You're explaining it well, and I get most of your argument. But your model assumes equivalent photos, don't you see the problem here? If you cancel out the difference in photos at the start of your equations then your model only looks at one aspect, you reduce it to just what the camera does:
But it isn't another matter. Your theory has already cancelled out:
The differences in how we use these cameras; the places where we take them, how we set up and compose, time it takes, serious or more relaxed and informal, our attitude or mindset when using them, the goals of what we are aiming to achieve with the photography.
How we react and respond in front of these cameras, how relaxed and comfortable we are with both photographers we know and ones we don't.
How our understanding of how we create photos impacts how we interpret and view other photos.
How viewing photos within the context that they were taken and outside that context, or how the difference from our expectation of what we expect to see alters our response. You already touched on this, take a photo with a smart phone and print it big you're "oh, ok.", say it's from a smart phone and it's "WOW!" Do the opposite and put your Z9 photo on your phone and it's "WOW" explain what it is and it's "oh, ok."
The length of time we take to view these photos, and the context in which they're presented, from social media to art gallery.
It's a very neat line you draw, that the photo is a constant entity because it can be measured as such and so equal IQ is just matching those measurements (not necessarily the content). But it also assumes that within limited parameters that the human eye will see an identical photo as identical. Which simply doesn'thold true in any situation and only holds mostly true if you view them together and flick between them. Viewed independantly all the above considerations come into play, with different content then there is no equivalence and their assumptions of which camera was used will be based on the considerations above and not a comparison of IQ.
Equivalent photo again... Why on earth would you want to use all that technology to reproduce something you can already easily do? All you do is reduce the value of the photograph and with it the desire for photographers to take that photo, it will no longer stand out.
Why not use that tech to produce that sports action photo that appear to have been taken at a much closer distance with a wider lens? Change the perception rather than the IQ. Better still you could combine this with a "montage of moments" clips encompassing the scoring of the goal/touchdown from the pass to the shot/touchdown. And being video all IQ equivalents are completely moot.
You see the problem of limiting the discussion to "equivalent photos"? Basically you are only comparing what the camera already does.
i agree with you to a certain extent, ive seen images taken with my daughters iphone of us out to dinner and it leaves my FF images for dead. its not from AI but from stacking multiple images. but a phone with not replace fine skin tone detail no matter how many images it stacks. because small pixels have very low DR and tonal data that cant be generated realistically.
I simply replied that you wouldn't choose one format over another due to the "relationship between shutter/aperture/field of view (the compromise between DoF and shutter speed if you like) and the visual effect this creates." Rather, you would choose a larger format over a smaller format for higher resolution, less noise, higher DR and/or the option for a more shallow DOF.
Exactly, which is why you don't choose one format over another to shoot equivalent photos. You choose one format over another because it can either take photos another format can't or because it can produce a higher quality photo than another. In short, assuming the differences in operation are negligible or unimportant (← that's a critical assumption), you choose a larger format for higher resolution, less noise, more DR, and/or the option for a more shallow DOF, and you choose a smaller format for smaller size, weight, and/or price. However, going back to my critical assumption, for sure, you might choose one camera over another because it has superior operation (e.g. faster and/or more accurate AF, higher frame rate, focus stacking, etc., etc., etc..), but that has nothing to do with the format, per se, but the particular camera bodies in question. But assuming, as I said, the differences in operation were negligible and/or unimportant, then it is just as I said.
Again, it did not. I never said (or implied) that you were obligated to take equivalent photos with different formats. However, lenses and differences in operation permitting, if the best photo using a larger format is equivalent to the best photo from a smaller format (same perspective, framing, DOF, and exposure time), then, of course, you would use those settings. You wouldn't, for example, use non-equivalent settings on the larger format if the photo were worse -- that would be ridiculous! But you absolutely would shoot non-equivalent photos on a larger format if the non-equivalent settings resulted in a better photo than you could get with a smaller format (and, indeed, this right here is the primary reason to choose a larger format over a smaller format).
We would only use the cameras differently if there were difference in operation between the cameras. For example, people use smartphones differently than dedicated cameras for exactly that reason: it is smaller, lighter, better connected, and has better built-in processing than a dedicated camera and we carry it with us regardless of whether we are going to take photos.
