April 4th, 2021

Looking at the EVF and LCD Features of the Fuji GFX 100S

Updated June 30, July 2,  August 8, 2021 and April 5, 2022

Including:

Zoom Ratios/Percentages for Live Viewing in the EVF and on the LCD;
EVF and LCD Viewable Regions While Zoomed In to Various Degrees;
Playback Mode Viewing Characteristics/System Features;
Suitability of Live View and Playback Modes for Precise Manual Focusing;
How to Select Focusing Aid Options;
Non-linear vs Linear By-Wire Focusing; Some Adapted Lenses;
And more…
(though I’ve generally preferred focussing and focussed, it seems that focusing and focused are considered dominant in American English, so I’ll give in to the latter, reluctantly…)

What started as a report for a few friends interested in this amazing camera has now evolved into this article.

Initial text of this article written based on GFX 100S firmware version 1.00. See notes on version 1.10 bug fix(-es) below.


The EVF is 3.690 million pixels at 4x3 shape, so that’s about 2219 x 1664 = 3,692,416

The Image Sensor = 11648 x 8736

The Sensor to EVF width ratio is then 11648 / 2219 = ~525% EVF image magnification is therefore required to reach 1 to 1, pixel for pixel viewing.


RULER TEST, Live Viewing and Playback, via EVF and LCD

I’ve done three experimental setups now where I position a 3-foot Starrett machinist’s rule with decimal inch markings about three feet in front of the camera, running horizontally through the middle of the frame, and look at it carefully in the EVF and on the LCD, as well as making exposures of RAW plus various sizes of JPEGs, reviewing them in Playback mode and viewing them on my desktop machine at 100% on a 100ppi monitor.

1) There are five zoomed-in levels for live viewing (EVF or LCD), and I’m calling the most zoomed in Level 5. To see the zoom percentage of each zoom level, relative to actual sensor pixels, see item 7) below.


2) The number of zoom levels for Playback mode reviewing of captures varies with the size and aspect ratio (JPEG pixel dimensions, “Large”, “Medium”, “Small” or RAW only and various cropped shapes) of JPEGs saved when making a capture. Playback zoom levels range from 12 to 31 in number. The more zoom levels, the further in you can zoom and the more detail you can see. Unlike Live View, in Playback mode the greatest available zoom level isn't excessive and is useful.

Small JPEG only (no RAW) 16x9 crop: 12 levels; Small JPEG Square crop: 16; RAW only or Small JPEG at 4x3: 19; Medium JPEG: 27; Large JPEG: 31. Notice the extra zooming allowed by Medium and Large JPEGs compared to RAWs Only. It’s only by saving Large JPEGs that you are allowed to see the full detail of your captures in Playback mode! But to prevent loading up your cards and computer with gobs of data, select “Normal” quality instead of “Fine” or “Super Fine”, resulting in ~12 to 20 MB JPEGs instead of 50 MB monster JPEGs like you get with Large+Super Fine. Small+Normal JPEGs are around 3 MB at full 4x3 shape, by the way, still a full 12 million pixels.


3) The regions shown in the EVF and on the LCD in Live View are identical to each other for all zoom levels, including not zoomed in, but the regions which are viewable by moving around while zoomed in change as the zoom level changes. None allow you to see all the way to the corners and all extreme edges, but zoom level three comes the closest to letting you see to the corners. The overall appearances of the EVF and LCD are consistently very close to one another.


4) The visible region of the image while zoomed in is different for each of the five Live View zoom levels. Level 3 reveals the most, at 96.72% of the full width of the raw file and comes close across the top as well. Zoom Level 4 reveals 90.60% of the width but less of the height. But max zoom, Level 5, reveals only 88.60% of the full width. I find Levels 1 and 2 to be useless, but for what it’s worth, Level 2 is the second best in this respect, at 94.89%. So if you want to see nearer to the edges, use Level 3. I have not examined this coverage in the vertical (short) direction, and I can see that it’s not necessarily proportional.

Here are my notes on that. I could see the ruler to better than a hundredth of an inch accuracy. Note that the first two inch dimensions shown in each case are the actual points on the ruler that I could see.  The difference between them is the third dimension (e.g. 17.01").

