionally, although not directly related to the brightness, I may want to ensure I get more frames after dark without having to have the timelapse run for longer than necessary. Due to the longer shutter speeds after dark my guess is that a 2 hour timelapse before and after sunset results in a video which is 80% daylight and 20% night (of course with a transition). Yes setting the maximum shutterspeed can reduce the object blur but overall not having the night image brighter than necessary is a good thing.
So my suggestion how to handle the night brightness is to have a percentage value factored into the algorithm calculating exposure. I can set it as a percentage of the initial captured image. What percentage to set may not initially be intuitive but it is something that would come with experience.
There are also some constants that can help. Sunset time is known, golden hour is known and blue hour is known. This information is available for any location online and therefore these times could be used to calculate when to introduce the percentage value in the exposure calculation. Additionally if I am shooting into the sun or away from the sun will have an impact on the sky brightness for a period of time after sunset. These timings are generally fairly constant and would be known prior to timelapse running.
So in summary the factors would be:
Golden Hour Start Time
Sunset Time
Blue Hour End Time
Percentage of Brightness of Reference Image.
Thoughts and comments welcome.