Nikon D4s: Bayer processing breakthru?

Started 2 months ago | Discussions
Tom Schum
Senior MemberPosts: 3,376Gear list
Like?
Nikon D4s: Bayer processing breakthru?
2 months ago

This month's print edition of Shutterbug magazine (July 2014) has a review of the D4s, and on page 40 they say, "The resolution of 3280 lines per picture height given by the image sensor is transferred into images with a maximum of 3258 lines per picture height.... details and fine structures in these images look very crisp, but still natural." I've left out a couple sentences in which they speculate about how this might be done, and complain ever so slightly about some artifacts at hard contrast edges, but if you are interested the magazine is worth buying.

Typically, from what I have seen so far, it is very good to get even 75% of the sensor resolution out of a bayer array (60% more typically), and here we are at 99% with the D4s!

Once the DPR review comes out, it will be very interesting to compare image quality between the D4s (16 mp) and any Merrill (15mp). I'm guessing the D4s images will compare quite favorably.

-- hide signature --

Tom Schum

 Tom Schum's gear list:Tom Schum's gear list
Sigma DP2 Sigma DP3 Merrill Fujifilm X-E1 Sigma 50mm F2.8 EX DG Macro Samyang 85mm F1.4 Aspherical IF +5 more
Nikon D4s Sigma DP3 Merrill
If you believe there are incorrect tags, please send us this post using our feedback form.
unknown member
(unknown member)
Like?
So what?
In reply to Tom Schum, 2 months ago

Tom Schum wrote:

This month's print edition of Shutterbug magazine (July 2014) has a review of the D4s, and on page 40 they say, "The resolution of 3280 lines per picture height given by the image sensor is transferred into images with a maximum of 3258 lines per picture height.... details and fine structures in these images look very crisp, but still natural." I've left out a couple sentences in which they speculate about how this might be done, and complain ever so slightly about some artifacts at hard contrast edges, but if you are interested the magazine is worth buying.

Typically, from what I have seen so far, it is very good to get even 75% of the sensor resolution out of a bayer array (60% more typically), and here we are at 99% with the D4s!

Once the DPR review comes out, it will be very interesting to compare image quality between the D4s (16 mp) and any Merrill (15mp). I'm guessing the D4s images will compare quite favorably.

-- hide signature --

Tom Schum

Reply   Reply with quote   Complain
Tom Schum
Senior MemberPosts: 3,376Gear list
Like?
Re: So what?
In reply to mroy, 2 months ago

Isn't a big part of Sigma's marketing the equivalence theory, that the Foveon image is equivalent to a bayer image with twice as many pixels?

If so, Sigma might need to re-think.

-- hide signature --

Tom Schum

 Tom Schum's gear list:Tom Schum's gear list
Sigma DP2 Sigma DP3 Merrill Fujifilm X-E1 Sigma 50mm F2.8 EX DG Macro Samyang 85mm F1.4 Aspherical IF +5 more
Reply   Reply with quote   Complain
unknown member
(unknown member)
Like?
No, it isn't.
In reply to Tom Schum, 2 months ago

Tom Schum wrote:

Isn't a big part of Sigma's marketing the equivalence theory, that the Foveon image is equivalent to a bayer image with twice as many pixels?

If so, Sigma might need to re-think.

-- hide signature --

Tom Schum

It's about capturing colour and detail as best as possible - and being able to deliver outstanding B&W, to boot.

Marketing is irrelevant.

Reply   Reply with quote   Complain
Tom Schum
Senior MemberPosts: 3,376Gear list
Like?
Re: No, it isn't.
In reply to mroy, 2 months ago

mroy wrote:

It's about capturing colour and detail as best as possible - and being able to deliver outstanding B&W, to boot.

But every camera wants to do that.  It's a common goal and maybe Nikon has achieved something like parity with the Merrill now, in terms of image quality.   I can't imagine how one might process bayer sensor output to get this kind of resolution.

Sigma has a big price advantage at the moment.  At this point one could buy ten DP3Ms for the price of one D4s, but maybe not next year.