A) Full Screen Live View: 16.47” to 33.48” = 17.01” = 100.07% of RAW file (close to a perfect match but showing a trivial 1 in 1400 pixels too much if I measured right)

B) Full Screen Playback: 16.45” to 33.42” = 16.97” = 99.83% of RAW file (close enough)

C) Live View Zoom Level 1: 17.31” to 32.48” = 15.17” = 89.25% of RAW file viewable by scrolling

D) Live View Zoom Level 2: 16.86” to 32.99” = 16.13” = 94.89% of RAW file viewable by scrolling

E) Live View Zoom Level 3: 16.68” to 33.12” = 16.44” = 96.72% of RAW file viewable by scrolling

F) Live View Zoom Level 4: 17.18” to 32.58” = 15.40” = 90.60% of RAW file viewable by scrolling

G) Live View Zoom Level 5: 17.36” to 32.42” = 15.06” = 88.60% of RAW file viewable by scrolling

H) RAW File in Iridient Developer: 16.428” to 33.426” = 16.998” = 100% Actual Width

I) RAW File in Adobe Camera Raw: 16.428” to 33.426” = 16.998” = 100% Actual Width

I have always wished that we could scroll to the real edges of the view when zoomed in, and I still wonder whether there is actually any good reason for cutting off some of the image in Live View. It may have to do with a rounding error in the discreet steps of zoomed-in movement, but I’m still guessing this could be done so as to show the entire area, even if it meant that the last section shown wasn’t the full 4x3 shape.

In Playback mode, you do get to to see 100% of the image area at all zoom levels. I have tentatively concluded that the most detail you can make out in either Playback or Live View modes is about equal to what you can see on a 100 ppi monitor, viewing the RAW file on the computer, at 50% magnification. In other words, not as much as routine viewing of the RAW file at 100% yields. I also find that in Playback mode, I prefer to use a zoom level which is not more than three steps below the max.


5) The factory default setting for the “Performance” function, which is supposed to be able to benefit how the EVF works in either of two ways, is “Boost > EVF Frame Rate Priority”. With this Frame Rate boost selected, however, zoom Level 3 looks absolutely awful, because it becomes merely a 2X scaled up version of zoom Level 2’s detail! This could be a firmware bug, but if it is, it was not fixed with firmware version 1.10, released June 30th, 2021. This doubling of Level 2 data for Level 3 display does not occur on the LCD — only in the EVF — so the LCD always looks great at Level 3.

Note: Firmware version 1.31 was introduced on March 30, 2022. Here is how Fujifilm describes it:

"Revised specification from Ver.1.30 to Ver.1.31

The firmware bugs have been fixed, which caused a EVF display failure when the “FRAME RATE PRIORITY” was selected in the [BOOST] menu by operating through the following menu flow.

[POWER MANAGEMENT] ➔ [PERFORMANCE] ➔ [BOOST]"

The only EVF display failure I've ever been able to detect is that described above, so when I saw this I assumed it was what they were talking about. Unfortunately, after installing the update from 1.30 to 1.31 there was no change at all to the problem. Level 3 is still utterly useless when Frame Rate boost is enabled. It is also still the case that when you use Boost > Resolution Priority there is no improvement to the resolution of the EVF. Nothing. I still have no idea what they mean to accomplish with this feature. It remains true, however, that boosting the frame rate actually does work to boost the frame rate — you just can't use it at what is probably the best overall zoom level, equaling 150% of the pixel-for-pixel magnification of the full-res image in the EVF, i.e. Level 3. Level 4 is still good though (for manual focusing and for checking AF focusing results). Level 5 is way over-magnified.

The F/W version for the body was 1.00 until 6/30/2021. The secret method for checking F/W versions of the body and an attached lens is to power the camera on while holding down the Disp/Back button. It’s weird not to be able to find the F/W versions in the menus, but this info has been moved to where one would proceed with a F/W upgrade.

