Fujifilm Color, Image Quality, Color Science, Simulations, and More Interview

DC.Watch published part one of a very lengthy interview with Kosuke Irie, the Technical Manager, Optical and Electronic Video Product Development Center, and Shinya Fujiwara, the Optical and Electronic Video Product Development Center. Below is a summary of the translated interview:

What is the core of image quality that Fujifilm thinks about?

  • Image quality is about four elements: Color, Sharpness, Noise, and Graduation that you can break down further
  • These elements must be in balance for good results
  • When making a new simulation Fujifilm thinks “Are each color and each gradation pointing in the same direction?”
  • Eg during classic chromes development the dark red channel didn’t look as vivid as other channels to the human eye so it was pointing in a different direction of being designed to look vivid for documentaries
  • It’s difficult to set goals for developing looks especially since the cameras differ so much with new sensors and processors or even different IR/UV cut filter differences.
  • For example, it is impossible to make the X-T2 and X-T3 produce exactly the same film simulation results so all they can do is make users comfortable with the look
  • Fujifilm tried to design from the hardware up to provide a consistent comfortable look
  • Tonality is becoming more consistent due to market demands, but some simulations like Provia have changed how they handle shadow areas such as 16 or less between 0 and 255 to make them less likely to collapse

Origin of film simulation

  • The FinePix S100FS came out in 2008 and was the first camera with film simulations under real names, but even back to the FinePix S3 Pro in 2004 Fujifilm had standard and F Chrome Mode for film simulations, which was the origin of film simulation
  • Fujifilm developed simulations because they were criticized for their images having a digital feel and customers didn’t expect that from a film company
  • The silver halide film people helped develop a comfortable reproduction with little digital feeling in the FinePix F810 which was released Aug 2004 which had the first digital Provia look
  • The PROVIA look has been used since 2004 for Fujifilm’s standard film look and how it was designed has remained Fujifilm’s simulation theory largely without major changes up until now
  • They couldn’t use the PROVIA name back then so F1b and F2 were used until 2008 in when they started calling them film simulations

Is there a difference in image creation by sensor format?

  • X-Trans and Bayer have a different spatial frequency and are tuned differently, but ideals like sharpness are the same
  • The 102MP of the GFX100 allows for a state closer to the data since things like noise are buried in the pixels which allows for less noise reduction
  • “Comparing the images of the GFX series and the X series with the same print size, the GFX series has a smaller magnification. Therefore, the grain naturally looks fine. When printed at the same magnification, the GFX series and X series are granular in the same size. Of course, in this case, the print size will be significantly different.”
  • There are two ways to do digital grain, one is dithering, which uses random dots, but it gives off a digital order so Fujifilm makes it look natural by applying a process that looks grainy based on the sensors noise that gives a natural organic feeling like film
  • Fujifilm has two-grain looks that make it look comfortable, one is image processing to make noise look grainy and the other is the film simulation grain effect, which has different nuances
  • The grain effect from the simulation is randomly generated using a pseudo-random number generator circuit designed with great care mathematically and visually, which is what makes it different from the noise looking like grain effect
  • The thought is that humans find images that are too uniform to be artificial and unnatural so adding a little noise can have a positive effect since the human eye forms perceptions from the noise

About film simulation

  • Irie would like to find a better word than film simulation for Fujifilm’s film profiles because they aren’t necessarily trying to replicate film as much as creating good images so to him the label isn’t necessarily accurate
  • Film simulation is not a copy of film, but to achieve an image quality goal that has existed since the era of film of overall image quality
  • For example, ASTIA is aimed at achieving the same goals as silver halide film ASTIA, but digitally there are no restrictions on silver-halide and there are other disadvantages which have lead to simulations like digital ASTIA realizing what silver salt could not do
  • Irie doesn’t like to call ASTIA software, but that’s the terminology used internally, but its design intent is to reproduce soft skin with the intent of reproducing bright skin tones with smooth skin tones, but silver halide film ASTIA sometimes looks softer overall.
  • The “software” branch of ASTIA should just be considered ASTIA without calling it soft because it is designed to be closer to the ideals ASTIA film was based on and starting with the X-S10 Fujifilm is trying to better explain its simulations
  • We can do anything digitally, but Irie and others think film might have an advantage in terms of color continuity
  • Film is superior for capturing color information, but digital is better for gray linearity since it can be expressed as a complete grayscale
  • It’s very easy to do White, Black, and Grey with digital with great levels of accuracy because of the way the senor captures light
  • With silver halide film the shadows start to turn blue while highlights turn yellow
  • Film can obtain ambiguity and express color continuity with good tone connection that provides a beautiful reproduction that is difficult with digital
  • For example, there is nothing quite light positive film shot underwater with its very deep beautiful blue
  • We still have a lot to learn from film
  • Color Chrome is based on Fujifilm’s efforts to reproduce the deep color expression of films like positive film and it works by suppressing the development of some colors to create deep color reproduction

Approaching color reproduction of film simulation

  • Film simulations try to ensure a degree of accuracy and colorimetric that is tuned for comfort and natural reproduction
  • Faithful color reproduction is necessary for skin tones or the color of a car
  • Faithful colors must be perceived as the same color when compared to the actual scene so this is where memory color comes into play because you need color reproduction to fall between numeric accuracy and memory color for it to be accepted, which makes the image feel like “the same color” when they are viewed alone
  • PRO Neg. is the closest to colorimetry in film simulations
  • PRO Neg. Hi and Pro Neg. Std are mainly different in the tone curve with high having tight shadows and highlight, but the colors have the same design
  • “The design intention of PRO Neg.Hi and PRO Neg.Std is to consider the difference in studio lighting.”
  • PROVIA, Velvia, and ASTIA have a magenta tendency that helps bring out blue skies, which is more effective than colorimetry
  • Classic Chrome skin tones are created by part of a cool image that is shifting to warm
  • ETERNA is based on cinema ETERNA film which can make the sky look a little greenish-blue until it is projected on the big screen
  • A little Magenta should be added to ETERNA when using it for photography to make it look good and impressive, but in video you need the opposite
  • “The difference between movies and photographs is that not only the pictures move, but there are also “sounds” such as dialogue and music.”
  • So if an image is too impressive it can throw off the balance between the film, dialog, music, etc…
  • In cinema, the sky is deemphasized to with greenish-blue to make it less of a distraction
  • Classic Chrome also deemphasized the sky and can be used as a film simulation for cinema, but it is more expressive than cyan
  • Classic Negative is a special film simulation, and it is designed so that the appearance of colors changes depending on the brightness, so dark tones get cyan boosted and bright tones get magenta boosted like film
  • Once again digital allows Fujifilm to realize their ideas by allowing for this kind of adjustment, which film couldn’t achieve
  • Classic Negative might be a little too bold
  • Classic Negative gives you the feeling you get when you look at a color negative after more than 10 years
  • It’s not hard to do these kinds of auto adjustments, but it is hard to do without failure
  • “For example, even if you take a picture of a person’s face in one scene, there are many tones from highlight to shadow, so if you just turn the color according to the brightness, the skin tone is neutral in the halftone, but the shadow is green. The highlight becomes magenta, something like that happens in one face and it becomes a collapsed reproduction. It was really difficult to arrange such “directions for each color” and express film-like nuances within a range that does not break down.”
  • Irie was worried up until the end and he thinks that the color reproduction of classic negatives has a slightly ridiculous side
  • The deciding factor about releasing Classic Negative was that it was coming on the X-Pro3 which can be a little sharp
  • The non-missing feeling of classic negative expression comes from the balance of color and tone curve

Believe it or not, there is a part 2 coming for this very lengthy interview and I will cover it here too. If you want to read the whole interview it can be found here at DC.Watch.

Please don’t forget to follow us on FacebookTwitterInstagramYouTube

Plus our owners’ groups
Fujifilm GFX Owners Group
Fujifilm X-H Owners Group
Fujifilm X-T Owners Group
Fujifilm X-S Owners Group
Fujifilm X-Pro Owners Group
Fujifilm X-E Owners Group
Fujifilm X-A Owners Group
Fujifilm X100 Owners Group

This entry was posted in Interview. Bookmark the permalink. Trackbacks are closed, but you can post a comment.
  • FCC disclosure statement: this post may contain affiliate links or promotions that do not cost readers anything but help keep this website alive. As an Amazon Associate I earn from qualifying purchases. Thanks for your support!