K-3 compares well to D800

bossa · 01-09-2014, 06:13 PM

Here's an OOF dud shot that shows pink/purple glow in the darks. This is Amp Glow on my D800E. 1/25th sec F/1.4 ISO 6400.

The effect seen here is mitigated by the white background colour of the server page as it's much worse on my monitor when looking at the original file in LR.

In order to see how bad this glow is I suggest changing the forum background colour to grey, black or just look through a frame made up of both your hands to mask out the white of the background on the page.

Attached Images

View Picture EXIF

Quicksand · 01-09-2014, 07:37 PM

Wow. Reminiscent of my old K20d.

dosdan · 01-09-2014, 09:20 PM - **1 Like**

Originally posted by bossa

Here's an OOF dud shot that shows pink/purple glow in the darks. This is Amp Glow on my D800E.

That looks like thermally-increased leakage current in the sensor. "Amp glow", at least in the old days, was asymmetric due to the close proximity of hot elements, e.g. readout amps, ADC(s) or the image processor, to the sensor, causing non-linear hotspot gradients on parts of the sensor. See Fig. 6 here: Noise, Dynamic Range and Bit Depth in Digital SLRs

Whereas the "amp glow" in your night-time shot is more generalised. Perhaps, better to refer to this as "dark-current noise" : http://www.truesenseimaging.com/component/docman/doc_download/75-ccd-image-sensor-noise-sources
.
Whatever, doesn't your raw processor offer a dark colour deflection (Green-Magenta) adjustment to handle this?

Here's the section from the SilkyPix manual that explains what causes the coloured dark regions and how to fix them:

Why do color casts occur in the dark portion?

This is caused by current leakage of an image sensor. Because of the leakage current, the intense black is not recorded as zero value in RAW data. SILKYPIX® Developer Studio Pro 5 is subtracting such leakage current data from RAW data (optical black correction) to develop an image.

However, a higher temperature increases leakage current of an image sensor, and vice versa. In most cases, a camera records black with constant level, regardless of the temperature, but some conditions or photographing environment may affect the level (optical black level). According to the color sensitivity of each image sensor, as the optical black level becomes larger, the color of the dark area becomes magentish, and as the level becomes smaller, the color becomes greenish. The "dark adjustment" function reduces coloring of the dark portion in those cases. It also corrects the white balance of the dark portion when a photograph is taken under severe conditions such as extremely low temperature, high-sensitive photography, and long exposure.

You can make corrections without affecting brighter areas because dark adjustment is performed after exposure bias in SILKYPIX® Developer Studio Pro 5.

Dan.

bossa · 01-10-2014, 01:38 AM

Thanks Dan.

Both my D800E's do exactly the same thing btw.

The noise is asymmetric Dan as it populates the edges of the sensor.

Last edited by bossa; 01-10-2014 at 02:05 AM.

dosdan · 01-10-2014, 03:48 AM

Originally posted by bossa

Thanks Dan.

Both my D800E's do exactly the same thing btw.

The noise is asymmetric Dan as it populates the edges of the sensor.

The sensor in the D800/D800E is a Sony Exmor sensor, (Confirmed: the sensor inside the Nikon D800 is made by Sony | Nikon Rumors) even if Nikon added their own toppings (microlenses & CFA, and in the case of the D800, an AA filter). Like other Exmors, it will have a separate ADC, Analogue CDS & Digital CDS processing stage for every column in the sensor (approx. 4912 columns). See: http://pro.sony.com/bbsccms/assets/files/cat/camsec/solutions/E_CMOS_Sensor_WP_110427.pdf

These elements are arranged along the edges: http://default.media.ipcdigital.co.uk/11133%257C00000604b%257C5bc7_D800-Cmos-1.jpg).

Even though these are low-power elements, it seems they may be generating enough heat to caused localised temperature gradients.

Dan.

KDAFA · 01-10-2014, 03:48 AM

Originally posted by bossa

... This thread has made me appreciate that the DA*60-250 and DA*300 on a K-3 can get close to, or outperform, something costing 8+ grand from Nikon.

Wow, taking that price into consideration makes the comparison all the more remarkable!

jani80 · 01-11-2014, 03:25 AM

Originally posted by bossa

I had a great copy of the 60-250... what a shame the K-3 wasn't around when I sold everything.

I as well sold all my pentax gear and bought Nikon D800 with 5 new lenses and I must admit that I have not been extremely happy. Where was the K-3 then...

normhead · 01-11-2014, 07:41 AM

Originally posted by dtmateojr

I'm just stating a fact that at f8, the D800 only loses 1/5 of its resolution but the K3 has lost half of its capability. The fact that you are noticing not much difference is testament to the fact that 12Mp is more than good enough for almost anything but the observation is hardly objective.