By selecting any of the other four Performance options (Normal; Boost EVF Resolution Priority; AF Priority - Normal; or AF Priority - Low Light) this Level 3 wipe out is solved. Go to: Wrench > Power Management > Performance > then select either Normal (supposedly gives you the full 460 shot battery life) or one of the three settings which is not the Frame Rate one (and supposedly then get only 400 shot battery life). Try as I might, I have not been able to see any difference in the EVF resolution as a result of enabling Boost > EVF Resolution Priority. This is most puzzling, and it remains true after the camera firmware being updated to version 1.10 (and still through version 1.31, as I just mentioned). No difference at all. So I’ll be leaving my Boost setting at Normal (meaning Off) for now. The fact that Boost Resolution seems to do nothing could also be a firmware bug. Again, if so, it has not been fixed by version 1.31 as far as I can see. Take a look at the Owners Manual, page 240, for a somewhat unclear explanation of the Boost features. One thing that’s missing from this explanation is just how and when resolution is boosted. As it is, it feels like a case of The Emperor’s New Clothes. The Frame Rate boost definitely does increase the frame rate and does look nice and smooth (smoother) when waving the camera around. For relatively non-athletic camera handling, I see no need for it.


6) My visual sense of looking at the ruler is that the system is actually very sharp overall, meaning that it’s resolving a tremendous amount of detail in the subject (of course, the camera does precisely that). As for how that is presented in the EVF, keep reading. I have found that with some subject matter, the “Focus Peak Highlight (Contrast Outline)” focus check option for finding the best focus does provide a little bit more precision, at least with that feature set to its factory defaults, which have struck me as optimal for this tool’s setup and which I have not yet altered. The peaking pixels are white and black and they generally do not overwhelm the subject at their peak when zoomed in, and sometimes they don’t show up enough, so these factory settings seem good for me. Mostly I don’t feel the need to use it though, so I stick with "Normal" as the focus check option. To test the options for Focus Peak Highlight go to AF/MF > Page2of3 > MF Assist > Focus Peak Highlight > then choose one of the eight options. White (High) is the default.

In general, I have found zoom Levels 3 and 4 to be the most useful by far, especially 4. Level 5 is so over-zoomed as to be superfluous for me (see item 7 below) but it may be helpful for some or at times.

Note: The use of the rear command dial for Live View zooming to check focus does not work when the little focus mode lever just to the right of the EVF is set to C (continuous AF).  It must be set to either S (single AF) or M (Manual focus). You can still use the AF button just to the right of that lever on the back of the camera when in Manual focus mode, to use AF to focus on a chosen AF point. I find AF to be highly accurate and very useful, but I've also found that in order for it to work well I have to use the smallest size focusing square, and of course it matters greatly what you aim it at. I also noticed when testing with larger squares that the point focused on was at the lower edge of the focusing square, not the center.

The default access to Live View zooming is via pressing in the rear command dial. But if you press and hold instead, the camera cycles one step through the group of four manual focusing check modes (assuming this default Function choice is assigned to the rear command dial's press function): Digital Split Image, Digital Microprism, Focus Peak Highlight, and Standard (no gadgets). When in those first two modes, you can’t zoom in at all. Press and hold four times to move through all four modes in a loop. It takes some getting used to, to avoid pressing and holding when you just meant to zoom in. You can also program one of the function buttons to initiate zoomed-in viewing, as I have. I initially chose the forward of the two function buttons to the right of the top LCD but later decided to use that button for accessing the self timer choices. To set this up, go to Wrench > Button/Dial Setting > Function (Fn) Setting > Fn 2, then choose Focus Check (or whatever) from the eight pages of options! Once zoomed in you control the amount of zoom (1, 2, 3, 4, 5) by turning the rear command dial left or right and you move around using the joystick (“focusing stick”). You can de-zoom all the way by pressing the command dial again or a zoom-programed function button or half-pressing the shutter release and probably other ways as well, which will leave the zoom level where you left it for quicker access next time or you can make yet another choice so that every time you zoom in you go straight to max zoom at whichever spot you have the AF square positioned. Currently that’s my favorite.