It is true that for people photography, the physical size of the camera and lens may provoke a different response from the people we are taking photos of. But, again, this is a difference in operation. But I can't see anyone reacting differently to someone with, say, an OM3 + 12-40 / 2.8 and someone else with a Z6.2 + 24-70 / 4. However, they absolutely would react differently to a photo from either of those vs 5x4 on a tripod or a smartphone.
Sure -- if you're a photographer talking with a prospective client and trying to sell your work to them, and going into all the details of how you created the photo etc., etc., etc., this will have an impact on how someone views the photo (the family friend I visited at the art show has explained how much work it is trying to sell the photo to people who "have more money than they know what to do with" and why they should spend the money for his photo).
But in the case of my story of showing the photos on my smartphone and the person being surprised at them thinking I took the photos with a smartphone, this plays no role, whatsoever. I was only showing them the photos because I was talking about a place I had been to and was going through my online gallery on the phone for the sole purpose of showing that place to them. The discussion was not even in the least bit about photography until they unexpectedly brought it up.
I really am not communicating well! You keep assuming, over and over, that I'm talking about equivalent photos. I am not. I am saying that there is a reason that I use a modern 45 MP FF camera with nice lenses rather than a smartphone (not that I don't also use my smartphone to take pics, of course, since I don't always have my camera with me!). I am saying that, in the near future, it is my opinion that more advanced AI software than is available now will severely diminish, or even completely cancel out, the reasons I do use a modern 45 MP FF camera (other than my personal connection to the photos that I take is from an actual photo, not an AI reconstruction, no matter how "realistic" or "pleasing" that AI reconstruction may be), which, aside from "the shooting experience", forensically accurate photography, and/or "authentic" photography, remove any "need" for dedicated camera systems.
Any and all! I think all of that will happen, and happen soon.
I don't know what I said that makes you think I'm talking about equivalent photos -- I'm doing nothing of the sort! The only reason I even brought it up is to rebut your statement of using 5x4 and [35mm] FF different based on "relationship between shutter/aperture/field of view (the compromise between DoF and shutter speed if you like) and the visual effect this creates" as opposed to difference based on resolution, noise, DR, and/or [potentially] more shallow DOF. And, of course, I should spell out that there are obvious differences in operation between a 5x4 film camera and a FF camera, which all by itself, makes the smaller format the superior choice for most people, just as differences in operation (which includes size and weight) make the smartphone the camera of choice for so many over dedicated camera systems.
But nothing I am talking about (aside from the above paragraph) has anything to do with Equivalence, whatsoever. It does, however, have everything to do with resolution, noise, DR, and/or the option for more shallow DOF, and, again, I'm saying that if more advanced AI software can completely close the gap between the smartphone and dedicated camera (by introducing "realistic" and pleasing detail, albeit fake, into the photo, making for an end result that looks like you had taken the photo with that super "high IQ" dedicated camera system).
I also want to say, again, that I am not saying that "high IQ" or "realism" is necessarily the end goal. I'm saying that "high IQ" (and/or the option for more shallow DOF) is why one chooses a larger format over a smaller format. If "high IQ" (and/or the option for a more shallow DOF) did not matter, why would anyone shoot with a more expensive larger format over the least expensive / most convenient option out there that gave the operation that was needed? Why would anyone buy an f/2.8 zoom over an f/4 zoom when the latter is so much smaller, lighter, and less expensive than the former? Why would anyone use an f/1.2 prime over an f/1.8 prime when, again, the latter is so much smaller, lighter, and less expensive than the former?
Are you saying it's simply ego and bling that people choose FF with f/1.2 primes and f/2.8 zooms over mFT with f/4 zooms? I mean, operationally, the OM1 is superior to my R5 (albeit not in ways that matter to my photography), and an OM1 system would be smaller, lighter, and less expensive than my FF system. What about a Sony R10, then? Surely that's even better still, right? If differences in resolution, noise, DR, and/or the option for a more shallow DOF don't matter, that is. And finally, why the R10 when a smartphone can do it all, too, with advanced AI software that will make its photos looks every bit as good as what I can get with any of the aforementioned photos? Because people will react differently to my AI enhanced smartphone than they will react to my dedicated camera? I will have to disagree.
