What is advantage and disadvantage between X-trans vs Bayer array

Started Sep 7, 2013 | Questions
Charles2
Senior MemberPosts: 2,125
Like?
Re: Comparison - before you say...
In reply to dotborg, Sep 8, 2013

Before you say which is which, you might provide the RAF file.

Reply   Reply with quote   Complain
bowportes
Senior MemberPosts: 1,452Gear list
Like?
Re: Comparison - before you say...
In reply to Charles2, Sep 8, 2013

And it's interesting that a native micro four-thirds lens is not being used in comparison to the native Fuji lens.

The shots are also quite different compositions -- the threads are in different locations, etc.

Why not simply provide a link to both raw files?

 bowportes's gear list:bowportes's gear list
Panasonic Lumix DMC-G5 Fujifilm X-M1 Fujifilm X-T1 Panasonic Lumix G 20mm F1.7 ASPH Olympus M.Zuiko Digital 45mm 1:1.8 +9 more
Reply   Reply with quote   Complain
ulrikem
Regular MemberPosts: 211Gear list
Like?
Frequency selection demosaicking: A review and a look ahead (research paper)
In reply to ealvarez, Sep 8, 2013

You may want to read the following research paper, possibly that's what Fuji drove to use the 6x6 CFA:

http://david.alleysson.free.fr/Publications/VCIP08.pdf

Link was taken from the very interesting thread at another forum. Well worth reading!!

 ulrikem's gear list:ulrikem's gear list
Fujifilm X100S Fujifilm X-Pro1 Fujifilm XF 18mm F2 R Fujifilm XF 35mm F1.4 R Fujifilm XF 14mm F2.8 R +1 more
selected answer This post was selected as the answer by the original poster.
Reply   Reply with quote   Complain
Randy Benter
Senior MemberPosts: 1,997Gear list
Like?
Re: Comparison - before you say...
In reply to bowportes, Sep 8, 2013

Charles2 wrote:

Randy Benter wrote:

In this comparison, the E-M5 is shot with the Olympus 50mm f/2 and the X-E1 is paired with the Fuji 35mm f/1.4.

You know -- subgroup with OP? Multiple user accounts? Hidden info in the file?

This sounds more like an accusation than a question. The lenses are identified in the EXIF data of the image file.

bowportes wrote:

Why not simply provide a link to both raw files?

These are the standard DPR studio scene test shots. You can download the raw files from the comparison page of any camera review. http://www.dpreview.com/reviews/fujifilm-x100s/18

 Randy Benter's gear list:Randy Benter's gear list
Leica X2 Fujifilm X100S Leica X Vario Leica M8 Fujifilm X-Pro1 +34 more
Reply   Reply with quote   Complain
Charles2
Senior MemberPosts: 2,125
Like?
Re: Comparison - before you say...
In reply to Randy Benter, Sep 8, 2013

Randy Benter wrote:

... This sounds more like an accusation than a question. The lenses are identified in the EXIF data of the image file. ...
These are the standard DPR studio scene test shots. You can download the raw files from the comparison page of any camera review. http://www.dpreview.com/reviews/fujifilm-x100s/18

No accusation intended. Thanks for the link. The scene was familiar, but you remembered the source while I did not.

Reply   Reply with quote   Complain
umanemo
Regular MemberPosts: 161Gear list
Like?
X-Trans is Simply More Added Value, In Our Favor!
In reply to ulrikem, Sep 8, 2013

From this extensive read, it seems that many, including the OP, could learn enough from Ulri's references to finally end these cyclical inquiries about the usefulness of the 6 x 6 PR-CFA that FUJI has chosen to implement and refine as X-Trans.

Before anyone posts again on this subject they should have a full and comprehensive understanding of the cited documents that I suspect FUJI is already ammending with their own advancements in research and development.

Cudo's to Ulri for His/Her diligence in the understanding of Image Sensor Technology and sharing that with us, Cudo's to FUJI for their pioneering spirit that benefits not only us but the advancement of imaging technology exponentially!

Everyone touts (no offense to Zeiss) the evolutionary development of the dated algorithms long since summarized for the 2 x 2 Bayer design sensor array mostly due to the longevity of the array pattern's presence in the world of imaging. Yet the 6 x 6 PR-CFA clearly shows a mathematical advantage. Sluggish program development of the RAW conversion algorithms that maximize that advantage is inevitable just as it was with the Bayer. The 6 x 6 will progress evolutionally and in the end we will all benefit from the X-Trans sensor's extended capability.