Access via the menus to the four manual focus check modes is here: AF/MF Setting > Page2of3 > MF Assist

One big surprise about zoomed-in movement during live viewing only was that Fuji enables jumping to the opposite side of the image by going one jump too far. Once you get used to this, it’s a great time saver for examining all four corners, for example. Zoomed-in movement via the joystick during live viewing is usually too fast by perhaps 50 to 100 percent. It takes about 1.5 seconds to traverse the screen the long way and about 1 second in the short direction, but the movement tends to be jumpy, sputtering while pressing the joystick continuously. Moving around while zoomed in during playback, in contrast, is much too slow. It varies but it often takes about 18 seconds to drive all the way across the screen the long way. Two+ seconds would be about right for both modes. Live View screen crossings at zoom Levels 3, 4 and 5 all involve 12 discrete jumps — it's odd that they’re not different from one another. It’s difficult to move in controllably discreet steps, as a brief touch is frequently enough to cause a double jump. The repeat rate is simply too fast.

If you have the touch screen feature of the LCD enabled, you can use it during live viewing by pulling your eye away from the EVF, then double-tapping on a part of the image to zoom into that region, then resuming viewing via the EVF, as an alternative to scrolling around with the joystick ("focus lever") while using the EVF. You can also tap once to select a spot, then use a press of the rear command dial to zoom in at that spot. If you are in Playback mode, unfortunately the touchscreen functions don't allow you to choose where to zoom in with a tap or to just double-tap to zoom in directly. You can set up playback zoom at various default zoom levels and to either go to the center to start or to where you have an autofocus point set. Since scrolling in Playback mode is so slow, it would be helpful if we were allowed to use the touch screen capability to more quickly move around, even though we have to disengage the sensor that turns the EVF on and off so as to flip to LCD viewing before the LCD will respond to a touch (we wouldn't want to be moving around accidentally because our cheek touched the LCD). But just having the scroll speed while zoomed in be right would be much better.

In the Pentax 645Z, playback zoomed-in movement is too slow by a factor which increases depending on the zoom level, up to about 8X too slow at maximum zoom. In the Sony Alphas, both live viewing and playback viewing while zoomed in are just right, about 2 seconds to cross the entire frame, regardless of zoom level, if memory serves.

When you press and hold the joystick of the GFX 100S sideways (firmware versions 1.00 and 1.10), you get a continuous series of jump commands, like pressing and holding a letter key on a computer keyboard. They are so closely spaced in time that controlling where you stop moving is quite difficult. The repeat rate is simply too high and should, I think, be cut by roughly 40%. Requiring that a fresh command be issued (a subsequent push on the joystick) after reaching an edge, to continue by flipping over to the opposite side of the image, also would make it much easier to land at the edge, as one often wishes to, rather than accidentally winding up lost, unexpectedly on the other side of the screen. One can get used to this, but it’s less than optimal as configured in the initial shipping firmware and the version 1.10 which has now followed it.


7) Here are the five magnifications of the EVF relative to 1 to 1 pixel scaling from the sensor. When the non-zoomed EVF and LCD image widths show 15.50” of my ruler set down in front of the camera, running horizontally right through the middle of the frame, on a tripod (call that 100% view, or 100% magnification) the five zoom levels show these amounts of the ruler:

Zoom 1: 6.33” = 40.84% view = 245% magnification. 245/525 = 46.7% of 1 to 1 pixel for pixel (nominally 50%)

Zoom 2: 3.95” = 25.48% view = 392% magnification. 392/525 = 74.7% of 1 to 1 pixel for pixel (nominally 75%)

Zoom 3: 1.98” = 12.77% view = 783% magnification. 783/525 = 149% of 1 to 1 pixel for pixel (nominally 150%)

Zoom 4: 0.99” = 6.39% view = 1,566% magnification. 1566/525 = 298% of 1 to 1 pixel for pixel (nominally 300%)

Zoom 5: 0.66” = 4.26% view = 2,348% magnification. 2348/525 = 447% of 1 to 1 pixel for pixel (nominally 450%)

Remember, the EVF needs to be at ~525% magnification to equal 1 to 1 pixel for pixel viewing of sensor data, as I calculated above. You can see that zoom Level 3 (when it’s not being not ruined by enabling Boost > EVF Frame Rate Priority with the initial firmware version 1.00 and the newer 1.10 version) is the one that then seems to make the most sense for best focusing, being as it’s approximately a 150% zoom level already. My prior experience has been that 100%, pixel for pixel EVF viewing is optimal for an EVF. Still, Level 3 looks pretty crisp despite being 149%. Zoom Levels 4 and 5 thus calculate out as simply being more and more scaled-up versions of sensor data, and sure enough, that’s how they look. The image is somewhat soft and much softer still when using Levels 4 and 5, respectively. I suspect this may be why some have characterized GFX EVF performance as not very good, but I’ve not experienced any of the other three GFX EVFs, and then there’s also the possibility that the EVFs got criticized because of this weird Level 3 upscaling from Level 2 business, or some other faux pas. Nevertheless, I do find that Level 4 is a little more revealing of detail than Level 3. Go figure. Perhaps that happens because the optics of the viewfinder impose a limitation that extra magnification can help to overcome.

In any case, this series of magnification steps strikes me as unusual for an EVF but also thoughtful, for having provided extra zoom at the top. But I do suspect that because the zoom percentages are not labelled in the EVF or on the LCD that this has led people to think the Live View system is not as sharp as it ought to be. Naturally, if you think you’re zoomed in all the way (i.e. to 100%) and the image looks very soft, but you’re really zoomed in to about 300% or about 450%, you’re going to think there’s something wrong with the EVF. Including such labels telling us the real zoom percentage for a moment, upon initially zooming in, might make us all understand a lot better what’s going on and therefore be more favorably impressed with the Live View system.

In Sony Alpha’s (at least the A7r II), the “Styles” menu settings affect the appearance of images. You might guess that these settings (sharpening, contrast and saturation) would only affect in-camera JPEG appearance, but they also affect EVF and LCD appearance during live viewing in the Sony's. Surprise, surprise! They do not affect raw file appearance however. Since this sharpening is simple unsharp masking (for fast computation), it tends to look pretty bad, as sharpening goes, but using the smallest amount of it, by choosing -2 on the Sony scale of -3 to +3, I find the EVF of the A7r II looks its best, enabling best precision when manually focusing. (I also set the contrast to the lowest setting, which opens up the shadows and highlights during live viewing considerably.)

In the GFX 100S, the Sharpness setting at I.Q. > Page 2of3 > Sharpness (-4 to +4, where -4 equals none or nearly none) does affect saved JPEGs and definitely does not affect saved RAW files, paralleling the Sony behavior. Sharpening settings do also effect what you see in Playback mode in the EVF and on the LCD, but the pattern in Playback mode is more complicated than with the Sony. When you've set the camera to record both a RAW file and a JPEG, the JPEG is used for Playback mode, so you can see what the actual saved JPEG looks like. A +4 setting results in a gross sharpening effect. I find the JPEGs look their best with sharpening at -4. But when you've set the camera to record only a RAW file, a lot changes. The preview you get in Playback mode is apparently a roughly ⅓ or 3/8ths scale (at max zoom) temporary file which does have a seemingly quite modest sharpening applied, in proportion to your sharpening setting, such that you would likely not even notice any effect of a sharpening setting higher than -4. As I mentioned elsewhere, to see full detail in Playback mode, you must save a JPEG at Large size! The JPEG compression can be the strongest, i.e. "Normal" quality, to save space on your card(s). At first it seemed as though one could save RAWs on one card and JPEGs on the other, and easily delete the JPEGs that way, while still having them for best focus verification while working with the camera, but that’s not working correctly now, so that option is a no go in testing so far.

The effect of the Sharpness setting on Live View images is apparently zero, however, so unlike the Sony, it seems there can be no tweaking of the EVF's ability to work best for critical manual focusing by adjusting the Sharpness control up a bit from the minimum. Nor does it strike me that this would necessarily be helpful, given that Fujifilm's range of zoom ratios favors higher zoom levels than Sony's system does. On the other hand, the I.Q. > Page 2of3 > Tone Curve settings for Highlights and Shadows do affect the appearance of Live View previews and JPEGs both! I'm keeping the Shadows setting at its lowest setting, -2, so as to have more open shadows but leaving the highlights at 0. Opening up the shadows all the way to -2 results in a big increase in colorfulness if you’ve also chosen the Astia film simulation, however, so you can offset that by using I.Q. > Page 2of3 > Color set to -4, the minimum, to balance out the effect of using the most revealing Shadows setting. But if you’ve chosen Pro Neg Standard as I have now, the Color is best left at 0. When I had chosen the Astia simulation, I preferred the highlights to also be at -2.


