If you were to say what has been the hottest camera brand on social media in recent years, I think the answer would undoubtedly beFuji (Japanese company)In 2024, when the Fujifilm X100VI was launched, the first batch of goods was snapped up by millions of reservations, with a premium of up to $5,000 or $6,000, enough to buy a new camera, which was called “electronic Maotai”.
Fuji’s secret sauce is the unique “film simulation” built into its cameras.
Many people think that Fuji’s film simulations are no big deal, and that they can perfectly replicate Fuji’s filters by pulling curves in Photoshop or applying a proven filter.
In reality, however, even if you don’t consider doing post in the computermore trouble than it’s worthThe thing is, Fuji’s film simulation is also far from simple; it’s based on Fuji’s experience with film color tuning, playing withColor ScienceIt can be said to be the only one of its kind.
After all, so many online analog Fujifilm parameters and filters, from cameras to cell phones all have, but still did not shake Fuji’s market position, hard to let Fuji from the roll parameters of the camera market, killed a special track. If you do not believe, you can also rent a Fuji back to try.
I tried it and it really doesn’t work ……
To figure out what’s so unique about this color science, or why others can’t reproduce it, you have to start with Fuji’s old standby, which is film.
How film records images
Film is one of those things that you look at as something that’s almost obsolete, but actually to this day it is stillTop technology products. Especially color film, the world can produce and manufacture this thing countries, than can make nuclear bombs are less. China only mastered the color film manufacturing technology in 1985, following the United States, Germany and Japan.Fourth country to have this technology. Presided over the development of Zou Jing is China’s first batch of academicians of the Academy of Engineering, she died, the obituary deliberately used a color photo, is to commemorate her achievements.
That’s a bit far, let’s get back to film and Fuji.
Film is used to image and develop color through chemical reactions. On color film, there are light-sensitive coatings that sense blue, green, and red, respectively, and each layer is coated with an emulsion that containsSilver halide crystals responsible for light sensitivity和Coupling agent responsible for color development, these component particles are suspended in gelatin. When rinsed, the silver halide is washed out, presenting light and dark information, and the coupling agent reacts to form different colored dyes.
Specifically in actual film products, color films tend to have more light-sensitive layers in order to better absorb light. For example, Kodak’s Kodacolor 400 color film, which came out in 1977, had six photopolymer coats. There are only more photopolymer coatings on the color films you can buy today.
! [Diagram of the structure of Kodak Kodacolor 400 color film](https://cdn.jsdelivr.net/gh/p0werdown/imgpost@master/uPic/柯达Kodacolor 400 Color Film Structure Diagram.jpg)
It’s logical that the light-sensitive coating for each color should ideally absorb only the color light of the corresponding color, but the reality is not that perfect. Take, for example, Fujifilm’s two famous color films, Provia 100 and Velvia 100. These are their “Spectral-Dye-Density Curves,” which show how the film absorbs different wavelengths of light.
Theoretically, on a perfect film, the curves of the three colors will not overlap. But, as we said earlier, in reality there are always errors, which can cause the colors of a photograph to be less than “pure” and to “contaminate” each other, and the degree of this “contamination” becomes, in a way, a feature of the film; for example, there is much more “contamination” between the two color dyes on the Provia 100 than there is on the Velvia 100.
In addition to its own characteristics, the film should be developed after shooting and color development, this process also involves a very complex chemical reaction, the content of each component, the ratio, reaction temperature, time and so on have very high requirements, a little carelessness will be off-color. This point, played the color film enthusiasts should have a deep understanding.
And this, still only from the macro level to analyze the principle of film imaging, already feel a little head; if we narrow the perspective, from the micro components to analyze the role of a specific kind of material, you will find …… more complex.
Complex film composition
We spoke earlier in the emulsion coupling agent, silver halide, gelatin, etc., in fact, they are a large class of chemical substances collectively, look into it, can be broken out of several or even dozens of different substances, each one a little more or less, will affect the final imaging of the color. Film manufacturers may also be mixed with some less than humane special chemical ingredients to adjust the color of the photo. It can be said that the formula of emulsion is the secret recipe and unique skill of each film factory.