Is seems quite excess-profit oriented to remain on the proven side of the fence. (Yes, that means you, SoCaNikon). There is no design evolution without the expense of R&D therefore FUJI should be supported for their pioneering spirit and the added value they are including in their X-Trans products.

We are simply the benefactors. I like to believe I am invested in receiving "Red Dot" IQ for a lot less "M"oney! Thanks FUJI for X-Trans, though I have yet to challenge it, I am confident that with what has been provided me, it will all be there when I need it!

Reply   Reply with quote   Complain
57even
Senior MemberPosts: 5,831Gear list
Like?
Re: X-Trans is Simply More Added Value, In Our Favor!
In reply to umanemo, Sep 8, 2013

umanemo wrote:

Many, including the OP could learn enough from Ulri's references to finally end these cyclical inquiries about the usefulness of the 6 x 6 PR-CFA that FUJI has chosen to implement and refine.

Before anyone posts again on this subject they should have a full and comprehensive understanding of the cited documents that I suspect FUJI is already ammending with their own advancements in research and development.

Cudo's to Ulri for His/Her diligence in the understanding of Image Sensor Technology and sharing that with us, Cudo's to FUJI for their pioneering spirit that benefits not only us but the advancement of imaging technology exponentially!

Everyone touts (no offense to Zeiss) the advanced development of the dated algorithms long since summarized for the 2 x 2 Bayer design array mostly due to the longevity of the array pattern's presence in the world of imaging. Yet the 6 x 6 PR-CFA clearly shows a mathematical advantage?

Is seems excess-profit oriented to remain on the proven side of the fence. (Yes, that means you, SoCaNikon). There is no design evolution without the expense of R&D therefore FUJI should be supported for their pioneering spirit and the added value they are including in their X-Trans products.

We are simply the benefactors. I like to believe I am receiving "Red Dot" IQ for a lot less "M"oney! Thanks FUJI for X-Trans, though I have yet to challenge it, I am confident that with what has been provided me, it will all be there when I need it!

I agree.

 57even's gear list:57even's gear list
Fujifilm X-Pro1 Nikon D800 Fujifilm X-E1
Reply   Reply with quote   Complain
DonSantos
Senior MemberPosts: 1,122Gear list
Like?
Wait and see
In reply to ealvarez, Sep 9, 2013

When the x-a1 comes out you'll see directly the advantage and disadvantage of the x-trans.

If they are indeed using the sony sensor then they are in trouble of a entry level camera giving better IQ then their flagship.  Only time will tell.

 DonSantos's gear list:DonSantos's gear list
Nikon D600 Sony Alpha 7 Voigtlander 40mm F1.4 Nokton Classic Nikon AF-S Nikkor 85mm f/1.8G Sigma 35mm F1.4 DG HSM +1 more
Reply   Reply with quote   Complain
dotborg
Veteran MemberPosts: 7,555
Like?
Re: Comparison
In reply to Dave Luttmann, Sep 9, 2013

Dave Luttmann wrote:

dotborg wrote:

Some people really want to attribute some mystical greatness to Fujifilm's X-Tranny sensor. Well here's a comparison. The colors aren't exactly the same as I didn't apply any profiles but who can tell me based on sharpness, detail, cleanness, goodness, whatever, which one is the Fujifilm X-E1 and which one is the Olympus E-M5? No sharpening was applied at all. The files were demosaiced using dcraw then the color balance, contrast and levels were adjusted in RawTherapee.

As Olympus applies heavier OOC sharpening and saturation, it would be rather hard to make a judgement call. Based upon raw files I have worked on, the Fuji can pull more detail and has lower noIse.

They were raw files.

Reply   Reply with quote   Complain
dotborg
Veteran MemberPosts: 7,555
Like?
Comparison Mk II
In reply to dotborg, Sep 9, 2013

Some of you were comparing the images based on the colors. The comparison was to show that Fujifilm's X-Tranny sensor does not have any advantage for detail or sharpness over a Bayer CFA sensor, so I edited the files to make the colors much closer.

So, which one has more detail and/or is sharper?

Reply   Reply with quote   Complain
ulrikem
Regular MemberPosts: 211Gear list
Like?
Re: X-Trans is Simply More Added Value, In Our Favor!
In reply to umanemo, Sep 9, 2013
Cudo's to Ulri for His/Her diligence in the understanding of Image Sensor Technology and sharing that with us, Cudo's to FUJI for their pioneering spirit that benefits not only us but the advancement of imaging technology exponentially!

