50mm Prime with Resolution for a D800? Come on Nikon/Sigma!

Started 11 months ago | Questions
Scott McMorrow
Regular MemberPosts: 362
Like?
Re: 50mm Prime with Resolution for a D800? Come on Nikon/Sigma!
In reply to Reilly Diefenbach, 11 months ago

Actually, you are wrong.  3200 lw/ph for a D800 computes to approximately 15 Mpix.  Your "test" does not show black lines that are separated at the pixel pitch.  In fact, you can already see at the edges that pixels are blurring into the neighboring white space.  Place another black region one more pixel over, with only one pixel of white space, and you will see the "resolution" collapse, because of poor microcontrast due to blur radius.

This is why resolution measurements are made with a specific contrast level specified.  MTF50 being a common standard, since it seems to track well with the way our visual systems perceive detail. Without specifying the contrast metric that is being used, a resolution measurement is meaningless.  This is akin to measuring bandwidth in analog electrical systems.  A measurement amplitude must be used or the bandwidth measurement is meaningless.  For electronics 3 dB (70.7%) amplitude or 6 dB(50%) amplitude is often used.

Reply   Reply with quote   Complain
Scott McMorrow
Regular MemberPosts: 362
Like?
Re: 50mm Prime with Resolution for a D800? Come on Nikon/Sigma!
In reply to MisterHairy, 11 months ago

MisterHairy wrote:

If you are finding it difficult to exploit the resolution of the D800 with an existing Nikkor 50mm lens then you are doing something wrong as all will comfortably outresolve it. You may need to look at your technique before considering blowing that sort of cash on the Zeiss 55mm as it would likely be wasted.

Please define the technical term "out resolve" in a scientific manner, rather then just bandying it about like everyone else in the thread seems to.  What do you mean when you say that a lens out resolves a sensor?  What is the test method?

Reply   Reply with quote   Complain
olyflyer
Forum ProPosts: 23,335
Like?
Re: 50mm Prime with Resolution for a D800? Come on Nikon/Sigma!
In reply to MisterHairy, 11 months ago

MisterHairy wrote:

Reilly Diefenbach wrote:

MisterHairy wrote:

If you are finding it difficult to exploit the resolution of the D800 with an existing Nikkor 50mm lens then you are doing something wrong as all will comfortably outresolve it. You may need to look at your technique before considering blowing that sort of cash on the Zeiss 55mm as it would likely be wasted.

The glass mystique dies hard :^)

£169 for the image and £3000 to feel confident that one has the very best lens ever ever ever (Otus is £3169 over here)

...and then shoot hand held with a manual focus lens...

I am pretty sure it is a nice lens and it most probably outresolves the Nikon 50/1.4G but I suspect that's only valid in studio tests.

Reply   Reply with quote   Complain
olyflyer
Forum ProPosts: 23,335
Like?
Re: 50mm Prime with Resolution for a D800? Come on Nikon/Sigma!
In reply to Scott McMorrow, 11 months ago

Scott McMorrow wrote:

MisterHairy wrote:

If you are finding it difficult to exploit the resolution of the D800 with an existing Nikkor 50mm lens then you are doing something wrong as all will comfortably outresolve it. You may need to look at your technique before considering blowing that sort of cash on the Zeiss 55mm as it would likely be wasted.

Please define the technical term "out resolve" in a scientific manner, rather then just bandying it about like everyone else in the thread seems to. What do you mean when you say that a lens out resolves a sensor? What is the test method?

I think that should be easy to define.

A lens out resolves a sensor when the maximum sensor resolution is reached with that lens. You can assume that the same lens on a better sensor would still have something to give.

However, I think all, or almost all, Nikon lenses fit that definition concerning the D800, so I am not worried for the next update.

Reply   Reply with quote   Complain
Matsu
Contributing MemberPosts: 865Gear list
Like?
Almost any 50 will do.
In reply to light_bulb, 11 months ago

For your states uses. I like the convenience of AF, but if you don't mind manual focus, your options will be even wider.

