'There is No Such Thing as a Real Picture,' Says Samsung Exec (theverge.com) 158
How does Samsung defend itself against the notion that its phone cameras are spitting out fake AI photos of not only the Moon, but most anything else you'd care to aim them at these days? For starters, the company's head of product is saying that every photo is fake. The Verge: Samsung EVP Patrick Chomet told TechRadar recently: "There was a very nice video by Marques Brownlee last year on the moon picture. Everyone was like, 'Is it fake? Is it not fake?' There was a debate around what constitutes a real picture. And actually, there is no such thing as a real picture. As soon as you have sensors to capture something, you reproduce [what you're seeing], and it doesn't mean anything. There is no real picture. You can try to define a real picture by saying, 'I took that picture,' but if you used AI to optimize the zoom, the autofocus, the scene -- is it real? Or is it all filters? There is no real picture, full stop."
Oh, Please. . . (Score:4, Insightful)
That's the biggest load of Malarkey I've ever read!
Re:Oh, Please. . . (Score:4, Insightful)
No, he's right. Digital pictures, particularly with all the various enhancements they go through today...they aren't in the same class as film pictures of old, and can be said to be "not real". Sure, they're based on real light data, but that's been modified before you get the final product.
Which is not to say there isn't a gradient of "fake"; obviously some are more manipulated ( or fabricated ) than others. Doesn't change the underlying point, however.
Re:Oh, Please. . . (Score:4, Insightful)
Re: (Score:2, Insightful)
kinda, samsung IS technically correct in their claim, but it also wouldn't be hard draw an arbitrary line in the sand somewhere ("dumb pipe, no after-sensor modification" etc) and break up the idiot-friendly binary of "artificial/not artificial"
Re: Oh, Please. . . (Score:5, Insightful)
On the other hand, there are certain actions that render a picture absolutely fake, and certain sets of actions that render it absolutely real.
If you do nothing but read the data from the light sensors and store it in a raw file, it certainly is real. If you literally substitute a stock image in and replace part of the image completely ignoring the sensor data in that area, it is definitely fake.
Samsung clearly is falling into the fake category today.
Re: Oh, Please. . . (Score:4, Funny)
but my street light being turned into a high res moon is pretty awesome. who needs a depiction of reality when you can have AUGMENTED reality
Re: (Score:2)
On the other hand, there are certain actions that render a picture absolutely fake, and certain sets of actions that render it absolutely real.
If you do nothing but read the data from the light sensors and store it in a raw file, it certainly is real. If you literally substitute a stock image in and replace part of the image completely ignoring the sensor data in that area, it is definitely fake.
Samsung clearly is falling into the fake category today.
Exactly.
Re: (Score:3)
I think there is a still a 'binary'
when are putting in something that was not there.
Autofocus - does not make something 'fake', my own eyes do it, it might be bending light but its not adding something, its not removing.
A 'polarizer' - might depend on how you are using it. Generally even though it is removing something, you are still faithful representing what is seen, and its again something an in person observer might partially effect by shielding an eye.
A telephoto lens - I would say not fake, the effect
Re:Oh, Please. . . (Score:4, Informative)
A telephoto lens absolutely does not have the same effect as moving in. When you move in, the relative position and size of things in the background will change. The perspective changes. But when you zoom in, all you're doing is cropping and enlarging. That still doesn't make it any more fake than viewing the scene through a telescope or binoculars, though.
People sometimes wear polarizers directly over their eyes... they're called "good sunglasses". So seeing a polarized image is also not fake.
Re: (Score:2)
But when you zoom in, all you're doing is cropping and enlarging.
Zooming also narrows the depth of field, which moving closer does not.
Re: (Score:2)
Re: Oh, Please. . . (Score:2)
Yes, it does.
The exec is trying to avoid criticism of Samsungs far end of the "fake" spectrum by associating their extreme behavior with accepted, much less interfering methods of photo capture.
Saying their is a spectrum implies a recognition of some level of enhancement we accept, and some level we don't accept (as real). Which refutes the exec's philosophy.
Re: (Score:2)
I'm old enough to have started photography in the film days. Other than black and white film developed using B&W developers, color film was color corrected automatically and a standard tone curve was also applied during printing. Color films does not have a gamut that accurately reflect the color seen by the human eye.