Resolution in the photo. What I mean is, for example, that a photo taken at 50mm cropped to the same framing as 100mm will have half the resolution as a photo taken from the same position at 100mm, given lenses of comparable quality used on the same camera.
Yes -- that internal stacking ability of smartphones give them a substantial advantage over a single exposure from larger sensor cameras in some situations. Same may apply to how they stitch together panoramas. That is, a panorama photo made with a smartphone may well have "higher IQ" than a single exposure from a dedicated camera. That said, the sunset photo I posted in the thread was a panorama taken with my phone, and I'm thinking a single exposure from my FF camera would have produced a "higher IQ" result.
Not now, of course. I'm saying that in the near future, I'm thinking they will be able to. Not by stacking more photos, but by using AI to analyze the photo taken and "enhance" it with fake, but very "realistic" detail, that will be even "better" than what you could get with any currently available consumer camera. That's the entire premise of this thread!
maybe shooting multiple exposure images and stacking them could get closer . i know everytime i look at my daughters 12 meg phone files from the past 8 years they are very good. and even my cheap phone does a better job than my xz1 can.
Sure. Assuming motion blur is a non-issue, then you can just keep stacking until you reach any desired level of "IQ" (although, with regards to resolution, you will be limited to the resolution of the lens and sensor). For example, the iPhone 16 has a 48 MP sensor with a diagonal of 10mm, resulting in a 4.3x crop factor relative to FF. Assuming the lens resolves relatively as well on the sensor as a FF lens resolves on a FF sensor (i.e., the iPhone lens is 4.3x as sharp as the FF lens), then if you stack 4.3² = 18 photos on the iPhone 16, it will produce a photo that is "as good as" a single exposure from a 48 MP FF sensor (all else equal, of course!). However, that's only for 26mm FF equivalent (obviously, if you use a 50mm lens on FF, the iPhone will have to crop 2x, so the resolution will be half, and you'll also have to merge 4 x 18 = 72 exposures to match the amount of light making up the crop to match the noise/DR).
That's very doable for lots of scenes, by the way, if, of course, you shoot lots of scenes around 26mm f/5.6-8 FF equivalent, not the least of which are landscapes. I don't know if the smartphones allow you to do this now (I should look into it, as I'd use it a lot!) or if they only stack for certain scenes automatically outside user control. In other words, the technology exists right now, for smartphones to match FF level IQ in certain circumstances.
However, what I'm talking about is AI processing that is significantly more advanced than what we have now that will let smartphones match the IQ of FF in any and all circumstances, even without stacking/stitching, albeit not forensically accurate, but "realistically accurate" in that, looking at the photo, you would not be able to tell if it were from a smartphone or FF camera, regardless of the scene or display size. As you note, in some circumstances, that is true right now, but it is only in a limited number of circumstances. I'm talking this happening for pretty much every circumstance in the near future, perhaps as soon as 5-10 years.
This is what I mean by "photography might die soon". Right now, with my FF camera, I have to think about what lens I am using, what aperture and exposure time I am using, what ISO setting I'm using, and where my focal point is. In this future I'm talking about, I won't need to be concerned with any of that. I will be able to decide the DOF and framing after the fact -- the software will make for convincing shallow DOF (if that's what I want) and crop to any framing I want filling in with "realistic" detail as needed to produce an output file of any "resolution" I want.
The artistic portion of taking the photo might change a bit, because I would no longer need to "fear" resolution loss due to cropping. So, I'd probably frame much wider than I normally do (use the widest angle the smartphone allows) and crop after the fact. That could result in my choosing different compositions rather than being limited to compositions close to the focal lengths available with the lens I am using. However, the processing will be no different from how I do it now with my RAW converter and photo editor, should I want a different look than what the camera gives me by default. Indeed, as I was saying earlier, I could even just shoot video, pluck out a frame from the video that I liked, and do the same with that frame as I would do with a single photo with no "IQ penalty". What I'm saying is that I could take a photo with a smartphone, say I used 5x4, process the photo accordingly, and no one would be the wiser.
