Author Topic: histogram stretch using curve from STF  (Read 4373 times)

Offline yock1960

  • Newcomer
  • Posts: 29
histogram stretch using curve from STF
« on: 2011 April 22 16:35:25 »
Back when I was just a babe in processing, before I had even heard of PI, I read and watched some of Warren Keller's tutorials on image processing where small levels & curves adjustments were made in PS to strectch the image. Recently, I saw a method to stretch using HistogramTransformation by doing a screen stretch with STF and then dragging & dropping it to HistogramTransform. It's neat and fast, you can use the sliders to quickly try and find the spot you think looks best. But are there any drawbacks to using this method? It's like it's too good to be true.

Steve

Offline Philip de Louraille

  • PixInsight Addict
  • ***
  • Posts: 289
Re: histogram stretch using curve from STF
« Reply #1 on: 2011 April 22 16:55:35 »
I would say that as long as you keep the original data, nothing is too good to be true.
Seriously... if your goal is astrophotos, no harm done but stretching will also stretch the noise so you have to calibrate and clean your data carefully if you really want to be pleased.
If your goal is scientific study, then stretching removes the linearity of your data and then you'll have to be careful with how you interpret that data even more so if you are going to compare different images.
And yes, doing an STF and then passing the coefficients from the STF to the HistogramTransformation function in one move is what makes PI outstanding.
Philip de Louraille

Offline sreilly

  • PixInsight Padawan
  • ****
  • Posts: 791
    • Imaging at Dogwood Ridge Observatory
Re: histogram stretch using curve from STF
« Reply #2 on: 2011 April 22 17:00:06 »
Is this referring to the process that Vicent showed us at the NEAIC workshop?

Steve
Steve
www.astral-imaging.com
AP1200
OGS 12.5" RC
Tak FSQ-106ED
ST10XME/CFW8/AO8
STL-11000M/FW8/AO-L
Pyxis 3" Rotator
Baader LRGBHa Filters
PixInsight/MaxIm/ACP/Registar/Mira AP/PS CS5

Offline Niall Saunders

  • PTeam Member
  • PixInsight Jedi Knight
  • *****
  • Posts: 1456
  • We have cookies? Where ?
Re: histogram stretch using curve from STF
« Reply #3 on: 2011 April 24 07:40:20 »
Hi Steve,

No, it isn't 'too good to be true' - it is just 'a method', just like all the other processing methods that are available.

What you must understand are the implications of 'applying' the methid, and how the application of the method will affect your image data thereafter.

Using STF and HistogramTRansforms 'together' is what I do all the time - but ONLY because the 'alternative' is to use Histo only, and to constantly repeat a cycle of 'slider adjust' and 'apply process' steps to a preview of the image. Tweaking the sliders in STF, and seeing the result 'in real time' is far more intuitive, and certainly gets me 'closer' to the result I am after 'quicker'.

However, even after I have transferred the STF settings to the Histo process, I will stop and see what the Histo window is telling me - especially when it comes to the number of pixels being 'clipped'. Sometimes the most favourable results can involve passing the parameter setup back and forth between STF and Histo, before finally 'applying' the Histo Transform 'permanently'.

At least the two tools offer you various methods, and a lot of flexibility as well. In fact, the level of flexibility can often mean that there is little reason to resort to Curves to try and achieve the same results - which is something you could not say about other programs. (I usually only need to turn to Curves to tweak overall Saturation or to globally adjust colour, i.e. Hue)

As others have said, just be aware of 'how' a particular process is going to affect your image data, and keep your adjustments within your acceptable limits.

Finally, don't forget to 'zoom in' when you are making STF slider adjustments - and also consider stretching the STF window to the full width of your Monitor, both of which give you much finer control of the end result. Similarly, decide when you want the R, G and B sliders to be 'locked' to each other, and when to 'unlock' them - again, for much greater control (for example, when I make an initial 'AutoSTF', I try it both 'locked' and 'unlocked' - 'unlocked' can usually help balance severe colour cast issues right from the outset, although this is no substitute for a careful DBE, BackgroundNeutralisation and ColourCalibration sequence).
Cheers,
Niall Saunders
Clinterty Observatories
Aberdeen, UK

Altair Astro GSO 10" f/8 Ritchey Chrétien CF OTA on EQ8 mount with homebrew 3D Balance and Pier
Moonfish ED80 APO & Celestron Omni XLT 120
QHY10 CCD & QHY5L-II Colour
9mm TS-OAG and Meade DSI-IIC

Offline yock1960

  • Newcomer
  • Posts: 29
Re: histogram stretch using curve from STF
« Reply #4 on: 2011 April 25 08:10:46 »
Thanks for the replies!

Philip - mostly I'm a 'pretty' picture maker. I try to make it as accurate (whole other subject there  :) ) as possible....to the best of my ability at least.

Steve - maybe....I'm pretty sure I saw this on one of Harry's later videos.

Niall - It's certainly fast! I guess what everyone is saying is kind of what I was thinking, I just was wondering if there was something I wasn't considering.

PI continues to impress, even after almost 2 years now!

Steve