Shoot both of them at f5.6 with a good lens if you want a real comparison.

FYI I shoot at f22 if needed. Sharpness is overrated but a comparison needs to be factual.

AM I missing something here...

Less than 5%, drop in resolution. I need to see something factual here, to support your statement.

**MJSfoto1956** · 01-11-2014, 08:59 AM

Originally posted by normhead

AM I missing something here... Less than 5%, drop in resolution. I need to see something factual here, to support your statement.

I agree with your assessment: his statement is nonsense especially given that I was attempting to test the "systems" not the lenses. More importantly, shooting at f/8 has no effect at all on corner sharpness. Diffraction really only starts to affect the ENTIRE image around f/11 and even then it is still quite good and worthy of comparison since it still makes sense to judiciously use f/11 when warranted. Or in other words: shooting at f/8 or f/11 (or any aperture for that matter) is indeed a "real world test".

Stagnant · 01-11-2014, 09:38 AM

Why did you not downsize the D-800 files to 24 MP, instead of performing upscaling of K-3 files ? Strangely enough in the last set of photos the corner sharpness of Pentax file is higher than centre sharpness. In the corners, around the tree Pentax is considerably sharper than Nikon. I wonder what caused such a result.

[Edit] I am referring to these shots :

**MJSfoto1956** · 01-11-2014, 10:20 AM

Originally posted by Stagnant

Why did you not downsize the D-800 files to 24 MP, instead of performing upscaling of K-3 files ? Strangely enough in the last set of photos the corner sharpness of Pentax file is higher than centre sharpness. In the corners, around the tree Pentax is considerably sharper than Nikon. I wonder what caused such a result.]

the quick answer: I wasn't particularly interested in pixel-peeping. I am primarily interested in how these two systems print at 20"x30" -- of which the images I posted are basically "ready to print". I too sense that this lens seems to appear "sharper" in the corners, somewhat strange behavior that is counter to how other lenses typically behave. But that is perhaps due to the fact that it is near the diffraction limit which affects the center more than the edges? Or perhaps we have developed a bias towards thinking the center should be sharper? I don't know. Still, when printed both images look great. That's all I really care about here and it showed me that the K-3 + 60-250mm combination is a serious photographic tool.

Cheers.
Michael

P.S. one minor detail: BOTH images were upsampled to 20"x30" directly in DxO. Trust me, the D800, good as it is, can't produce an image this size without upsampling!!!

J MICHAEL SULLIVAN Photography
Join my October 2014 Photo Workshop in Tuscany! 10% Discount until March 1st!

normhead · 01-11-2014, 10:49 AM

Originally posted by MJSfoto1956

the quick answer: I wasn't particularly interested in pixel-peeping. I am primarily interested in how these two systems print at 20"x30" -- of which the images I posted are basically "ready to print". I too sense that this lens seems to appear "sharper" in the corners, somewhat strange behavior that is counter to how other lenses typically behave. But that is perhaps due to the fact that it is near the diffraction limit which affects the center more than the edges? Or perhaps we have developed a bias towards thinking the center should be sharper? I don't know. Still, when printed both images look great. That's all I really care about here and it showed me that the K-3 + 60-250mm combination is a serious photographic tool.

Cheers.
Michael

P.S. one minor detail: BOTH images were upsampled to 20"x30" directly in DxO. Trust me, the D800, good as it is, can't produce an image this size without upsampling!!!

J MICHAEL SULLIVAN Photography
Join my October 2014 Photo Workshop in Tuscany! 10% Discount until March 1st!

^ +1 Now here's a guy I understand.