By the your logic and that moronic exec, even film photos are fake.
Re:Oh, Please. . . (Score:4, Insightful)
I'll go one step further and say what we see with our eyes is fake, using his logic. Our brain processes signals received from the eyes. If he's saying because sensors are involved to produce the image, then our brain is the same as a sensor and making its best guess at what it's being told.
Re: Oh, Please. . . (Score:5, Informative)
I see slightly different colors from each eye. My brain actively picks one over the other, but I can "see" the difference by closing one eye or the other. Your mind does all kinds of "post-processing".
Re: (Score:2)
Re: (Score:2)
Re: (Score:3)
They give us those nice, bright colors / They give us the greens of summers / Makes you think all the world's a sunny day, oh yeah!
Now I'm going to have Simon and Garfunkle in my head all day.
Re: (Score:2)
Ah, memories. The summer that song was a hit, I was working in a big photo development and printing lab owned by Fujifilm handling consumer photos of mostly Kodak 126, 110, and 35mm cartridges and rolls. We found it amusing to hear that song repeating on the radio at work all day.
Re: (Score:3)
Even with film, it was hard to capture a rainbow. It was interesting using different types of film too. Kodachrome, Etkachrome, Fujichome all captured colours differently.
Re: (Score:2)
Re: (Score:2)
you can still get fuji film and it is stunningly good.
I always felt like it was a bit too saturated and "pinkish", compared with Kodachrome. Ektachrome was "cooler" and less saturated than Kodachrohohohme, but still not carnival-colors like Fuji, or the almost-no-color Agfa...
Re: (Score:2)
It's more like "fake" is a matter of degree -- how close the picture is to the actual reality. If the full fidelity of reality is 1.0, meaning all of reality, and 0.0 is nothing to do with reality whatsoever, a film was maybe 1.0/inf, a digital picture was 1.0/inf/2, an "AI" enhanced photo is anywhere from 1.0/inf/5 to 1.0/inf/5000 etc.
Maybe the scale would work better not normalized for infinity though. : -)
Re: (Score:2)
That's arguably why it deserves to be classified as 'malarkey'. He's responding to accusations that his just-hallucinate-in-details-the-optics-can't-gather system is faking by making the (true) statement that all photos are fake in order to change the subject from whether all photos are fake in the same way and to the same degree(which is
Re: (Score:3)
No, he's right. Digital pictures, particularly with all the various enhancements they go through today...they aren't in the same class as film pictures of old, and can be said to be "not real".
I disagree. If you want to get waaaaaaay into the weeds on camera film production, check out the "Smarter Every Day" series where the host takes a tour through the Kodak factory. From memory I believe part 2 gets into the chemistry of the photoreactive film coating (link [youtube.com]). There are two more hour-long videos in his main series on different parts of production
Re: (Score:2)
Re: (Score:3)
No, he's right. Digital pictures, particularly with all the various enhancements they go through today...they aren't in the same class as film pictures of old, and can be said to be "not real".
That's somewhat true, but what Samsung is accused of doing is having its camera's image manipulation software add in details that aren't actually there.
That's not just manipulating contrast and sharpening edges. That's making stuff up.
Re: (Score:2)
Digital pictures, particularly with all the various enhancements they go through today...they aren't in the same class as film pictures of old, and can be said to be "not real".
Heh. Ever developed film? Not trying to be snarky, just remembering some stuff I heard decades ago. This isn't exactly a fresh topic.
Which is not to say there isn't a gradient of "fake"; obviously some are more manipulated ( or fabricated ) than others. Doesn't change the underlying point, however.
To me the threshold is fairly distinct and easy to see: Is the work being done on the pixels, or is it being done on the image? Are you looking at a curve of values and applying a transformation, or are you looking at a 2-dimensional set of pixels and re-positioning or re-constructing them?
Or.. in film terms- Are you playing with the intensities of light (or the balance
This is not what we are talking about here (Score:2)
> Sure, they're based on real light data
That's not what Samsung was doing though. They use the real light data to detect whether you are trying to take a photo of the moon. Then, they would superimpose a highly detailed image of the moon with detail that was not in your original data (because you were not using a telescope) within your photo.
