Ed Applying luma curve to photos and footage

[JayUtah]

Yes, that last bit. From what I can tell, Nt1 has created anomalies is images that I suppose could be due to manipulation by shadowing forces or could just as easily be created by nt1's only manipulations.

The inverted paraboloid luminance map is 100% responsible for both the aliasing he observes and the "halo."

Imagine a luminance (brightness) channel with values from 0 to 100. His curve would render brightness Y=45 and Y=55 as exactly the same brightness in the adjusted image, even though in the original image data those would differ by 10% of the brightness gamut. Further, since the first derivative of the curve is smallest at the apex, there will be very little difference in adjusted luminous intensity between all the values 45 ≤ Y ≤ 55. This is where these bright bands are coming from that he wants to attribute to "alpha channel halo."

Toy image manipulation programs generally don't handle alpha channels well. In an RGBA color space, a few will simply ignore the A (transparency) channel and apply full RGB to the composition. Thus rather than a smooth blending between foreground and background pixels as indicated by the alpha (A, brightness) channel, you would get a blocky, pixelated composition. Now in 2001, all the Adobe products obeyed alpha channels. So most of the handwaving about alpha channels is nonsense. But the real nail in the coffin is that in such compositions the alpha channel is simply not preserved. It's used in the operation of compositing and then discarded when the compositional layers are mixed down. None of the images @nt1 is using actually has an alpha channel. The notion that some fiddling with sliders can recover some "hidden" alpha-obscured data after composition is categorically absurd.

 

[Cosmic Yak]

Has nt1 said yet what he thinks the significance of this supposed effect is?

 

[JayUtah]

Am i a country or judicial system? Am i the one that should be investigating a crime properly?

It was investigated as a crime. If you're claiming the criminal investigation was done improperly, this suggests that you have some expertise in law and forensic investigation that would form the foundation of that claim. Are you making such a claim?

 

[Norman Alexander]

And since you asked

RUOK?

 

[Norman Alexander]

From further back and image technicalities aside, nt1 is simply trying to shift the burden of proof. All that "prove me wrong" stuff. We know how that goes.

I also have concerns that the responses we get here may be AI generated too. If this was a real person posting some of this snark, I would be worried about the level of unhealthy paranoia. That said, I am not going to waste my time delving into a clearly disturbed mental state.

In summary, there is no claim to argue, the arguer is arguing in bad faith and with a lack of technical understanding of their own tools, and their posting behaviour indicates we can't expect rational responses. It's not much at all to work with, is it.

 

[JayUtah]

Has nt1 said yet what he thinks the significance of this supposed effect is?

All I can glean is that he considers it anomalous and, possibly, evidence of image tampering. He has said :—

I can show the same halo through separate and different generations of iteration.
* * *
It is an alpha halo hidden in the photo. Exposed by bending light.

But I cannot find where he fits the "alpha halo" hypothesis into any enclosing narrative or hypothesis.

Now if we follow @nt1's formulation of the hypothetic-deductive method, here's what we come up with...

1. Question – Identify a specific, testable question.

The question appears to be, "What causes the aliasing and banding observed in the images after a particular luminance map is applied?"

2. Hypothesis – Form a prediction based on existing knowledge.

The hypothesis appears to be, "It is an 'alpha halo' caused by improper compositing of digital images that contain an alpha channel."

In Kuhn's reformulation of the scientific method, we would just test this against the null hypothesis. In the original formula proposed by Popper, we would develop a hypothesis set and run them all against each other. Among the hypotheses in the Popper formulation would be, "The aliasing and banding are artifacts from the application of the luminance map." And of course we could think of others.

3. Experiment – Design and conduct a controlled test, changing one variable at a time.

This step is entirely missing from @nt1's presentation so far. A controlled test of the "alpha halo" process would mean controlling for such effects as artifaction from the adjustments used to develop the observation we wish to explain. We're not explaining something that occurs unbidden in the real world, but something that emerges only when the observation is developed in a certain way. The characteristics of "that certain way" need to be made part of the experiment control.

4. Observation/Data Collection – Record what happens during the experiment.

Since no experiment actually tests the "alpha halo" process in this demonstration, this step is moot. The method has already been detailed impermissibly.

5. Analysis – Interpret the data to look for patterns or outcomes.

The analysis seems to be simply to note a superficial similarity between the observed effect and what is purported to be the effect of the "alpha halo."

If we were testing the "artifact of the luminance map" hypothesis empirically, we would apply it to synthetic images of various kinds. And I may do that if there is any actual interest. Right now, @nt1 seems to be yanking everyone's chain. But absent empirical evidence, the effect follows analytically. This is math, and math is comfortably deterministic. We can say with mathematical certainty that in the luminance map applied, Y=0.45 and Y=0.55 will map to the same output brightness, thus resulting in the observed effect. At worst this fact would go into the design of the empricism for other hypotheses.

6. Conclusion – Determine whether the hypothesis was supported or refuted.

Since the hypothesis was never tested, it is merely being asserted here as a jumped-to conclusion.

