In this post I will try describe tools and process of filming UFO, it's based on FarSight method but simplified. I will try to explain step by step, and will go into details where neede, my main goal is to show that there is no single way of filming them, but if you understand general idea - you will be able to adapt it the way it fits you better.

Lets start from the most basic example how to film UFO in the sky, all you need is:

  1. Camera that can film 4k resolution at least 60 fps (frames per second)
  2. Infrared filter
  3. Computer to view filmed footage.

NOTE: Always check my equipment list for recommendations. I try to keep this list up to date with my current equipment and equipment that I tried but decided no to use.

Filming

This is the easiest step, attach lens and film at 4k 60fps for 1 - 10 minutes. Longer the video - better chanses to catch something. Also highly recommend to use some kind of stand to place your phone while filming, hands might get tired and start shaking - this will make video processing very difficult.

  1. Attach lens, select 4k ant at least 60fps on your device
  2. Point camera to the sky
    1. Make sure other objects are not in the frame, for example: trees, buildings ...
    2. Use stand. Shaking hands will make searching UFO much harder.
    3. Try to find location with least amount of clouds, most UFO objects that I filmed are above clouds.
  3. Film for ~10 minutes

Video processing

If you filmed for 10 minutes at 60 fps, then you will have:

  • 10 minutes X 60 seconds = 600 seconds of video
  • 600 seconds X 60 frames per second = 36 000 frames
  • Each frames contains around 3 840 horizontal pixels and 2 160 vertical pixels

Now that you have 36000 frames (still images) at 4K resolution - you should go through each of them, comparing current frame to previous searching for some UFO. And UFO are usually very small and fast, they sometimes appear in video for 10 frames - it's usually faster then blinking.

Below are three randomly choosen frames of my video

  • One frame has nothing
  • One frame has UFO
  • One frame has bird

These frames are not in the same order as the list, use stopwatch and see how long it takes for you to figure our which frame is which.

So I beleave this is the main reason why UFO filming didn't gain any traction - it is very time consuming.

Video processing with software

When I first viewed video frame by frame - I realized, that this is insanity. Nothing really changes between frames, frames are so larger, and I don't even know what I suppose to look for, can my iPhone 15 Pro even capture them, does my IR680 lens is good enough to see them... it was frustrating experience. Being software develope, I started looking for a solution, and my solution was to write application that would detect movement in video, general idea of my application:

graph TD A[Original video file] --> B[Read frame by frame] subgraph processing_script.py B --> D[Convert frame to grayscale] D --> E[Compare current frame with previous frame] E --> G{Check if compared frames have any differences} G -->|Frames are different| J[Draw precise timestamp to that frame] G -->|Frames are identical| K[Skip frame, do nothing] end E -->|Write all frames to file| F[Full difference file] J -->|Write only different frames| H[Only different frames file]

One thing worth mentioning is that during frame comparison, you can and should specify treshold, what does that mean. My current setup has treshold range from 0 to 255,

  • 0 - nothing changed, pixels are identical between frames
  • 255 - everything changed and pixels have opposite colors

I'm still searching for perfect treshold value, it really depends on the weather, cloud movement, camera.

Identical frames as above, but this time only differences are shown:

File sizes: Original video - 2.3gb (10:07 length) Full length video in black and white, showing differences - 388.7mb (10:07 length) Video with only different frames - 74.2mb (1:42 length)

It takes almost an hour to process 10 minutes video on my machine.

Instead of investigating 10 minutes video, comparing each pixel - you only need to review 1:42 minutes length video, and it takes around 10 minutes to do it.

Summary

With the help of some scripts we can sideload hardest work to computer, leaving as a reviewers of final results. Script itself is no where near finished, currently it is just a prototype, it is definetly possible to have real time processing.

With this technique I managed to find UFO in almost all videos that I took, some of them have one or two frames that have some anomalies, but due to video quality difficult to say for sure.