Jman13
Offline
• • • • • •
Upload & Sell: On
|
 p.4 #3 · Samsung NX1 Rolling Impressions | |
Ok. Well, When I'm wrong I"m wrong.  I did a setup with two cards I made up, with 10 stops of light (actual perceptual stops based on a study I read of RGB values vs. rendered exposure on digital sensors). I did discrete blocks, each with the text of each camera written inside a block of gray. In each case, the text is exactly one stop darker (for the bright patches) and one stop lighter (for the dark patches)...when the text isn't readable, that's the limit, see?) Each card is printed with my Pixma Pro 100 on baryta paper (deep blacks and good contrast).
I put one card right in front of an incandescent lamp, and the other divided by a black divider on black felt in deep shadow. I then took 1/3 stop increments at base ISO (200 for the Fuji, 100 for the a6000 and NX1) at 35mm and f/8 (Fuji 35/1.4, Sony 35/1.8 and the 16-50 on the NX).
I then looked at each exposure and pulled highlights as much as I could until I got to the exposure where the text on the 255 patch was blown out and blended into the 255 patch and no more could be recovered. . I then went to the previous darker exposure as the exposure to test, which maxes out the highlights. (Note that the cameras clipped the red channel first, and at different times, but I used a neutral white balance and only cared about luminance DR here...it's just too complex evaluating each color channel, though the X-T1 clipped the red first)
Regardless of ISO 'fudging', each sensor is within 1/3 stop of each other on the EV for the exposure that just barely doesn't clip. The Fuji was 1/4 sec at ISO 200, the Sony 1/2 sec at ISO 100, the Samsung 0.4s at ISO 100. (Fuji and Sony Identical, the Samsung was 1/3 stop darker in EV before clipping).
One thing I am not wrong on: the RAW file has more 'headroom' in the a6000 files, with true 'clipping' occurring 1.5 stops away from when it first 'clips' at default exposure. The NX1 only gets about 2/3 stop before it blows out. However, as noted above, this is a metering in the RAW file thing and not a true DR issue, since the max exposure for both is within 1/3 stop of each other.
Here's the bright part of the card at 'max' exposure for each. For readability, I continued pulling down until the 255 patch was at 230 instead of bright white.
Fuji X-T1:
http://www.jordansteele.com/2015/xt1_dr_bright.jpg
Sony a6000:
http://www.jordansteele.com/2015/a6000_dr_bright.jpg
Samsung NX1:
http://www.jordansteele.com/2015/nx1_dr_bright.jpg
I then disregarded the bright half and went to the dark card. I pushed the file 5 full stops in lightroom to pull as much shadow detail as humanly possible out, to see how dark on the dark card I could read.
Interestingly enough, the darkest patch where the text was visible (though in all cases, barely readable) was the same - the third darkest patch (second from the left on the second row...the 0 patch is the first on the third row).
While the Samsung of course has a resolution advantage here, the X-T1 produces the cleanest pushed shadows at the pixel level, with the Samsung very close behind. The X-T1 also maintains neutral color.
Normalized for resolution, the Samsung takes a slim lead in noise. The a6000, however, is massively behind. The same level is visible, but there is an awful green cast, and it's super noisy and shows banding. (By the way, this is what I mean when I've said that the a6000 files don't handle punishment like the Fuji's do).
NOTE: Changing order to put the normalized Samsung near the X-T1:
Samsung NX1:
http://www.jordansteele.com/2015/nx1_dr_dark.jpg
Samsung NX1 (Normalized to 16MP):
http://www.jordansteele.com/2015/nx1_dr_dark_normal.jpg
Fuji X-T1:
http://www.jordansteele.com/2015/xt1_dr_dark.jpg
Sony a6000:
http://www.jordansteele.com/2015/a6000_dr_dark.jpg
|
Tho Fuji lens wouldnt handle 28 mpix APS-C.. tad bit too much resolution for them. That "kit" lens for NX-1 is very good, even tho it might not seem as such..