01-09-2014, 09:20 PM - 1 Like	#93
dosdan Pentaxian Join Date: Dec 2007 Location: Brisbane, Australia Photos: Gallery \| Albums Posts: 1,741	Originally posted by bossa Here's an OOF dud shot that shows pink/purple glow in the darks. This is Amp Glow on my D800E. That looks like thermally-increased leakage current in the sensor. "Amp glow", at least in the old days, was asymmetric due to the close proximity of hot elements, e.g. readout amps, ADC(s) or the image processor, to the sensor, causing non-linear hotspot gradients on parts of the sensor. See Fig. 6 here: Noise, Dynamic Range and Bit Depth in Digital SLRs Whereas the "amp glow" in your night-time shot is more generalised. Perhaps, better to refer to this as "dark-current noise" : http://www.truesenseimaging.com/component/docman/doc_download/75-ccd-image-sensor-noise-sources . Whatever, doesn't your raw processor offer a dark colour deflection (Green-Magenta) adjustment to handle this? Here's the section from the SilkyPix manual that explains what causes the coloured dark regions and how to fix them: *Why do color casts occur in the dark portion?* This is caused by current leakage of an image sensor. Because of the leakage current, the intense black is not recorded as zero value in RAW data. SILKYPIX® Developer Studio Pro 5 is subtracting such leakage current data from RAW data (optical black correction) to develop an image. However, a higher temperature increases leakage current of an image sensor, and vice versa. In most cases, a camera records black with constant level, regardless of the temperature, but some conditions or photographing environment may affect the level (optical black level). According to the color sensitivity of each image sensor, as the optical black level becomes larger, the color of the dark area becomes magentish, and as the level becomes smaller, the color becomes greenish. The "dark adjustment" function reduces coloring of the dark portion in those cases. It also corrects the white balance of the dark portion when a photograph is taken under severe conditions such as extremely low temperature, high-sensitive photography, and long exposure. You can make corrections without affecting brighter areas because dark adjustment is performed after exposure bias in SILKYPIX® Developer Studio Pro 5. Dan. Last edited by dosdan; 01-09-2014 at 09:58 PM.

01-11-2014, 09:38 AM	#100
Stagnant Veteran Member Join Date: Dec 2013 Photos: Gallery Posts: 603	Why did you not downsize the D-800 files to 24 MP, instead of performing upscaling of K-3 files ? Strangely enough in the last set of photos the corner sharpness of Pentax file is higher than centre sharpness. In the corners, around the tree Pentax is considerably sharper than Nikon. I wonder what caused such a result. [Edit] I am referring to these shots : http://www.haywood-sullivan.com/MAGNAchrom/articleimages/_D8A7291_9000_DxO.layers.jpg http://www.haywood-sullivan.com/MAGNAchrom/articleimages/IMGP1360_9000_DxO.layers.jpg

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
K mount lenses to D800 - good or bad idea?	Fontan	Photographic Industry and Professionals	8	06-04-2012 07:40 PM
How does Nikon 85/1.8 compares to FA77?	yusuf	Canon, Nikon, Sony, and Other Camera Brands	42	06-02-2012 05:05 PM
5D mark 3 beating Nikon D800...hands down - Video	D4rknezz	Canon, Nikon, Sony, and Other Camera Brands	17	03-31-2012 11:09 AM
5D Mk 3 Vs D800	D4rknezz	Canon, Nikon, Sony, and Other Camera Brands	5	03-02-2012 10:42 PM
Hank Williams Jr. Compares Obama to Hitler.- Fox & Friends - 10/3/11	jogiba	General Talk	7	10-04-2011 08:49 AM

01-09-2014, 07:37 PM	#92
Quicksand Pentaxian Join Date: Apr 2008 Location: Central Ohio (formerly SF Bay Area) Posts: 1,519	Wow. Reminiscent of my old K20d.

01-10-2014, 01:38 AM	#94
bossa Veteran Member Join Date: Jul 2011 Location: Adelaide Posts: 4,546	Thanks Dan. Both my D800E's do exactly the same thing btw. The noise is asymmetric Dan as it populates the edges of the sensor. Last edited by bossa; 01-10-2014 at 02:05 AM.

01-10-2014, 03:48 AM	#95
dosdan Pentaxian Join Date: Dec 2007 Location: Brisbane, Australia Photos: Gallery \| Albums Posts: 1,741	Originally posted by bossa Thanks Dan. Both my D800E's do exactly the same thing btw. The noise is asymmetric Dan as it populates the edges of the sensor. The sensor in the D800/D800E is a Sony Exmor sensor, (Confirmed: the sensor inside the Nikon D800 is made by Sony \| Nikon Rumors) even if Nikon added their own toppings (microlenses & CFA, and in the case of the D800, an AA filter). Like other Exmors, it will have a separate ADC, Analogue CDS & Digital CDS processing stage for every column in the sensor (approx. 4912 columns). See: http://pro.sony.com/bbsccms/assets/files/cat/camsec/solutions/E_CMOS_Sensor_WP_110427.pdf These elements are arranged along the edges: http://default.media.ipcdigital.co.uk/11133%257C00000604b%257C5bc7_D800-Cmos-1.jpg). Even though these are low-power elements, it seems they may be generating enough heat to caused localised temperature gradients. Dan.