Once you get past the shallow depth of field fad, you will notice that in a studio taking full body and head and shoulders portraits you will be at a maximum of f/4, and more likely at f/5.6-f/8 depending on what kind of lighting you set up.

They're all already too sharp at those apertures, even for the most meticulous make-up.

You can buy more resolution, and there are uses for that, but what really concerns you is rendering, and sometimes you pay big bucks for those subtleties.

Your lighting makes a bigger difference than your lens will under those conditions. Buy the 50G. It's cheap, it's good, it's a little boring, but there's nothing offensive in its character or rendition. You will have resolution to spare.

 Matsu's gear list:Matsu's gear list
Nikon D7000 Nikon D800 Nikon AF-S Nikkor 70-200mm f/2.8G ED VR Nikon AF-S Nikkor 24-70mm f/2.8G ED Nikon AF-S Nikkor 50mm f/1.8G +6 more
Reply   Reply with quote   Complain
Scott McMorrow
Regular MemberPosts: 362
Like?
Re: 50mm Prime with Resolution for a D800? Come on Nikon/Sigma!
In reply to olyflyer, 11 months ago

what does it mean for a lens to meet the maximum sensor resolution?  How is this determined?

Reply   Reply with quote   Complain
michaeladawson
Veteran MemberPosts: 7,942Gear list
Like?
Re: 50mm Prime with Resolution for a D800? Come on Nikon/Sigma!
In reply to Scott McMorrow, 11 months ago

Scott McMorrow wrote:

Nexu1 wrote:

Nikon 50mm f1.8G: http://www.photozone.de/nikon_ff/631-nikkorafs5018ff?start=1

Sigma 35mm f1.4: http://www.photozone.de/nikon_ff/792-sigma3514dgfx?start=1

Both peak at 3900-4000.

Which are both about 20% below the maximum resolution of 4912 lw/ph of the D800 sensor.

One of the serious problems with the photozone.de resolution measurements is the way in which they measure border and corner resolution. What they do is to optimize focus in each zone, and then measure resolution. This neglects field curvature. Imatest MTF50 measurements made of the corners, when the center is in focus will be much lower than they report.

Aren't the Photozone MTF numbers determined from tests on a D3X camera?  And if so, then the capability of the lens hasn't been reached yet.

-- hide signature --

Mike Dawson

 michaeladawson's gear list:michaeladawson's gear list
Panasonic Lumix DMC-LX3 Nikon D200 Nikon D4 Nikon D800E Nikon AF-S Nikkor 70-200mm f/2.8G ED VR +17 more
Reply   Reply with quote   Complain
Reilly Diefenbach
Senior MemberPosts: 8,387Gear list
Like?
Re: 50mm Prime with Resolution for a D800? Come on Nikon/Sigma!
In reply to Scott McMorrow, 11 months ago

Scott McMorrow wrote:

Actually, you are wrong. 3200 lw/ph for a D800 computes to approximately 15 Mpix.

No.  Take any decent 50 and start counting actual pixels from side to side.  They'll all be there.

70-300VR

Your "test" does not show black lines that are separated at the pixel pitch. In fact, you can already see at the edges that pixels are blurring into the neighboring white space. Place another black region one more pixel over, with only one pixel of white space, and you will see the "resolution" collapse, because of poor microcontrast due to blur radius.

Uh, this was taken by the Sigma 70 in residence at Imaging Resource, their sharpest lens.  This is as good as it gets.

This is why resolution measurements are made with a specific contrast level specified. MTF50 being a common standard, since it seems to track well with the way our visual systems perceive detail. Without specifying the contrast metric that is being used, a resolution measurement is meaningless. This is akin to measuring bandwidth in analog electrical systems. A measurement amplitude must be used or the bandwidth measurement is meaningless. For electronics 3 dB (70.7%) amplitude or 6 dB(50%) amplitude is often used.