7. Report – Share results for review or replication.

Sharing the results here doesn't seem to be having the outcome @nt1 expected or desires. Note how the scientific method doesn't include belittling the reviewers when they point out the obvious flaws in the experiment.

 

[nt1]

I've explained how your manipulation of the luminance channel is producing these effects. They aren't in the photo. They are 100% caused by you fiddling with the luminance in a way you don't seem to understand. In this latest example, your method—which eliminates the difference between adjacent midtones—is causing the aliasing.


I did.


I do care, which is why I'm engaging your purported method.

They are not just from my fiddling.

I also agreed previously that it can create false results.

It also can identify physical impossibility.

There is no need to treat it like a caveman and fire. Or an inquisition or crusade.

It is. By simple fact. A tool that can expose cause for concern for a photographic image. That is a fact.

I have supplied limited examples. I have others. I would also accept any submissions of footage and show my cause for concern.


As a civilian. I am not an expert. I am not an authority. I do not dismiss anything in the manner i have recieved here. Which speaks volumes.

I like your wig.

 

[nt1]

Yes, that last bit. From what I can tell, Nt1 has created anomalies is images that I suppose could be due to manipulation by shadowing forces or could just as easily be created by nt1's only manipulations. Anyone that questions that is met by...snark. I'm not especially convinced that nt1 has found anything worthy of more investigation. This is especially true because nt1 has not suggested a falsifiable claim. What can we debunk. Yes, those images appear to have been manipulated and they appear to have artifacts as a result. Well, we know for a fact that nt1 has manipulated them. Anything else....I don't know.

Created by my own manipulations.... that are available to anyone to replicate and judge.

Look big guy. Take your judgement towards me and apply it to the same event. Are you comfortable with it? Then say that. Just say "there is no reason to investigate any information because it is accurate enough to have brought all to justice."

Post a sign at the front door.

If not.

State them here, any questions you have remaining. Ill go work on those.

 

[nt1]

The inverted paraboloid luminance map is 100% responsible for both the aliasing he observes and the "halo."

Imagine a luminance (brightness) channel with values from 0 to 100. His curve would render brightness Y=45 and Y=55 as exactly the same brightness in the adjusted image, even though in the original image data those would differ by 10% of the brightness gamut. Further, since the first derivative of the curve is smallest at the apex, there will be very little difference in adjusted luminous intensity between all the values 45 ≤ Y ≤ 55. This is where these bright bands are coming from that he wants to attribute to "alpha channel halo."

Toy image manipulation programs generally don't handle alpha channels well. In an RGBA color space, a few will simply ignore the A (transparency) channel and apply full RGB to the composition. Thus rather than a smooth blending between foreground and background pixels as indicated by the alpha (A, brightness) channel, you would get a blocky, pixelated composition. Now in 2001, all the Adobe products obeyed alpha channels. So most of the handwaving about alpha channels is nonsense. But the real nail in the coffin is that in such compositions the alpha channel is simply not preserved. It's used in the operation of compositing and then discarded when the compositional layers are mixed down. None of the images @nt1 is using actually has an alpha channel. The notion that some fiddling with sliders can recover some "hidden" alpha-obscured data after composition is categorically absurd.

Toy image apps.

Ok. Please name a non toy. Ill try that.

Ill go try others now. Thank you for the advice

 

[nt1]

It was investigated as a crime. If you're claiming the criminal investigation was done improperly, this suggests that you have some expertise in law and forensic investigation that would form the foundation of that claim. Are you making such a claim?

Incorrect. And hilariously lazy in my opinion. Similar to the same level of opinion you have applied to me..

 

[JayUtah]

They are not just from my fiddling.

They are, as I explained. You have not engaged with that explanation.

I also agreed previously that it can create false results.

It will create false results. It is a mathematical inevitability. You don't seem able to engage with the math.

It also can identify physical impossibility.

You have not shown this. Nor have you explained how your method distinguishes false positives from what you think constitutes actual data of interest.

There is no need to treat it like a caveman and fire. Or an inquisition or crusade.

Straw man. No one is frightened or mystified by your method. I am amused. I know exactly what it is doing and how it is failing. I have explained this to you several times. You are either unwilling or unable to address that.

Your formulation of the scientific method (which is pretty much verbatim copy pasted from ChatGPT) includes a review step in which qualified reviewers study your work and identify weaknesses in it. This is what is happening to you now. If you are uncomfortable with having your work reviewed by experts, then you are probably not cut out to be a scientist.

It is. By simple fact. A tool that can expose cause for concern for a photographic image. That is a fact.

No, it is not. You have not shown that you understand what the tool actually does. You seem to be regarding it as some magic wand that gives you a bottom-line answer to some particular question touching upon authenticity.

I have supplied limited examples. I have others.

Irrelevant. You are simply making the same mistake over and over.

As a civilian. I am not an expert. I am not an authority. I do not dismiss anything in the manner i have recieved here. Which speaks volumes.

I am an expert in the field you are dabbling in. I'm telling you what you're doing wrong. You are not amenable to any such suggestion.

 

[nt1]

RUOK?

Textese. Thats what level i am dealing with?

 

[nt1]

They are, as I explained. You have not engaged with that explanation.