Re: (Score:2)
No, he's right. Digital pictures, particularly with all the various enhancements they go through today...they aren't in the same class as film pictures of old, and can be said to be "not real". Sure, they're based on real light data, but that's been modified before you get the final product.
The same is true of film, actually. With film, the photons strike the film and cause chemical changes, that much is "real", much like the electrical charges produced when light hits a CCD. But in both cases, after that initial sampling of reality, a lot of post-processing happens before an image is printed/displayed that people can look at -- and good photographers have always manipulated every stage of that processing to get the effects they wanted. With digital the space for manipulation is a little more
Re: Oh, Please. . . (Score:2)
No, it's total bullshit. Why? Because it is not necessarily any more true for a digital camera than for a film camera. Film has grain, sensors have cells. Film and digital cameras both have lens aberration and so on. Film is processed during developing, digital is processed on the device and then maybe more later. In both cases your goal can either be fidelity or fantasy. A digital camera can be designed to do only necessary processing, or more.
A digital camera which intentionally does only minimal processi
Re: (Score:2)
Re: (Score:2)
No, he's right. Digital pictures, particularly with all the various enhancements they go through today...they aren't in the same class as film pictures of old, and can be said to be "not real". Sure, they're based on real light data, but that's been modified before you get the final product.
Physical photos can be manipulated too, both in the camera and during development as well as by choice of film and paper. I'm not an expert, but I would guess that just about anything one could digitally one could do physically, albeit with more work. I also guess that many (most?) digital photo manipulations are based on, or originally stem from, older physical manipulations.
Re: (Score:2)
My samsung smartphone can save raw images just fine, thank you.
Re: (Score:2)
The readout of the sensor's pixels mapped to RGB i guess. So yeah, something like that.
Smartphone (Score:2)
>> Now if you start talking about digital cameras and RAW images..
Today, those are called "Smartphones"
Re: (Score:2)
Re: (Score:2)
> Doesn't change the underlying point, however.
His underlying point is if you take images on a cellphone, especially on an iPhone or Samsung variants of Android phones, they apply extremely heavy filtering - Huawei phones even simply replace parts of the image with their 'AI engine'.
Now if you start talking about digital cameras and RAW images..
The latter is where I would posit that most people would solidly defne the image (or parts thereof) as "fake".
Re:Oh, Please. . . (Score:4, Interesting)
That's the biggest load of Malarkey I've ever read!
Far from it in my opinion. I suspect you don't like it because he is not taking a scientific argument, but a philosophical one. It's the same argument as Plato's Allegory of the Cave, and Descartes's Discourse of a Method.
It seem his argument is we're living in a simulation, [scientificamerican.com] so it doesn't matter if our software captures something different that what our eyes perceive. Claiming the camera produces a "fake" image will not win this argument, because there is no reality.
To win this argument, it should be stated that the human eye is considered the standard by which all cameras should be measured. This technology does not move closer to reproducing the images the human eye does, but instead moves farther away. By that standard, this technology produces low quality images.
Re: (Score:2)
Re: (Score:2)
It's post-truth nonsense for photography. Discrediting everything until the least trustworthy viewpoints (photos) are on par with everything else. Instead of selling us politics based on lies he's selling us a photo app based on lies.
Re: (Score:2)
From a computational perspective, it's sort of true, in a way. A less sensational and more precise statement of the principle would be "In the most general case there is no really reliable way to tell just by looking at a picture whether it is a photograph or an artist's rendering." Which shouldn't be news to anyone here, but there *are* people out there who are unaware of this concept, and sometimes they cause actual trouble. For a solid level-headed and detailed examination of this issue, see the "W
Re: (Score:2)
However, while both a photo and an image can be faked, it is easier to do so with images. Photo manipulation takes a darkroom and a lot of time. Anybody with a computer can doctor images these days very easily.
Perhaps there is no such thing (Score:4, Funny)
He's not wrong, but... (Score:2)
Sure, every image produced by a camera is the processed output of the sensor, it's not the same as a human eye, or your individual eye, or whatever. Any phone with a decent camera does a lot of work to overcome the limits of the optics and so forth.
That's different from just replacing the data from the camera with a stored image though. While it's probably mostly harmless for the Moon, we know that it can do a lot of damage when applied to human beings. Literally impossible standards of beauty.
