From: Milton Aupperle <milton@outcastsoft.com>

Date: June 11, 2010 11:07:11 AM MDT

To: Astro_IIDC@yahoogroups.com

Subject: Re: [Astro_IIDC] Re: NGC 4565 Gradient removal

Hi John;

On 11-Jun-10, at 9:59 AM, John wrote:

Hi Milton,

Very nicely done.

Thanks. I reprocessed most of my recent DSO images and will be replacing the old ones soon with the new ones later today.

The reprocessed image looks much better than the previous version. I am interested to see the nature of the original gradient. Do you have a gradient challenged "before" to go with your after?

The darker image was the gradient challenged "before" image. I had to darken it to hide as much of the gradient in the background as I could. I don't have a copy of the LRGB images without gradient removal to show how much it was. But if you use Apple's "DigitalColor Meter" tool, you can see that red green and blue can vary from 1800 to 100 across the image in the before shot. One corner may have red at 1800 with green blue less than 300, whereas other corners may be low in red and high in green. The new image has a consistent RGB ratio across the image.

If your imaging the same exposures, same sky conditions and similar targets you can pretty much use the same settings over and over without having any issues. But to find the right levels requires a bit of trial and error. I just run 2 to 4 stacks of the same movie varying the gradient by 10% each time to get it dialed in. And it is a "trial and error thing" because it depends on how much curve stretching your doing. You basically stack your frames with the cutoff, then stretch the image as appropriate to get the right levels of bright / dark and then measure the edges to see how level it is. The more you push the processing, the more likely

You have to be very careful with gradient removal, especially for things like large nebula. It's very easy to remove the nebula while balancing the image :) I shot the cone nebula under hazy conditions when it wasn't dark enough (last attempt before it drifts too far west and is blocked) and using anything more than 10% gradient removal would remove a large portion of the nebula.

The only reason I need to do this is because my LRGB images are all different exposures and at different binning too. To get the colors balanced properly initially, I need to vary my exposure times, and that

OT:

Tonight it's supposed to clear off with low turbulence (and stop raining / blowing / hailing / snowing) for the first time in a 10 days, which is really fortuitous timing. The Exoplant "TrES-3b" (http://en.wikipedia.org/wiki/TrES-3b ) orbiting Mag 12.4 star USNO J1752071+373244 in Hercules is transiting between 2:12 am and 3:29 am MST tonight. That happens to fit right into my darkness "window of opportunity "which is about 2.5 hours now and also within the viewing area I can see (70 degrees east to 230 degrees west). Its a maximum dip of 0.029 magnitudes over 77 minutes, but if the turbulence is low enough, I likely can detect it. I have other opportunities on June 15, June 28, July 1st and July 15th too.

TTYL..

Milton Aupperle

Thanks,

John B

--- In Astro_IIDC@yahoogroups.com, "milton_aupperle" <milton@...> wrote:

I've been reprocessing a bunch of DSO images to test out the new Gradient removal code, and also improve them.

As an example, here is my original version of NGC 4565:

http://www.outcastsoft.com/AstroImages/NGC4565_MJA_20100307.jpg

Because of the gradient, I lowered the background to hide it, which darkens everything out.

Here is the new version which has gradient removal applied:

http://groups.yahoo.com/group/Astro_IIDC/files/DSO/NGC4565_20100306.jpg

Because this suppressed the gradient to minimal amounts, I could stretch the image a lot further.