I guess this technique is pretty interesting and intuitive in the sense that if the noise distribution is bell-shaped and centered around the true pixel value, by the law of large numbers, the more samples we take, the lower the variance and the more the mean will converge to the true value. Therefore, we simply need to take more frames for the noise to "cancel each other out" and converge. However, this has its drawbacks because it would also require both the subject and camera to stay in place or else you would just get a bunch of overlaps and the mean frame would be blurry.
colinsteidtmann
Is there a similar thing for videos? I'd imagine it would be harder because you're already having to capture ever ms, so if you're averaging them then you might just end up wit reduced fps which isn't good.
brianqch
I think Colin is right about having reduced fps when trying to use this for videos. Although you can break down the video into individual frames, there might be some unwanted effects when averaging. Just like what jayc809 said, we might have blurry shots in the form of motion blur.
eugenek07
This feels very similar to long exposure shots where you take more shots to increase the information within your image. On the other hand, long exposure seems additive, while this averages the frames.
I guess this technique is pretty interesting and intuitive in the sense that if the noise distribution is bell-shaped and centered around the true pixel value, by the law of large numbers, the more samples we take, the lower the variance and the more the mean will converge to the true value. Therefore, we simply need to take more frames for the noise to "cancel each other out" and converge. However, this has its drawbacks because it would also require both the subject and camera to stay in place or else you would just get a bunch of overlaps and the mean frame would be blurry.
Is there a similar thing for videos? I'd imagine it would be harder because you're already having to capture ever ms, so if you're averaging them then you might just end up wit reduced fps which isn't good.
I think Colin is right about having reduced fps when trying to use this for videos. Although you can break down the video into individual frames, there might be some unwanted effects when averaging. Just like what jayc809 said, we might have blurry shots in the form of motion blur.
This feels very similar to long exposure shots where you take more shots to increase the information within your image. On the other hand, long exposure seems additive, while this averages the frames.