That was really cool fldave, thanks for sharing. I guess you masked out the surf, it gave it a cool effect.

So did you set the exposure to a manual setting of what it would be when the sun was brightest, or was it on auto?

Since it was a sunset, it was going to get darker, so I set it to manual to zebra stripes 100, then didn't touch anything. Definitely easier than guessing what the brightest exposure will be 30 minutes after sunrise.

I set the lower surf to darken about the same rate as the sky was darkening.

OK, so we're talking about BOTH film and video here. And we're talking about a film movie camera with a rotating shutter, which means that the shutter speed is directly related to film speed. (Obviously there were (are) movie cameras with non-rotating shutters -- the Beaulieu MR8, I think -- but they are not common.)

I guess I was only thinking about video where the shutter speed (in a modern video camera) can be set to just about anything you want, regardless of whether you capture in 24p, 30p, 60p, 50i, 60i or any other speed.

I would think that any timelapse or stop motion with film would be done with a standard still camera, loaded with film stock, or else a standard movie camera fitted with some sort of lens containing a standard leaf shutter. Otherwise, how can you take just one frame and get good exposure? I do not have experience with Panavision or anything like that, but I think most of these require several frames to come up to speed, so that would not work for anything where you wanted to take just one frame and then wait a long time before the next frame.

"Otherwise, how can you take just one frame and get good exposure? "
A serious ND filter. We do have a couple, rather expensive.

But aside from that I was only using the film camera as an example because it's easy enough to understand.

Let me try a different tack.
Timelapse is a way of speeding up reality. How would reality look if it was sped up in front of you and a video camera. It's easy enough to guess without getting into any physics. Just look at how fast and slow moving objects appear when shot with our video cameras. With timelapse we're increasing the speed that everything moves at, the amount of motion blur should do the same to mimic what would happen if they really were moving that fast.

Bob.

With timelapse we're increasing the speed that everything moves at, the amount of motion blur should do the same to mimic what would happen if they really were moving that fast.OK, I [at last] get your point. You want MORE blur because anything moving that fast would be blurred. I am a little slow sometimes.

I guess the reason I didn't understand at first is that the original post was about a sunrise, and even when sped up a lot, it still moves slowly. By contrast, I think you were looking at those ships zipping across the harbor in that video posted earlier in the thread, or perhaps thinking of time lapse construction videos where people are zipping along, or sunset videos they show on the weather broadcasts here in the states, where they'll show not only the sky, but some highway down below, with cars zooming along at 1,000 mph.

The only counter to this -- and I'm now speaking from an aesthetic standpoint -- is that when things get sped up to the point of unreality, I'm not sure what the eye "expects" and therefore using a slower shutter to blur things will just make the picture less sharp, but not really make it look any more "real." In other words, we don't have a basis in reality for normal objects moving that fast.

Now if I were trying to speed things up a little by undercranking the camera, as was often done in Keystone Cops and other similar sequences, then I think what you describe would be very important. Of course there are many other problems with undercranking. The Keystone Cops, for instance, would have to run in order to move that fast, yet instead, since they aren't, it just looks like they are walking really, really fast, which looks fake (and funny).

So, as I think about this, is there ever actually a situation where you are trying to make something look real (i.e., like some reality with which we are all familiar) when you change its speed, and therefore is there ever a time when you would actually want to change the shutter speed? Or do speed changes always put us in the realm of the unreal?

I think you may be confusing yourself with what happens in high speed cameras. In that case, when a camera -- through whatever process is used -- takes hundreds or thousands of frames per second then, by definition, each frame has to be exposed for far less than 1/24 or 1/30 or 1/60 of a second. Otherwise, there wouldn't be time to advance and expose the next frame of film.

Quite right! Highspeed cameras don't even use a shutter, they use rotating prisms and there's no intermittent motion, the film slides behind the rotating prism in one continuous motion. And that's why the effective exposure time is so short and why a lot of light must be used to illuminate the subject.

Since it was a sunset, it was going to get darker, so I set it to manual to zebra stripes 100, then didn't touch anything. Definitely easier than guessing what the brightest exposure will be 30 minutes after sunrise.