-- hide signature --

Tom Schum

 Tom Schum's gear list:Tom Schum's gear list
Sigma DP2 Sigma DP3 Merrill Fujifilm X-E1 Sigma 50mm F2.8 EX DG Macro Samyang 85mm F1.4 Aspherical IF +5 more
Reply   Reply with quote   Complain
unknown member
(unknown member)
Like?
They do indeed look
In reply to Tom Schum, 2 months ago

Tom Schum wrote:

mroy wrote:

It's about capturing colour and detail as best as possible - and being able to deliver outstanding B&W, to boot.

But every camera wants to do that. It's a common goal and maybe Nikon has achieved something like parity with the Merrill now, in terms of image quality. I can't imagine how one might process bayer sensor output to get this kind of resolution.

Sigma has a big price advantage at the moment. At this point one could buy ten DP3Ms for the price of one D4s, but maybe not next year.

-- hide signature --

Tom Schum

very similar:

http://camerasize.com/compare/#243,532

Reply   Reply with quote   Complain
ChristianHass
Senior MemberPosts: 1,873Gear list
Like?
Re: They do indeed look
In reply to mroy, 2 months ago

mroy wrote:

very similar:

http://camerasize.com/compare/#243,532

Clearly the Sigma is much wider 

But I agree with Tom that it's very interesting if the D4s sensor in some way has improved on what was considered the theoretical limit for Bayer sensors, and as such it will be interesting to see how it compares with the Merrills.

-- hide signature --
 ChristianHass's gear list:ChristianHass's gear list
Sigma DP1x Sigma DP1 Merrill Sigma DP3 Merrill Sony RX100 II FastStone Image Viewer +1 more
Reply   Reply with quote   Complain
unknown member
(unknown member)
Like?
Fine by me
In reply to ChristianHass, 2 months ago

ChristianHass wrote:

mroy wrote:

very similar:

http://camerasize.com/compare/#243,532

Clearly the Sigma is much wider

But I agree with Tom that it's very interesting if the D4s sensor in some way has improved on what was considered the theoretical limit for Bayer sensors, and as such it will be interesting to see how it compares with the Merrills.

-- hide signature --

if your photography subjects are primarily lots of naturally occuring pairs of lines, neatly ordered.

I wonder how both cameras fare with bricks - a far more popular subject, I believe?

Reply   Reply with quote   Complain
Tom Schum
Senior MemberPosts: 3,376Gear list
Like?
Re: Fine by me
In reply to mroy, 2 months ago

mroy wrote:

if your photography subjects are primarily lots of naturally occuring pairs of lines, neatly ordered.

I wonder how both cameras fare with bricks - a far more popular subject, I believe?

You've mentioned one of my favorite photographic subcultures!

Here is a wonderful thread:

http://www.dpreview.com/forums/post/53698118

Now that you mention it, personally I prefer lots of naturally occuring pairs of lines in my photos, but I think you're right about bricks being a more popular subject.

-- hide signature --

Tom Schum

 Tom Schum's gear list:Tom Schum's gear list
Sigma DP2 Sigma DP3 Merrill Fujifilm X-E1 Sigma 50mm F2.8 EX DG Macro Samyang 85mm F1.4 Aspherical IF +5 more
Reply   Reply with quote   Complain
Tom Schum
Senior MemberPosts: 3,376Gear list
Like?
Re: They do indeed look
In reply to ChristianHass, 2 months ago

ChristianHass wrote:

But I agree with Tom that it's very interesting if the D4s sensor in some way has improved on what was considered the theoretical limit for Bayer sensors, and as such it will be interesting to see how it compares with the Merrills.

Thank you for validating my own opinions!  I don't seem to be getting much of that lately and it feels good.

