Hacker News Reader

Comment by dylan604 | Hacker News Reader

I often use long exposure (with heavy ND filters) during the day for the specific purpose of removing people from the image. Cityscapes in a busy area is nearly impossible to not have people in the image which makes it difficult to sale the image. I'm only taking 20-30 seconds exposures vs the 7 minutes mentioned in the article, but it really makes eliminating the crowds a much more manageable task.

View on HN

Have you seen the technique where you take many quick photos over the course of a few minutes, and then use photoshop to create a single image of all the parts that don’t change?

https://photoshoptrainingchannel.com/remove-tourists-stack-m...

I am familiar with this technique, and it works well for a single still image. My end goal is rarely a single image as I'm usually shooting timelapse. Since it's timelapse, the motion blur still shows motion, but it also pretty much ensures nobody is identifiable.

Come Fly With Me did a really fantastic timelapse effect https://www.youtube.com/watch?v=BFnhCeNkMO8 where the people in the foreground didn't run at regular speed.

Ah! I'd seen something like this but instead of photoshop they used Python+Pillow to XOR a bunch of images together.

This makes me wonder. Is there an algorithm that allows taking multiple short-exposure photos, and eliminating the moving parts by copy+pasting areas from different photos in the set?

just take the median

Photoshop does this. See your sister commenter.

do u have a gallery online?

Reply on news.ycombinator.com