Any 50mm, 60G, 24-85VR at center, all produce the exact same stuff when zoomed in, as I have verified a thousand times.  I would defy you or anyone else to come up with pixel level crops from 50mm lenses at f8 that look any better from one to the next, regardless of the technobabble.  Have at it, Scott.  You're obviously on a mission.  Let's see what you've got!  No charts, please, just pics.

Reply   Reply with quote   Complain
light_bulb
Contributing MemberPosts: 514
Like?
Re: 50mm Prime with Resolution for a D800? Come on Nikon/Sigma!
In reply to Reilly Diefenbach, 11 months ago

Reilly Diefenbach wrote:

light_bulb wrote:

Sorry, but you must be kidding.

Nope, far from it. The break point is about 3200 lwph at most, which most lenses can do without breaking a sweat, at least center to mid stopped down a bit. Download any of the full size jpgs available all over the net or use one of your own. Clearly visible pixels will be the result zoomed in, which you can count one by one in Photoshop on the pixel grid from one side to the other. Want higher resolution? You'll have to buy a 50MP camera. Not what you wanted to hear, but pixels don't lie:

Portland, OR

24-85VR

Corner sharpness and overall contrast, color, etc. are all still in play, but not resolution with any 50mm stopped down. If you have any 50mm D800e full size jpgs which will not show pixels at 400%, by all means put a few of them up so we can see them.

Well, centre to mid is not sufficient anyway. However I will check out whether I do have an issue with my sample.

BTW, it is quite some time ago that I was in Portland OR on a trip all the way down from Vancouver to LA.

Cheers Andreas

Reply   Reply with quote   Complain
light_bulb
Contributing MemberPosts: 514
Like?
Re: The 50 1.8 out-resolves the D800
In reply to moving_comfort, 11 months ago

moving_comfort wrote:

light_bulb wrote:

Sorry, but you must be kidding.

The 50 1.8 G is fairly limited stopped down even in the centre not to talk about the borders/corners where it visibly suffers from field curvature.

This became very obvious to me once I used the D800 50 1.8 G and D800E 85 1.8 G side by side in studio. Since I currently also have a D800E, my further tries with the 50 1.8 G on the D800E confirmed that it never allows to achieve the crisp appearence that the 85 1.8 G can deliver.

.

??

Reilly was right. I suspect you have a bad 50 1.8G?

.

-- hide signature --

Here are a few of my favorite things...
---> http://www.flickr.com/photos/95095968@N00/sets/72157626171532197/

Probably. However my impression was that my results were consistent with reviews I have come accross.

Reply   Reply with quote   Complain
light_bulb
Contributing MemberPosts: 514
Like?
Re: I only know that the...
In reply to olyflyer, 11 months ago

olyflyer wrote:

light_bulb wrote:

Unfortunately there is quite a bit of atmospheric haze that limited what might have been achieved.

...but no lens or camera can cut through haze and magically remove it.

Yes indeed and this is why contrast and perceived sharpness suffers and such pictures cannot demonstrate what is possible under first best conditions.

Reply   Reply with quote   Complain
light_bulb
Contributing MemberPosts: 514
Like?
Re: 50mm Prime with Resolution for a D800? Come on Nikon/Sigma!
In reply to Maji, 11 months ago

Maji wrote:

I guess you want a 50mm lens for around $500 but want the sharpness of a lens of similar FL but costs around $2500 or more. There is a reason why the Otus and the Leicas cost so much. I am sure Nikon and Sigma can design and build one, but the market will be so small that it may not be economically feasible for them to do it.

I would like something reasonably better at no premium price (could be 500$). With growing pixel counts there will be a further upgrading of lenses or it will be entirely meaningless.

Reply   Reply with quote   Complain
light_bulb
Contributing MemberPosts: 514
Like?
Re: 50mm Prime with Resolution for a D800? Come on Nikon/Sigma!
In reply to Nexu1, 11 months ago

Nexu1 wrote:

Nikon 50mm f1.8G: http://www.photozone.de/nikon_ff/631-nikkorafs5018ff?start=1

Sigma 35mm f1.4: http://www.photozone.de/nikon_ff/792-sigma3514dgfx?start=1

Both peak at 3900-4000.

