4.69
#...
·
 |
|---|
|
Can I just check my GHS technique with those of you who are more expert than I? I tend to do about three stretches. 1. place the cursor in the peak of the histo, send to the value to the SP, set b to 8, stretch until the peak is about 25% from the left, apply 2. place the cursor in the peak of the histo, send to the value to the SP, set b to 5, stretch until the peak is about 25% from the left, apply 3. place the cursor in the peak of the histo, send to the value to the SP, set b to 2, stretch until the peak is about 25% from the left, apply Obviously there is variation from image to image, but my basic question is whether the overall procedure of successive steps with gradually reducing values of b is appropriate. Thanks |
1.91
#...
·
·
1
like
|
|---|
|
Hi Stuart, Responding to your question just to subscribe to this topic. I started using GHS recently. Now I reached the stage when I really appreciate how little I know about its capabilities. As you say, each image will require different approach. But so far, I "discovered" my routine, which is slightly different from yours: First I set the SP bar at approximately bottom 1/3 of the left slope of the histogram (I am doing only monochrom images so far). Then I set intensity to 12 and apply strech, until the peak of the histogram moves approxinately to 15% point from the left. For the next stretch, I set the SP at around bottom 1/3 of the right slope of the histogram, set intensity to 6 and apply strech, till the peak approaches, but does not reach, 25% point. After that I put SP at around 0.3 point, reduce intensity to 3 and apply strech just enough, so the transofrmation curve starts to acuqire "S" shape (so the image gets a bit more contrast). Usually the strehc is no more than 0.5 at this stage. If necessary, I will do the same step, setting SP at 0.3, but reducing intencity to 1. Then apply stretch just a little so the histogram peak sits exactly at 25% (easier this way to match stretch on other monochrome channels). Sometimes I need to adjust Shadow Point as well, to achieve this position. But, I realize there must be much more precise scinetific approach, how to analyze each image and apply necessary streching by this tool. CS David |
6.80
#...
·
·
1
like
|
|---|
|
I am a new adopter of GHS, so my thoughts here may be a bit off, but I'll share in hopes that if I'm way off, someone can help me. To know where to start the stretch, you need to understand something about the general makeup of the stretched image. In galaxy images with a lot of background pixels, the histogram position for symmetry would definitely be right of peak (and sometimes quite a bit right). In a nebula image, the signal is quite often much more evenly spread in ratio between light/dark, so the histogram symmetry location of interest is nearer the peak (but usually still barely right of peak). I use the standard HT stretch as a first indicator to understanding the distribution of signal in the image and likely histogram location for where symmetry needs to be. I select that rough location, send to the symmetry parameter, and then use the slider to explore around that location. In the first stretch, a (8-10) B value is used with symmetry and D (2-6) in play. In the second stretch, a lower B (1-3) with lower D (1-3) often works, with symmetry being determined (and set) from any humps seen in the down sloping log view of the histogram. Exploring the slider with B & D after first stretch seems key. Keep the shadows control from pushing the histogram leftward (darkening the image) as you adjust D. I find I don't often need to use bright protection. To finish the stretch(s), iterate on beating down humps in log view until achieving a relatively straight downslope result (usually 3-4 total stretches works). I'm still learning, but this approach to learning GHS seems to work for me. |
4.62
#...
·
·
3
likes
|
|---|
| Personally my favorite use of GHS is in star stretching. After I separate the stars from the calibrated stack I'll almost always use GHS with a Color stretch, RGBBlend clipping at 0.8 (many times) I'll set the stretch to 7 and the local contrast to 3. Almost always a perfect result for me with only tiny bit of tweaking needed. |
