Fujifilm X-T3 Review

Fujirumors says no...

IMX571

if so why use first in a fuji camera
make sense

Same thing happened with the X-T2 and A6300. X-T2 launched first. A6300 lauched a little later with the same sensor, better image quality, and nearly all the features that have taken years for Fuji to add via firmware.

Only one brand makes X-Trans sensors. And not for any significant amount of time. It's not Adobe's fault. Then again, C1 not supporting the camera's built-in profiles like Adobe does is a pretty big letdown.

Thanks so much for your valuable contribution to this discussion. I’m sure Fuji will call you immediately to take your valued option into consideration and change their entire sensor tech based on your feedback.

Don't worry, they've already started (GFX-50S, GFX-50R). People don't seem to have a problem with those in industry standard software. Even Fuji is smart enough not to play games with pro stuff. ;)

Then the X-T3 is perfect. In a fairly lighted room (not completely illuminated) you will have remarkable results. So we agree.

At 6400 ISO the XT-2 is still better than the XT-3, from what I can see in these studio shots...

I shoot ISO 12k equivalent fairly frequently. And I don't like the noise levels pb the best sensor.

So, yeah, if you shoot anything indoors that's moving, you need this ISO.

@Karroly I still do not understand if I see the same images ...
HIGH LIGHT https://imageshack.com/a/img923/7687/yG3M66.png
LOW LOGHT https://imageshack.com/a/img923/1774/2WGL2H.png

Gianluca82,
I do not know where you got those screenshots. They definitely do not match what I can see on the DPR studio scene comparison :
https://www.dpreview.com/galleries/3570392623/albums/dpr

Why are you using ImageShack? It appears to require an account to see images.

@Karroly Set 6400 ISO for both, low light, central position of the frame. Do you really see the images of the X-t2 better?

@UncoyDP If you do not have an accont set 6400 ISO for both, low light, central position of the frame.

Thank you for the information. I get lots of highlights recovery from Nikon d500 raw files but maybe ARC does a good job and not the camera.

Thanks Richard

I have never had good results recovering from highlights. Shadows - definitely.

I remember reading something about how film sensitivity lowers as it receives more light, allowing good highlight recovery. Hopefully digital will get there one day. Individual photo site read based on individual exposure levels. Sounds like fantasy? I see no reason it can't be feasible.

While the continual increase in dynamic range for camera sensors will make it easier and easier to avoid having clipped highlights in a photo, clipped is clipped. Once you go all the way to white, there's absolutely zero image information to work with and therefore nothing to recover. AI-based tools like Content-Aware Fill can make a surprisingly good guess as to what *should* have been there—and I'm sure those types of tools will get better and more commonplace over time—but it will never be more than a guess, just as if you had a human illustrator come in and draw over the clipped areas in Photoshop. What's gone is gone.

They are now dedicated to that 5 stops shadows raising nonsense.

It's a real life test, they say...

The reason that shadow recovery is so important is precisely *because* highlight clipping is impossible to fix. If you have a high-contrast scene, it often makes sense to dramatically underexpose the shot in order to save the highlights, knowing that later on in post you can pull the shadows back up. Therr have certainly been times when I've underexposed by four or five stops; my camera doesn't have 5 stops of shadow recovery though, so I do it as an exposure bracket. Sunrises and sunsets are the classic application, but anytime you're shooting into the sun you're going to want to majorly underexpose and then recover the shadows.

I shoot sunset and sunrises quite often and never need to push the shadows 5 stops... and i rarely do HDR...

@xiao_xiang

"I remember reading something about how film sensitivity lowers as it receives more light, allowing good highlight recovery. Hopefully digital will get there one day."

Check out the Fuji S5 Pro. It's actually the reverse of most sensors, having a TON of highlight recovery (due to dedicated highlight pixels!), with decent shadow recovery. At 1:1 pixel view, it actually beats the D850 and A7R III in DR! Too bad that was the last of that line of sensors from Fuji. Imagine how much DR they could get out of that design with today's tech! O.o

Yeah, really up to 6400, I think Panasonic is making a mistake moving to Full Frame. That have a decent piece of the market with m43.

