When I use Kagi Translate for Firefox, the YouTube integration for translating subtitles updates word by word, and the result is dynamically centered, which means that individual words are effectively almost always moving. This is completely unusable, as the effort needed to track words that move in a stop-start fashion, at a random, unpredictable cadence, is significantly higher than what it takes to read static subtitles.
I expected the subtitles to show up in either of these two ways:
Show all the subtitle words at once, so that all the translated subtitles covering a certain time period (usually one sentence, or at least as close as possible) appear at the same time. This is how 90% of all subtitles work. The individual word, once printed, never moves with respect to the edges of the screen or the centre of the viewport. This method is preferable because most fluent readers read significantly faster than a human can speak, meaning that once the subtitle is read, the viewer can glance at the actual visuals of the video while the narration catches up.
This is strongly preferred.
If you have to display the results on a word-by-word basis (nobody does this except YouTube), at least make sure that all the words currently on screen are not dynamically re-centred, meaning the word, once printed, does not move relative to the screen size or the centre of the viewport. This way, even though the appearance cadence of the words is still slow and random (which impairs fast readers), at least the reader doesn't have to visually track a moving word across the screen.
If you won't consider changing the default, please, at a minimum, consider adding these alternate styles as an option.