What is the impact of this change in composition? For example, in 2021, the U.S. Environmental Protection Agency (EPA) announced that it would ban the manufacture and sale of chemical products containing the substance “isopropylated triphenyl phosphate (PIP (3:1))” in the United States. A Fuji film called Velvia 100 just happens to contain this substance, butVery low percentage, less than 0.0003 per centThis is roughly equivalent toOne PM2.5 particle in a small watermelon 10 centimeters in diameter.
It stands to reason that such a small percentage shouldn’t have any impact on the environment, and it doesn’t seem like it would be a big deal even if it were removed. But Fuji is still very serious about complying with the U.S. bill, and is rather on theThis film has been taken off the shelves all over the United States.alsoNo recipe modification. Also, Fuji has more than one film in their Velvia line, but only this one was eventually taken off the shelves, suggesting that the other models do not contain this substance.
It follows thatEven if this is less than 0.0003%, it will have a visible effect on the final imaging result.But only Fuji knows exactly what the impact is.
And that’s the key to Fuji’s ability to play with film analog.
Adobe doesn’t know chemistry, but Fuji does.
Fuji is an old player in the film industry, it was founded in 1934 and has been making film for 90 years to this day, with a deep understanding of the principles and intricacies of film imaging.
On the one hand, the complexity of film imaging itself makes it difficult for post-production software such as Photoshop to reproduce it by applying filters and pulling curves.
The principle of the filter is very simple. It is essentially a color chart that maps color A to color B according to some algorithm, thus changing the color of the whole photo.
Photoshop’s color curves are a little trickier, but essentially they’re just adjusting the red, green, and blue tones separately. But as we mentioned earlier, theWhen film is imaged, the three colors “interfere” with each other.Photoshop is a software, the software runs according to the algorithm, and the algorithm is designed by people. If the person who wrote the algorithm does not understand the characteristics of the light-sensitive layers of imaging and color development, then how to tell the software how to adjust it?
On the other hand, film emulsion formulations are reallyIt’s too complicated., each component affects the final imaging result. Like the example we talked about earlier, theLess than 0.0003% compositionIt could make Fuji choose to give up a country’s market outright. Even if we did, by some means, obtain a detailed recipe and formulation process for Fujifilm emulsions, it would be very difficult to work backwards from the results to the relationship between the different components and the color – after all, sensitometry and chemistry are still both complex disciplines, and it’s difficult to manufacture color film even positively, let alone backwards.
So it comes back to the question above: how can the people who write the software and make the filters use the algorithms to simulate the effects of film when they don’t understand chemistry and don’t know what each ingredient is for and how it will affect the final image?
Adobe doesn’t get it, but Fuji does! And Fuji doesn’t just understand chemistry, it understands algorithms.
Fuji’s approach to film analog started with CMOS. Instead of the usual Bayer array of pixels on that CMOS, Fuji designed its own specialized pixel arrangement.
The photodiode on the CMOS converts light signals into electrical signals and is responsible for sensing light and darkness, and above this layer of photodiodes, there is a layer of color filters that are responsible for capturing color.
By pixel alignment, I mean the arrangement of the red, green and blue colors on this layer of color filter.
Simply put, the smallest cell of the Bayer Array CMOS is a 2×2 matrix with a pixel arrangement of RGBG (red, green, green, blue), that is, a 2×2 quadrangle with two green pixels occupying the opposite corners and the other two pixels being red and blue. This arrangement was invented by Kodak employee Bryce Bayer, hence the name.
The most common problem with Bayer arrays is that they are prone to moire. Moire pattern is actually two high-frequency signals of similar frequency, together, before each other to produce interference when the grain. For example, when you take a picture of a computer screen with a camera or cell phone, or some very dense patterns, the strange patterns that appear are moire patterns.
Many after 90’s childhood memories of “Adventure Tigers” decryption card, is used to design the principle of Moore pattern.