"whoopty do" ?

I think the m43 suffers from being an 'old fashioned' aspect ratio - it's just so 1960's. 3:2 seems to be a nice compromise between square and panoramic; and is probably the preferred shape for most older photographers who started out with film.

For me full frame sensors hit the sweet spot of the right (intuitive) aspect ratio combined with the light-gathering properties of a large sensor. Nothing to do with noise or extreme iso's, but everything to do with the option of large-scale prints.

Cheers.

LiOm, "4:3 (1.33:1) (generally read as "Four-Three", "Four-by-Three", or "Four-to-Three") for standard television has been in use since the invention of moving picture cameras and many computer monitors used to employ the same aspect ratio. 4:3 was the aspect ratio used for 35 mm films in the silent era"
https://en.wikipedia.org/wiki/Aspect_ratio_(image)

Cheers

Am I wrong or you can actually shoot raw at 3:2 with those m43 cameras?

Hi Causio. RAW files will always be 4:3 in m43 cameras. I think only Panasonic LX100 and LX100 II have multi-aspect sensors with which you can shoot in 4:3 or 3:2 without much loss. LX100 is 12MP and LX100 II is 17MP (becomes 16MP in 3:2 mode).

PostModernBloke,
"4:3 was the aspect ratio used for 35 mm films in the silent era"
For films, ie movies, but not for stills.
4:3 was choosen for the first digital cameras because it was the TV and computer screen format ratio at that time. And this time is over...and was much shorter than the reign of the 3:2 ratio for still photography... Looking at 4:3 pictures on a 16:9 screen sucks, and being obliged to crop and lose resolution also sucks... It is also the case when starting from a 3:2 picture, but to a lesser extent.

Karroly, do you really look at photos on your TV?

ozturet,
Yes, I do. I never print pictures and I look at them on today screens, mainly on a UHD TV screen for the highest resolution. And whatever the screen size (TV, computer screen or smartphone), their ratio is 16:9 or close...
4:3 is dead...

Ahhh the good old days.

Should we also refer to excellent fuel economy by the standard of a 1950's American car?

@cosinaphile yes to shoot at 12800 you do need a highly specialized piece of equipment ..... its called a camera.

Some people make interesting and thoughtful comments to help, enlighten or otherwise discuss. Others, like to demonstrate their inability to interact like an ordinary person.

Xiao Oh dear... Are you as boring as this in person?

Or good high ISO performance and with a fast prime.

DxOMark has never tested X-Trans sensors and has ignored Fujifilm in general.
Recently they seem to only be testing mobile phone cameras.
PhotonsToPhotos has always tested Fujifilm including X-Trans.
For example:
www.photonstophotos.net/Charts/PDR.htm#FujiFilm%20X-T2,FujiFilm%20X-T3
and:
www.photonstophotos.net/Charts/RN_ADU.htm#FujiFilm%20X-T2_14,FujiFilm%20X-T3_14

@808_freedive: It also led the various brand marketeers into competing on synthetic benchmarking to keep the smaller chip cams appear competitive

huh? where did you see that? DxO is the land of big sensors, high mp count and upgrade path myth! A smaller sensor will never have any chance to score there...

Colour is not subjective and not a matter of opinion. It is measurable, quantifiable and achievable. That means it is comparable. In many industries colours are defined quite clearly. But the amateurish reviewing of digital cameras avoids the issue like a cat avoids the hot milk.

Color, color accuracy, color reproducability are absolutely measurable and quantifiable. Arguing against that is absolute nonsense.

Anyone involved in product development, production, marketing, printing, packaging, and many more would laugh at that assertion. You photographers are so full it of sometimes.

Appreciation and preference for colors is subjective and opinion based.

The two are very separate.

An opinion about things measurable, quantifiable, and factual is worthless. Evidence is what you use to argue against facts. The moon is not made of green cheese however many times you vote for lying trump.

