Examples of effect of rotation variables

Discussion forum for Tawbaware's Star Tracer and Image Stacker software
Post Reply
dsjtecserv
Posts: 551
Joined: Sat May 05, 2007 3:25 pm
Location: Northern Virginia
Contact:

Examples of effect of rotation variables

Post by dsjtecserv » Fri May 21, 2010 9:51 pm

I'm also trying to learn how to use the controls in Star Tracer, but I think this will help illustrate and explain the effect of the "Seconds" and "Count" settings. Star Tracer can only work with the original image, and that is what is being rotated to create the extended or filled in trails. Thus, if you are trying to extend the small arc that you get from a relatively short exposure, the "Seconds" field should not be any longer than the original exposure; in fact, it should be a little shorter so that the rotated trails overlap a bit. Otherwise they will look like a nice string of sausage links, or be completely separated from each other.

The Count indicates the number or rotations (in each direction) indicates how many copies will be rotated and overlaid. If the Seconds is close to the number of seconds of the original exposure, the star trail arcs will be extend an amount that roughly corresponds to the Count -- if the count is 10, the trails will be about 10 times as long. But since there normally will be some overlap, the actual length of the combined trails will be something less than the original trail length times the Count.

Here is the original image, which had an exposure of 241 seconds.

Image

This one had the rotation specified as 60,5,-60,5. In other words each is rotated the angular equivalent of 60 seconds from the previous position, and this is repeated 5 times in each direction.

Image

I held the 60 seconds the same, but increased the number of rotations (Count) to 10.

Image

...Count of 30...

Image

...and finally, just for kicks, a Count of 100:

Image

These were created from an original to which I had applied a rather sloppy alpha mask, and this becomes obvious in the last two, when bits and pieces of things that aren't stars were replicated many times. Good masking technique is clearly going to be a useful skill when using Star Tracer.

Also note the black "gores" at the edges, which apparently results from rotation of the mask at the edges into the image itself. I think this is related to the fact that I had Correct Distortion left on in the Options menu. With distortion correction turned off, the masked edges go away.

There is obviously still a lot to learn and a lot of options to try out, but I think this shows how the rotation options work.

Dave

luigi
Posts: 21
Joined: Mon Apr 06, 2009 6:34 pm

Post by luigi » Fri May 21, 2010 10:15 pm

Thanks now I got it :)

Luis

dsjtecserv
Posts: 551
Joined: Sat May 05, 2007 3:25 pm
Location: Northern Virginia
Contact:

Follow-up: Effect of "Seconds"

Post by dsjtecserv » Fri May 21, 2010 10:21 pm

In the first series I held the Seconds constant at 60. Since the original image was shot at 241 seconds, it took about four rotations before the rotated images began to actually extend the trails, and each extension after that wasn't as long as it could have been. In this series I changed the Seconds to 200, to be less than the original exposure, to avoid creating any gaps. Note that it takes much fewer Counts to create the same or longer extension of the trails, compared to the first series.

Rotation of 200, 10, -200, 10:

Image

Rotation of 200, 30, -200, 30:

Image

Rotation of 200, 100, -200, 100:

Image

On the last one, the arc of the trails is nearly 180 degrees. That would be a long night's journey into day for a real exposure!

Also note that I turned off distortion correction, which eliminated the masked edges. Now to practice cleaning up my alpha mask so that I don't rotate other garbage into the picture.

Dave

luigi
Posts: 21
Joined: Mon Apr 06, 2009 6:34 pm

Post by luigi » Sat May 22, 2010 12:23 am

The new Refine Edge function in CS5 is good to mask well, anyway yours is a difficult image to mask.

dsjtecserv
Posts: 551
Joined: Sat May 05, 2007 3:25 pm
Location: Northern Virginia
Contact:

Post by dsjtecserv » Sun May 23, 2010 2:04 pm

The alpha mask needs to be absolutely black or white. Any intermediate density in the mask results in unpredictable artifacts of the mask itself getting placed into non-masked parts of the image. I learned that by experience!

So far I'm finding it best to manually paint a mask into a layer mask on a duplicate layer of the image. Refine Edge doesn't help much because most of its functions end up feathering the edge in some way. The mask also help in doing fairly aggressive nose reduction on just the sky itself, to minimize Star Trails rotating the noise.

After the mask is done I Load Selection and use that to create the new alpha mask. Just to be sure about the absolute black/white thing, I do a threshold adjustment on the alpha mask. I then flatten the layers and save the file.

Dave

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest