"The best" X-Trans sensor

Omar Gonzalez made a video "Which Fujifilm X-Trans Sensor is really the BEST?", to which Ritchie Roesch responded with a text "Comparing X-Trans Sensors".

"The best" for whoom, for what, when, where etc. - the question is improperly formulated as is always the case with "the best" questions, but anyway the question of differences between the X-Trans sensor generations (I, II, III, IV) is interesting. IMHO the generation I (X-Pro1, X-E1) is my favourite for its (analogue) film and grain like look, the generation IV (X-T3 & 4, X-S10) gives punchier images, the generations II & III maybe offer nicer colours. Of course, it all depends on many other variables, uses and habits.

What is your take, and why-when-where?

I've watched the video and found it interesting, although I find the discussion as such rather tiring. So long as image quality is sufficient, the video also illustrates that you can reasonably replicate any one sensor's character in post, should it be necessary.

I'd always choose the camera with the capabilities and features I like, and tweak (where necessary) the images to look the way I want them to look.

To that end, I would neither stick with an old camera that no longer meets my needs for its sensor's character, nor would I move to a newer generation if my IQ and functional needs are still met by a less than current generation.

Omar is entertaining and usually very accurate. His conclusions in this instance pretty well reflect what the consensus of opinion among members here has indicated over time and in piecemeal fashion.

Essentially, the 16mp, 24mp and 26mp sensors (X-trans II, III and IV) are all pretty close to each other in IQ, but the color temps and processing differ in arriving at default raw settings and JPEG output. Essentially, the older sensors are warmer and the new more color accurate. The 16mp sensor tends to show a lot of red channel saturation (and I would add that this can easily result in red channel blow out - it certainly was a problem with the Pentax implementation of the same Sony sensor design).

Perhaps the most important take away from Omar (and mentioned in some postings here) is that the JPEG processing on the 24mp sensors seem to be cleaner than the new sensors. Possibly, the greater emphasis on video performance is related to the relatively minor compromise on the JPEG stills side.

I think Omar might have missed one important factor - and that is the 160 ISO setting on the new sensor is yielding a cleaner base result. If I recall correctly, he simply standardized to 200 ISO for comparison purposes... and that isn't quite fair.

Another point not made by Omar relates to the more crop-able aspect of the denser sensors. Omar noted that for real-world viewing, 16 mp looks pretty much as sharp as the denser sensors. True, but you wouldn't want to crop much at all before you run into trouble - especially so with the newer higher-resolution computer monitors. Not to mention that some of us still like to make large prints on occasion.

The key takeaway here is that the older sensors have less dynamic range.

In Omar’s video, the older sensors produce more saturated colors SOOC. I think what has happened is in each successive sensor generation, the DR and noise performance increased, meaning truer Color reproduction. Someone correct me if I’m wrong but that’s my take on it.

When it comes to AF performance, the newer the better. Moreover old sensors work worse or not at all with AF adapters (Fringer/Viltrox)

I have XT20 and XT3 and, to be honest, can't see any advantage of X-Trans III over IV (I always shoot RAW).

Cheers,

Artur.

Are you sure that for AF the newer the better really holds? Evidence?

Emile15 wrote:

Are you sure that for AF the newer the better really holds? Evidence?

I can't provide any evidence regarding 1st and 2nd X-Trans but here are few facts from my experience:

• When I had Viltrox EF-FX1, AF worked significantly better on X-T3 than on X-T20.
• Viltrox EF-FX1 didn't work on X-Pro I (1st X-Trans) and Fringer doesn't support PDAF on 1st X-Trans.
• When I switched from X-T20 to X-T3 (for BIF) I noticed significant AF-C improvement 
• According my experience (didn't do any tests) low light AF performace was better on X-T3 even with FW 3.21.
• The 4-th X-trans on X-T3/X-T4  supported by the latest FW 4+ has really good AF especially  AF-C and eye tracking (compared to X-T20)

Well, I bought X-T3 instead X-H1 due to one reason - AF (based only on what I read -also on this forum). I believe that despite X-T20 and X-H1 are sharing the same sensor, overall AF performance is better on X-H1 due to better software.

There are a lot of opinions here on the forum, saying that even old sluggish lenses (56/1.2, 60/2.4) work much better in terms of AF on newer bodies/sensors.

It seems natural, that newer sensors shouldn't be worse (at least in terms of AF) than previous (let's see what we get with new high density 36-40 Mpix sensor in X-H2). There are more AF points, better software etc - I can confirm this on my two bodies.

BTW I also read about better low light performance (less noise) on 3rd vs 4th generation, but even if so - I didn't noticed significant difference.

No offence, but I wonder if you have any evidence/use-case when older sensor is better in terms of AF. If so - please enlighten me

I can understad that the old 1st gen has different nice rendering (according personal preferences) - but it's hard to believe that AF performace was better.

Cheers,

Artur

Pocket Lint wrote:

The key takeaway here is that the older sensors have less dynamic range.

The review indicated the newer the sensor the less contrast produced. This referenced the processing and was not intended as an indication of dynamic range. I do believe there have been very minor DR improvements in each version related to added detail due to increased pixels... and the lower ISO floor on the newest sensor. I don't think Omar measured DR, though.

In Omar’s video, the older sensors produce more saturated colors SOOC. I think what has happened is in each successive sensor generation, the DR and noise performance increased, meaning truer Color reproduction. Someone correct me if I’m wrong but that’s my take on it.

Saturation and color precision are a function of processing parameters whereas DR and noise performance are measured without regard to the color characteristics applied. When processing, the red channel can blow out (due to over-saturation) and lower detail... but the native DR of the sensor should be measured before processing.