There is also
Re: (Score:3)
That's different from just replacing the data from the camera with a stored image though.
the ceo is right in that that's just perspective. that gadget is a tool just like a kodak instamatic, rembrandt's brushes or my eyes and wild imagination and it is true: every picture is an interpretation.
the ceo is dead wrong, though, if he thinks that this fixes the real issue, which was labelling as "image enhancing" a bag of tricks that isn't what you would expect from classic image enhancing at the time. although technically correct because the concept of enhancing is so broad, it was clearly misleadin
Of course it's not "AI" (Score:2)
On the other hand I could totally see a cell phone company wanting to jump on the AI hype train, which is why he's being so evasively silly. He thinks it adds mystique to the phone. It doesn't, it just encourages nutters to think Samsung faked the moon landing.
Re:Of course it's not "AI" (Score:5, Informative)
I worked with an ex-Kodak engineer who said that even pre-“AI” as we know it, many consumer digital camera images were modified in between the sensor capture and the user seeing it based on user expectations. He would say when you think of “green grass,” “blue sky,” and “yellow sun,” you have a cartoon high-saturation image in your head of what that looks like. When the photo you took has comparatively dull grasses and skies, you are disappointed. So the consumer digital cameras would “fix” those colors with transforms.
These “non-AI” transforms (convolutional blur/sharpen filters, nonlinear lookup tables) often inspired and informed AI imaging models down the road, including ones that do sophisticated upscaling/inpainting/deblurring of consumer camera images. So I can see why someone with a lot of experience in the digital imaging industry would say there’s “no such thing as a real picture.” They’ve been modifying raw sensor measurements for as long as there have been digital cameras.
Re: (Score:2)
That's usually done in the processing from sensor to jpeg. When you shoot raw, you get raw - bland, unsaturated, and unprocessed.
I'd draw the line at the substitution of pixels, following the same guidelines as photography competitions & journalism. You can burn, dodge, saturate, and crop all you want. But if you manipulate the image to add or remove specific artefacts, then it's no longer "real".
People forget that burning, dodging, and saturation are also key variables intrinsic to developing film phot
Re: (Score:2)
Some sophisticated arguments are also ridiculous (Score:5, Insightful)
In the phone world he isn't wrong (Score:5, Insightful)
In the DSLR/Mirrorless world he most definitely is wrong. RAW data captured on a sensor represents the photons which hit that portion of the sensor through the bayer filter. It is possible from a real camera to see this in all it's black and white bayer filtered glory.
A phone on the other hand has been spitting out fake images for over half a decade now. Very little of what comes out is representative of the sensor, be it artificially adjusting the image colour depending on the analysed style of the image (something we've been doing for nearly a decade), clever noise reduction schemes, extracting subjects and creating fake bokeh on the background (phones have sensors way to small for this to be a real optical effect), to today's AI madness.
But honestly who cares. Unless you start seeing CSI taking phone snaps of a crimescene for use in a court the entire topic is moot. Photography has always had an element of creative endeavour in it, and I've got a picture of a scuba diver stepping out of a puddle on a footpath from the 1930s that shows this isn't even limited to modern computers.
Re:In the phone world he isn't wrong (Score:5, Informative)
In the DSLR/Mirrorless world he most definitely is wrong. RAW data captured on a sensor represents the photons which hit that portion of the sensor through the bayer filter. It is possible from a real camera to see this in all it's black and white bayer filtered glory.
Now compare that raw image to the final picture that gets generated. It won't look anything alike. The biggest difference is likely the lens correction designed to adjust for the distortion caused by the lens itself. Then you have the various filters designed to map that raw sensor data to color data, which can be tweaked to adjust the white point and the overall contrast and brightness of the scene.
The final photo will be based on the raw sensor data, but it won't be the raw sensor data. What most people would call the "picture" a DSLR takes isn't what the sensor sees. It's the sensor data that's been sent through a whole lot of post processing to produce a final image, not entirely unlike what a phone does.
Re: (Score:3)
I suspect that thegarbz knows this: most people who shoot RAW do so because they want to control the post-processing, so they see all the filters that go between the sensor data and the final image. It still feels like a refutation of
when I can expose the sensor and process the data it records using filters with only local effects and w
Re: (Score:2)
The same thing is true of any film print. The only time you see exactly what's on the film is projected slides from slide film. Any printed photograph involves a complex process to convert the original film image into what winds up on paper. Not to mention that a lot of changes can happen by careful choice of development process. Again you have an image that's based on the light that came in through the lens, but there's a complicated process for converting that light into the final image.