It depends. If you have a fast prime lens (e.g. The 23mm), you don't need IBIS nor OIS as long as your shutter speed isn't slower than e.g. 1/60-1/80.
If you go slower (e.g. 1/30) I would recommend a tripod.

Most Fujifilm zoom lenses have OIS, which is seen by many often to be better than IBIS.

For video, I would recommend a gimbal. IBIS and OIS might help a bit, but only a gimbal can really eliminate shaky images. Especially when there is a lot of movement.

RvEL

You've fallen into the old my camera doesn't have it so I don't need it trap we all do on occasion.

@BluebomberTurbo - yes there’s a little more blue noise on the older camera in that specific scenario, but switch to jpeg and not only does that noise disappear (obviously) but the XT2 has more detail, suggesting that blue noise is easily dealt with and the underlying image data is still superior. It’s even more obvious as you increase ISO further.

I am not biased, i'd love an XT-3. I couldn't give a penny about the slight high ISO step-back. However, it IS there and quite interesting just as it was interesting on the 6D. These things go cleaner not noisier. So it is something to mention, a break in a pattern.

@marc petzold

"Capture One does give much better Color Output - than Adobe does."

Depends on the camera. For my A6300, the color is horrific. Green skin and sunrises. Switch to the A6500 profile, EXACT same sensor, and things look normal again. No reason for that. This is especially noticeable in red, where the A6300's C1 profile leans heavily towards yellow. Not sure where they got the idea it should, since the OOC JPGs between the two cameras (and later ones) are all similar:

https://www.dpreview.com/reviews/image-comparison?attr18=daylight&attr13_0=sony_a6300&attr13_1=sony_a6500&attr13_2=sony_a9&attr13_3=sony_a7iii&attr15_0=jpeg&attr15_1=jpeg&attr15_2=jpeg&attr15_3=jpeg&attr16_0=100&attr16_1=100&attr16_2=100&attr16_3=100&attr171_0=1&normalization=compare&widget=1&x=-0.032586919444676385&y=0.6077722840767557

Continued...

There are many reasons I don't like C1, and this is probably #1. The fact that it's VERY hard to create ICC profiles to correct issues like this is easily #2, as it's VERY easy to create basic profiles for Adobe, and have things look correct in just a couple of steps.

@Threaded

"suggesting that blue noise is easily dealt with and the underlying image data is still superior. "

Well yeah, internal NR is tuned for the specific camera. Just like the Nikon D5 and Sony A9 do some amazing things with high ISO JPGs. But use RAW, and all of those optimizations go out the door. You can probably eventually get to that level, but it'll be a lot of work. And each individual image can have a different level of noise, so there's no real universal preset.

BBT, i use the X-E1. And second, i can't stand LR. And the stupid, silly catalogue all feature - further, X-Trans IQ from LR isn't good enough. 3rd, LR is still being sluggish from handling. If one wants to get the most out of the X-Trans Sensor, use Iridient Developer, or on Windows X-Transformer, Capture One is also a great idea, instead of Lightroom. Capture One does also have better tools, than LR from everyone's infamous monopolist -Adobe.

Ebrahim, the difference between the X-T2/T3 DR is only at the highest ISO, and therefore suble. The difference between the 6D & 6D II is a complete different story, it affects the whole DR, at the full ISO step, not only on ISO 25600/51.200, and also, it's neglible, not even +/- 0.5 EV, you can see it for yourself here:

https://www.dxomark.com/Cameras/Compare/Side-by-side/Canon-EOS-6D-Mark-II-versus-Canon-EOS-6D___1170_836

No offense. ;) Good Light!

@tbcass - RvEL hasn't necessarily fallen into anything, some of us actually prefer not to have IBIS. It adds cost, weight, size, heat, battery load, etc. It is not needed for all kinds of photography, and in fact, some say it can be detrimental to image quality depending on the shooting scenario (Panasonic has even admitted this).

