This may or may not be a Hover Fly. But it surely is colourful

[JBW]

ISO 100, F/8, Aperture priority, 55mm, 1/40.
This might be as much as I can get out of my setup.

[Shadowman]

Nice.

[JBW]

Nice.

thanks I worked at this one. I must have tried 300+ different shots to get this one.

[Dave Humphries]

This might be as much as I can get out of my setup.

Not quite Brian, not quite.

If you can apply some Local Contrast Enhancement* just to the fly (or parts of it), this could 'pop' a little more.
I would demonstrate, but have a meeting this morning to attend - perhaps later.

* USM (UnSharp Mask) try: Amount 25%, Radius 100px, Threshold 0


With regard to the lens, you may be correct.

It would be helpful (to me at least) to see the original as captured full frame, so I (we) can see how much enlargement by cropping you are doing.

Cheers, Dave

[rudi]

Good shot Brian,
but it appears to me that the focus is a bit behind the eyes (happens to me very often), or it can be my older eyes .

[JBW]

Not quite Brian, not quite.

If you can apply some Local Contrast Enhancement* just to the fly (or parts of it), this could 'pop' a little more.
I would demonstrate, but have a meeting this morning to attend - perhaps later.

* USM (UnSharp Mask) try: Amount 25%, Radius 100px, Threshold 0


With regard to the lens, you may be correct.

It would be helpful (to me at least) to see the original as captured full frame, so I (we) can see how much enlargement by cropping you are doing.

Cheers, Dave

It was give or take 5500px on the long side and that would take me a long time to upload. I did use local sharpening on the fly and ended up with an unsharpen mask amount 40, radius 50 threshhold was up around 30 if memory serves.

Here is a screen shot you and others may find helpful

[JBW]

Good shot Brian,
but it appears to me that the focus is a bit behind the eyes (happens to me very often), or it can be my older eyes .

your eyes are fine it is a bit behind but the back is very nice?

[JBW]

Not quite Brian, not quite.

If you can apply some Local Contrast Enhancement* just to the fly (or parts of it), this could 'pop' a little more.
I would demonstrate, but have a meeting this morning to attend - perhaps later.

* USM (UnSharp Mask) try: Amount 25%, Radius 100px, Threshold 0


With regard to the lens, you may be correct.

It would be helpful (to me at least) to see the original as captured full frame, so I (we) can see how much enlargement by cropping you are doing.

Cheers, Dave

Hi Dave I tried the unsharpen mask in Gimp on the fly only from 10 to 100 and to me it certainly made the fly pop but it also make it too hard? If you have the tine and can work it to your liking I am curious what the difference will be.

[Dave Humphries]

Hi Brian,

Apologies in advance - there's a lot of good stuff here I have learnt the hard way and if you or anyone liked my butterfly pictures, I must being doing it right - oh dear, that sounds arrogant, but isn't meant to.

I did use local sharpening on the fly and ended up with an unsharpen mask amount 40, radius 50 threshhold was up around 30 if memory serves.

Those USM settings are not any I'd ever use.

I learned that for sharpening, a
high Amount (e.g. 100), combined with
low Radius (e.g. 0.3-0.4 for output sharpening and no more than 10 for process sharpening), combined with a
low Threshold that just avoids sharpening noise, would be the are to go.

If the edges or fine detail looks to 'crispy', reduce the Amount.
If there are visible halos, you have the radius too wide, so reduce that.
If noise is being sharpened, increase the Threshold, but I'd never go above 10 - if the noise is that bad, it needs dealing with* before ANY sharpening is applied.
* Using a third party system such as Neat Image or Topaz DeNoise, then sharpen afterwards, preferably, if image is for on line viewing only, do not sharpen before the downsize to show here (or on your blog) - because downsizing, ideally by a factor** of 2 or more, will reduce noise (by averaging), leaving less to be picked up by the output sharpening.


Whereas for Local Contrast Enhancement (LCE), the settings should be
low Amount (e.g. 10 - 30), combined with
high Radius (e.g. 80 - 200), combined with
zero Threshold (e.g. 0).

Use a lower Radius and/or Amount to reduce the 'pop' effect.
There's no point in having Threshold any higher than zero because (with such a large Radius) LCE will not latch on to noise, all it might do is mess up what you're trying to achieve.

What you did is a mix of the two 'systems' and I'm not surprised it didn't give a good result, especially if you then applied what I suggested 'on top'.


** Downsize factor (for on-line display, not printing!)
I always aim to downsize by a factor of 2 or more (so the image looks 'sharp enough') - occasionally, as with your shot here, this isn't possible.
e.g. So ideally; if I intend to display here at 950px image height, I would ensure that any compositional cropping left an image sized at least twice that; e.g. 1900px high, but often it is more, up to 4,000 with my 24MP sensor (6000w x 4000h) if I don't crop.

As I say, it sometimes isn't possible, the bird or bug is just too far away for some reason, but the shot is worth showing anyway.

