Annoyance at Halo.....

[Ollokot]

Hi Folks,
If you look at the rock in the right of the foreground in this image a halo is clearly visible,it suddenly appears when I apply Local Contrast Enhancement,an USM:20%,50px.
Would it be caused by the contrast already present between the rock,sky,and water when LCH is applied or is something else amiss.
Best Wishes,Pat.

[Colin Southern]

Hi Folks,
If you look at the rock in the right of the foreground in this image a halo is clearly visible,it suddenly appears when I apply Local Contrast Enhancement,an USM:20%,50px.
Would it be caused by the contrast already present between the rock,sky,and water when LCH is applied or is something else amiss.
Best Wishes,Pat.

Hi Pat,

A USM of 20% over a 50px radius is a BIG adjustment ... I'm not surprised it's showing. Personally I use more along the lines of 40% @ 4 pixels for content/creative sharpening. You could always just mask it out (on a duplicate layer) or use the history brush if you wanted though.

[Dave Humphries]

Or try 10% over 250px

[Ollokot]

Hi Colin,
Thank you for your reply,I think I could be a bit heavy handed in trying to get images as crisp as possible,with this image I was wondering why the halo hit just that part.With more practice and better inittial capture hopefully my photography will improve.
Thanks again for your input.
Best Wishes,Pat.

[Ollokot]

Thanks Dave,will try that,
Best Wishes Pat.

[FrankMi]

Hi Pat, if you are not shooting in RAW, check carefully (pixel peep) to see if the in-camera JPEG processing is adding a halo. If it is, then additional processing can enhance the existing halo significantly. Either way, it is far more likely that the excessive contrast settings are creating it as pointed out by Colin.

[revi]

Afaik, 20% @ 50 px is in between sharpening and LCE (local contrast enhancement)...
and 20% seems rather high for a 50px radius.

I'd suggest trying several radii and amounts to see the influence of the two parameters (the 10% @ 250 pixels, mentioned by Dave, is a fairly typical value for LCE). And yes, too high an amount gives you a halo, but there's a relation between radius and amount, in that a smaller radius supports a larger value for amount before you get a visible halo.

As for why the haloing only hits the right hand rock:
it doesn't, but it's there that it shows the most, probably due to the colour densities you have there (if you look closely you can also see haloing around the rock on the left)

Remco

[Dave Humphries]

For me 20% @ 50 px is LCE and I often use those, or quite similar figures, BUT they'll only be good for a fairly busy scene; one where there is no high contrast edges followed by smooth (ish) tones - just like the silhouetted rocks

I think of it like this: Basically what the LCE is doing is applying a 20% exposure difference spread across 50 pixels band around the edge - now that's going to be noticeable on a smooth(ish) tone, so: reduce the exposure difference (to 10%) and spread it across many more pixels (by increasing the radius - to the max of 250 when necessary) to avoid the problem.

Update:
Another thing to be wary of is that in-camera dynamic range enhancements (called Active D-Lighting on Nikon bodies, but something else on Canon, etc.) work in a very similar way to LCE with these values, a classic example is a swan against a smooth mid-tone river - the swan can get a dark 'halo' around it on the water's surface
Turn off that feature when necessary.

Cheers,

[FrankMi]

For me 20% @ 50 px is LC

When would it be better to use the USM to do LCE instead of using an Adjustment Layer, Brightness/Contrast, contrast slider?

What are the strengths/weaknesses of each?

[Dave Humphries]

When would it be better to use the USM to do LCE instead of using an Ajustment Layer, Brightness/Contrast, contrast slider?

That's easy; one I know how to do and I understand the process in my mind, so why do I need the other

What can I say; it works for me (for now).

Therefore I cannot answer the second part of the question, sorry.

Expanding a little; I have a theory that one can obtain, at pixel 'level', the same finished result by more than one method using Photoshop - e.g. to set white and black points one could do it with the Levels or Curves dialog (and possibly a layer and multiplication), whichever method/dialog you use is a personal choice and I have something that, for now, works for me. One day I may invest time to learn more about PS and do it another way.

(yes I am lazy)

Cheers,

[arith]

The rocks edges are to sharp. Soften it a bit and I'll hang it on my wall, if I could afford it.

[Ollokot]

Hi Folks,
I would have replied earlier but my ageing pc is giving me a lot of trouble especially when working with photoshop elements 7.
Here is the same image with a usm of 10%,250px and output sharpening of usm 80%,0.6px. The halo seems to be reduced and the overall image a bit softer,does it look better.

[arith]

Yep, it is better. I've got an ageing pc, I built it myself years ago and I can't even remember how I did it, except one step at a time. I expected it to last a year before I needed to upgrade bits of it, but it has been going over three years now and I haven't got any money for an upgrade.
A new one would cost at least £500 in parts not counting externals or software, but when this is gone, if it can't be repaired I'm stuffed.

Cheered me up just by saying ageing, nice photo though.

[Dave Humphries]

Here is the same image with a usm of 10%,250px and output sharpening of usm 80%,0.6px. The halo seems to be reduced and the overall image a bit softer,does it look better.

Hi Pat,

I can see some CA (Chromatic Abberation), those red/cyan-green edges, that needs to be dealt with first (before sharpening), then I'd use a maximum radius when final sharpening the reduced image of 0.3 or 0.4px.

I'd offer to have a go if you could get a RAW to me (I'd use Elements on it myself so you know it is do-able), but I won't be around until tomorrow afternoon/evening now - as I'm hoping to get out early for some 'atmospheric' photography in the morning.

Cheers,

[Dave Humphries]

Hi Pat,

Your second for comparison

Well, I have had a go at the RAW, but the file does seem very susceptible to a halo developing if local contrast is used, I have it too, despite different attempts to avoid it.



That pic took a lot of recovery of highlight detail (which I see you got back too), the over exposure probably added to the CA, I had to clone lots out towards both sides, but I left one bit in, on a vertical on the left and even after a downsize from 4688 to 700px, it is still visible.

Unfortnately, I didn't refer to your choice of colour temp while RAW and had to fix it later on.

As I have not achieved anything spectacularly better than your results, there seems little point going into the details, since you did ok by yourself. I will say I did the LCE at 250px radius after the downsize to 700px, so it really shouldn't have given a halo of a few tens of pixels wide like it has. Maybe a 'blended layer' approach would be better - I've never seen the point in learning before

Hope that helps, if only to confirm we both don't know the answer

Cheers,

[Ollokot]

Hi Dave,
Many,Many Thanks for your time and efforts which have indeed helped.I have learned a lot from this thread and once again have been enlightened here at CiC.
My Best Wishes to you and to all at CiC,

Pat,