8) Both linear and non-linear by wire focus work well. The non-linear has finer motion when turning the ring slowly than the linear does by at least 2 to 1. It does go wildly non-linear when you move the focus ring very quickly, but the non-linear mode rate of focus change is far more consistent at slow ring rotation rates than other by-wire focusing systems I’ve used, making it pretty feasible to use rocking focus back and forth to find the precise optimum. The non-linear mode at least comes close to maintaining one, very slow, constant rate of focus change throughout a range of very slow rotational speeds, which is just what one would want. So overall, I like the non-linear better. Also, with focus peaking enabled (reached easily by pressing and holding the rear command dial once to cycle through each of the four modes, one of which is peaking — if that function is selected for the rear command dial at Wrench > Button/Dial Setting > Function Setting > R-Dial > then choose Focus Check) depending on the subject matter it’s easy to nail the finest whiff of best focus, more repeatably I think than without it. Like a half-inch accuracy at 27 feet vs an inch, something like that (at best, which was a thing I had to do to assess the sensor swing and tilt of my camera as well as I could). Find the Linear/Non-Linear setting at Wrench > Button/Dial Setting > Page 2of3 > Focus Ring Operation.

Accurate manual focusing is just not a problem. I didn’t autofocus with the camera one single time, for my first 1,500 test exposures for the process of carefully examining 12 different lenses. These included four GF 32-64’s, six Pentax 645 lenses and two Mamiya 645 lenses. Happily though, I have since found the AF to be extremely accurate as well as fast, at least if the little movable focusing and zoom square in the EVF and on the LCD is set to its smallest size, because it seems as though the AF system focuses on whatever is at the bottom edge of the little square, rather than e.g. an average of what’s in it or on whatever is in the center. At least that’s what I’ve noticed in a cursory look.

The Mamiya manual focus 105-210 is a big winner, especially at 105 to 140mm, not as spectacular at 210. As in sharp corner to corner wide open.

The Pentax 150A is surprisingly excellent, considering what a simple formula it has. The Pentax 45-85A is the non-native lens that’s most important to me, as it can form the mainstay of a tilt-capable setup for tilt focus in landscape work, and it does fine if stopped down enough, e.g. f/11 I’d say from memory. My 35A is good. So is the 75A. The 120A has too much curvature of field, which was a surprising disappointment. The 80-160 FA is, as usual, a solid performer, though I think I will prefer the Mamiya 105-210, even though I won’t be able to operate it with tilt, given my adapter setup being limited to tilt with the Pentax 645 lenses. I am fairly happy with the 32-64 performance at the shorter end, 32 to 39mm or so is pretty good, but with curvature taking a significant bite out of the DOF. My rule will be that even with careful compromise focus, I'll need f/8 to get only a distant plane all into good focus. Were there no curvature of field, this lens could be sharp wide open on a plane at infinity. At 45mm, the lens is manageable for getting everything in focus at f/11 to f/16 in some landscape situations (distant, fairly flat scenes not requiring much depth of field) but in this range and longer the lens has a good deal stronger effective curvature of field than is optimal for landscape work. At 64mm, my copy and two of the other three are not so hot, with curvature eating up about ¼ of the available f/11 DOF. One of the four copies, that for a friend, is substantially better at 64mm, with curvature eating up only about 1/8th of the available DOF at f/11. When a primary goal of a photograph is to have everything well in focus, as is typical of landscape work, a curving plane of focus can be especially pernicious, and I think this GF lens, the 32-64, has the most of any of the GF lenses, which is a real shame. When it's in focus, it tends to be very well corrected. Curvature is apt to cause top corners of distant subject matter to be badly out of focus or inversely cause the central portion of the lower, foreground edge to be out of focus, or both, when this lens is near its long end. Medium Format magazine has just published a long and detailed look at the 32-64 and 45-100 GF lenses that I wrote. It is available by subscription only. I have been very impressed by the quality of the magazine and it’s new sister magazine, Elements, which features landscape work only.