I note from your sensor width of roughly 5500, after cropping, you had an image 1218px wide, which you then downsized to 1000 wide for display - this small downsize factor, plus the weird sharpening settings go a fair way to explain why this shot looks less good.

If it had been mine; from the 1218px wide cropped image, I'd have downsized to say 700px wide (to fit thread), or perhaps even exactly by two (e.g. to 609px), then done the USM at 100% Amount, 0.3px Radius and Threshold 1 or 0 (if noise allows). Yes, the image would be smaller, but people would just say "that's nice and fairly sharp Brian, shame it isn't bigger" and you could say "Yes, I am hoping to get a Macro lens soon to allow me to shoot these critters much closer"

The downsizing (by at least 2x) and output sharpening also help improve Depth of Field perception, so more of fly would appear sharp.

Also, FWIW, I almost never input or process sharpen (apart from LCE, which I almost always do), I rely on the output sharpening only.

The only exception to this is if I have a seriously soft image that needs rescuing, I may try this (I think it is a technique known as 'Octet' sharpening), I do it manually with repeated passes of USM, but I think some programs may automate the process.
First pass: Amount (A) = 20%, Radius (R) = 4px, Threshold (T) = 3. (last depends on noise)
Second pass: A = 25%, R = 2px, T = 4.
Third pass: A = 35%, R = 1px, T = 5.
Fourth pass: A = 45%, R = 0.5px, T = 6.
I might vary those Amounts or skip a pass, depending on how it appears to be 'coming on' as I go through the passes, half the time, I'll back track (history) and start over with different values of A, R and T if it is not having the desired effect.
Note the Radius was reduced by two on each pass, giving the name (think 'octaves' in music). The a Amount is increased each pass, but not by as much. Threshold I couldn't decide, might be OK to leave that at single value throughout, depends on image noise; more noise, higher number, if low/no noise, then zero (so that sharpening is applied to fine details).

Sharpening is an area where everyone has 'their way' that works for them, this works for me - but I'm not saying it right for everyone and I'm sure some may disagree with me.

HTH, Dave

If you have the tine and can work it to your liking I am curious what the difference will be.

PS
If you can get the full size RAW, or an unsharpened 5500 wide jpg to me somehow (DropBox?), I could have a go at it, but I have told you above almost exactly how I'd process it.

[JBW]

You're right there is a lot of good stuff in your comment. You're wrong it isn't nor does it sound arrogant. it is simply a true statement. I will work my way through it and see what happens. What's 'drop box'?

[Dave Humphries]

Thanks Brian,

What's 'drop box'?

A place you can use to pass me large RAW file without necessarily making it publicly accessible.

See www.dropbox.com

Cheers, Dave

[JBW]

Thanks Brian,



A place you can use to pass me large RAW file without necessarily making it publicly accessible.

See www.dropbox.com

Cheers, Dave

I tried to apply your suggestions to the shot as it came out of Sony Pro. My blog uses 650 as the standard photo size so I scaled down to 650 and then sharpened before doing any of the final tweaks. I think it's better.

[Dave Humphries]

I tried to apply your suggestions to the shot as it came out of Sony Pro. My blog uses 650 as the standard photo size so I scaled down to 650 and then sharpened before doing any of the final tweaks. I think it's better.

Me too

[Dave Humphries]

Hi Brian,

You sent me the RAW file and I have processed in Photoshop CC as I normally would, although with regard to the crop, I tried to match yours.
Yours 1218 x 975px, mine 1201 x 936px, hopefully it won't jump too much when switching between them in LyteBox here at CiC.

As a learning experience/comparison for my own benefit as much as yours - I saved four versions;

The first sharpened as I normally do, with simple USM in a single pass of:
A=100%, R=0.3 px and T=1



The second with the Octet sharpening; 3 passes as follows;
A=20%, R=2px, T=1
A=30%, R=1px, T=1
A=50%, R=0.5px, T=1


This version (above) has a slight halo noticeable on the edge of the wings.

I felt this could be improved if the brightness of the halo were reduced as follows:
A=15%, R=2px, T=1
A=25%, R=1px, T=1
A=60%, R=0.5px, T=1


Or alternatively (to make the existing halo narrower), we could try:
A=20%, R=1.2px, T=1
A=30%, R=0.6px, T=1
A=50%, R=0.3px, T=1



Discussing the effects of those changes may help people understand the effects of what's happening.
However, as can be seen, the differences are the proverbial "gnat's whisker"

I haven't paid much attention to the noise, it wasn't too bad as you shot at 100 iso.

The Octet sharpening is definitely better than simple USM, which version is almost indistinguishable, I probably slightly prefer the final version with less radius, but the first Octet looks sharpest, if you can live with the halo.

My apologies if my versions are a bit bright for you though

I lifted the exposure (0.75 stop) and the shadows in ACR, also gave Clarity a whack at 50%, then, since I still had a little tonal range in hand, applied LCE with USM at Amount 30% Radius 100px and Threshold 0 before trying the different sharpenings on separate layers.

Cheers, Dave

[JBW]

I can live with the bright.