This dead horse offends me, so I will beat it a little more.I ran a program called "stopwatch" on my laptop... and video taped it. Then looked at it frame by frame to come up with the actual time.
View attachment 565723
What does that total time include? The time on YouTube's progress bar is about 12 seconds (there's only so accurately I can hit pause) from the D12's first hiss to the B6's ejection charge, so obviously it's not that whole. By the same "pause the video" method, the fuse burn event seems to be one second.
Right from the start I found one thing in your numbers odd, and didn't bother to post about it until @BABAR's post, because it is too small to matter.
The second row of the burn rate calculations says that 100 frames times 0.033 s/frame is 3.34 s, rather than 3.3 or 3.33. But 100 frames at truly 30 fps is 3.33 s. A tenth of a percent different and unimportant. Today I remembered that in the NTSC spec (old analog TV) the rate is not exactly 30 fps, but ever so slightly off; I looked it up and learned it's 30/1.001, i.e. a tenth of a percent. So I checked and found that it's the same for HD. That would make 100 frames 3.3376 s, rounded off to 3.34. So your numbers are correct on that point. But that leads to 24 frames being 0.8008 s, or rounding to 0.801 s, not 0.802. So, as you say: . That changes my other answers to 48.701 in/s or 4.058 ft/s.
But all this is fiddling around in the fourth or fifth significant digit, and I feel comfortably certain that the fuse's burn rate must vary by more than that not only from batch to batch but from foot to foot off the same roll.
It's all BABAR's fault; I wouldn't have weighed in if he hadn't first.