timelapse.. the river, the mulberry tree and the lake

[iceman (Mike)]

Short exposure times make a huge difference.

My 40D is nowhere near as sensitive as the 5D Mark II, not to mention my fastest lens is only f/2.8 @ 17mm.

That means I need 30 second exposures @ ISO1600 or 3200, giving much noisier and grainier images than Phil will get.

Also the timelapses are much smoother with short exposures, not to mention no little trails either.
But he'll also take 3x as many photos to process

[Paul Haese]

Alex, I think this depends on whether you are using full frame or not. On my D3 with this lens I got seagulls at the corners and edges. No lines. With my stars over the murray shot I used F3.5 got lines in the edges but no seagulls like at 2.8. I figure the extra real estate makes for the extra distortions at wide open. Are you using yours on a full frame or crop?

Sorry Phil for the hi-jack.

[alexch (Alex)]

Originally Posted by Paul Haese

Alex, I think this depends on whether you are using full frame or not. On my D3 with this lens I got seagulls at the corners and edges. No lines. With my stars over the murray shot I used F3.5 got lines in the edges but no seagulls like at 2.8. I figure the extra real estate makes for the extra distortions at wide open. Are you using yours on a full frame or crop?

Sorry Phil for the hi-jack.

Full frame Nikon D700.

Attached are two unedited top corners of the full frame from a recent image at 14mm, f/2.8, 30 sec, ISO32000. Note that in "Corner 1" the stars are more elongated because it is further away from SCP.

Cheers,
Alex

[Phil Hart]

Originally Posted by Paul Haese

Sorry Phil for the hi-jack.

On the contrary, this is very much on topic! You've got straight to one of the thorniest issues for timelapse and it's a very real trade-off (and some very pricey lens decisions as well!).

there's a couple of reviews that suggest the 24mm mkII may do even better where the mkI already does pretty well considering how fast it is:

http://www.the-digital-picture.com/R...ns-Review.aspx

http://www.modeemi.fi/~leopold/Photo/CanonEF24mmF14L2/

Particularly the next time I do a moonlit sequence, i might try stopping down to f2.5 as the sacrifice may be worth it, so thanks for prompting me to think more about it

Phil

[alexch (Alex)]

Originally Posted by philiphart

On the contrary, this is very much on topic! You've got straight to one of the thorniest issues for timelapse and it's a very real trade-off (and some very pricey lens decisions as well!).

there's a couple of reviews that suggest the 24mm mkII may do even better where the mkI already does pretty well considering how fast it is:

Indeed, I am always on the lookout for a coma-free fast (f/1.4) wideangle lens, but it seems none exist at the moment.

Phil, you might find this review useful, lenstip is the only site that tests coma:
http://www.lenstip.com/245.7-Lens_re...tigmatism.html

They don't have a review for V1 of this lens to compare though.

[Paul Haese]

Hmmm, that is interesting. By way of example I have attached two sections of one of my images. Taken with D3 and 14-24 f2.8. There is some field curvature (PA related) but the seagulling is really evident. When I stopped this down I found I got much better results. MInd you this is 5 minutes on a guided mount. Some of my more recent images I have just the lines such as you do Alex.

I am assuming shorter subs even on a guided mount will show less of this particular problem. It is quite odd that your images wide open show no seagulling. I may have to investigate this a little more. Perhaps my lens needs some adjustment work done to it.

[alexch (Alex)]

Originally Posted by Paul Haese

Hmmm, that is interesting. By way of example I have attached two sections of one of my images. Taken with D3 and 14-24 f2.8. There is some field curvature (PA related) but the seagulling is really evident.

What focal length was that? I get similar seagulls at 24mm but not at 14mm.

[stevous67 (Steve M)]

What nice experience to watch this video, thanks Phil.

Steve

[Bassnut (Fred)]

Originally Posted by iceman

Short exposure times make a huge difference.
Also the timelapses are much smoother with short exposures, not to mention no little trails either.

Not necessarily, well number of exposures for smoothness anyway. After a huge amount of investigation and testing, I found a way of extrapolating frames in After effects to smooth timelapse motion after being plauged by stutter.

Here is a short test time lapse of the Milky way of only 150 frames extrapolated to 510 frames at 30fps. The exposure times were 20 secs with a 30 sec delay which was very stuttery in real time but rendered quite smoothy (scuse the low quality).

Trails are still evident of course due to the exposure time, but I now think very long smooth timelapse is possible with relatively few (short) exposures.