Kudos should all go to forum fujix-forum.com user MikeS. He's the man who understands/explains all the details. (I was barely able to follow...).

Everyone touts (no offense to Zeiss) the evolutionary development of the dated algorithms long since summarized for the 2 x 2 Bayer design sensor array mostly due to the longevity of the array pattern's presence in the world of imaging. Yet the 6 x 6 PR-CFA clearly shows a mathematical advantage. Sluggish program development of the RAW conversion algorithms that maximize that advantage is inevitable just as it was with the Bayer. The 6 x 6 will progress evolutionally and in the end we will all benefit from the X-Trans sensor's extended capability.

Progress has been made and more progress will be made. But note that decoding 6x6 seems to be much more difficult and hence take much longer. Also, parametrization of the decoding process is more complex. Decoding raw-files already takes several seconds on my not-too-old machine. I don't fancy waiting longer than this just to get a few more detail pixels.

 ulrikem's gear list:ulrikem's gear list
Fujifilm X100S Fujifilm X-Pro1 Fujifilm XF 18mm F2 R Fujifilm XF 35mm F1.4 R Fujifilm XF 14mm F2.8 R +1 more
Reply   Reply with quote   Complain
DMillier
Forum ProPosts: 17,640
Like?
Re: X-Trans is Simply More Added Value, In Our Favor!
In reply to ulrikem, Sep 9, 2013

I'm not sure why people expect more detail because the OPLF filter has been omitted.  Technically, an AA filter is required to avoid aliasing.   Fuji and other manufacturers are playing with fire by omitting it.

Aliasing is always present if nothing interferes with the sharpness of the optical image presented to the sensor. Usually that is the AA OLPF.  I'm used to the effects of this omission because I own a Kodak 14n and a couple of Foveon Sigma DSLRs.  The Sigmas are immune to colour moire because of their layered sensor approach but they luminance alias like crazy.  The Kodak has a regular Bayer CFA and it luminance and colour aliases like crazy. It isn't pretty.

The X-trans approach claims to reduce the risk of colour aliasing/moire effects but it can't do anything about luminance aliasing (jaggies, false diagonal striped patterns and the like) unfortunately.

And the benefits of omitting the AA filter seem tenuous to me. It doesn't really improve detail unless some kind of ridiculously heavy filter is being used, really what it does is improve acutance.  Acutance (a.k.a. microcontrast) can be restored in post with the judicious application of deconvolution sharpening.

I suspect that the positive image quality aspects of xtrans have little to do with it being xtrans, it's just a good package.

Anyway, the postman has just this minute handed me a box with my new XE1 in it, so I'll find out whether theory meets practice!

Cheers

Dave

-- hide signature --
Reply   Reply with quote   Complain
unknown member
(unknown member)
Like?
Re: Comparison Mk II
In reply to dotborg, Sep 9, 2013

To me, the leaves on the bottle at the top of the photo are noticeably different in each photo. They are cleaner in the bottom photo. I could be wrong, but that's what I noticed.

Can you PM me the answer?

 cantanima's gear list:cantanima's gear list
Fujifilm FinePix 2600 Zoom Fujifilm X10
Reply   Reply with quote   Complain
DonSantos
Senior MemberPosts: 1,122Gear list
Like?
Re: Comparison Mk II
In reply to cantanima, Sep 9, 2013

cantanima wrote:

To me, the leaves on the bottle at the top of the photo are noticeably different in each photo. They are cleaner in the bottom photo. I could be wrong, but that's what I noticed.

Can you PM me the answer?

First set the bottom one is oly.  Second set the top one is oly

 DonSantos's gear list:DonSantos's gear list
Nikon D600 Sony Alpha 7 Voigtlander 40mm F1.4 Nokton Classic Nikon AF-S Nikkor 85mm f/1.8G Sigma 35mm F1.4 DG HSM +1 more
Reply   Reply with quote   Complain
Randy Benter
Senior MemberPosts: 1,997Gear list
Like?
Re: Comparison Mk II
In reply to DonSantos, Sep 9, 2013

DonSantos wrote:

cantanima wrote:

To me, the leaves on the bottle at the top of the photo are noticeably different in each photo. They are cleaner in the bottom photo. I could be wrong, but that's what I noticed.

Can you PM me the answer?