The issue with theses reviews is that they are based on what the lenses can deliver on a D3X. Any comparisons may therefore be limited by what that camera's sensor can resolve.

This approach is ok because you cannot use a lens on its own. However dfferences in lenses can remain hidden by the D3X combos but be visible with the D800.

Reply   Reply with quote   Complain
Scott McMorrow
Regular MemberPosts: 362
Like?
Re: 50mm Prime with Resolution for a D800? Come on Nikon/Sigma!
In reply to michaeladawson, 11 months ago

You are right, they are tested with a D3X.  As a result, i have some real issues with the validity of these measurements.  They are much too close to the theoretical resolution of the D3X, and are supposedly based on pictures imported into ACR with no sharpening applied.  If that is the case, it is pretty much impossible to get an MTF50 result that is close to half the theoretical resolution of the sensor with Imatest, due to the rolloff of the AA filter.

They could be reporting MTF10, which is not what they say they are reporting.  Or Photozone could be using Imatest to compute MTF50 with optimal sharpening of the filtered image.  Without more information about which Imatest measurement is being used, and what the processing in ACR is, it is not possible to fully interpret these results.

Reply   Reply with quote   Complain
MisterHairy
Senior MemberPosts: 1,454Gear list
Like?
Re: 50mm Prime with Resolution for a D800? Come on Nikon/Sigma!
In reply to Scott McMorrow, 11 months ago

Scott McMorrow wrote:

MisterHairy wrote:

If you are finding it difficult to exploit the resolution of the D800 with an existing Nikkor 50mm lens then you are doing something wrong as all will comfortably outresolve it. You may need to look at your technique before considering blowing that sort of cash on the Zeiss 55mm as it would likely be wasted.

Please define the technical term "out resolve" in a scientific manner, rather then just bandying it about like everyone else in the thread seems to. What do you mean when you say that a lens out resolves a sensor? What is the test method?

Oh honestly! Do the words "Nyquist limit" and "aliasing" mean nothing to you? Maybe, if you are having to ask me to clarify some (not very) "technical terms". Look them up, have a think and figure out how it is that a lens can induce the latter if it cannot resolve beyond the former.

Some people!

 MisterHairy's gear list:MisterHairy's gear list
Nikon D700 Nikon Df Nikon D810 Nikon AF-S Nikkor 70-200mm f/2.8G ED VR II Nikon AF-S Nikkor 24mm f/1.4G ED +14 more
Reply   Reply with quote   Complain
light_bulb
Contributing MemberPosts: 514
Like?
Re: 50mm Prime with Resolution for a D800? Come on Nikon/Sigma!
In reply to KLO82, 11 months ago

KLO82 wrote:

From dxomark:
50mm f1.8G --> D800: 16P-Mpix , D610: 16P-Mpix

50mm f1.4G --> D800: 16P-Mpix, D610: 16P-Mpix

85mm f1.4G --> D800: 19P-Mpix , D610: 19P-Mpix
100 2.0 Macro Planar --> D800: 17P-Mpix , D610: 17P-Mpix

135 2.8 APO Sonnar --> D800: 28P-Mpix , D610: 23P-Mpix

55 Otus --> D800: 29 P-Mpix , D610: 23 P-Mpix
It seems that even 100mm macro Planar is not very sharp either. But 135 APO Sonnar and 55 Otus are really very sharp.

Thanks for this. Yes I was aware of these figures even though I do not fully understand the dxo metric behind them. I would not take them too serious if I not had the impression that pictures look somewhat mushy.

Reply   Reply with quote   Complain
Scott McMorrow
Regular MemberPosts: 362
Like?
Re: 50mm Prime with Resolution for a D800? Come on Nikon/Sigma!
In reply to Reilly Diefenbach, 11 months ago

At f/8, sure.  But that is pretty arbitrary, isn't it, and not what the OP stated.  Diffraction limits all lenses with a D800 sensor at f/8.  You are right, they all look the same in the center.(at least reasonably decent lenses do).  That is not necessarily the case across the entire frame.  At wider apertures there are significant differences in lenses.

