Latest Fuji X-trans vs Bayer, has X-trans lost it already?

This wasn't the case, say 5-7 years ago, but I recently checked Fuji OOC jpegs on DPR comparison studio and I see the Fuji Bayer sensor outresolving X-trans at camera defaults.

I know, I know, some of you might be thinking: "Not again, it was already proven beyond any reasonable doubt that Abobe ACR is of the the worst X-trans engines...", right? No, I am talking about the Fuji own Jpeg engine for Bayer that seems to outperform Fuji's bread-and-butter engine for X-trans at least at camera default settings.

So I am happy to give you a credit of doubt, yes you can squeeze extra details extra details with your Iridium from raws, so can do I with my old copy of LR and amazing X-A5 raws. But let's meet in the middle and agree that Bayer and X-trans offer similar performance today.

Point is that if even the mighty Fuji themselves can't show OOC X-trans advantages over their own entry level Bayer cameras, doesn't it prove that the days of the X-trans sensor advantages are over?

1st link: https://www.dpreview.com/reviews/image-comparison/fullscreen?attr18=daylight&attr13_0=fujifilm_xa5&attr13_1=fujifilm_xpro3&attr13_2=fujifilm_xt100&attr13_3=fujifilm_xt30&attr15_0=jpeg&attr15_1=jpeg&attr15_2=jpeg&attr15_3=jpeg&attr16_0=200&attr16_1=200&attr16_2=200&attr16_3=200&normalization=full&widget=1&x=-0.2122792937399679&y=-0.6203029880394706

2nd link: https://www.dpreview.com/reviews/image-comparison/fullscreen?attr18=daylight&attr13_0=fujifilm_xa5&attr13_1=fujifilm_xpro3&attr13_2=fujifilm_xt100&attr13_3=fujifilm_xt30&attr15_0=jpeg&attr15_1=jpeg&attr15_2=jpeg&attr15_3=jpeg&attr16_0=200&attr16_1=200&attr16_2=200&attr16_3=200&normalization=full&widget=1&x=0.3286136587998473&y=-0.167603984861482

PS. I'm quite surprised how X-Pro3 are indistinguishable from X-T30. Looks like exactly the same sensor, the same electronics around it and the same image processing algorithms.

And if someone is going to start the "it's out of focus, dah", I'm going to debate this by claiming that it's quite unlikely to see a consistent focusing error with X-T30, X-T2, X-T3 and not having it with X-A5 and X-T100.

The lower end cameras add significantly more sharpening by default than the higher end ones. That said, there is no good reason why the Bayer sensor cameras shouldn’t be every bit as as sharp.

kiriak wrote:

different jpeg sharpening could explain it,

but the overall impression is the same when I look at the RAW samples of the above links

Well, the default LR RAW processing is still far from optimal for X-Trans files.

kiriak wrote:

but the overall impression is the same when I look at the RAW samples of the above links

Agree, but let's not go that way. There will be a thousand of inputs that we should all switch to Iridium as it is known to outperform the Adobe ACR and our observations will be easily dismissed.

I prefer to focus on the native Fuji processing engine instead, if they can't show significant advantages, then again, what's the point of having a different Sony sensor for them today?

il_alexk wrote:

kiriak wrote:

but the overall impression is the same when I look at the RAW samples of the above links

Agree, but let's not go that way. There will be a thousand of inputs that we should all switch to Iridium as it is known to outperform the Adobe ACR and our observations will be easily dismissed.

I prefer to focus on the native Fuji processing engine instead, if they can't show significant advantages, then again, what's the point of having a different Sony sensor for them today?

X-Trans does still have a small advantage in the moire department and with optimal demosaicing I might be able to get a tiny bit more fine detail out of an X-Trans file, but I agree, with high MP AA filterless sensors it’s advantages hardly outweigh its drawbacks (for Lightroom users, anyway). I’d very much like to see a premium Bayer model.

Erik Baumgartner wrote:

il_alexk wrote:

kiriak wrote:

but the overall impression is the same when I look at the RAW samples of the above links

Agree, but let's not go that way. There will be a thousand of inputs that we should all switch to Iridium as it is known to outperform the Adobe ACR and our observations will be easily dismissed.

I prefer to focus on the native Fuji processing engine instead, if they can't show significant advantages, then again, what's the point of having a different Sony sensor for them today?

X-Trans does still have a small advantage in the moire department and with optimal demosaicing I might be able to get a tiny bit more fine detail out of an X-Trans file.

