Fujifilm’s new X-H2 finally pushes its APS-C camera system to higher resolution
theverge.com
theverge.com
The X-H handles nice, but as a still (amateur) photographer, the X-Ts are where it's at. Having separate, physical controls for exposure compensation, speed, ISO, and aperture is a delight. The tilting screen on the X-T line is also a better fit for still photography than the articulated one on the X-H lines. Here's hoping they keep it.
I have an X-T3 now, I'm waiting for the X-T5. It should come out in a few months, and cost a bit less than the X-H. The main advantages should be the 40 MP sensor (vs 26) and sensor-based image stabilization (introduced in the X-T4). They also offer a pixel-combining mode now that shifts the sensor very small, precise amounts and then combines images to increase resolution up to 4x in the final result. You need a tripod and some time, but I'm looking forward to playing with it.
It's hard to explain exactly how, but the color rendition straight out of the camera on Fuji kit is also second to none. Other posters commenting on RAW files are right, though - they're a pain because of the non-standard sensor. Only Fuji's terrible software and commercial offerings like Lightroom do a decent job. Darktable, for example, doesn't. I stopped shooting RAW partly for this reason, but also because the JPG engine on Fuji kit is outstanding. It makes the best JPGs straight out of the camera I've ever seen.
The APS-C sensor size across the entire line also means that high-quality glass is a bit more affordable, smaller, and lighter than some of the competition.
I guess I became a huge Fuji fanboy along the way.
I currently own a pen-f, which is tiny. But I figure if I'm going to haul around a brick, might as well go for full-frame. Especially since Sony came out with the a7c. However, Sony doesn't seem to have small lenses and their tone rendition doesn't have the same effect on me.
Guess I'll just have to wait and hope for a pen-f mark 2...
> They also offer a pixel-combining mode now that shifts the sensor very small, precise amounts and then combines images to increase resolution up to 4x in the final result. You need a tripod and some time, but I'm looking forward to playing with it.
This is great. My pen-f has this, and the images it produces are wonderful. But there are some failure modes, like leaves moving, and the water has a high likelihood of looking weird.
The thing is that I want enough of an upgrade from my m43 (with which I'm happy enough) for it to make sense, and the xt4 and its lenses are too close to full frame to make it worth it.
The current smaller Fuji bodies are still fairly bigger than a pen-f, and, for my use, a net downgrade (no in-body stabilization). The only line-up that compares is Sony's, with the rangefinder format. But as I've said above, their tonal rendering isn't exactly what I like.
On an old Sony a65 I never use the exposure compensation or ISO or aperture preview physical-controls. There's a control wheel for shutter speed and aperture depending on PASM mode, which I use all the time.
But then I never use out-of-camera JPG, just raw at lowest ISO — to me, that seems to provide more possibilities.
(For example, converting a single raw file multiple times into an exposure series and then using the old Enfuse software to combine the "best" exposed pixels.)
What trouble have you had with Fuji RAW + Darktable? I've been using it without trouble, but I wonder if I'm doing something wrong?
But seriously, the out-of-the-camera JPGs are _ridiculously_ good. Sure, you're throwing away information, but I found I very rarely used that information.
If I'm doing something particularly intricate, complex, or delicate I might shoot raw. For walking around the city taking snapshots? JPG is more than fine.
Glad they're keeping the X-Trans. I love the film grain like look.
Sounds interesting. Having dropped out of the pro-digital-camera loop many several years now, I am only guessing that perhaps the "X-Trans" has digital (compression?) artifacts have an analog "feel"?
EDIT: did some Googling, sounds like it's the Bayer pattern the CCD uses that is unique — interesting: https://petapixel.com/2017/03/03/x-trans-vs-bayer-sensors-fa...
I moved up the the X100V when that came out to try and get even better AF performance (most pictures are of my busy kids) and it has been a worthy upgrade.
Since then I’ve tried an X-T3 and an XPro2. I fairly quickly sold both. The XT3 is just too big for me to want to take out. The XPro2 is smaller but still much bigger than the X100.
As hard as I tried to justify spending more money, the X100v is basically perfect for my needs.
A Leica Q2 is on my wishlist but I wonder if it too would feel too big to take with me most of the time.
It’s true that the best camera is the one you have with you and because of the size of the X100, you can basically have it with you all the time.
I have a 24MP A7-III. The very lenses I use can resolve at 60MP siblings, too.
Modern glass foundries of Nikon, Canon, Sony and Fuji can create pretty impressive glass nowadays.
I wonder if the future would not be the opposite? Store low-res images and upscale those you need with the help of AI.
Maybe in the future we can have just one super high resolution image for a shoot, store all the other images as super low res, and use the high-res image as a reference when upscaling. That would be very cool I think.
A 2TB drive is very sufficient for any amateur photographer using a camera in 24MP-40MP range.
We store raw images to archive what we got. AI interpolates. It's not acceptable.
No. Maybe for you, but not for me. I have a 4TB drive and it's almost full. I need to upgrade to 6 or 8 TB and it's a pain.
Source: I develop raw processing software.
It also has a downside in that it reduces fine detail somewhat compared to Bayer. In practical terms it also has a downside in being less supported by software, maybe because raw processing software developers consider it a big pain. ;)
I'm happy they made X-trans also for the 40mpx, it should improve the details (which is a weak spot due to the 3x3) significantly and still maintain the benefits of a 50% larger color sampling area.
X-Trans has less color sampling per unit area, so you get more chroma noise.
So the demosaicing either gets lots of color artifacts, or it has lots of baked in chroma noise reduction.
