Fuji X-A1 or X-M1 Part 1: To X-Trans or not X-Trans

Fuji X-A1 or X-M1 Part 1: To X-Trans or not X-Trans

To Bayer or not to Bayer, that is the question one must ask oneself when deciding between the X-A1 or X-M1
Does one shed the glass cover for clarity, or save a Franklin and obscure thine eyes from the world?

The main difference between the X-A1 and X-M1 is the sensor sub-system. The X-M1 is X-Trans and costs roughly $100 to $200 more than the X-A1 which is a traditional Bayer sensor. The X-M1 also has no optical low pass filter, and the X-A1 filter has a weak OLPF. The only other minor difference I have noticed is grip material. The X-A1 is a rubberized material and the X-M1 faux leather.

The Images were processed from RAW in Adobe Lightroom RC5.3, and used the following settings:

X-M1 ISO 200 X-A1 ISO 200
Exposure 0 + 0.10
Sharpening 35 33
Radius 1 1
Detail 29 30
Masking 30 33
Noise Reduction 12 17
X-M1 ISO 3200 X-A1 ISO 3200
Exposure – 0.35 0
Sharpening 35 33
Radius 1 1
Detail 29 30
Masking 30 33
Noise Reduction 30 30

I picked the above settings based on what was pleasing to me. Without any NR the X-A1 might have had slightly more detail in the fabric at base ISO, but also more image noise, even at base ISO I think it needs some NR. I also applied sharpening until I noticed offensive artifacts, then dialed it back slightly.

I think overall I prefer the output from the X-Trans sensor for this comparison. The noise at ISO 3200 is a bit more blotchy and there is less noise. The colors from the Bayer sensor might be more accurate, but I prefer the X-Trans output here. Some more work with white balance and camera profiles might be able to get them more similar.

Next up I will run dynamic range tests when I have some time. Is it worth $200 more for X-M1? Maybe, maybe not. X-A1 is a steal for $500 right now.
Fujifilm X-M1 Compact System 16MP Digital Camera Kit with 16-50mm Lens and 3-Inch LCD Screen (Black)
Fujifilm X-A1 Kit with 16-50mm Lens (Black)





Here are the sample shots for you to compare:

DSCF0015 DSCF2056
X-M1 ISO 200 f/7.1 X-A1 ISO 200 f/7.1
DSCF0015-2 DSCF2056-2
X-M1 ISO 200 f/7.1 (100%) X-A1 ISO 200 f/7.1 (100%)
DSCF0015-3 DSCF2056-3
X-M1 ISO 200 f/7.1 (100%) X-A1 ISO 200 f/7.1 (100%)
DSCF0016 DSCF2057
X-M1 ISO 3200 f/7.1 X-A1 ISO 3200 f/7.1
DSCF0016-2 DSCF2057-2
X-M1 ISO 3200 f/7.1 (100%) X-A1 ISO 3200 f/7.1 (100%)
DSCF0016-3 DSCF2057-3
X-M1 ISO 3200 f/7.1 (100%) X-A1 ISO 3200 f/7.1 (100%)

6 Comments

  1. Does seem to be a bit more low-ISO detail in the X-trans sensor sample. And a bit more high-ISO detail (but more objectionable noise) in the Bayer sensor camera. In fact the more noise in the Bayer sensor camera is enough to make the X-trans high ISO sample “overall on the average typically better” than the Bayer sensor Fuji. Just as you observe in your article.

    I.e. at high ISO would rather have less X-trans sensor noise to struggle with in post-processing, than getting an extra 2% or whatever more finest Bayer sensor detail.

    Suppose there is some tiny bit of difference in some of the color rendering, but who would ever notice this. And scene and post-processing variations and luck would be much more significant in making 2 different photos “look different” than would this example’s sensor color rendering differences.

  2. admin says:

    Yup, I think I have changed my opinion about X-Trans to some extent. Will shoot a landscape comparison with both now to be sure. Adobe has improved a lot in processing of X-Trans as well (when I had my XP1 it was not good).

    Eric

  3. For the sheer thrill of not having done enough work yet, wonder if another way to compare the sensors is with “equal final noise.” In other words, how much less sharp would the non-X-trans sensor image look than the X-trans image, if the non-X-trans image suffered enough noise reduction to appear “as smooth as” the X-trans image. Did a few seconds of playing around with the non-X-trans image clip you posted,

    https://dl.dropboxusercontent.com/u/15387676/ERphotoReviewNonXtransFujiSample-extraLuminanceNoiseReduction.jpg

    simply som luminance noise reduction, then further re-sharpening. Wonder if anyone would notice the sharpness difference between final images from both cameras, where both cam images had somewhat difference noise reduction treatment…to get to the final display level of apparent noise.

  4. giliath says:

    Hi, Eric.
    Where can I download the original photos? Or maybe you can compare the IQ of the border? I heard that X-trans has something to improve the IQ of the border. (The short flange of X mount could cause some border IQ trouble). I wonder if it is true.

  5. admin says:

    These photos don’t really have any border detail (and this lens has a fair amount of corrections applied in SW). Let me see if I can’t take an outdoor shot, but might have to wait until after Christmas.

  6. giliath says:

    Thank you! Even though I’m afraid I give up the plan of changing to fuji system. All the lens I’m interested in (10-24 & 23mm) are too expensive. I don’t think it is a good idea to buy so many expensive apsc lens now, as ff mirrorless appears, and its price is not bad.