I will not shy away from using f/16 on this sensor, and sometimes smaller. F/22 does look pretty terrible though, beyond the reach of remediation with sharpening, unless achieving sufficient DOF is worth the tradeoff of losing a solid chunk of resolution to diffraction, as it may well be at times. On my Phase One P45+ back, I used the Mamiya 105-210 at either f/32 or very close to it for a very important picture and the results are spectacularly sharp. The sensels in that 39 MP back were quite a bit larger, 6.8 microns, versus the 3.76 of this sensor. So if f/32 were just OK with the older sensor, then (3.76/6.8) x 32 = 17.8 should be just OK with this sensor, and I’d say that’s roughly the case. The smaller the sensels get, the less you can stop down before you see a given amount of diffraction on a per pixel basis… But at the same time, we do have more pixels and more detail to play around with.


9) The appearance of the LCD versus the EVF is virtually identical (I don’t see any meaningful differences apart from the effects of looking at the EVF panel through optics and the EVF pixels being smaller).

Fuji’s UI for this camera is the most detailed I’ve ever seen by a substantial margin. By and large I’d say they have built a superior feature set, with the caveat that it's quite difficult to remember where everything is buried, and some of the important stuff is buried a few layers deep. I was quite fond of the UI Pentax built as found in the 645Z. Features like enabling a 2-shot exposure bracket instead of starting at 3 shots (also a Fuji feature). There is a lot to like in Pentax’s UI. Fuji’s has more features and includes many clever advances, but does also require more learning from us. Also I would like it very much if every camera maker would add the very long overdue feature which has finally appeared in the Nikon Z cameras — automatic expose to the right metering, rather than only various flavors of averaging metering. The absence of this feature means that we need to stick to Manual exposure control if we hold that clipped highlights are a no-no. The combination of a well-implemented, automatic ETTR capability and an automated 2nd exposure favoring the shadows would be a welcome addition.


10) Comparing the Playback image quality to the EVF image quality, with EVF at Zoom 4 (showing 0.99” of ruler) and with Playback of a Large JPEG showing 0.97” of ruler (closest possible match at high magnification) the two views appear equally sharp. So if there’s any truth to the rumor that image playback is sharper than EVF Live View, it would be either because A) Zoom 5 is being used, and that’s more magnified than Playback can go to, or B) Zoom 3 is being used with Boost > EVF Frame Rate Priority enabled.

I think that about does it for EVF questions. I did wiggle the camera around just a bit with the Boost - Frame Rate no longer enabled and I didn’t see anything to be bothered by, just some jumpiness here and there, but I imagine if you were photographing skateboarders at 70 mph six feet away you’d want to boost the frame rate if you could! Conceivably for less extreme portraiture or bird photography also, but I sort of doubt it would make a difference for most portraiture. I’ll be keeping the Boost setting at Normal for now as I said (Wrench > Power Management > Performance > Normal).


11) The camera’s low battery warning works nicely. It had shown the easy to notice red battery icon low power warning dozens of shots before I voluntarily stopped using it in my first experience with it. I have yet to have the camera lock up from reaching the end of the battery’s power. Having a sufficiently generous amount of leeway is a nice feature. Charge percentage climbs at nearly 1% per minute with either one or two batteries in the twin charger, plugged into my 27 watt-rated at 9 volts, 15 watts-rated at 5 volts, Motorola Turbo Power AC adapter that I had purchased recently for my phone. Plenty fast. An early attempt to charge two Fuji brand batteries in the twin charger at once when it was powered by a lowly, 5-watt Apple older iPhone USB AC adapter resulted in only one of the two batteries charging (not enough power for both and the charger knew it). I had to switch the batteries around in the charger to get the other one to charge. Using the more modern and more powerful AC adapter, with a USB-C port on it, requiring that I purchase a 1.5 meter USB-C to USB-C cable to optimize the setup, charging two batteries at once and at normal full speed (about 90 minutes typically) worked just fine. I like the Fuji charger. It’s smaller than it seems in pictures. Also, the rate of charging, readout percent vs time, appears to be quite linear from a 0% readout all the way to a 100% readout. I haven’t used either the twin charger or the camera with the AC to USB adapter or cable supplied with the camera yet, but I’m sure that adapter will also allow charging of both batteries at once. I used my other adapters and cables initially while preparing for the camera's arrival.