-- hide signature --

Tom Schum

 Tom Schum's gear list:Tom Schum's gear list
Sigma DP2 Sigma DP3 Merrill Fujifilm X-E1 Sigma 50mm F2.8 EX DG Macro Samyang 85mm F1.4 Aspherical IF +5 more
Reply   Reply with quote   Complain
Lin Evans
Forum ProPosts: 15,629Gear list
Like?
Maybe, but I don't think so.....
In reply to Tom Schum, 2 months ago

Hi Tom,

I think you might want to actually look at actual resolution charts... Tech Radar has done a pretty thorough examination of this camera and it doesn't look like it does all that much better than my D7000 and not close to my D5300 in terms of linear resolution... Probably best to actually look at the resolution charts and not trust the TechRadar's numbers because at times it appears what they say and what the charts actually reveal are at odds. As an example, they rated the D4S at 28 and the SD1 at 24 in terms of resolution. Looking at the actual resolution charts you will scratch your head on these numbers. Bottom line is that the D4S does very well for a Bayer 16 megapixel sensor, but as I look at the resolution charts - it doesn't get that near the Merrill's.

http://www.techradar.com/us/reviews/cameras-and-camcorders/cameras/digital-slrs-hybrids/nikon-d4s-1212962/review

Best regards,

Lin

Tom Schum wrote:

This month's print edition of Shutterbug magazine (July 2014) has a review of the D4s, and on page 40 they say, "The resolution of 3280 lines per picture height given by the image sensor is transferred into images with a maximum of 3258 lines per picture height.... details and fine structures in these images look very crisp, but still natural." I've left out a couple sentences in which they speculate about how this might be done, and complain ever so slightly about some artifacts at hard contrast edges, but if you are interested the magazine is worth buying.

Typically, from what I have seen so far, it is very good to get even 75% of the sensor resolution out of a bayer array (60% more typically), and here we are at 99% with the D4s!

Once the DPR review comes out, it will be very interesting to compare image quality between the D4s (16 mp) and any Merrill (15mp). I'm guessing the D4s images will compare quite favorably.

-- hide signature --

Tom Schum

Reply   Reply with quote   Complain
Lin Evans
Forum ProPosts: 15,629Gear list
Like?
Re: Nikon D4s: Bayer processing breakthru?
In reply to Tom Schum, 2 months ago

Hi Tom,

There are quite a few images from the D4s available around the web. They look decent, but not much different from my Nikons to me. I'm beginning to wonder if these resolution chart photos have much relevance to real life photography?  It make me wonder if manufacturers are optimizing their sensors to look good on b&w resolution charts and how that translates to everyday photography? This could be one of the issues which plague this forum when people get results which conflict with the "numbers" from testing.

I'm reminded of when Mike Chaney (Qimage Author) compared his 4.6 mp Sigma to the 12 MP Canon 5D and found the results very interesting. As a long-time user of Nikon dSLR's, I've always found them very decent, but not spectacular as far as IQ goes. Shooting them side-by-side with my Sigma's they just don't have that something special I see from the Foveon sensors. Neither do my Canon dSLR's nor my E3 Olympus but the Olympus produces stunning jpgs right out of the camera.

Best regards,

Lin

Tom Schum wrote:

This month's print edition of Shutterbug magazine (July 2014) has a review of the D4s, and on page 40 they say, "The resolution of 3280 lines per picture height given by the image sensor is transferred into images with a maximum of 3258 lines per picture height.... details and fine structures in these images look very crisp, but still natural." I've left out a couple sentences in which they speculate about how this might be done, and complain ever so slightly about some artifacts at hard contrast edges, but if you are interested the magazine is worth buying.

Typically, from what I have seen so far, it is very good to get even 75% of the sensor resolution out of a bayer array (60% more typically), and here we are at 99% with the D4s!

Once the DPR review comes out, it will be very interesting to compare image quality between the D4s (16 mp) and any Merrill (15mp). I'm guessing the D4s images will compare quite favorably.