This used to be true in the past when we compared X-trans to Sony, Olympus or Nikon sensors without AA filters. But the XT100/XA5 Jpeg engine from Fuji is doing an amazing job to reduce the moire with X-A5. IMO it's one of the best jpeg engine I've seen.

, but I agree, with high MP AA filterless sensors it’s advantages hardly outweigh its drawbacks (for Lightroom users, anyway). I’d very much like to see a premium Bayer model.

If you switch to RAW and then compare, you'll see that results for both sensor types are similar, which means Fuji's JPEG engine in Bayer-sensor cameras likely applies more contrast and sharpening by default.

If you switch to RAW and ISO 6400 you will see that X-Trans still has an advantage with lower chroma noise.

But in general, I agree, there's little reason for Fuji to continue using X-Trans in the future. Modern Bayer sensors are excellent so staying with X-Trans is probably a political and marketing statement.

This is one of the times I wish some know-how engineer from Fuji is reading these posts and is willing to chime in to explain things. Why would a lower end camera have better JPEG output than a higher end one? I have an X-T3 and I almost exclusively shoot JPEG.

SpeedyNeo wrote:

This is one of the times I wish some know-how engineer from Fuji is reading these posts and is willing to chime in to explain things. Why would a lower end camera have better JPEG output than a higher end one? I have an X-T3 and I almost exclusively shoot JPEG.

Define "better" - they are more contrasty and higher perceived sharpness. I think the first reply answered that question. Typically, more consumer level products have increased sharpening, as they know those types of users are less likely to edit them in post.

Dr_J wrote:

If you switch to RAW and ISO 6400 you will see that X-Trans still has an advantage with lower chroma noise.

The chroma noise issue is apparent even in JPEG:

X-Trans clearly has an chroma noise advantage over the T100 sensor, and the X-Trans III cameras overall noise performance imo.

I think we can all agree that there isn't much in it either way.

Working with Capture One (I left LR several years ago - before switching to Fuji - due mostly to the precision dynamic color control unavailable in the Adobe products)... I have come to the conclusion that the Bayer sensor might have a very slight advantage at base ISO, and the randomness of the X-trans array noise is a slightly greater advantage at higher ISOs.

Even after years with X-trans now, I find it somewhat more challenging to get the best out individual shots due to inconsistencies involved in relatively smooth image areas vs. busier portions of the image. The Bayer pattern is better for smooth areas.

I was lurking around Fuji for a long time, but stayed away until reports of how Capture One had finally really nailed support for X-trans roughly 2-3 years ago. With so many LR users still not entirely satisfied, I have to wonder if Fuji is sticking with X-trans mostly out of the desire to be different. However, the alliance with Phase One / Capture One probably has proven out to work well for both companies.

Acrill wrote:

Dr_J wrote:

If you switch to RAW and ISO 6400 you will see that X-Trans still has an advantage with lower chroma noise.

The chroma noise issue is apparent even in JPEG:

Never saw chroma noise as an issue, it's easy to fix. Loss of details and loss of DR at high iso usually more difficult to compensate. And this is where Bayer seems to be better.

X-Trans clearly has an chroma noise advantage over the T100 sensor, and the X-Trans III cameras overall noise performance imo.

This could be subjective, but I prefer the Fuji Bayer jpegs over X-trans in the DPR studio shots even at high ISO. The difference in amount of details Fuji pulls out of X-A5 and X-T100 is amazing compared to washed-out X-Pro3 jpegs.

https://www.dpreview.com/reviews/image-comparison/fullscreen?attr18=lowlight&attr13_0=fujifilm_xa5&attr13_1=fujifilm_xpro3&attr13_2=fujifilm_xt100&attr13_3=fujifilm_xt30&attr15_0=jpeg&attr15_1=jpeg&attr15_2=jpeg&attr15_3=jpeg&attr16_0=6400&attr16_1=6400&attr16_2=6400&attr16_3=6400&normalization=full&widget=1&x=0.6168829959811521&y=-0.1610622729443509

No. X Trans hasn't lost it. It has still the same advantages/disadvantages. For me. absolutely nothing wrong with X-trans. Stunning resolution, low chroma noise, almost no moire.

As already pointed out, no conclusions can be drawn about the default JPEG output, until said default parameters are spelled out for each camera.

P.S. It is Iridient, not Iridium:)

Acrill wrote:

Dr_J wrote:

If you switch to RAW and ISO 6400 you will see that X-Trans still has an advantage with lower chroma noise.

The chroma noise issue is apparent even in JPEG:

X-Trans clearly has an chroma noise advantage over the T100 sensor, and the X-Trans III cameras overall noise performance imo.

