Absolute Difference

Moderator: jsachs

Marpel
Posts: 553
Joined: September 13th, 2009, 3:19 pm
What is the make/model of your primary camera?: Nikon D810
Location: Port Coquitlam, British Columbia

Absolute Difference

Post by Marpel »

Over the years, I have been using Composite with Absolute Difference (then a Levels and Color, full dynamic range) to compare two generations of an image to confirm the difference between the two.

I often have multiple derivations of an image, some very subtle, even invisible to my eyes, and if if I run out of time, will save some derivations, then re-open later to carry on. As a result, I find myself doing the above process, just to confirm what I have done to a particular image.

So, I just wish to confirm, if I use this process and see no visible change (the image appears to remain black), that there is absolutely no difference between the two images, even to the smallest degree. I looked at the help explanation of Absolute Difference and it was not explicit in this regard.

Thanks,

Marv

jsachs
Posts: 3303
Joined: January 22nd, 2009, 11:03 pm

Re: Absolute Difference

Post by jsachs »

Yes -- to make sure, you can Autorange the result to exaggerate any minute differences.

Also, if you use Blend instead of Composite and choose Absolute Difference, you will get a Strength slider you can use to magnify any differences.
Jonathan Sachs
Digital Light & Color

Marpel
Posts: 553
Joined: September 13th, 2009, 3:19 pm
What is the make/model of your primary camera?: Nikon D810
Location: Port Coquitlam, British Columbia

Re: Absolute Difference

Post by Marpel »

Excellent,

Thanks very much.

tomczak
Posts: 1140
Joined: April 25th, 2009, 12:56 am
What is the make/model of your primary camera?: Fuji X-E2
Contact:

Re: Absolute Difference

Post by tomczak »

Also, one option in Compare is offset difference, which also has a strength slider - instead of black, 50% grey indicates no difference.
Maciej Tomczak
Phototramp.com

Marpel
Posts: 553
Joined: September 13th, 2009, 3:19 pm
What is the make/model of your primary camera?: Nikon D810
Location: Port Coquitlam, British Columbia

Re: Absolute Difference

Post by Marpel »

Thanks tomczak for the further suggestion.

As an aside, an unexpected benefit of my initial post is the discovery of Blend, rather than using Composite, especially for straight forward composites (without the need for alignment points etc), and in particular the added Blend Modes. I have always been curious why Composite was lacking some of those modes, and as I often experiment with different Blend Modes, Blend will likely be my go-to Composite option.

Jonathan, I'd like to also say, even though this thread has nothing to do with this particular point, that I have repeatedly compared Color Balance in PWP, Photoshop (CS6) and even Nikon's Capture NXD. And I have found PWP consistently gives me better results. Every now and then the others will perform as good, but PWP seems to hit it way more often. Just thought I would mention, and didn't want to start a new thread just to say so.

Marv

doug
Posts: 111
Joined: April 24th, 2009, 10:06 am
What is the make/model of your primary camera?: Nikon D-500
Location: Toledo, Ohio USA

Re: Absolute Difference

Post by doug »

When starting out in digital photography over 15 years ago, I used this feature of PWP to verify the reported horrors of jpeg degredation through multiple saves. I ran a test in which I created multiple generations of the same jpeg image by saving each successive generation anew, thus creating a succession of saved generations (with zero intervening edits). The absolute difference comparison with generation zero remained black up through at least 8-10 generations of saves.

I concluded that the reported deleterious effect of multiple jpeg saves was exaggerated -- at least through a handful of cycles. Was this a wrong conclusion? I don't recall if I applied Autorange of brightness which might have revealed differences sooner, eh?

jsachs
Posts: 3303
Joined: January 22nd, 2009, 11:03 pm

Re: Absolute Difference

Post by jsachs »

I would not be surprised if the effects of multiple JPEG saves to depend on the JPEG quality settings, but I have not tried the experiment. A small number of saves at high quality is usually benign.
Jonathan Sachs
Digital Light & Color

Marpel
Posts: 553
Joined: September 13th, 2009, 3:19 pm
What is the make/model of your primary camera?: Nikon D810
Location: Port Coquitlam, British Columbia

Re: Absolute Difference

Post by Marpel »

Doug,

Although I cannot confirm the point at which repeated saves of a jpeg start to show degradation, I can say that a few years back when I started to use Absolute Difference to confirm the difference between two version of an image, initially I neglected to use Autorange.

After a few repeats of a black image, suggesting to me at the time that there was no difference, I clued in and tried Autorange, which did reveal some difference.

Point being, I was surprised what could actually be revealed with Autorange, as compared to the first step of A.D. which may give a seemingly black image.

Marv

doug
Posts: 111
Joined: April 24th, 2009, 10:06 am
What is the make/model of your primary camera?: Nikon D-500
Location: Toledo, Ohio USA

Re: Absolute Difference

Post by doug »

Thanks, Marpel
Since my "jpeg degradation" test was over 15 years ago, I can't recall if I was smart enough to use auto range. I suspect not. And as a result I may have misled myself into believing that the black result after multiple saves indicated no change in the file whatsoever.

Doug

HanSch
Posts: 63
Joined: May 20th, 2009, 12:30 pm
What is the make/model of your primary camera?: Pentax K5
Location: Venlo, Netherlands
Contact:

Re: Absolute Difference

Post by HanSch »

I tested it with jpeg quality set to 95. After say 5-6 saves there is definitely a difference, the file size changes a bit.
Autorange is required to see the differences after the Absolute Difference transformation. The question is, is it a relevant difference in practice?

Post Reply