Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - topboxman

Pages: [1] 2 3 ... 8
1
Just to be on the same page as you, do you actually combine LRGB right after stacking LRGB separately but before any kind of post processing? In other words do you combine LRGB while images are still linear (unstretched)? If so, I have seen several tutorials suggesting it's better to process L and RGB separately and do the LRGB combination later while LRGB components are non-linear. PixInsight tutorial says LRGB combination requires non-linear image components. Processing L and RGB should be as similiar as possible like same DBE samples and similar Histogram Transformation.

For DBE, personally I always run DBE after RGB combination instead running DBE on individual RGB components. PixInsight tutorial showed running DBE after RGB combination. I have seen others prefer running DBE on separate RGB components. Maybe I should try running DBE on separate RGB components to compare.

Peter

2
There's nothing wrong with calibrating with another software. I use Nebulosity for calibration because it's quicker and cannot find any differences with PixInsight calibration. Saving Nebulosity's calibrated files in 32 bit floating point works fine with PixInsight. From then on it's all PixInsight processing starting with Star Alignment, Image Integration, RGB combination, and post processing.

Nebulosity cannot read PixInsight's FIT files. Unfortunately there don't seem to be a standard protocol for FIT.

Peter

3
Hi Mike,

I find SilverCup's method very good. I was just curious about HaRVB-AIP script that's part of RC5. All I am asking is which is the preferred method and how to use HaRVB-AIP script. My understanding from other posts that SilverCup's "Adding Ha to RGB" script was supposed to be part of version 1.8 but I cannot find it.

After I captured LRGB and Ha of M82 from my light polluted backyard, I tried many different processing methods to get the best results. Ha captured the red star burst at the core very nicely but Luminance did not. When I applied Luminance to (HaR)GB, it made it worse because it removed most of the colorful red star burst at the core. So I ended up extracting the Luminance from (HaR)GB and processed this Luminance as a pseudo-L and then combined psuedo-L with (HaR)GB and got pretty decent result.

I am too lazy to go to a dark site.

Peter

4
No one capture Ha and adds Ha to RGB images? I find this method much better than LRGB under heavy light pollution. Capturing images with Luminance under heavy light pollution is extremely difficult. Ha filter does an awesome job of blocking light pollution and I can image at 30 minutes sub-exposures.

Here is a sample of my most recent image of M82 Cigar Galaxy using Ha and RGB. This uses SilverCup's script for adding Ha to RGB.

http://peternagy.smugmug.com/Telescopes/Galaxies/15823423_dS2M3V#!i=2430678873&k=TsVxrFw&lb=1&s=A

Peter

5
General / Adding Ha to RGB: SilverCup's method or HaRVB-AIP?
« on: 2013 March 30 00:06:48 »
Which "Adding Ha to RGB" script is better?

SilverCup or HaRVB-AIP found in RC5?

I have used SilverCup's script with pretty good success in version 1.7. I had a heck of a time putting this script in RC5 but it works.

Anyone can give tips on how to use HaRVB-AIP script?

I do not understand the following HaRVB-AIP parameters:

Ha % in R Channel: defaults to 33%
R % in Ha Channel: defaults to 33%
Noise It: defaults to 3

If I use defaults of 33% for both, it looks strange.

If I use 100% for both, it looks better and closer to SilverCup's script.


So, which script do you like better.

Thanks,
Peter

6
General / Re: Strange issue with HDR
« on: 2013 March 28 01:56:31 »
It's happened again. This time it was combining Luminance with RGB. When I apply Luminance image to RGB image, the whole image got dim. I had to run "Reset PixInsight Settings" again. Now it's fine but for how long?

Anyone else seen this before? It seems that PixInsight settings are getting corrupted. My machine is Windows 7 64 bits. I have been using PixInsight for a long time and never had issues till now. I am still using version 1.7.

I didn't want to migrate to version 1.8 until all the bugs are ironed out. I wasn't interested in being an early adopter for a new and untested version.

Peter

7
General / Re: Help with flats
« on: 2013 March 27 17:15:31 »
I find it best to expose the flats until the peak is at or near the center of histogram. I used to pay attention to max value but it can be mis-leading because the max value could come from hot pixel rather than useful pixel from each flat.

Peter

8
General / Strange issue with HDR
« on: 2013 March 27 14:27:11 »
I had a strange experience with PixInsight 1.7 when I was processing an image. When I was using HDR tool, it would remove or strip out lots of data from my image. Also, after using HDR, I extracted Luminance from color image for Luminance mask, the Luminance background was horribly noisy and grey/white probably because of bad HDR processing.

So I ran "Reset PixInsight settings" and everything is back to normal. Is this a known issue?

Thanks,
Peter


9
Gallery / Re: IC405/410/417
« on: 2013 February 12 21:31:37 »
Hi Nigel,

This is very nice.

Did you know there is a script based on Harry's Adding Ha to a Galaxy tutorial?

http://pixinsight.com/forum/index.php?topic=3446.60

Look at third post of page #5 and download "HaRGBCombinationV1.4.zip".

It allows you to quickly input numbers and preview the results. It's very handy.

I am not sure if this script works with version 1.8 but works with version 1.7. Read on this thread and it discusses about the script and version 1.8.