First set the bottom one is oly. Second set the top one is oly

No, the opposite is true. You can easily determine which is which by the location of the colored strings/laces.

 Randy Benter's gear list:Randy Benter's gear list
Leica X2 Fujifilm X100S Leica X Vario Leica M8 Fujifilm X-Pro1 +34 more
Reply   Reply with quote   Complain
DonSantos
Senior MemberPosts: 1,122Gear list
Like?
Re: Comparison Mk II
In reply to Randy Benter, Sep 9, 2013

It looks like you are correct.  Funny I preferred the fuji all along but at least half preferred the oly.

No, the opposite is true. You can easily determine which is which by the location of the colored strings/laces.

 DonSantos's gear list:DonSantos's gear list
Nikon D600 Sony Alpha 7 Voigtlander 40mm F1.4 Nokton Classic Nikon AF-S Nikkor 85mm f/1.8G Sigma 35mm F1.4 DG HSM +1 more
Reply   Reply with quote   Complain
Charles2
Senior MemberPosts: 2,125
Like?
Re: Frequency selection demosaicking: A review and a look ahead (research paper)
In reply to ulrikem, Sep 9, 2013

ulrikem wrote:

You may want to read the following research paper, possibly that's what Fuji drove to use the 6x6 CFA:

http://david.alleysson.free.fr/Publications/VCIP08.pdf

Link was taken from the very interesting thread at another forum. Well worth reading!!

The essence of the asserted X-Trans improvement is this:

Reply   Reply with quote   Complain
VisualFX
Contributing MemberPosts: 631
Like?
Re: X-Trans is Simply More Added Value, In Our Favor!
In reply to DMillier, Sep 10, 2013

DMillier wrote:

I'm not sure why people expect more detail because the OPLF filter has been omitted. Technically, an AA filter is required to avoid aliasing. Fuji and other manufacturers are playing with fire by omitting it.

Aliasing is always present if nothing interferes with the sharpness of the optical image presented to the sensor. Usually that is the AA OLPF. I'm used to the effects of this omission because I own a Kodak 14n and a couple of Foveon Sigma DSLRs. The Sigmas are immune to colour moire because of their layered sensor approach but they luminance alias like crazy. The Kodak has a regular Bayer CFA and it luminance and colour aliases like crazy. It isn't pretty.

The X-trans approach claims to reduce the risk of colour aliasing/moire effects but it can't do anything about luminance aliasing (jaggies, false diagonal striped patterns and the like) unfortunately.

And the benefits of omitting the AA filter seem tenuous to me. It doesn't really improve detail unless some kind of ridiculously heavy filter is being used, really what it does is improve acutance. Acutance (a.k.a. microcontrast) can be restored in post with the judicious application of deconvolution sharpening.

I suspect that the positive image quality aspects of xtrans have little to do with it being xtrans, it's just a good package.

Anyway, the postman has just this minute handed me a box with my new XE1 in it, so I'll find out whether theory meets practice!

Cheers

Dave

-- hide signature --

Please show evidence of X-Trans luminance jaggies and false diagonal striped patterns, etc. I have yet to see one X-Trans photo that exhibits any issues like what you say.

Reply   Reply with quote   Complain
dotborg
Veteran MemberPosts: 7,555
Like?
Re: Comparison Mk II
In reply to DonSantos, Sep 10, 2013

DonSantos wrote:

cantanima wrote:

To me, the leaves on the bottle at the top of the photo are noticeably different in each photo. They are cleaner in the bottom photo. I could be wrong, but that's what I noticed.

Can you PM me the answer?

First set the bottom one is oly. Second set the top one is oly

Exactly the opposite.

Reply   Reply with quote   Complain
dotborg
Veteran MemberPosts: 7,555
Like?
Re: Comparison Mk II
In reply to Randy Benter, Sep 10, 2013

Randy Benter wrote:

DonSantos wrote:

cantanima wrote:

To me, the leaves on the bottle at the top of the photo are noticeably different in each photo. They are cleaner in the bottom photo. I could be wrong, but that's what I noticed.

Can you PM me the answer?

First set the bottom one is oly. Second set the top one is oly

No, the opposite is true. You can easily determine which is which by the location of the colored strings/laces.

Another clue is the deep shadows. The Fujifilm sensor has about a stop better noise, which makes sense since it has about twice the surface area.

Reply   Reply with quote   Complain
Keyboard shortcuts:
FForum MMy threads