You should r
Re: (Score:2)
Generative AI and AI in general allows you to manipulate photos in a way to create "new reality". For example,
Re: (Score:2)
Even an AI doesn't know what Taylor Swift looks like unclothed, so the best it can really do is substitute in the body of someone it thinks is similar and stick her head on it -- which is pretty much exactly what a human would do to fulfill Rule 34.
Re: (Score:2)
Re: (Score:2)
Yeah, right! (Score:2)
And you can keep going with that line of devolving too. Our eyes are making filtered images. And our brains also interpret the optical nerves. Colour is an illusion, which is true. Making the argument that everything is filtered therefore nothing is real is one of those silly rabbit holes.
A straight photo is real.
Re: (Score:2)
It is, indeed, a "silly rabbit hole". But it's also valid. The idea of "real picture" is based on film with negatives, but even then there was lots of pre- and post- processing. There really *is* no such thing as a "real picture" except in the sense of "this is a picture, and it exists".
Perhaps the appropriate thing to do is let go the idea of "real picture", as it was always a gross oversimplification, and with electronics based cameras doesn't have a real touchstone anywhere. "Seeing is believing" and
Re: (Score:3)
It is, indeed, a "silly rabbit hole". But it's also valid. The idea of "real picture" is based on film with negatives, but even then there was lots of pre- and post- processing. There really *is* no such thing as a "real picture" except in the sense of "this is a picture, and it exists".
Perhaps the appropriate thing to do is let go the idea of "real picture", as it was always a gross oversimplification, and with electronics based cameras doesn't have a real touchstone anywhere. "Seeing is believing" and "Pics or it didn't happen" were always more about human psychology than objective reality (though of course they were about both).
OTOH, this is rather good grounds for rejecting any photo from a Samsung camera as evidence.
There is the matter of images that were never photographs however. Think of an analog of Dali's persistence of memory. So we get another turd in the punchbowl of "what is this AI art thing?"
In fact, that these images exist merely means we have to admit that they exist. (now that was profound, eh?)
New technology and imaging's part in it has always been an issue. At one time, photographs which were all created by chemical processes at the time, was not considered to ever be art. And then there was surre
guy is talking out his arse. (Score:2)
This is a loss (Score:4, Insightful)
What about an analog picture? (Score:3)
Would that be considered a "real" picture? (analog meaning you shoot the picture onto film, and let's assume no auto-focus or auto-exposure)
Is that the limit of what makes a picture real?
What if you use a digital camera but turn off all assistive features? Is that a real picture?
Sounds like this guy is trying to blur the picture (sic) of what is acceptable, so that Samsung can feel more comfortable using AI (or whatever) to automate as much as possible.
I wonder how long it will be before we talk to our phone/camera and say (my dog) and it seeks out your dog in the frame, takes a picture, manipulates to some absurd extend, and presents it as reality.
Re: (Score:3)
The thing which strikes me as weird about the quote
is that the first half seems to suggest that if you don't use AI at all then it is real, and then the last sentence seems to walk that back. Is the AI part of where he draws the line or not?
Re: (Score:2)
>Would that be considered a "real" picture? (analog meaning you shoot the picture onto film, and let's assume no auto-focus or auto-exposure)
If you want to be extremely technical, no. Film is designed to only capture the visible light portion of the EM spectrum, unless you go for specialty films like the ones that capture Infra Red... and don't capture parts of the visible light spectrum.
Even RAW files are not "real", as he says. They are not just sensor dumps straight to the SD card / NAND. There are al
Re: (Score:2)
--So the image is already changed from the real world before it even reaches the initial sensor.--
OK, so is there such a thing as a "real picture" and, if so, how is that defined?
I tend to go with the interpretation that "real" implies minimal manipulation of the original capture, though I suppose 'original capture' can now become abstracted from the original human intent.
So is it about purity? Intent? Not ceding control to another 'entity'?
Messy, to be sure
Re: (Score:2)
Did you put a filter on your lens? Either for UV purposes or for colour adjustment? Did you go black and white, or even sepia? Did you go with a high shutter speed and fast lens? Did you use a fisheye lens?