The X-H series has IBIS, nobody's suggesting that shouldn't. Some of us are just happy there's a non-IBIS flagship option. What's so wrong with that/why does every model need to have all the same features as each other?

People are bananas around here...

@BlueBomberTurbo - so did they just not bother tuning the XT3’s in camera conversion? If it’s raw is genuinely superior, why aren’t it’s own jpegs? The fact is if the XT2 can deal with blue noise internally and get a better image, then so can your raw editor. Bottom line is blue noise or not, it has the advantage over the XT3 here as well - and really, why wouldn’t it, when it clearly has the advantage in daylight too.

The XT-3 is definitely more of a step-back in noise performance than the 6DII was. Here: https://www.dpreview.com/reviews/image-comparison/fullscreen?attr18=daylight&attr13_0=fujifilm_xt3&attr13_1=fujifilm_xt2&attr13_2=canon_eos6dmkii&attr13_3=canon_eos6d&attr15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&attr16_0=12800&attr16_1=12800&attr16_2=12800&attr16_3=12800&normalization=full&widget=628&x=-0.10143652212560715&y=-0.6257770623719088

Gosh, just get a Life! Now, you're clearly biased - how does it feel, being busted? :> I don't take you for serious anymore. Someone, who compares APS-C vs FF Sensors into Noise & DR performance - really, with comparsions like this, you've lost all credibility. It's a apples vs. oranges comparsion.

Just for the record, you'd also compare a Toyota Yaris with a Ferrari, and then moan about that the latter is faster. Keep complaining, and posting about Testshot noise comparsions at ISO 12.800 and have Fun!

Other People are just buying and enjoying this camera, and doing real photographs with it. I am really bored from this, you're exactly that type, which is fishing for the hairs into the soup. Just for the record, 95% of my photography is between ISO 100-800, max. up to ISO 1600, if ever.

Good Light!

I sm not comparing the XT-3 with the 6D at all. I am comparing the extent of the step-back in noise performance.

@marc petzold

"If one wants to get the most out of the X-Trans Sensor, use Iridient Developer, or on Windows X-Transformer, Capture One is also a great idea, instead of Lightroom."

One of the reasons I'll never buy a Fuji X-Trans camera. They dug themselves a hole by trying to be different in a meaningless way. I'm not changing my workflow just because of Fuji's arrogance.

And no, I don't use LR, I use ACR. Faster, and easier to organize. Don't need the DAM part of the system. And at least Adobe offers its own profiling software, free of charge. Though there are WAY more powerful programs out there.

@Threaded

"The fact is if the XT2 can deal with blue noise internally and get a better image, then so can your raw editor. "

Funny, because not even Fuji can create external editing software that can do that. They rely on the camera's internal processing, instead of writing dedicated (and potentially superior) software, to be able to edit RAWs on the desktop. You literally have to have your camera plugged into the computer to be able to edit anything. If Fuji, the company that created the hardware/software running the camera, can't even do it, why should anyone else be expected to?

@BlueBomberTurbo - Fuji offer Silkypix, Capture One, and yes their own innovative software using the camera’s own jpeg engine - so what? Outsourcing raw software is pretty common. None of this changes the fact that the XT2 sensor is still outperforming the XT3’s at high ISO’s, even in the example you gave.

BBT, one of the reasons, i dont use Adobe anymore - worse IQ than the competition, and i personally hate monopolists, capiche? If one isn't being able, or into the mood, to use an alternative software to get the most IQ out of it - then you're simply not the userbase, for which the X-T3 was being developed for. So many people are just like lemmings, going mainstream, into every way...boring...

Good Light!

@Threaded

"Fuji offer Silkypix, Capture One, and yes their own innovative software using the camera’s own jpeg engine "

Just saying that neither Silkypix or C1 can compare to the output of the camera directly. It's been said as much by Fuji users. Nikon and Canon have software that duplicates the look of their SOOC JPGs. I believe Sony does, too, despite offering free C1.