It will create false results. It is a mathematical inevitability. You don't seem able to engage with the math.


You have not shown this. Nor have you explained how your method distinguishes false positives from what you think constitutes actual data of interest.


Straw man. No one is frightened or mystified by your method. I am amused. I know exactly what it is doing and how it is failing. I have explained this to you several times. You are either unwilling or unable to address that.

Your formulation of the scientific method (which is pretty much verbatim copy pasted from ChatGPT) includes a review step in which qualified reviewers study your work and identify weaknesses in it. This is what is happening to you now. If you are uncomfortable with having your work reviewed by experts, then you are probably not cut out to be a scientist.


No, it is not. You have not shown that you understand what the tool actually does. You seem to be regarding it as some magic wand that gives you a bottom-line answer to some particular question touching upon authenticity.


Irrelevant. You are simply making the same mistake over and over.


I am an expert in the field you are dabbling in. I'm telling you what you're doing wrong. You are not amenable to any such suggestion.

You. Are an expert.

Ok. Great. That is good news.

List what credentials you* can publicly without comprising identity. Be concise* and specific.

List your professional application in the manner possible without revealing identity

What is your expertise in manipulating photos to reveal unnatural light.

I am being genuine.

 

[JayUtah]

Created by my own manipulations.... that are available to anyone to replicate and judge.

I have replicated them. I adjudicate them to be false positives, artifacts of the misuse of luminance mapping. I have explained in suitable detail how the artifacts arise. Why are you seemingly unable or unwilling to consider those explanations?

Take your judgement towards me and apply it to the same event. Are you comfortable with it? Then say that. Just say "there is no reason to investigate any information because it is accurate enough to have brought all to justice."

No, there is no reason to judge another claim in order to identify the flaws in your method. There is no need to investigate your claims further because the "anomaly" you are trying to explain has a trivial explanation.

 

[nt1]

I have replicated them. I adjudicate them to be false positives, artifacts of the misuse of luminance mapping. I have explained in suitable detail how the artifacts arise. Why are you seemingly unable or unwilling to consider those explanations?


No, there is no reason to judge another claim in order to identify the flaws in your method. There is no need to investigate your claims further because the "anomaly" you are trying to explain has a trivial explanation.

Hey wig? I got to be honest. Your quoting and responding has lost its merit. It did with the judge wig avatar. But the shaming tactics by quoting sections with ridicule is of no value to me. If you care, one way or the other. Quote what you want and say it

I have replicated them. I adjudicate them to be false positives, artifacts of the misuse of luminance mapping. I have explained in suitable detail how the artifacts arise. Why are you seemingly unable or unwilling to consider those explanations?


No, there is no reason to judge another claim in order to identify the flaws in your method. There is no need to investigate your claims further because the "anomaly" you are trying to explain has a trivial explanation.

Serious question where did you replicate and show work?

 

[nt1]

I have replicated them. I adjudicate them to be false positives, artifacts of the misuse of luminance mapping. I have explained in suitable detail how the artifacts arise. Why are you seemingly unable or unwilling to consider those explanations?


No, there is no reason to judge another claim in order to identify the flaws in your method. There is no need to investigate your claims further because the "anomaly" you are trying to explain has a trivial explanation.

That is a very poor answer and explaination.

Besides your words. What is your evidence?

 

[JayUtah]

You. Are an expert.

Yes.

List what credentials you* can publicly without comprising identity. Be concise* and specific.

B.S. in engineering. M.S. in computer science. PhD work in engineering. Professionally licensed as an engineer.

List your professional application in the manner possible without revealing identity

I am part owner of an engineering company. Our work includes forensic engineering investigation. Our forensic investigations including the analysis of images, including detecting possible forgeries or tampering in the images.

What is your expertise in manipulating photos to reveal unnatural light.

Leading question. Before starting a company with my colleagues, I taught engineering and computer science at one of the top five universities in the United States for computer graphics and image analysis. I am also published in Science for image analysis, specifically for the purpose of determining authenticity.

I am being genuine.

No, you really aren't.

 

[nt1]

Well, I wasn't actually asking you, and it's a thing called "humour", which you may have heard of.
Given your apparent inability to answer direct questions, it's all that's left.

Oh. Humor. Like what i used? Was my response not to the acceptable level of humor as yours?

So you are a judge of 9/11 and humor.

Anything I need to know you sre the authority of? Ill do my best.

 

[JayUtah]

But the shaming tactics by quoting sections with ridicule is of no value to me.

If you believe you are being personally attacked, report the post for moderation. Otherwise do not claim persecution as a rhetorical stunt.

Serious question where did you replicate and show work?

I replicated them on my own personal computer. I have explained the mathematical principles behind your error. Address them or admit you cannot.

 

[Thermal]

@nt1, have you tried falsifying your result? Take a still from.a genuine videoed image from the same time frame (around 2001) which has been digitized and downloaded through YouTube, then run it through your editor, and see if the masking effect shows? You claimed this earlier with an airplane image but im.not sure if the original image was digitized from 25 yr old video et al.

Alternatively, could anyone else here familiar with nt1's process duplicate it?