That all said, of course not all photos would be equal. I already mentioned the ability to frame wide and crop after the fact with no resolution/noise/DR penalty might lead to me choosing different compositions. With regards to the processing of the photo to get the look that I want, nothing would change (other than having more options with regards to framing and more options with regards to DOF).
What I'm saying, is that here's a compilation of photos I've take over the years. If I had the tech I'm talking about, I could have taken all of those with my smartphone. I mean, not exactly the same photo, mind you, but if I showed you the original gallery and the new gallery from when I time traveled back with this new smartphone tech, you would not be able to tell me which gallery of photos was from the dedicated camera and which was from the smartphone better than a 50-50 guess in the same way that if I time traveled back and used the same equipment I used back then, you wouldn't be able to tell which gallery was the original and which was the time traveled gallery better than a 50-50 guess, despite the photos being different (unless, of course, my "photographic eye" has changed over time more than I think it has, or the "framing freedom" had more effect on the photos I take than I thought it would).
In this you describe what cameras are in an logical way, not how we as humans use them in an illogical way. What you say is true if you look at what a camera is, but again you fail to allow for what happens when you place them in human hands as being a relevant part of defining the resultant photo. Explain the selfie if all cameras are equal, it certainly had more impact on the direction op popular photography than anything I've done... 😀
Besides, if your advanced understanding is that all cameras are capable of taking the same photo then it is in you mindset to confirm that with the photos you take, not disprove it. You can't tell me this has no impact on the visual look of your photos.
The whole argument here is a comparison that is equivalent to what exists, a future photo that is equivalent to one that exists today and the yardstick is an understanding of IQ as limited by current sensor performance.
Yet when Ai can duplicate the IQ of a Z9 then we are already at the point where photography is free of the constraints of camera design, image IQ will no longer be relevant, or a yardstick by which to compare performance or even have any relevance to photography at all. And the photos you produce with the Z9 and the way it abstracts will become nostalgia as "The Digital Look" much in the same way film is viewed today. It seems illogical to me that we would develop new technology to replicate what we already do other than at some point develop it as a nostalgic filter in the same way current smart phones imitate film.
Can't you see that in your OP, by using the current limits of large sensor camera IQ to frame the shortcomings of current smart phones you're already restricting the development of the technology. Because you keep it within the equivalence framework you are using the old technology to define and to some extent predict the capabilities of the new. To use a restricted example it's like the film photographer explaining the film ISO can be applied directly to digital whist insisting that 35mm (the largest format that could be realistically portable and handheld) didn't create the explosion of travel shots, or the limitations of fixed ISO film didn't promote the use of shallow DoF action, panned racing cars, gritty pushed B&W realism. Or that the arrival of 100ASA roll film didn't create the whole genre of amateur family snapshots.
I think it's worth a bold but don't want to shout:
If a phone can match any sensor's performance then image IQ will simply cease to be a relevant measure. Photography will cease to be rooted by the reality that a current sensor can record.
I'll leave it there as I can't add anymore, but I will read any response because I'm genuinely interested it in a considered opinion that's so different to my own.
it all sounds great but you missed on perspective distortion which AI will never be able do especially on faces and people. i also noticed my daughters iphone 6 was using iso 25. i also have noticed that AI is "fake" its not intelligent its copying and not creating. there are no images like my extreme macro anywhere in the world so AI has nothing to copy.
Ah! Yes, I am presuming that people use cameras as tools with an understanding of their abilities and limitations. 😅 And I can explain the "explosion of the selfie", too: it's the differences in operation, not IQ, that made it a thing. And I've long said that the differences in operation between cameras is far more important than the difference in IQ between cameras for most people in most situations. I'm now adding to this statement by saying that I believe the difference in IQ between cameras may soon become irrelevant for pretty much everyone in every circumstance, aside from those who seek forensic accuracy and/or "authentic" photography.,
YES!!! That's what I'm trying to say!
I think we've reached an understanding, actually! I think the whole disconnect was that I was primarily discussing cameras from the standpoint of IQ by emphasizing the assumption that differences in operation were negligible and/or unimportant, and that the people using the cameras were using them with an understanding of what a camera actually does.