Smells like red herring (Score:3)
I take real photos with my Pentax K3-II DSLR (Score:3)
No wonder.. (Score:2)
With executives like him, I can see why Samsung profits dropped by 95%.
Unintended consequenses (Score:3)
In a post-truth era, nothing is real. (Score:4, Interesting)
He's not wrong, on a certain level, but he's taken it to an extreme that would have been laughable twenty years ago. And now? It's grounds for serious debate.
I think we're living in a new era. An era that's going to be hard to define in coming generations. We've separated into different realities, yet still live among those who "subscribe" to different interpretations of the same subjects. Pictures aren't real, facts aren't real, opinions are, and fantasies are so much more than real.
I wonder what historians will say about this era, if there are any? Will they admire our imaginations while pitying our inability to disengage them long enough to form a consensus? Or will they simply pity us outright, for being unable to engage in the reality we all share at a strong enough level to maintain stability?
I remember growing up wanting there to be something to happen. Something to shake things up. Something different. Something Interesting. Be careful what you wish for. When the world grants your wishes, it rarely works out the way you'd want it to. These are interesting times. Painfully so.
Re: (Score:2)
Re: (Score:2)
>He's not wrong, on a certain level, but he's taken it to an extreme that would have been laughable twenty years ago
I think you mean he's using the the exact same arguments the old blowhard film photographers used when Digital first came out? You must be too young to remember that I guess.
No, I was there for that as well. But, somewhat rightly so, people railed against it at the time. I witnessed the same with the development of the DAW (Digital Audio Workstation) and recording via computers vs. tape. Same rage. Some of it warranted, some of it not, much like with photography.
Now? It's more a matter of philosophical debate. We've moved the goalposts so far away from reality that reality itself is now open for debate.
And I am sure (Score:2)
Wait for it ... (Score:2)
So what if (Score:2)
Re: (Score:2)
I was playing with the AI upscalers and thought at a glance they did pretty good, but when I pushed them too far, they all looked more like paintings or cartoons rather than plausible content when given photorealistic input.
Thought it was interesting because I didn't recall AI output being described as 'cartoony' before your commennt but it resonates with my experience where AI upscaling at some level turns into cartoon looking detail even today.
Re: (Score:2)
AI upscaling and denoising works best on line art/anime art, because it's much simpler than actual photos, so it's easier to predict how the lines and curves will extend.
remember photocopiers changing numbers? (Score:3)
https://www.theverge.com/2013/... [theverge.com]
"the underlying compression technology used by the machines is to blame. "The Xerox design utilizes the recognized industry standard JBIG2 compressor which creates extremely small file sizes with good image quality, but with inherent tradeoffs under low resolution and quality settings," explains principal engineer Francis Tse.
Further, he points out that Xerox already advises customers of this possibility in its software, describing results from normal compression settings as "acceptable" while warning that "text quality degradation and character substitution errors may occur." The company says upping the quality level to "higher" should guarantee the copy accuracy its customers expect."
Exciting! (Score:2)
I can't wait to try this while purchasing a phone from Samsung.
"That will be $1100, sir"
"Yeah, but what is money anyway, but a promissory note of value delivery? Here, take this cocktail napkin I've written 'IOU' on. It's the same thing, see?"
I don't know how you can disagree. (Score:5, Insightful)
I know it's philosophy more than logic, but sure. I was sitting at a campfire late one night last summer, and pointed my iPhone camera straight up into what I only saw as black trees, with a field of stars above them. When I looked at the picture, the branches were lit up subtly and beautifully. The phone had picked out the details and shown them wonderfully. That's NOT what my eyes could see. I didn't do anything special. Point and click.
The phone is much better at making pleasing images than representing reality... but very seldom do you want the picture to represent true reality. Certainly not when it's your phone. Reality usually sucks in comparison.
Re: (Score:2)
Re: (Score:3)
Why is it sad? I spend considerably more time in nature than the average person here, and almost all of it without a camera (or a phone) within reach. May to October, I'm out of cell coverage about 40 hours per week.
The idea that your memories travel with you till death is, frankly, rubbish. They morph over time, and can't be trusted almost immediately. A great many people will see the memories vanish far before their bodies pack it in.