In order to solve the problem of moiré, the Bayer array will cover the CMOS with another layer of low-pass filter to filter out some of the high-frequency signals that will cause moiré, which can be improved a little bit, but at the cost of possibly affecting the sharpness of the image, as well as some of the very small details on the screen.
By the way, this is just fromdoctrinalThe last thing I want to talk about is the fact that a lot of cameras that use Bayer Array CMOS nowadays also remove the low-pass filter.
Fuji’s own design of X-Trans CMOS is a new pixel arrangement, the smallest cell of this pixel arrangement is 6×6, and in this 6×6 matrix, the three color arrangements also look more “random”, unlike the Bayer array that repeats itself, which actually reduces the frequency of the pixel arrangement on the CMOS, which suppresses moiré and removes low-pass filters, resulting in better sharpness. This actually reduces the frequency of pixel alignment on the CMOS, which suppresses moire, and eliminates the low-pass filter for better sharpness.
Also, the Bayer array’s CMOS has only two colors, green and red, or green and blue, on each row or column, whereas on the X-Trans CMOS, there are three colors, red, green, and blue, on each row or column, which, according to Fuji, allows theBetter color reproduction–though I find this doubtful, since Fuji’s medium format line uses Bayer array CMOS.
Whether or not it’s “better reproduction,” it’s clear that Fuji’s CMOS has its owndistinctivenessThe X-Trans is a 6×6 array with 9 red pixels and 9 blue pixels. For example, also on a 6×6 array, the Bayer array has 9 red pixels and 9 blue pixels, while the X-Trans has only 8, which also means that it is 11% less sensitive to red and blue, and the distance between two red or blue pixels is a bit farther.
Perhaps it’s this particular arrangement of pixels that gives Fuji photos their colors, which are distinctive at their roots.
In addition, this complex arrangement of pixels requires a better processor to decode it. So Fuji also got its own X-Processor series of processors to handle the complex color signals of the X-Trans and slapped on its own film simulation algorithms. This X-Processor processor is also optimized for film simulation, and Fuji has deliberately kept all of these algorithms written to death in the in-camera processor, and is determined not to leak them, as evidenced by another incident, which we’ll talk about later.
From an algorithmic point of view, the essence of film simulation is that the拟合In the same way that you add water if there’s too much flour, or flour if there’s too much water, there are dozens or hundreds of parameters in the final algorithm that represent the various chemical compositions in the film emulsions and developers’ fluids. Engineers design film simulation algorithms by adjusting the size of the impact of these parameters on imaging results. A larger parameter or a smaller parameter can result in a very different simulation. To summarize this large number of parameters and to specify the effect of each parameter requires a good understanding of the film’s formulation.
Then Fuji knows all too well what substances are in their film formulas and how they affect imaging color.Engineers working on film can draw lessons from this and then talk to engineers developing cameras to summarize algorithms for simulating various types of film.
Perhaps you might think thatIsn’t this just like the filter mentioned earlier, replacing color A with color B?
From a certain point of view, it is correct to understand it this way, but this transformation process must be very complex, with a particularly large number of parameter factors and a very resource-intensive computational process.
Otherwise, you’d think that there are already more filters on this market for film simulation than there are real films in this world.
Fuji is very conservative in this matter. Keep in mind that Fuji’s first camera with film emulation was the FinePix F700, which came out in 2003, more than 20 years ago now, with decades of film development experience ahead of it. That’s it, Fuji’s own camera only has theOnly 20 modelsThe film analogs are just that, less than one a year on average.
This kind of restraint and seriousness is not something that can be compared to pulling a few curves or applying a filter. People have been developing something for twenty years, and it’s hard to stay on top of it, so how can they let you catch up or bend the curve so easily?
The software can’t do it either.
I wrote earlier that Fuji is careful to save the film simulation algorithms to the X-Processor processor in the camera and is determined not to leak them. I’ll explain it here.
In fact, Lightroom and Capture One have built-in Fujifilm analog filters, and they can also open RAW files taken by Fujifilm cameras and apply film analog to these RAWs. There are even many tutorials on the Internet that teach users to change the model information of the RAW files, so that old Fujifilm cameras, or even photos taken by other brands, can also be applied to the new Fujifilm film simulation.