-- hide signature --

Tom Schum

Reply   Reply with quote   Complain
PrebenR
Senior MemberPosts: 2,664
Like?
Re: Nikon D4s: Bayer processing breakthru?
In reply to Tom Schum, 2 months ago

Tom Schum wrote:

This month's print edition of Shutterbug magazine (July 2014) has a review of the D4s, and on page 40 they say, "The resolution of 3280 lines per picture height given by the image sensor is transferred into images with a maximum of 3258 lines per picture height.... details and fine structures in these images look very crisp, but still natural." I've left out a couple sentences in which they speculate about how this might be done, and complain ever so slightly about some artifacts at hard contrast edges, but if you are interested the magazine is worth buying.

Typically, from what I have seen so far, it is very good to get even 75% of the sensor resolution out of a bayer array (60% more typically), and here we are at 99% with the D4s!

Once the DPR review comes out, it will be very interesting to compare image quality between the D4s (16 mp) and any Merrill (15mp). I'm guessing the D4s images will compare quite favorably.

-- hide signature --

Tom Schum

Don't believe all that you read

-- hide signature --

Lightwriting with Sigma
«Bad reputation»

Reply   Reply with quote   Complain
mypic
Forum MemberPosts: 67Gear list
Like?
Re: Nikon D4s: Bayer processing breakthru?
In reply to PrebenR, 2 months ago

PrebenR wrote:

Tom Schum wrote:

This month's print edition of Shutterbug magazine (July 2014) has a review of the D4s, and on page 40 they say, "The resolution of 3280 lines per picture height given by the image sensor is transferred into images with a maximum of 3258 lines per picture height.... details and fine structures in these images look very crisp, but still natural." I've left out a couple sentences in which they speculate about how this might be done, and complain ever so slightly about some artifacts at hard contrast edges, but if you are interested the magazine is worth buying.

Typically, from what I have seen so far, it is very good to get even 75% of the sensor resolution out of a bayer array (60% more typically), and here we are at 99% with the D4s!

Once the DPR review comes out, it will be very interesting to compare image quality between the D4s (16 mp) and any Merrill (15mp). I'm guessing the D4s images will compare quite favorably.

-- hide signature --

Tom Schum

Don't believe all that you read

-- hide signature --

Lightwriting with Sigma
«Bad reputation»

That goes for both sides.

 mypic's gear list:mypic's gear list
Sigma DP2 Merrill Sony Alpha 7
Reply   Reply with quote   Complain
mypic
Forum MemberPosts: 67Gear list
Like?
Re: They do indeed look
In reply to Tom Schum, 2 months ago

Are you surprised?

 mypic's gear list:mypic's gear list
Sigma DP2 Merrill Sony Alpha 7
Reply   Reply with quote   Complain
DMillier
Forum ProPosts: 17,499
Like?
Re: Maybe, but I don't think so.....
In reply to Lin Evans, 2 months ago

Interpretation of resolution charts often seems to depend on opinions about aliasing.

If you take an ultra conservative view, you would say that the resolution limit occurs at the point there is the faintest trace of an alias.  If you are more moderate, you might overlook small amounts of aliasing on the grounds that the integrity of the result immediately recovers.  If you are extremely liberal/optimistic you might simply close your eyes to a blatantly obvious collapse into aliasing and say "look, ignore that and that and that because there is a tiny spot in between those aliases where it looks like the nine lines have become nine again" so resolution really is much higher!

A lot of review sites seem to more towards the conservative end of the spectrum.

-- hide signature --

"...while I am tempted to bludgeon you, I would rather have you come away with an improved understanding of how these sensors work" ---- Eric Fossum
Galleries and website: http://www.whisperingcat.co.uk/
Flickr: http://www.flickr.com/photos/davidmillier/

Reply   Reply with quote   Complain
DMillier
Forum ProPosts: 17,499
Like?
Re: Nikon D4s: Bayer processing breakthru?
In reply to Lin Evans, 2 months ago

Optimising the sensor doesn't sound very likely, how would it be done? Optimising processing and raw conversion, maybe.