Not sure, X-T100 here looks cleaner and sharper to me than X-Pro 3. The noise is much finer and eventually is easier to clean in PP.

biza43 wrote:

As already pointed out, no conclusions can be drawn about the default JPEG output, until said default parameters are spelled out for each camera.

I understand that these are the conclusions X-trans proponents would like to keep, however this is what was discussed at this thread, so these are the conclusions from what was really discussed.

• Fuji X-A5/X-T100 OOC jpegs are "more contrasty and higher perceived sharpness"
• Fuji X-A5/X-T100 OOC high ISO jpegs have more chroma noise, but still show significantly more details and contrast. I definitely prefer OOC Jpegs of X-A5 over X-Pro3
• Some users here tend to agree that Bayer and X-trans provide a similar IQ, with differences so minor that choosing a winner wouldn't make sense. I see this as an acceptance of the fact that the days of clear advantages of X-trans are over.
• There are claims that moving to a proper raw processing SW would change the balance in favour of X-trans, yet no proof has been given.

My personal problems with the last claim are:

• There is a fundamentally incorrect assumption that X-trans sensor will benefit from proper RAW processing more than Bayer, it will show more details with better colors and less noise. C1 is quite good for processing Bayer sensors too, just in case anyone missed it.
• There is a fundamental assumption that Fuji Raw Studio can't show any advantages of X-trans over Bayer, yet C1 or RT or Iridient will somehow magically do it. Please do keep in mind that if you process raws in Fuji Raw Studio you are very likely to end up with a similar demosaicing results as with OOC jpegs.
  If anything I can accept that LR handling of Xtrans raws is not good, but it doesn't mean that C1 can beat Fuji or LR or even C1 with Bayer raws.

Another problem with the last claim is that nobody bothers to prove it. It sounds like a mantra to me, and I personally don't believe in it, but hey, why wouldn't someone take DPR raws and use Iridient, C1 or any other SW and try to show the magical advantages of X-trans sensors.

So here is the challenge.

Take these raws and show what you can do with your best RAW engine. If you can beat fuji's OOC X-A5 Jpegs, I promise to do the same with X-A5 raws so we can see if there is any advantage for X-trans or how significant it is.

Rules are:

• All settings must be applied to the whole image, no local adjustments (except for demosaicing)
• Just a basic demosaicing, moire, contrast and denouncing are allowed
• Standard Fuji profile, WB "as shot", no messing up with custom LUT/contrast tables, let's keep vibrance/saturation under control, so we can compare apples to apples rather than our creative skills.

1. Low light, high ISO

https://www.dpreview.com/reviews/image-comparison/fullscreen?attr18=lowlight&attr13_0=fujifilm_xa5&attr13_1=fujifilm_xpro3&attr13_2=fujifilm_xt100&attr13_3=fujifilm_xt30&attr15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&attr16_0=6400&attr16_1=6400&attr16_2=6400&attr16_3=6400&normalization=full&widget=1&x=0.6202582464220795&y=-0.15956978761016444

2. Good light, low ISO

https://www.dpreview.com/reviews/image-comparison/fullscreen?attr18=daylight&attr13_0=fujifilm_xa5&attr13_1=fujifilm_xpro3&attr13_2=fujifilm_xt100&attr13_3=fujifilm_xt30&attr15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&attr16_0=200&attr16_1=200&attr16_2=200&attr16_3=200&attr126_2=1&normalization=full&widget=1&x=-0.21297495298157976&y=-0.6218702714186282

Anyone? Let's see how much your beloved X-trans is better than Bayer today.

P.S. It is Iridient, not Iridium:)

Thanks, my fingers can't type this, the Iridium automatically comes out when I type it. At least it's not Polonium

The X-Trans design was applied to 16 MP sensors in Fujifilm's first generation mirrorless cameras (e.g. X-Pro, X-T1).  At that time X-Trans cameras had two benefits; increased resolution due to removal of the AA filter and reduced color moire due to the different CFA pattern.

At 24+ MP on an APS-C camera most competitors have removed the AA filter from their Bayer equipped bodies.  So as a practical matter there should be little difference in resolution between Bayer and X-Trans sensors with the same MP count.

There still seems to be a small advantage with color moire reduction at 24MP for X-Trans compared with a Bayer equipped sensor.  The advantage should be further reduced if competitors start introducing 30+ MP cameras.

So the question that is often debated, with no definitive answer, is whether the X-Trans CFA is worth it due to the extra processing power needed and the different handling required for raw developer software.

The X-Trans CFA also makes it much more difficult for Fujifilm to implement things like pixel shifting.  That doesn't matter to me but it does to some.