Processing without losing precision.

[sheeprug]

Thus far I've been using DSS and photoshop without really thinking too much about what happens to the data.  However, with the current rubbish weather it seems a good time to think a bit more about what's happening.  My processing chain thus far has been:

1. Stack images using DSS. 
2. Open the autosave001.tif file in photoshop
3. Stretch using the levels tool.
4. Convert to 16bit mode using the local adaptation method - default settings. 
5. Adjust contrast using the curves tool until it looks good.
6. Bump up the saturation until it looks better.

Now since I start off with a 14bit sensor (Canon 750),  and each time I double the number of frames I increase the precision by 1 bit, if for example I stack 16 frames the dynamic range is increased by 4 bits so I have a maximum of 18 bits/pixel  (though obviously it might not all contain good data).   So the question is therefore,  do i  need modify my processing to ensure I'm not throwing away precision?  And if so how?      

Best, SR

[Galen Gilmore]

When converting to 16 bits I use the exposure and gamma setting, I don't know how much difference that makes though.

[ollypenrice]

Stage three is critical with a vast range of refinements available. If you stretch in Levels by moving the mid point you'll get a standard logarithmic stretch which will give a fairly natural result and avoid catastrophes - but it will certainly not give you the best possible image. This is a very big subject but things you might look up on the net would include...

- Digital Development Process (a 'bought in' sophisticated stretch which, personally, I don't use.)

- Star masking.

- Inverse masking.

- Layer masking (of two different stretches using the same data or of shorter and longer exposures.)

Stage Five I'd avoid the contrast tool. And if you're going to use Curves (and you should!) don't play with it randomly. Be sure to think through which parts of the curve are affecting which parts of the image. Understand the tool before using it.

No global processing - or at least very little! This is a crucial insight. Different parts of the image need different kinds of processing. Photoshop layers (or horrible masks in unfathomable Pixinsight ) allow you to carry out diametrically opposed processing techniques to different parts of the image. You can noise reduce areas of faint signal, sharpen areas of strong signal, isolate stars, etc etc. Learn about the Ps selection tools (Colour Select is very useful) and the use of Layers and the eraser to keep or delete changes between layers.

Steve Richards' new book (on FLO) would be a good place to start.

Olly

[sheeprug]

Thanks Olly, 

Some useful insights there.   I think for starters I'm going to have to get a better understanding of how Photoshop layers work. I can do simple stuff, but there's clearly a whole lot more. I wasn't aware of Steves new book - I'll take a look.

SR

[wimvb]

On 2018-02-27 at 21:57, ollypenrice said:

- Star masking.

- Inverse masking.

- Layer masking (of two different stretches using the same data or of shorter and longer exposures.)

Photoshop layers (or horrible masks in unfathomable Pixinsight )

Hmm, layers AND masks. What should we think of that? PixInsight is soo much easier; just masks to deal with. No double processing needed.  

Joking aside, different parts of an image do need to be treated in different ways, however you tackle that. But it is worth spending time on processing, rather than rush to get the image done. The good thing is that the data doesn't go anywhere, so you'll have plenty of time to experiment. Since, unfortunately, the clouds don't go anywhere either.

On 2018-02-27 at 21:57, ollypenrice said:

Steve Richards' new book (on FLO) would be a good place to start.

If it's anywhere near the quality of his "old" book, I couldn't agree more.

[ollypenrice]

Just now, wimvb said:

Hmm, layers AND masks. What should we think of that? PixInsight is soo much easier; just masks to deal with. No double processing needed.  

Joking aside, different parts of an image do need to be treated in different ways, however you tackle that. But it is worth spending time on processing, rather than rush to get the image done. The good thing is that the data doesn't go anywhere, so you'll have plenty of time to experiment. Since, unfortunately, the clouds don't go anywhere either.

If it's anywhere near the quality of his "old" book, I couldn't agree more.

It is. I'm working my way through it at the moment. I do some things differently and have the odd preference here and there for alternative approaches but that's entirely normal in processing and stops us all from making the same pictures. Given the minefield of appalling 'How to' processing videos on the net, some of which make your eyes water, this book is much needed.

Olly

[Stub Mandrel]

I've surrendered and order Steve's book :-)

It's the hint that it explains approaches using advanced masking techniques that appeals. Anything to wind up the PI users

[Rush]

Can you upload an Image which was done in PS as mentioned above.

Just for curiosty and if its ok with U.  Maybe there R ways to improve and to share with others who r not rich in processing.

Was once a PS junky , Now im clean:). Different Softwares have different precision and advantage.

So lets have look at it. its going to be fun anyway.

CS

Rush