First, let me set up a bit of background about what I already know.
I know the N64 (and I think the SNES too) use a nonstandard method for generating a video signal. They output video frames at the same rate that most equipment generate video fields. So instead of 30fps it's 60fps (actually it's not exactly 30 and 60, but rather 29.97 and 59.94, but for this discussion I'm calling it 30 and 60 for simplicity), but at half the vertical resolution, so that each pair of N64 frames lasts the same duration as a normal NTSC video frame. Well that's not exactly the case. A normal NTSC video frame is 525 lines. The length of a field would be half of that, at 262.5 lines. A N64 doesn't break the last line of its frame on a half line, so a N64 frame actually has a vertical resolution of exactly 263 lines, not 262.5 lines. So this leaves a synchronization problem. If you send the signal at exactly 60fps, due to there being more lines in the frame than a normal NTSC field, you end up needing to increase the speed the lines are sent by a factor of 1.0019 in order make sure that each pair of N64 frames lines up with exactly with a single NTSC frame. However, this would slightly increase the bandwidth of the signal, or decrease the horizontal resolution of the image if you keep the bandwidth the same with a lowpass filter. Alternatively, you could send the lines at the same rate specified by the NTSC standard and thus keep the same bandwidth and horizontal resolution, but you would increase the duration of the frame by a factor of 1.0019, and this might cause a synchronization issue on some TV sets (which would result in a rolling image on those TV sets).

And here's my question.
I'm not sure which technique the N64 uses to handle the extra line. Increase bandwidth? Decrease horizontal resolution? Or increase frame duration? If anybody here knows these technical details, please let me know. Thanks in advance.