Nikon D5 !!

Samaistuin and BlueBomberTurbo

If you use the Comp choice where all are viewed at the same size (resolution) the Sony A7iii is the best. Try it and you'll see.

https://www.dpreview.com/reviews/image-comparison/fullscreen?attr18=lowlight&attr13_0=fujifilm_xt3&attr13_1=sony_a7sii&attr13_2=fujifilm_xt3&attr13_3=sony_a7iii&attr15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&attr16_0=25600&attr16_1=51200&attr16_2=25600&attr16_3=51200&normalization=compare&widget=628&x=-0.0934833204034135&y=-0.9496206647398844

To be honest, I was really against IBIS if it meant Fuji had to compromise IQ. I preferred to bring two bodies when travelling, a Fuji and a M43 with IBIS for low light shots. But yet I bought a X-H1 because it meant having only one, albeit big and heavy, body to travel with.
Then, last holidays, I brought the XF 16mm f1.4 with me, and I didn't really *need* IBIS, f1.4 1/15s and 1600 ISO were enough. Without IBIS, I imagine I could have had good results with 1/30s and 3200 ISO. Not really a problem with current cameras.

Also, since the days I used XM1 + 10-24mm I never needed a tripod anymore even in churches or castles. It's not about questioning the usefulness of stabilization, it's about questioning the usefulness of IBIS, there's a difference.

How many Fuji primes have OIS again? Zero? Is that the right number? That's the usefulness of IBIS, all of your lenses are stabilized. Heck, even that Fuji 2.8 zoom lacks OIS.

This talk about "compromising IQ" is total nonsense. It was marketing speak for "we didn't plan for this and now people want it but we don't have the tech". Fuji spouts some story about how IBIS won't give optimal image quality (to try and placate existing users), which is patently false as Olympus, Panasonic, Sony, and Pentax (Nikon now too) have shown with countless IBIS cameras. If it had any truth to it, Fuji wouldn't have added IBIS to the XH1. Fuji is late to the game here, that's the only issue at play.

If you want to keep beating the "IBIS can't be done without degrading image quality" horse, feel free, but you're only fooling yourself. Modern IBIS systems work exceptionally well and have zero disadvantages. Oly/Pana put IBIS into tiny, cheap, entry level bodies. Surely Fuji can figure this out.

Honestly I would love to use Fuji cameras. I love the lenses, I love the design and controls, but I'm not going to switch to a larger format system and get worse image quality in low light because there is no IBIS (or OIS with most of the lenses I would use in low light). IBIS is the only clear thing the Fuji X system is missing, and I hope they figure it out.

@earthquake: "IBIS is the only clear thing the Fuji X system is missing, and I hope they figure it out."

so what's this thing called on the X-H1??? I don't understand your post, or maybe you didn't read anything I wrote. I was against it, per Fuji's explanation. They released the X-H1 with IBIS, I'm happy with it, there are no issues with IQ.

But should IBIS be on each and every camera and make them bigger and possibly more expensive? I don't think so. Should they work on reducing the size of the mechanism? definitely. It looks like there are only engineers in this forum who are saying I'd do this in my sleep.
M43 is using a smaller sensor and when talking about tiny cheap entry level, the IBIS on the E-PM2 doesn't help much with wide angle lenses, the E-PL series has IBIS but again performance is far from E-M1 II IBIS and their bodies are getting bigger as well. Panasonic has no IBIS for anything below GX series. And how long did it take them to have decent IBIS??

Don't get me wrong, I love the ability to shoot 1s at 100mm on my E-M1 II + 12-100mm f4 but this kind of performance remains unique. And Olympus has been working with IBIS for quite a while.

