PIX 2015
zirco77

zirco77

Lives in Canada Canada
Joined on Oct 14, 2012

Comments

Total: 3, showing: 1 – 3
On Nikon 1 J5 and Samsung NX500 added to studio test scene article (131 comments in total)
In reply to:

Menneisyys: I've imported both ISO6400 NX1 & NX500 low-light RAW's to ACR9.

The NX1 RAW is around 10% brighter (on the white background of the portraits on the left, 12%; in the center area, some 4%. Both measured in the red channel.). Assuming exactly the same lighting, this may mean the NX1 is more sensitive to light. (Actually, in DPR's previous reviews, they did state the NX1 is more sensitive to light than most other cameras.)

Nevertheless, 10% difference in in-RAW brightness wouldn't result in such a huge (1EV in higher ISO) difference. There is something fundamentally wrong with the NX500 - again, it produces even noisier results than the A5100 equipped with Sony's over a year-old 24 Mpixel sensor.

As for 3. above, it looks like color matrices, tone curves, etc. are all the same. However, tags related to "RGGB white balance" have different values. (for the technically inclined, see tags 0xa021 to 0xa024 in http://www.exiv2.org/tags-samsung.html) I can't tell if those are from the taken image or from the sensor calibration. I've seen those values used by software to read SRW files.

Strangely enough though, all sets of 4 values have the same middle values between NX1 and NX500. That is, assuming RGGB order (like the sensor bayer pattern), green values are the same, but others (red and blue) are higher on the NX500, whatever they mean.

This is somewhat in line with what I observed in ACR9: after adjusting RAWs for white balance (average on center target and grey patches) and illumination +0,35EV for NX500), red/blue patches are more saturated on NX500 (and histogram shows wider values), and red/blue noise is more visible too.

Knowing where WB RGGB values are from could help.

Direct link | Posted on Jun 10, 2015 at 03:36 UTC
On Nikon 1 J5 and Samsung NX500 added to studio test scene article (131 comments in total)
In reply to:

Menneisyys: I've imported both ISO6400 NX1 & NX500 low-light RAW's to ACR9.

The NX1 RAW is around 10% brighter (on the white background of the portraits on the left, 12%; in the center area, some 4%. Both measured in the red channel.). Assuming exactly the same lighting, this may mean the NX1 is more sensitive to light. (Actually, in DPR's previous reviews, they did state the NX1 is more sensitive to light than most other cameras.)

Nevertheless, 10% difference in in-RAW brightness wouldn't result in such a huge (1EV in higher ISO) difference. There is something fundamentally wrong with the NX500 - again, it produces even noisier results than the A5100 equipped with Sony's over a year-old 24 Mpixel sensor.

There are quite a few places to look for. I did a few tests with 100 and 3200 ISO RAWs from NX1 and NX500 (the "day light scene) in ACR9.

1. Consistent lightning. Illumination is down by about 0.35EV on the NX500, but white balance is off too. But I'd assume that they use a very reliable light setup.

2. Factory sensor binning. Given its the same part number (S5KVB2) I would think Samsung doesn't bin them at that level.

3. Sensor (and any attached electronic common to both NX1 and NX500) calibration. Every such "system" surely have to go through some calibration, for QA and to get an individual profile: "black frame" noise, ISO amplification, sensor to other color-space matrix, etc. Perhaps the NX1 gets away with the best samples. That matrix is probably embedded withing SRW files, I'll try to have a look...

4. Various electro-magnetic interference. Might be higher in a smaller body. Likewise, were they using live-view or the viewfinder for the NX1 scene shots?

Direct link | Posted on Jun 10, 2015 at 02:31 UTC
On Just Posted: Sony Cyber-shot DSC-RX1 sample images article (288 comments in total)

Anybody noticed how hard it tries to remove blue fringes, probably in internal jpeg conversion? Have a look at the "Blue Ribbon" neon sign in the gallery. Other signs have "proper" (assuming they're expected) color glow, but the blue sign has a darkened and desaturated fringe. Makes me wonder if the lenses have blue chromatic aberration issues so severe that they felt the need to aggressively remove it afterwards.

Direct link | Posted on Oct 14, 2012 at 02:56 UTC as 5th comment
Total: 3, showing: 1 – 3