The background noise in an image is the "grain" or "static", the random pattern of variations in brightness you can see here (contrast set to 5 in Sitrec)

I was looking for repeating patterns in the clouds when one of my bots noticed there was an unusual pattern of repeating noise in the background, repeating 32 frames apart and separated by 512 pixels.

I generated a "Noise Pattern Detection" tool to further find and visualize this.
https://www.metabunk.org/sitrec/?custom=1/MH370 Thermal/20261017_113826_d2a0eb1f6b55d8032489098a94f3b4f0.js

Turn it on, and you'll immediate get a very distinctive graph with peaks every 32 frames. This graph shows how much similarity is detected between frames a certain distance apart. So this is showing every 32 frames, then 64, 96, etc. A very artificial signal of every 32 frames.

It's not perfect because the image often contains clouds, or a plane and orbs, so the noise is obscured.

Click "Max Contrast" to see the noise better, and then "Show on video". The "Where the Repeat is" timeline shows green where the repeat signal is greater. Click on a green region.

Then click on the Source or Target. These show which columns match between two frames, color-coding the match. The pattern consistently maps source strip = target strip XOR (some number)

You can also click on a column in the main image. Either way, it will show the duplicated columns with their frame numbers.

Note the last column, with the plane in, is different.

I also added a "NPD Folder Scan" button that scans all the videos in a folder (recursively) for repeating noise patterns. I ran it on my "Investigating" folder, which has about 4000 video files collected over many years. Only the MH370 hoax videos have it. (There were some false positives with a strongly repeating texture, but that only showed it needed an artificial pattern to exist)

This unique pattern of repeating noise must be artificial. I think this is yet another demonstration that the video is fake.