Re: (Score:2)
Your visual system does not give you a perfect rendition of reality. It has limitations as far as brightness and contrast ranges, that, as you experienced, contemporary digital imaging has been able to expand beyond.
But, oddly, we don't think of the myriad of transformations and distortions that our eyes and brain impose on our view as misrepresenting reality, even though we can clearly demonstrate cases where that is true (search for visual illusions and you'll have examples by the dozen). We understand
Re: (Score:3)
That's a great example. If you'd spent ten minutes looking up into that black sky and not looking into the fire, you would have been able to see the branches lit up.
Is the photo not real because it didn't represent what your eyes saw at that moment? Is what your eyes see after ten minutes not real because it more closely resembles what the camera saw? Did you notice that your brain edited out your nose and blind spots?
That family photo album? (Score:2)
That family photo album?
"it doesn't mean anything" - Samsung EVP Patrick Chomet
Heartless bastard.
Re: (Score:2)
Re:That family photo album? (Score:4, Interesting)
Well, let's extend the argument a bit. You gather your family together. Dad lays off the booze for 24 hours, mom carefully applies makeup to hide the crow's feet, and puts on a long sleeved shirt to make sure the bruise on her forearm isn't visible. The son combs his hair for the first time in six months. The daughter takes off the goth makeup and braids her hair. Everybody huddles close together despite the fact that they haven't had a real conversation in months, and they all smile. A picture is taken.
Is that picture real?
If so, what do you care if the colours pop better and the skin looks better than the lighting should have allowed?
No True Scotsman (Score:2)
There is no true Scotsman. ( https://en.wikipedia.org/wiki/... [wikipedia.org] )
This is a classic.
I have worked with raw sensors... (Score:2)
...not RAW files produced by cameras, but the electrical signals from sensors going into custom electronics
There is a lot of processing required to make the signals into a picture
It's not a yes/no question, it's a spectrum, from the minimum processing required to make an image to full on artistic manipulation
Very Seldom Do You Want the RAW image. (Score:4, Interesting)
Weddings, sports, nature... if you do any of these things professionally, you have your preferred set of tools. I like Capture One myself. But everybody who takes digital photography seriously processes the output. Nobody wants the unprocessed raw - especially if they're buying it from you. That's because that little window of time where the place you're standing is lit perfectly is fleeting (if it exists at all). The golden hour isn't guaranteed, and it certainly isn't an hour.
Go ahead and try it if you like. Turn all the automatic settings off on your camera of choice and go straight to RAW. See how that works out for you. Even Ansel Adams produced some duds - and his were expensive.
Basic rule (Score:5, Interesting)
If it isn't an optimization to more accurately represent the photons that struck the CCD, it's a degree of fakery.
Anything that removes information - compression, adjusting brightness or contrast for instance - isn't fake, it's a trade-off. In some cases this can be abused in a way that should be called out as fake, but as a general rule it's fine.
When you alter the image in a way that adds information, details that were never captured by the sensor in the first place, that's fake. And that's not a grey area, that's an absolute.
2024 Bingo Card (Score:2)
Gotta admit, I did not have "representative of a major corporation getting deeply philosophical as an attempted legal defense" on there.
Y'all are missing the point, it's not real vs fake (Score:2)
Talking about what is real and what is not real is exactly the diversion that Execute Vice President Patrick Chomet was trying to get you stuck on. If you take a picture of a light bulb or a white plate, and the camera shows you a picture of the moon, then that is a FRIEKING MORONIC STUPID BUG. His comment should not spark a debate about real -vs- fake, because it was neither real nor fake, it was just wrong.
The real discussion here should be around why we tolerate corporations lying and spinning, and why
Terrible! (Score:3)
CoPilot is arguing that Chomet's claim that reality is subjective is wrong because our interpretation of reality is subjective. Awesome!
Re: (Score:2)
Certainly gives pause to a serious hobbyist.
Re: (Score:2)
Then he'll get into "yes, the AI is trying to represent reality as best it can" It sees a tiny far off moon and recognizes "moon" and substitutes actual moon imagery oriented to be consistent with the vague shapes it can latch on to. Yes, it's not specifically the moon precisely as it was when you took the picture, but close enough to match all details it could see and all added detail is consistent with known reference images. In this philosophy, it did fail as it dramatically changed the color of the i