YouTube seems to be full of these videos at the moment.  I like Omar but I’m not sure how helpful comparisons like this are; at the end of the day all these “filmic” labels are hugely subjective and personal and there is more to all of these sensors than simply the look of the JPEG’s; each one comes packaged in a different camera with different abilities and failings, all of which also need to be factored in.

Having used every X-Trans generation and been well down the rabbit hole of which look I prefer, the conclusion I’ve always come to is it’s a fools errand to look back.  The current cameras and sensors do a better job in more situations.  X-Trans I cameras were a pig to use, X-Trans II bodies did terrible things to skin tones at higher ISOs, X-Trans III was great but had some issues with AWB, and X-Trans IV - honestly - is the best of the lot IMO, especially when you take classic neg into consideration.

Threaded wrote:

[...] each one comes packaged in a different camera with different abilities and failings, all of which also need to be factored in.

Exactly, as sensor is an integral element of the each camera model, it shouldn't be perceived/discussed as interchangeable part like CPUs in PC systems.

I believe Omar's idea was to focus on some common sensor's output attributes  specific for all implementations of each X-Trans generation. I don't want to justify whether is was helpful, useful or just casual informative material.

Having used every X-Trans generation and been well down the rabbit hole of which look I prefer, the conclusion I’ve always come to is it’s a fools errand to look back. The current cameras and sensors do a better job in more situations. X-Trans I cameras were a pig to use, X-Trans II bodies did terrible things to skin tones at higher ISOs, X-Trans III was great but had some issues with AWB, and X-Trans IV - honestly - is the best of the lot IMO, especially when you take classic neg into consideration.

I agree with your overall conclusion

Cheers,

Artur

The Davinator wrote:

One of the reasons I have never so.d my X-Pro 1 is the color. In the X-Pro 3, greens are greyish and not pleasing to me.

I can see preferring an older camera if you are strictly a jpeg SOOC shooter and really like the particular look you’re getting out of it. When it comes to which has the best sensor though, the color you’re seeing is almost entirely determined by the processing - you may or may not be able to reproduce the color and contrast from the older camera with the the limited jpeg setting options, but you can absolutely do it with the RAW file. This difference in greens could easily be corrected and a custom profile produced in a couple of minutes that will always deliver the greens you prefer. For RAW shooters, I just don’t get the preference for older sensors based on color alone. For jpeg shooters, maybe, but I think you can probably get the results you want (and then some) with some careful adjustment of the jpeg/WB settings.

/nt

Novim wrote:

Omar Gonzalez made a video "Which Fujifilm X-Trans Sensor is really the BEST?", to which Ritchie Roesch responded with a text "Comparing X-Trans Sensors".

"The best" for whoom, for what, when, where etc. - the question is improperly formulated as is always the case with "the best" questions, but anyway the question of differences between the X-Trans sensor generations (I, II, III, IV) is interesting. IMHO the generation I (X-Pro1, X-E1) is my favourite for its (analogue) film and grain like look, the generation IV (X-T3 & 4, X-S10) gives punchier images, the generations II & III maybe offer nicer colours. Of course, it all depends on many other variables, uses and habits.

What is your take, and why-when-where?

The best x trans sensor is the one you own.

EMs1950 wrote:

I shoot in RAW format and for me there is no good or even best X-Trans sensor.

I have worked with the X-E1, the X-E2S, the X-T1, X-T2 and X-T3.

And besides that with the X-A1 and currently with the X-T100.

My experience is that nature and landscape photos with an X-A1 and Bayer sensor beat the cameras with X-Trans sensor with the same megapixels showing much better structures in the details.

Also, the photos with my X-T100 (and certainly those of the newer X-T200) show more detail and better rendering compared to X-T2 or T-T3.

There are different methods developing the RAWs. The worst result for X-Trans comes from Lightroom, as it has for years. Even with the current LR CC Classic.

CaptureOne develops X-Trans much better. I still see a slight advantage for the Bayer sensors on my X-A1 and X-T100.

An almost artful effort to work Bayer preference into a conversation entirely devoted to the X-trans sensor comparisons. Whether we like it or not, all of the Bayer-sensors are placed on Fuji bodies that are otherwise essentially entry level in terms of performance (and in some instances build).

Yeah, I wish the higher level Fuji bodies would give us the option to decide between the two arrays. But it really doesn't make much difference - as long as the (now majority) of us here have finally left behind slavish devotion to Adobe incompetence and greed.

In any event, please let's not highjack the thread...

The Davinator wrote:

Erik Baumgartner wrote:

The Davinator wrote:

One of the reasons I have never so.d my X-Pro 1 is the color. In the X-Pro 3, greens are greyish and not pleasing to me.

I can see preferring an older camera if you are strictly a jpeg SOOC shooter and really like the particular look you’re getting out of it. When it comes to which has the best sensor though, the color you’re seeing is almost entirely determined by the processing - you may or may not be able to reproduce the color and contrast from the older camera with the the limited jpeg setting options, but you can absolutely do it with the RAW file. This difference in greens could easily be corrected and a custom profile produced in a couple of minutes that will always deliver the greens you prefer. For RAW shooters, I just don’t get the preference for older sensors based on color alone. For jpeg shooters, maybe, but I think you can probably get the results you want (and then some) with some careful adjustment of the jpeg/WB settings.

That is the issue. With the first gen, I don’t need to correct...it is just right out of camera. That is the point.

Well, you were comparing RAWs which aren't right out of the camera. If you compare the right out of the camera jpegs, Despite the obvious exposure (and significant sharpness) discrepancy, they're actually very close - nothing a slight EC tweak wouldn't fix. I consider any necessary changes of settings to reach a desired result "adjustments". Any tool with so many variables will require a certain amount of fine tuning and calibration for optimal results. Should we also accept our monitor's settings right out of the box?