Peter

10
Image Processing Challenges / Re: NGC2359
« on: 2013 February 03 19:57:31 »
Hi Julian,

Unsharp Mask can be useful, just try not to over kill it.

Peter

11
Image Processing Challenges / Re: NGC2359
« on: 2013 February 03 17:09:26 »
Hi Julian,

What are the exposure times for each channel? Please tell us the sub-exposures and total number of subs for each channel.

My processes are based on Harry's video tutorial ( http://www.harrysastroshed.com/pixinsighthome.html ) and PixInsight video tutorial ( http://www.pixinsight.com/videos/NGC1808LRGB-vperis/en.html ). I will try my best to explain what I did. My processing skill level is beginner/intermediate.

1) Combine Sii/Ha/Oiii using LRGB Combination. Uncheck "L" box.
2) Crop the image to remove black borders using Dynamic Crop.
3) DBE. This one is tricky because nebulosity was all over the image. It was hard to find background not covered by nebulosity. I had to make lots of guesses. I am not sure if this is a good way. The image looks better if you un-link RGB channel in STF after clicking on "A" (autostretch). But I linked RGB channel and it was easier to find darkest areas this way. I don't think this is the best way to do it. You could try Automatic Background Extractor (ABE) and use default values.
4) Background Neutralization. Create a small preview at area with little to no nebulosity coverage.
5) Color Calibration. Use same preview as #4 for Background reference. Create a new preview at Thor's helmet area for White reference. Uncheck Struture Detection. There are several ways to do Color Calibration, this is the most common way.
6) Multi Scale Median Transform (MMT) to reduce noise while image is still linear. Use Harry's tutorial at http://www.harrysastroshed.com/at%20noise.html . The tutorial uses Atrous Wavelet but it's very close to MMT. Watch it and you will get the idea.
7) Histogram Transformation. I use iterative process. I start with middle slider and move to left, apply to image. Then use left slider and move it to right (don't clip it!!!!) and apply to image. Repeat the process until it looks good. Don't get too greedy and over stretch too much.
8 ) Create star mask for HDR and Morphological Transformation. Increase threshold to remove noise, you only want to see stars. Sometimes you need to increase to 0.500. Also you may need to increase the scale if big stars don't show up on the map. Apply star mask to image.
9) HDR. Make sure that you invert the mask so that the stars are red in order to protect the stars. Test HDR with several different layers for HDR. I think 4 or 5 layers are good. Either use Linear Interpolation or B3 Spline (5) for Scaling functions.
10) Invert the mask again to protect everything except the stars. It's up to you to make the stars smaller using Morphological Transformation. I use Circular Structure, 2 iterations and adjust the amount to between 0.4 and 0.8. The less amount the less reduction.
11) Remove star mask. Extract Luminance and apply Luminance mask to image. Run Local Histogram Equalization. Adjust Contrast Limit to your taste. Play with between 0.2 and 2.5. I used 2.0. Remove Luminance mask.
12) ACDNR. I don't adjust Luminance mask for this. For Lightness, I check Lightness Mask, Std Dev. to 2.0. For Chrominance, Lightness Mask is checked and Std Dev. to 3.0. Apply to image.
13) Curves Transformation using "S". Extract Luminance and apply Luminance mask to image. Use preview. Click and drag at the center to upper left direction until it's to your taste. Apply to image. Remove Luminance mask.
14) Curves Transformation using "RGB/K". Create a "S" shaped curve. It's usually very small adjustment.
15) If the image looks too soft, you could try Unsharp Mask. I just discovered this recently. Play with Std Dev and Amount. Go very easy on this. I think it works better if you apply Luminance mask first. I normally do not use this tool because they can look too fakey.

I hope this is clear. I follow the video tutorial and not necessarily understand why it works. As time goes on, you will slowly learn why they work.

Attached is an updated image using Unsharp mask.

Peter

12
Image Processing Challenges / Re: NGC2359
« on: 2013 February 03 03:48:52 »
Here is my take. Very nice data. I not not well experienced with NB processing.

Thanks for sharing the data.

Peter

13
General / Re: Stretching in PI
« on: 2013 January 04 21:21:37 »
I understand and thanks for the explanation.

I guess this is a personal preference. I still prefer to manually stretch using Histogram Transformation. I use it iteratively by adjusting and applying HT to the image until it looks good.

Peter

14
General / Re: Stretching in PI
« on: 2013 January 04 18:29:15 »
What if L and RGB were exposed differently that if you stretch the same way, it may end up looking bad? I simply use HT for L and RGB until both look great regardless of whether they were stretched equally or not. I was not aware that you need to stretch both L and RGB near equally before combination. I thought it depends on the original content of L and RGB.

Peter

15
General / Re: Stretching in PI
« on: 2013 January 04 16:27:53 »
I never transfer STF to histogram. I only use STF just for viewing purpose to see what the linear image would look like when stretched. After I am done with processing linear data, then I use Histogram Transformation. I run HT iteratively until I am satisfied.

I find it somewhat difficult to control STF to transfer from STF to HT. It there an advantage of transferring STF to HT?

Peter

Pages: [1] 2 3 ... 8