Photography - Was Nikon's move to Toshiba a mistake?

On 22/02/2013 7:56 p.m., RichA wrote:
On Feb 21, 11:50 pm, Me wrote:
On 22/02/2013 5:09 p.m., RichA wrote:

Someone on Dpreview remarked that Toshiba lacks a processing step used
in the Sony sensors to remove banding. You can see it in this D7000
image in the sky, but the poster showed it very prominently in the
shadows of a 200 ISO shot from a D5200. This reminds me a bit of the
old Olympus CCD sensors in their original DSLR's. I don't know if the
D7000 has a Toshiba sensor, or a Sony one. It's hard to say when
Nikon began using them.

D7000

http://www.dpreview.com/galleries/re...735834/dsc_210...

D5200

http://forums.dpreview.com/files/t/0...871a3627054ed9

The D7000 uses a Sony fabricated sensor, IMX071.

The D7000 image is curious as the ISO setting doesn't seem to be in the
exif data. I don't see banding so much as blotchy but random chroma
noise, as if it was taken at high ISO.

The "D5200" sample is worthless, without exif data and information about
post-processing. There's certainly horizontal banding in lifted shadow
areas, but there's no proof this is even from a D5200.
Some base ISO D5200 raw samples might reveal if there's an issue or not.

Despite the obvious banding, the image is actually very good. I see
the EXIF data in Firefox's EXIF add-on.

snip
OK - I didn't look at full exif - didn't see that it was there.
Imaging resource has D5200 raw files available for d/l. They're only
their standard studio shots, and I can only open them in Nikon ViewNX at
the moment.
Anyway, I can't see any hint of banding or pattern noise in ISO 100 deep
shadows pushed two stops, only the same typical random luma noise I've
seen on every Nikon DSLR since the D3/300. IMO, pushing base ISO D5200
raws would be a doddle for prints any size - you can see the grain on
screen, but you almost certainly wouldn't in print.
So - I'm buggared if I know what the story could be with the shot in the
link above. I suspect someone is trolling on DPReview - it wouldn't be
the first time.

On DPR's site, the converted raw D5200 samples look the same as the D600
samples, at higher ISO allowing one stop for FX (ie the D5200 ISO 800
sample looks just the same as the D600 ISO 1600 sample - as you should
expect). Not just noise, but resolution, colour, everything. It's
enough to make you wonder about Fx format - given the huge cost and
weight penalty - unless you really need that "extra stop" (something
which has always killed the wallet). I suspect most of us don't, and of
those of us who think we do, most of us are probably wrong.
But you can keep the D5200 with pentaprism, no AF motor etc. What you
get for the extra $300 for the D7100 is a no-brainer - unless you
absolutely must have an articulating screen for macro or whatever.