The charger’s little LCD screen doesn’t shut off immediately upon the readout reaching 100%. This may just be a nicety for showing us the 100% reading for a while or it may be because 100% is initially a rounded up number, and only when the battery really goes from e.g. 99% to 100% does the screen go dark and presumably the charging circuit also stops. If that’s the case, then the last little bit of charge is not linear, as one sees so often with charging devices and their accompanying batteries.


12) I’ve done all of my testing with a 64 GB Sony Tough card, rated at 299/300 MB. I chose it for two reasons: A) It’s the fastest card available now and I wanted to give the camera every chance to feel fast (it does), and B) I like the toughness features, which are pretty remarkable. So far I’ve never had to wait for image buffering or to review an image after capture — the camera is fast in general, despite the immense resolution. But I’ve not done any testing to see what the maximum continuous sustained capture rate is either. I’m guessing from reading other’s tests that it’s faster than one frame per second, likely around .7 or .8 seconds between frames, but in practice I’ll never need to sustain such high rates so it’s not a priority for me to test that yet. [Later I ran a bunch of tests and the results were better than expected. The camera could sustain between 2 and 3 frames per second, depending, recording 14-bit raw lossy or lossless compressed plus Large Normal JPEGs to the card indefinitely! I never tried going beyond 140 frames in a row, but wow. I’m happy.]

One hiccup that occurred with the card was that on two occasions upon inserting the card into the lower slot (#1), the camera couldn’t read the card. I’m pretty sure it was because there were a couple of bits of fibrous dust on the contacts. Blowing and/or wiping them off solved the problem and it’s happened twice more after a couple of hundred more insertions.


13) I have done nearly all of my lens testing hand-held. This is a major departure from nearly all prior practice. I have found the IBIS to be apparently 100% effective at neutralizing camera movement down to something like a 2 to 1 ratio (e.g. lens 60mm, shutter speed 1/30th) at least. Needless to say, this changes a lot about ultra-high resolution photography. Also, the shutter in EFC mode is remarkably quiet and soft. I really like it. Later though I realized that for the range of shutter speeds and the kinds of subject matter that I'm interested in the electronic shutter is simply better. Were it not for the volume-configurable artificial sound, you'd not hear the shutter tripping. We can still choose any of the six variations of RAW files, 14- or 16-bit plus uncompressed (~210MB), losslessly compressed (~110MB) or lossy-compressed (~63MB). Choosing 16-bit slows the camera down quite a bit while providing an extremely small sliver of extra super-deep shadow quality, so I'm expecting to forego it nearly always.


14)
 I discovered a bug in firmware version 1.00: If you use the "A" setting on a GF lens barrel, with the front command dial (reconfigured in the menu) to set the aperture, never set it to f/32, or it will get stuck there! Hopefully this will be fixed shortly after my bug report is absorbed by Fujifilm. [After installing the new version 1.10 firmware I can happily report that this has been fixed.] If your camera should get stuck at f/32, you can resort to using only the manual aperture ring on the lens, or you can use Wrench > User Setting > Reset > Still Menu Reset and perform that one of three kinds of resets of the camera's settings. Having done that, manually re-enter the many settings that got nuked. Most camera settings don't get nuked. I made snapshots of many menu screens first with my phone so as to help me remember how I had everything configured. Using the Fujifilm X Acquire program to save nearly all of the camera settings and reload them after the reset didn't work to fix the f/32 bug, becuase it just re-initiated the stuck condition. Nevertheless the ability of Fujifilm X Acquire to store all your camera settings in a saved settings file, with which you can restore your camera after a reset or transfer settings to another GFX 100S is a very nice thing and good insurance, so I'd recommend getting familiar with it. One of its quirks is that it doesn't have a conventional U.I. — you interact with it via a menu in the macOS menu bar.


That’s all for now.


—Joseph Holmes