-- hide signature --

"...while I am tempted to bludgeon you, I would rather have you come away with an improved understanding of how these sensors work" ---- Eric Fossum
Galleries and website: http://www.whisperingcat.co.uk/
Flickr: http://www.flickr.com/photos/davidmillier/

Reply   Reply with quote   Complain
Roger
Senior MemberPosts: 1,878Gear list
Like?
It's in what you want or need Tom
In reply to Tom Schum, 2 months ago

Hi Tom

You know there is no doubt the D4s will be a great camera.  It is up to you or who ever to use what they want to use, what they enjoy.  I hear pro and cons for all cameras.  Use what works for you or what you enjoy using.  I wonder why every one gets so nasty over cameras.

I, like some, own a lot of different cameras and lenses, more than I'd like to admit and they all do something different, like different films.

Have fun

Roger J.

Reply   Reply with quote   Complain
DMillier
Forum ProPosts: 17,499
Like?
Re: Nikon D4s: Bayer processing breakthru?
In reply to Tom Schum, 2 months ago

Different algorithms can reconstruct different amounts of detail. Allegedly, 90% plus of Nyquist is possible for luminance details with the right processing approach (although often 70% is quoted). Colour resolution is lower, but will that matter that much. We'll find out with the Quattro!

There is an interesting thread here: http://www.luminous-landscape.com/forum/index.php?topic=58661.20 based around Bart van der wolf's special res chart that covers the subject in some detail if you are interested in that kind of thing.

-- hide signature --

"...while I am tempted to bludgeon you, I would rather have you come away with an improved understanding of how these sensors work" ---- Eric Fossum
Galleries and website: http://www.whisperingcat.co.uk/
Flickr: http://www.flickr.com/photos/davidmillier/

Reply   Reply with quote   Complain
docmaas
Veteran MemberPosts: 4,738
Like?
Re: Nikon D4s: Bayer processing breakthru?
In reply to Tom Schum, 2 months ago

This reminded me of some anti diffraction technology I had seen briefly mentioned in Sony's newest RX100 iii.

"Two new technologies contribute to remarkably realistic images and bring out the full depiction performance of α lenses. Detail reproduction technology depicts details more faithfully with a more natural sense of dimension by preventing the overemphasized outlines that plague most digital images. Diffraction-reducing technology suppresses the effects of diffraction, which causes points of light to appear blurred especially at small aperture settings (large F-numbers). By taking the aperture setting into account, this technology faithfully restores clarity to points of light and other fine details: It brings out the potential of the lens, 36-megapixel image sensor and optical low-pass filter-free design so you can enjoy deep focus shooting of landscapes and more with a higher sense of clarity."

It's reproduced completely here but you can find it under the bionz processor at: http://discover.store.sony.com/sony-technology-services-apps-NFC/tech_imaging.html

There is also this about selective noise reduction:

"Area-specific noise reduction is now more powerful than ever. It selectively divides the image into areas based on patterns (such as edges, textures and evenly colored areas like blue skies), then applies the most appropriate noise reduction for each area to markedly reduce noise and improve image quality — especially at high-sensitivity settings. Working together with detail reproduction technology, it delivers supremely clear images with exceptionally fine detail."

Since both these are in the Bionz chip I suspect the manipulations are already present in the raw image but I can't say so for sure.

Mike

Tom Schum wrote:

This month's print edition of Shutterbug magazine (July 2014) has a review of the D4s, and on page 40 they say, "The resolution of 3280 lines per picture height given by the image sensor is transferred into images with a maximum of 3258 lines per picture height.... details and fine structures in these images look very crisp, but still natural." I've left out a couple sentences in which they speculate about how this might be done, and complain ever so slightly about some artifacts at hard contrast edges, but if you are interested the magazine is worth buying.

Typically, from what I have seen so far, it is very good to get even 75% of the sensor resolution out of a bayer array (60% more typically), and here we are at 99% with the D4s!

Once the DPR review comes out, it will be very interesting to compare image quality between the D4s (16 mp) and any Merrill (15mp). I'm guessing the D4s images will compare quite favorably.

-- hide signature --

Tom Schum

Reply   Reply with quote   Complain
Keyboard shortcuts:
FForum MMy threads