@Shadow9d9: same deal again, where did I say there was a compromise? I said I wouldn't want it *if* I had to compromise. ISO 200 with my 16mm? means shutter speed would have been 1/2s. Even with M43 I would have less keepers at 1/2s. I just checked my library and there not that many pictures that are sharp at 100%, I think 1/4s is a sweet spot though.
And I think it is better to have some room with 1/15s ; on holidays you can't always stop and take the ideal shot, and at night, having walked all day, it's better to be safe and end up with good shots instead of a blurry picture at ISO 200. Maybe if I'm confident enough I'll go down one stop next time.

what is clearly needed is a monopod that looks like a walking cane, but taller to lean on or like some hiking canes, then you can place cam on it and steady it via techniques.

but surely 51200 is not needed ever unless in dark, but then what the heck are you doing, peeping tom or stalker or prevert ????? ;)

all the crap sold out there and no useful 'cane' monopods !!!! ;)

It wouldn't need to be at 1/2 in a majority of situations. And if it did, ypu'd simply do it a bit faster and bump iso. It'd still be a ton less iso overall.

Yes, ibis should be in every camera, like pretty much all panasonic, olympus, sony, and now nikon have been. Xh1 was released with older technology for full price...

That's not the link I asked (moreover, I asked *how* to get it, the one that sends you directly to the cameras you want, iso you want etc) and I know that raw converters are responsible for the colors, but thanks anyway :) Side note: all the converters I tried tend to give an orange tint to the red with X-T3 files. But perhaps C1 11 will fix this once it'll have the fuji profiles.

So XT3 iso6400 is not Canon's iso6400. Fuji iso6400 is actually Canon iso12800 (with same shutter speed and aperture values).

@Richard Butler

I see.

But doesn’t the Fuji X-Trans cameras still differ in their RAW’s, not just their JPEG’a? If you use the same lighting, SS, Aperture and ISO on the X-T3 and for example the Nikon D500 on the same scene, won’t either the X-T3 be darker or the D500 brighter (depending which one you expose for)?

ah this funny topic. I remember I tested I tested a Sony RX10 and Canon 6D with a f2.8 and with the same ISO, aperture and shutter speed, and the Sony was actually making brighter pictures. I never heard anyone complaining about Canon, the 6D was always praised for its high ISO performance. Maybe it was light transmission on the lens or something, my test was quick and dirty...

I think ultimately we shouldn't care about the numbers, only about what the camera produces. If a company cheats and it really matters, it will reflect on our ability to make low light picture with acceptable noise. So far I haven't tried other cameras and thought, here is that 2/3 EV that was missing on my Fuji.

Thank you. The difference observed between the X-T3 and X-T2 at highISO was for Adobe Standard Profile?

Thank you - I can see in the EXIF of the ACR developed JPG - Adobe Standard profile was used. Thanks

This is a good point, though, maybe not for this particular review, but maybe someone who owns an X-T3 and Adobe Raw could produce a comparison of that with SilkyPix?

Don't need to own an X-T3 as the RAW FILE CONVERTER EX 3.0 powered by SILKYPIX (Ver.8.0.0.0) is freely available for download.

Downloaded the ISO51200 RAW with simulated daylight lighting: DSCF2835.RAF and ran it through the Fuji RAW file converter 3.0 -
https://www.dpreview.com/galleries/7858448340/albums/fuji-x-t3-highiso-51200-raw
Two versions -
default setting -
XT3_51200_DSCF2835.JPG
custom setting with every parameter to do with noise reduction set to 0 -
XT3_51200_DSCF2835_NR0_95.JPG

FWIW - I've also added the <<X-T2>> ISO51200 RAW conversions for comparison to the same gallery album
https://www.dpreview.com/galleries/7858448340/albums/fuji-x-t3-highiso-51200-raw

For those of us reading on mobile, what is the outcome of your conversion in comparison to acr?

The Sony A9 has a slightly lower DR for reasons most likely unrelated to the stacked sensor - no such thing is visible on the 1" models, for one thing.

@Kharan

Sensor size plays a huge role when its FF vs 1". The bigger the sensor, the more data is recorded for each pixel. The RX100 has very tiny pixels, and only shoots 11+7-bit RAW, so there wasn't enough DR recorded to be impacted in the first place.