Using a custom CSV to manage the FOV

2 days ago 6

PROTECT YOUR DNA WITH QUANTUM TECHNOLOGY

Advertising by Adpathway

I've done a few TR-3Bs (aeroplanes) and Mystical Plasma Orbs (satellites) now, and Sitrec has been a great tool in building out the scenario. But, to date, one thing bothered me: how to handle camera zooms. The mystics that walk among us love their zoom button (I wish they'd love their focus ring just as much). My previous solution to this was to merely edit the video to clip out a single section at a constant FOV/zoom, which generally worked. It also had the advantage of reducing the length of the video loaded into Sitrec.

This worked until the latest video of a lovely TR-3B over Astoria, Oregon.

This video was relatively short, and the zoomed-in sections weren't good enough to be clipped out. So now was the time to learn more about the capabilities of Sitrec as I've seen some advanced ones done by others. So it was time to read the docs. Initially, the FOV was only mentioned in passing, and most track settings were about panning, tilting and tracking. I found the section on MISB data and saw that Sitrec had some control over the FOV value, but it wasn't immediately apparent how. But then I recalled reading that CSV importing was an option, but what could it import? Time to read the code. After a bit of a rummage, I eventually found this section on CSV type detection, which showed that using line1 columns 0 and 1 with "frame" and "fov" allowed for the FOV to be set for each frame.

JavaScript:

if ((csv[0][0].toLowerCase() === "frame" || csv[0][0].toLowerCase() === "time") && (csv[0][1].toLowerCase() === "fov" || csv[0][1].toLowerCase() === "zoom")) { return "FOV" }

A quick test showed it worked as expected, so now to recreate the zooming used in the video. The video is 402 frames long, which is a lot of frames for which to code an FOV. However, there are only two zoom-in movements and a single zoom-out back to the wide angle. The first was to acquire some basic information. The frame numbers the zoom begins and ends and for each frame use the Look View to find a suitable FOV value for those two frames. I then wrote the world's shortest and ugliest code to do a basic frame interpolation that spits out a CSV format series of frames.

#include <stdio.h> #include <stdlib.h> int main(int argc, char *argv[]) { if(argc < 5) { printf("Incorrect args\n"); exit(-1); } int start_frame = atoi(argv[1]); double start_fov = atof(argv[2]); int stop_frame = atoi(argv[3]); double stop_fov = atof(argv[4]); double step = (stop_fov-start_fov) / (stop_frame-start_frame) ; for(int i = start_frame; i <= stop_frame; i++) { printf("\"%d\",\"%f\"\n", i, start_fov); start_fov += step; } return 0; }

I repeated this for the two remaining zooms in the video, collected the CSV files, used a text editor to assemble all the CSV lines into a file with the required first line header, and dropped it Sitrec. The results came out okay for a reasonable match on my first attempt.

A few notes: I couldn't really get a good star match because the video was filmed on a potato, and the only location given was Astoria, Oregon so I just picked a central location there and placed Sitrec's camera there. But I was happy with the results, and another TR-3B is solved (to my satisfaction anyway, I did put a post on the original Reddit post but it got auto-deleted due to my lack of karma, go figure, they've automated ignorance now).

The techniques may well be known to more experienced Sitrec users, but I thought I'd post this for other noobs like me in the hope others find it useful.

The Sitch:-

https://www.metabunk.org/sitrec/?cu.../14452/2025_09_15_Astoria2/20251026_183117.js

Read Entire Article