Regarding the sections you mentioned, there are two situations I think it may refer to:
- showing these points only as some visual guideline
- allowing users to navigate only to these points
For the first situation, I think, it will be fine to display just the start (usually 0:00) and end time for the user to 'callibrate' the timeline in his/her mind. If you want, you can provide additional points on the timeline, so that users - while dragging the time slider, could better tell where (when?) they are - at least before they click the slider itself, because I suppose that there is going to be some quite obviously required elements, like slider control, current time and time under slider upon dragging.
The second is worse, and I believe it is more content related case. I did some of these, but if you fix the navigation to only some points on the video timeline, these need to be related to the content. In other words, the sections must reflect not time, but real 'chapters' of the video. So no default, time based splitting into sections is a good one in this case.
What you could consider, if you expect to deal with long videos and/or need high accuracy for selecting a point on the timeline, instead of splitting into sections, you could use some UX pattern. The best one for this, methinks, is the one provided by Apple in iOS. You use the slider as you would normally, but if you drag your finger away from the timeline slider (while dragging it), the accuracy of sliding increases. It's like zooming on some local part of the timeline to allow more precise selection. Of course this pattern works on iOS, so I don't think it's used anywhere online. To apply it, you would have to display some visual clue for the user upon dragging. This might be some text and some effect of zooming on the timeline (you know: time spans going apart, so that initially 1 minute is e.g. 100px, while upon dragging it the slider, some more precise scale appears, and if user drags his cursor up, the points on this scale get zoomed).