01-10-2014, 03:48 AM	#96
KDAFA Veteran Member Join Date: Dec 2012 Posts: 528	Originally posted by bossa ... This thread has made me appreciate that the DA60-250 and DA300 on a K-3 can get close to, or outperform, something costing 8+ grand from Nikon. Wow, taking that price into consideration makes the comparison all the more remarkable!

01-11-2014, 03:25 AM	#97
jani80 Inactive Account Join Date: May 2011 Location: Helsinki Photos: Gallery Posts: 64	Originally posted by bossa I had a great copy of the 60-250... what a shame the K-3 wasn't around when I sold everything. I as well sold all my pentax gear and bought Nikon D800 with 5 new lenses and I must admit that I have not been extremely happy. Where was the K-3 then...

01-11-2014, 07:41 AM	#98
normhead Pentaxian Join Date: Jun 2007 Location: Near Algonquin Park Photos: Gallery \| Albums Posts: 40,451	Originally posted by dtmateojr I'm just stating a fact that at f8, the D800 only loses 1/5 of its resolution but the K3 has lost half of its capability. The fact that you are noticing not much difference is testament to the fact that 12Mp is more than good enough for almost anything but the observation is hardly objective. Shoot both of them at f5.6 with a good lens if you want a real comparison. FYI I shoot at f22 if needed. Sharpness is overrated but a comparison needs to be factual. AM I missing something here... Less than 5%, drop in resolution. I need to see something factual here, to support your statement.

01-11-2014, 08:59 AM	#99
MJSfoto1956 Site Supporter Join Date: Dec 2013 Location: Boston Posts: 1,305 Original Poster	Originally posted by normhead AM I missing something here... Less than 5%, drop in resolution. I need to see something factual here, to support your statement. I agree with your assessment: his statement is nonsense especially given that I was attempting to test the "systems" not the lenses. More importantly, shooting at f/8 has no effect at all on corner sharpness. Diffraction really only starts to affect the ENTIRE image around f/11 and even then it is still quite good and worthy of comparison since it still makes sense to judiciously use f/11 when warranted. Or in other words: shooting at f/8 or f/11 (or any aperture for that matter) is indeed a "real world test".

01-11-2014, 10:20 AM	#101
MJSfoto1956 Site Supporter Join Date: Dec 2013 Location: Boston Posts: 1,305 Original Poster	Originally posted by Stagnant Why did you not downsize the D-800 files to 24 MP, instead of performing upscaling of K-3 files ? Strangely enough in the last set of photos the corner sharpness of Pentax file is higher than centre sharpness. In the corners, around the tree Pentax is considerably sharper than Nikon. I wonder what caused such a result.] the quick answer: I wasn't particularly interested in pixel-peeping. I am primarily interested in how these two systems print at 20"x30" -- of which the images I posted are basically "ready to print". I too sense that this lens seems to appear "sharper" in the corners, somewhat strange behavior that is counter to how other lenses typically behave. But that is perhaps due to the fact that it is near the diffraction limit which affects the center more than the edges? Or perhaps we have developed a bias towards thinking the center should be sharper? I don't know. Still, when printed both images look great. That's all I really care about here and it showed me that the K-3 + 60-250mm combination is a serious photographic tool. Cheers. Michael P.S. one minor detail: BOTH images were upsampled to 20"x30" directly in DxO. Trust me, the D800, good as it is, can't produce an image this size without upsampling!!! J MICHAEL SULLIVAN Photography Join my October 2014 Photo Workshop in Tuscany! 10% Discount until March 1st!

01-11-2014, 10:49 AM	#102
normhead Pentaxian Join Date: Jun 2007 Location: Near Algonquin Park Photos: Gallery \| Albums Posts: 40,451	Originally posted by MJSfoto1956 the quick answer: I wasn't particularly interested in pixel-peeping. I am primarily interested in how these two systems print at 20"x30" -- of which the images I posted are basically "ready to print". I too sense that this lens seems to appear "sharper" in the corners, somewhat strange behavior that is counter to how other lenses typically behave. But that is perhaps due to the fact that it is near the diffraction limit which affects the center more than the edges? Or perhaps we have developed a bias towards thinking the center should be sharper? I don't know. Still, when printed both images look great. That's all I really care about here and it showed me that the K-3 + 60-250mm combination is a serious photographic tool. Cheers. Michael P.S. one minor detail: BOTH images were upsampled to 20"x30" directly in DxO. Trust me, the D800, good as it is, can't produce an image this size without upsampling!!! J MICHAEL SULLIVAN Photography Join my October 2014 Photo Workshop in Tuscany! 10% Discount until March 1st! ^ +1 Now here's a guy I understand.