But in fact, right …… if you really changed it in these ways, you will find that the color of the photos that come out of the set of filters on these software is the same as the Fuji straight out will bemake a big difference。
Fuji also has its own computer software for solving RAW, called X RAW Studio, which is the only computer software that allows you to restore 100% of the in-camera effects of RAW filters.
However, the software is very demanding, you need to have your Fujifilm camera connected to your computer via USB at all times, and you can’t just use the software to open the RAW files in the card reader. In fact, X RAW Studio calls the X-Processor processor in the camera when applying the filters, you just do it on your computer.
Algorithmic simulation of reality is difficult
Finally, I’d like to leave Fuji behind and talk about how hard it is to simulate film from a different perspective.
In the abstract, algorithmic analog film is simulating a thing in our lives, or an aesthetic.
It’s very easy to use algorithms to generate things that we haven’t seen before, universes, planets, that kind of thing. Because we haven’t seen it anyway, so whatever you make it look like is what you make it look like.
But trying to use algorithms to simulate things we’re familiar with is immensely complicated; water, ice, fire, hair …… – things that are particularly common in our lives – are instead difficult for algorithms to simulate. Because weIt’s too familiar., the algorithm simulates something slightly different and we can immediately sense that something is wrong.
Those who are familiar with Pixar may know that Pixar especially likes to post some behind-the-scenes stories before and after the release of a movie for publicity, in which as long as it involves technology, it is something like earth, ice, water, hair, and other things that seem to be rare and commonplace. This shows how difficult it is to use algorithms to simulate well what people are familiar with in reality.
Again, film for quite some time was the one thatSomething people are quite familiar with. It may be true that today’s young people haven’t had much exposure to film, but young people didn’t pop out of a stone, they were born and exposed to a world that was already filled with countless classic photographs from the era of film photography. In the process of growing up, young people are more or less exposed to these film photos, and their aesthetics have been subconsciously influenced in the process. They have long since memorized this texture, and when they grow up and are exposed to cameras and film analogs, they may not be able to tell where it’s wrong if once the analogs don’t look like it, but theIt’s easy to see something’s wrong.。
This is probably better exemplified by the movie industry. Even in the age of digital cameras, the post-production of a movie will deliberately ‘age’ the picture to simulate the feel of film, and it’s even harder to simulate a movie. A two-hour movie with 24 frames is equivalent to a total of 172,800 photographs, and it is difficult to keep the color tones of these 170,000 pictures consistent and make them look like they were shot on a roll of film. …… If there is a slight mistake in one of the lenses, and the digital flavor comes out, the atmosphere of the whole film may be destroyed.
And, even when a movie is shot on film, it has to be scanned digitally during editing and post-production, and then ‘printed’ to film for projection. During this process, the post colorist even has to go so deep as to simulate the noise of a particular type of film, so that the whole film looks harmonious and beautiful, without losing the special beauty of film. The colorist for Star Trek has shared this experience online.
After writing 4,000 or 5,000 words, I think I’ve explained the whole ‘how hard is Fuji film simulation’ thing in a shallow way. The root of the problem is that many people simply think of Fujifilm’s proud film simulation as a simple set of filters, or a few parameters that seem to be set up in such a way that you can get the exact same effect.
But it’s actually possible to tell by purely logical reasoning: if Fuji’s film simulation is so easily ‘cracked’, then why has Fuji been so hard to find for so many years, with a serious premium, and discontinued second-hand phones costing more than when they’re new to the market? Is it really just hype? Why can’t other brands hype this up? Panasonic this year, a new camera LUMIX S9, you can set LUT straight out of the photo and video, it is reasonable to say that you can easily set a variety of film filters, kill Fuji several streets, but this machine listed not long after the hotness of the past, fell below the release price ……
This series of questions, in fact, points to one real reason:Fuji’s film simulation is truly unique in its own way, something that no other manufacturer can catch up with.That’s why I thought of exploring it to see if I could find this ‘uniqueness’.
I think I seem to be finding a bit of it.