Sorry for so many questions, but what do the Zebra Stripes do for it exactly? My camera does support it, I checked, the manual doesn't really explain it, and what I Googled I don't "get".

Yours is done at a sunset, mine will be a sunrise. So would I still use the Zebra 100 setting on it?

And although I'm not where I'm going to film the sunrise at right now, I checked this morning and made a mental note of where the manual setting would be at. So I could use that?

Zebra stripes visually show you areas that are overexposed, or which exceed (too bright) some set level of exposure. The default (100 or 100+) is designed to show where you have totally "blown out" the highlights, and therefore where you will have zero detail, no matter what you try to do in post. The detail is lost forever. That's why the zebra functionality is pretty much mandatory.

Now that we're on the same wavelength we're now really into questions of artistic choices. As you say cars don't travel at 1,000mph and people don't walk at 100mph etc.
My general view of anything like this is to go with what's closest to reality, that's why I obsess over things like motion blur in CGI but as you've said in this case we're so far removed from reality we don't have much of a reference so it becomes an aesthetic choice. I really like the look when everything has the correct motion blur but if you look at some of the work on Vimeo where this is used you soon seem some of the oddities that produces. Cars blur at 1,000mph and then stop in one frame at the traffic lights. Our brains immediately get put off by that. On the other hand having the waves on a beach blur is pretty pleasant to look at visually compared to having them stutter.
You raised timelapse of a building construction. That's a case where you probably wouldn't want the natural MB. It'd be very handy in this case to be able to pause playback and have a look at what was happening at that point in time. The MB would kill that off.

Bob.

Dave, I really like your idea of speeding up the sky but keeping the surf at normal speed. This was a DUH moment for me - I'll use that in the future.

Jerry

Jerry, the mask could be better, a few surf issues at the horizon, but that seems to be the only issue.

I'm back from vacation, and here is my first time-lapse sunrise:

http://www.vimeo.com/1854300Gatlinburg, TN Sunrise[/link]

It's not perfect, but I appreciate all the suggestions last week on how to do this. I got to my destination a little later than expected this morning, got rushed, and didn't set my camera to manual exposure... go figure!

So if I would have set my exposure to manual, the beginning of the video would have started out darker than it looks here right?

It was actually pretty dark when I started, so I guess being an automatic exposure the camera made it look lighter?

Pretty good for auto, I thought the fog was very cool. Yes, it would have been darker had you used manual. With my New Orleans sunrise, there was lots of night time activity to see on the HD version, lights, cars, elevators.

Dumb question; When I video a sunset for, lets say,12 minutes, the best I can do tis to compress the 12 minutes into a 3 minute clip(which is way too slow). How do I convert a 12 minute clip to a smaller interval of time?

Simple answer, do it it mulitple passes.
Go from 12 minutes to 3 minutes, take the 3 minute AVI and go from that to 1 minute.

Bob.

Thanks Bob, not quite clear. I should render the 3 minute version, then, compress again?

"I should render the 3 minute version, then, compress again? "

Yes.

That has to work!

Bob.

Actually, there is a better way that will avoid the dual compression.

First, speed up the clip by as much as you can. And, BTW, you can speed it up by MORE than 4x. To do this, first Ctrl-Drag the trailing edge of the event to the left as far as it will go. This will give you the 4x speed-up that you have already got. But then, add a velocity envelope (Insert -> Video Envelopes -> Event Velocity. Drag this all the way to the top of the event and you now have a 3x4 = 12 times speed-up.

Pretty nice, eh?

But wait, as they say in the late night commercials, there's more.

You can certainly do what Bob suggests, namely to render the result of this speed up out to another file. I suggest using a DV AVI file for SD video or a Cineform AVI file for HD video. You then bring this file back into another instance of Vegas and speed it up again.

But there is a better way: nested veg files.

What I would do is save the project which has your 12x sped-up video. This will give you a VEG project file. All you have to do is place that on the timeline in another instance of Vegas and then speed THAT up by another 12x (or whatever additional speed increase you need). This avoids the time and potential quality loss of an intermediate render. I just did this and got a very nice 144x speed increase.