3 ms·

would have liked to have seen a better explanation of the red lines in the image. it sounds like those are added by humans to outline the runners' chests; how d

would have liked to have seen a better explanation of the red lines in the image. it sounds like those are added by humans to outline the runners' chests; how do they ensure the lines are drawn on the frame exactly at the finish line?

grogenaut 2y ago
If you read the other comments and the article the camera is one pixel wide, each vertical column in the image IS a frame. It's ensured with physics. It's a bit subtle but makes sense once you grok that. So each pixel left is 1/40,000 later
- laweijfmvo 2y ago
  so every column is the finish line?
  - grogenaut 2y ago
    yes, there's literally only one column of pixels, the sensor is say 4k tall by 1 wide, eg a row of pixels. The x axis in the image (left right) is the time domain. Imagin instead of the people running from left to right, instead they were still and the pixels were moving right to left, or if that breaks your brain, think about if the track was a people mover and the runners were a static diorama moving left to right and the sensor is just it over time. But also none of that is real.
iainmerrick 2y ago
Does it matter if they're off by ±1 pixel? It's pretty easy to eyeball the rightmost part of each athlete's torso and the frame rate is high enough that there's plenty of space between them. It looks plenty reliable as long as the camera was positioned and calibrated correctly (which I assume was carefully tested beforehand).
dist-epoch 2y ago
There could be two sets of line, one line assigned exactly when each runner crosses to get his time, and another set assigned when the first runner crosses, and this picture is distributed.
jhayward 2y ago
> how do they ensure the lines are drawn on the frame exactly at the finish line? They don't. The horizontal axis is time. The red lines mark the time each runner's torso was counted as crossing the finish line.

grogenaut 2y ago

If you read the other comments and the article the camera is one pixel wide, each vertical column in the image IS a frame. It's ensured with physics. It's a bit subtle but makes sense once you grok that. So each pixel left is 1/40,000 later

laweijfmvo 2y ago
so every column is the finish line?
- grogenaut 2y ago
  yes, there's literally only one column of pixels, the sensor is say 4k tall by 1 wide, eg a row of pixels. The x axis in the image (left right) is the time domain. Imagin instead of the people running from left to right, instead they were still and the pixels were moving right to left, or if that breaks your brain, think about if the track was a people mover and the runners were a static diorama moving left to right and the sensor is just it over time. But also none of that is real.

iainmerrick 2y ago

Does it matter if they're off by ±1 pixel? It's pretty easy to eyeball the rightmost part of each athlete's torso and the frame rate is high enough that there's plenty of space between them. It looks plenty reliable as long as the camera was positioned and calibrated correctly (which I assume was carefully tested beforehand).

dist-epoch 2y ago

There could be two sets of line, one line assigned exactly when each runner crosses to get his time, and another set assigned when the first runner crosses, and this picture is distributed.

jhayward 2y ago

> how do they ensure the lines are drawn on the frame exactly at the finish line? They don't. The horizontal axis is time. The red lines mark the time each runner's torso was counted as crossing the finish line.