On the other hand, you were placing differences in operation and, more importantly, peoples' perception, often illogical, of what a camera is and what it does based on the particular camera, where both of these factors influence the types of photos a person would take (e.g., the explosion of selfies with the advent of the smartphone). To add to the selfie example, I think you are saying the very fact that you mount a 4x5 camera on a tripod and use a prime lens to photograph a scene [strongly] influences they types of photos you will take, whereas I was coming from the opinion that if a smartphone could take the same photo IQ-wise, then you'd take the same photo, anyway, amongst other photos you couldn't take with the 4x5 on a tripod.
Perspective distortion comes entirely from the distance between the subject and the camera. In other words, if you take two photos of a given scene with any camera-lens system, crop one or both photos to the same framing, the perspective distortion will be identical.
That said, I believe that AI, in the near future, will be able to make it look like a photo taken from 5 meters away with a 50mm lens, for example, will have the same perspective distortion as a photo of the same scene taken from 20 meters away with a 200mm lens, and vice-versa. Will the photos be identical? No. But displayed side-by-side, I am saying you will not be able to tell which is which better than a 50-50 guess.
Curiously enough, fora often bandy about the acronym "IQ" without qualification as to what kind of IQ is meant, instead vaguely meaning everything.
However, there is actually a Standard IEEE 1858 forCamera Phone Image Quality (CPIQ); but I know of no such comprehensive single Standard for "normal" cameras, only the many related expensive and inscrutable I.S.O. Standards e.g. "Dynamic Range".
So, with no common standards of quality how would one compare a phone to a 45 MP FF camera?
If one were rich, one could just buy and apply the CPIQ-equivalent I.S.O. Standards to the phone, which might not be all that easy ...
depends who is looking at the image. maybe i should set up a studio head shot with my 180mm and a 28mm lens, take the image into Ps and see if I can actually apply lens correction to the 28mm image to match the 180mm 🤔
I just set my studio up and shot a head and shoulders at 28 mm then at 180mm ,as i predicted its impossable to correct a distorted 28mm portrait to match the 180mm perspective. doesnt matter what tools you have in Ps . even a complex action does not work because the image taken is from a 3d subject . i will now take the image into portrait pro to see what i can do.
edit. just took the image into portrait pro, it has extensive tool range compared to Ps and is industry standard software. 😂 im afraid i could make myself look like
Pearce Bronson 😁but could not produce an original 180mm perspective .
When I say "IQ", I tend to mean resolution, noise, and DR as the main sources. However, there are other considerations, such as CA, flare, bokeh, etc., and there are elements of IQ that are tied to the operation of the camera, such as focus accuracy. So, kind of a weighted "vaguely meaning everything" definition. 😁
More seriously, I think higher resolution, lower noise, and higher DR are universally seen as good things, because if you want lower resolution, more noise, and less DR, you can get that from the "higher IQ" photo, but you can't do the opposite without "side effects". Some elements of IQ, such as flare and bokeh, can be subjective in that one photographers distracting flare, for example, is what makes the shot better for another photographer. Actually, since I bring it up, let me give a cool example of that:
Isn't that the coolest flare you've ever seen? None of that photo was edited in. It was using the Tamron 150 / 2.8 macro, which would do that in certain circumstances.
What's wrong with the whole "IQ means what I want it to mean -- no more, no less" definition? It makes for more lively discussions. 😇
When the actual pixels on screen are mostly software generated then comparisons between the kit will be totally redundant. The only measures will be subjective, which image you prefer. And I can guarantee that what we prefer will not be along the lines of any technical accuracy with the base data or follow the science of how things should be technically correct.
For instance:
Why not? And why would it have to follow the exact perspective that controls how current lenses render on sensors? Once free of that connection you can do what you want, or the public prefers. With faces everybody has a focal length that best suits their features, where the compression/extension works with their features rather than against it. So if you want to look your best then teach your phone how to do it, which will be a simple as taking photos at different distances and then; choose what you prefer/let Ai choose what is the closest to the currently trending look/invent your own look. Three distances and five angles at each distance should do it, or possibly just access to your photo library...
In other words instead of trying to duplicate how the perspective rendering of a face changes with distance you could instead cancel that out and use you most flattering perspective in all photos regardless of distance.