On Bayer, you can simply choose a different demosaic method (IGV or LMMSE, for example), or apply additional chroma noise reduction, and the playing field is leveled noise-wise.
However, since we are now at 40mpx I suspect the level of details will be limited by other things than the color filter anyway. If someone makes a similar camera with bayer (they haven't yet), it might not be able to resolve any more details.
If so, wouldn't that eliminate the downside of less details, while keeping the benefit of sampling from a larger area?
All should not be lost though. A lot of noise in a picture is independent of a pixel's neighbor. So, if you have more pixels, you can average out a neighboring pixel to get better low light sensitivity.
That said, all of us do love new toys : D
DSLRs with sensors the same size as 35mm film were called "full frame".
Canon uses 1.6x, and everyone else is on 1.5x bandwagon.
I love m43 systems, but their low light capabilities leave something to be desired. I shoot tango nights sometimes, and anything smaller than a FF system can’t keep up with the lack of photons, unfortunately.
Other than that, they’re brilliant little systems.
There is some stuff I shoot:
However, when you're trying to shoot movement with 1/60, f/2 @ ISO10K, nothing smaller than FF cuts it. It's almost pitch dark and flash is strictly prohibited.
I love more compact systems, but we can’t bend physics, for now.
Same for full frame. That’s equal to 35mm film.
So, why does it have the same processor? Is this camera fast enough for the kind of photos you’d make with that sensor? Is the market for these things not large enough for them to design the larger body that would be needed to house the larger battery needed to power that? Or does such a processor not exist, and is the market too small to create it? Or something else?
https://www.fujirumors.com/list-of-fujinon-xf-lenses-that-ge...
In any case, you're almost always capturing video in pre-processed 422, at which point the image processor has already debayered the image.
Anyway with used FF cameras now well under $1000, I have no idea why anybody would buy an APS-C camera. Even if you love some APS-C lens, you can just mount it on a FF camera and crop. Usually the camera will show you the cropped view in the viewfinder, but raw photos are taken at full resolution for you to manually crop later.
Although I tend to just shoot on auto-exposure and bracket 3 shots, 0EV, -2EV, +2EV. One of those will usually give me what I want, without any messing around with dials.
FF cameras are great, but a totally different thing.
I took a XT-3 with me to Tokyo. While it’s a nice camera, I still found it too bulky but without the benefits of FF.
I ended up going with a A7R3 combined with a RX100. I leave the hotel room with the former gets taken when I know there’s something important I want to capture, if it’s at night, or when I know bulk isn’t an issue. The latter is taken when bulk is of concern.
I have an A3 print in front of me, from a photo taken with 24Mpx a7iii, and cropped a little in post. You’d never look at it and think “it needs more resolution”.
It’s not very hard to get moiré on a 45mpx sensor with a modern lens, hence we most definitely need higher resolution cameras. But even with a lower resolving power than that there’s an advantage.
FF = full frame, refers to sensor size. They are about 1.5x - 1.6x[0] in surface area compared to APS-C sensors.
Sensor size has various benefits in how you frame your photos, and can have larger sensor pixels at the same resolution, which can result in less noisy photos.
One such explainer of the differences: https://expertphotography.com/full-frame-vs-aps-c/
[0] EDIT: See below, this was a mistake; this is the crop factor, not the difference in surface area. It's more like 2.5x (depending on the exact camera/sensor.)
Actually more like 2.3× in surface area.
For example, you might using a FF camera with a 52mm f/1.2 lens, in order to get the very shallow depth of field this gives and/or the very high light-gathering ability. A good 52mm f/1.2 lens is expensive, but quite feasible (they've existed for many decades).
To get the equivalent effect with an APS-C camera, you would need a 35mm f/0.8 lens. As far as I know, no such lens exists (though it's theoretically possible to construct a lens as wide as f/0.5).
In practice, FF may have other advantages, due to engineering considerations. For example (though I don't really know), it might be that electronic noise from the rest of the camera has more effect on noise during pixel readout when the pixels are smaller.
The Arri Alexa's Alev III sensor, which has won basically every Best Cinematography Oscar for the last decade, shoots at about 3k resolution (~7MP) yet holds up perfectly fine on huge cinema screens.
Majority benefit from shooting at a higher resolution is the ability to crop/zoom during processing. Sometimes, it's the ability to actually reduce the resolution to get rid of noise/grain.
Perhaps somebody is giving more priority to weight, size, price. Or lacks interest in shallow depth of field.
> Even if you love some APS-C lens, you can just mount it on a FF camera and crop.
Cropping a FF photo, especially from a cheap FF camera, may result in lower resolution than what an APS-C camera offers.
You can get a feel by how much post-processing is done by installing something like "opencamera" on a flagship phone, disable all filters and look at the raw file (if your phone allows it) and compare the results.
I went out with my android shitphone and someone brought their iphone 12 pro max. Here's a comparison (opencamera + rawtherapee vs iphone camera jpeg):
https://i.imgur.com/0E3KoBq.jpg
The iphone looks awful. It overbrightened/hdr-ed everything and made the sky blue (it was grey).
I have an iPhone and won’t likely switch to Android again any time soon, but it’s good to know that there are such flexible options now.
But, most people are happy with the default output from iPhone in most circumstances.
Now I wonder, why would people buy FF cameras (especially if they don't shoot 35mm film cameras)? They are still bulky. Pick APSC or MFT which is smaller. If you want superb image quality without compromise, then skip FF and join the medium format club.
:D