Again, show me a photograph with alternating dark and light lines at the pixel pitch, and then we can make a true comparison.  The 50mm f/1.8G is pretty darn good lens, especially for the price.  Most lenses will resolve all pixels in the center at some degree of contrast.  The question is, what is the contrast. (Often referred to as microcontrast.) You will find that a lens like the Otus will have significantly higher contrast than the 50 mm f/1.8G.

My point is that resolution alone is not enough to characterize a lens/camera system. It is resolution with contrast that is important.  Lower contrast indicates a larger blur circle.  When the blur circle is not dominated by diffraction, then lens aberrations come into play.

Until you can show a photograph or measurement you have made with a standardized method for measuring resolution, your statements and photographs are anecdotal.  Interesting at best.

Reply   Reply with quote   Complain
light_bulb
Contributing MemberPosts: 514
Like?
Re: 50mm Prime with Resolution for a D800? Come on Nikon/Sigma!
In reply to MisterHairy, 11 months ago

MisterHairy wrote:

If you are finding it difficult to exploit the resolution of the D800 with an existing Nikkor 50mm lens then you are doing something wrong as all will comfortably outresolve it. You may need to look at your technique before considering blowing that sort of cash on the Zeiss 55mm as it would likely be wasted.

For some the wisdom seems to be that if anything goes wrong using Nikon gear than the fault necessarily needs to reside behind the camera.

I have done test shots handheld with a Pentax 645D and a Leica S2 at not too fast shutter speeds with excellent results in terms of resolution. Probably I am spoiled by what I have seen.

Reply   Reply with quote   Complain
light_bulb
Contributing MemberPosts: 514
Like?
Re: 50mm Prime with Resolution for a D800? Come on Nikon/Sigma!
In reply to olyflyer, 11 months ago

olyflyer wrote:

MisterHairy wrote:

Reilly Diefenbach wrote:

MisterHairy wrote:

If you are finding it difficult to exploit the resolution of the D800 with an existing Nikkor 50mm lens then you are doing something wrong as all will comfortably outresolve it. You may need to look at your technique before considering blowing that sort of cash on the Zeiss 55mm as it would likely be wasted.

The glass mystique dies hard :^)

£169 for the image and £3000 to feel confident that one has the very best lens ever ever ever (Otus is £3169 over here)

...and then shoot hand held with a manual focus lens...

I am pretty sure it is a nice lens and it most probably outresolves the Nikon 50/1.4G but I suspect that's only valid in studio tests.

Indeed, there is quite a bit that might go wrong.

Reply   Reply with quote   Complain
light_bulb
Contributing MemberPosts: 514
Like?
Re: Almost any 50 will do.
In reply to Matsu, 11 months ago

Matsu wrote:

For your states uses. I like the convenience of AF, but if you don't mind manual focus, your options will be even wider.

Once you get past the shallow depth of field fad, you will notice that in a studio taking full body and head and shoulders portraits you will be at a maximum of f/4, and more likely at f/5.6-f/8 depending on what kind of lighting you set up.

They're all already too sharp at those apertures, even for the most meticulous make-up.

You can buy more resolution, and there are uses for that, but what really concerns you is rendering, and sometimes you pay big bucks for those subtleties.

Your lighting makes a bigger difference than your lens will under those conditions. Buy the 50G. It's cheap, it's good, it's a little boring, but there's nothing offensive in its character or rendition. You will have resolution to spare.

You are right with portraits resolution is a double-edged sword. You may not always want to see what a lens is able to resolve.

It may require a perfect skin, extra care for all detail, make-up or extra work in post. This is why Nikon has got its new 58 1.4.

I am also using the 50 1.8 for other purposes where resolution does not come at these kinds of 'cost'.

Reply   Reply with quote   Complain
Keyboard shortcuts:
FForum MMy threads