Velocity Reviews - Computer Hardware Reviews

Velocity Reviews > Newsgroups > Computing > Digital Photography > Nikon new release D7100

Reply
Thread Tools

Nikon new release D7100

 
 
Me
Guest
Posts: n/a
 
      03-12-2013
On 12/03/2013 12:23 p.m., Floyd L. Davidson wrote:
> Me <(E-Mail Removed)> wrote:
>> On 11/03/2013 11:35 p.m., Floyd L. Davidson wrote:
>> <snip>
>>>
>>> I don't think that is a correct evaluation of what
>>> Nikon is doing at all. They try to sell what people
>>> want to buy.
>>>
>>> That doesn't mean Nikon is not aware that a 150MP
>>> sensor will provide a better image in terms of aliasing.
>>>
>>>> http://imaging.nikon.com/lineup/dslr/d7100/sample.htm
>>>>
>>>> I await more sample images with interest.
>>>
>>> Do you actually expect Nikon marketing to provide images
>>> that demonstrate a fault???
>>>

>> Indeed they did when the released the D800 and D800E,
>> IIRC, to demonstrate they issue they put up some photos
>> on the Nikon Imaging site of a woman wearing a kimono,
>> with moire clearly visible in the D800E shot.
>>
>> There have been plenty of A:B comparison photos taken
>> with D800 and D800E put up on the 'net. With a tiny bit
>> of low radius USM applied to the D800 shots, they look
>> just like the D800E shots - I can't tell them apart at
>> 100% pixel view anyway. The AA filter isn't a cure-all
>> either -
>> you can still get moire with the D800.

>
> All of the above is exactly true.
>
> Of particular significance is that last line, "you can
> still get moire with the D800". That is with all these
> other effects that David Taylor believes will remove it
> for cameras without an AA filter *and* with an AA filter
> in place.
>

A worrying aspect to this, is that at present rate of progress, these
arguments are likely to to continue for a very long time indeed:
http://oi50.tinypic.com/10qgtjn.jpg
<g>
 
Reply With Quote
 
 
 
 
David Taylor
Guest
Posts: n/a
 
      03-12-2013
On 11/03/2013 23:23, Floyd L. Davidson wrote:
[]
> Of particular significance is that last line, "you can
> still get moire with the D800". That is with all these
> other effects that David Taylor believes will remove it
> for cameras without an AA filter *and* with an AA filter
> in place.


The optical anti-alias filters used with DSLRs are not the same as the
analogue or digital brick-wall filters used in audio, and do not have a
sharp cut-off just below the Nyquist frequency.

As you have now resorted to personal attacks, don't expect any further
response from me.
--
Cheers,
David
Web: http://www.satsignal.eu
 
Reply With Quote
 
 
 
 
David Taylor
Guest
Posts: n/a
 
      03-12-2013
On 12/03/2013 02:25, Me wrote:
[]
> A worrying aspect to this, is that at present rate of progress, these
> arguments are likely to to continue for a very long time indeed:
> http://oi50.tinypic.com/10qgtjn.jpg
> <g>


Nice one! There really should not be an argument, as the physics and
signal-processing theory are well understood. The discussion is really
about "when are aliasing artefacts sufficiently weak that they can be
ignored?". Different people will have different requirements.
--
Cheers,
David
Web: http://www.satsignal.eu
 
Reply With Quote
 
David Taylor
Guest
Posts: n/a
 
      03-12-2013
On 11/03/2013 19:48, Me wrote:
[]
> Indeed they did when the released the D800 and D800E, IIRC, to
> demonstrate they issue they put up some photos on the Nikon Imaging site
> of a woman wearing a kimono, with moire clearly visible in the D800E shot.
>
> There have been plenty of A:B comparison photos taken with D800 and
> D800E put up on the 'net. With a tiny bit of low radius USM applied to
> the D800 shots, they look just like the D800E shots - I can't tell them
> apart at 100% pixel view anyway. The AA filter isn't a cure-all either -
> you can still get moire with the D800.


Yes, you would expect this. Unlike the filters (analogue and digital)
used in audio, the optical low-pass filter used in cameras does not have
a particularly sharp cut-off. It's not a "brick-wall" filter (i.e. one
have near zero response just after Nyquist and near 100% response just
before Nyquist).

As there is not "correct" value for such a comprise filter, you find
some cameras having a slightly stronger AA filter (set at a lower
spatial frequency) and others compromise the other way, set the nominal
cut-off at a higher spatial frequency, and hence produce sharper images,
but with a great possibility of aliasing artefacts.

But if there is no information at or above the Nyquist frequency, there
can be no alias artefacts.
--
Cheers,
David
Web: http://www.satsignal.eu
 
Reply With Quote
 
Wolfgang Weisselberg
Guest
Posts: n/a
 
      03-12-2013
David Taylor <(E-Mail Removed)> wrote:
> On 09/03/2013 11:03, Wolfgang Weisselberg wrote:
>> David Taylor <(E-Mail Removed)> wrote:


>>> All I'm saying is that it's (excuse the pun) not black and white. Given
>>> a sufficiently high pixel density, the AA filter may not be required
>>> under certain circumstances, and as pixel density increases, those
>>> criteria are increasingly likely to be encountered.


>> I agree. I just disagree with you that current DSLR pixel
>> counts are close to a 'sufficiently high pixel density'.


> Again, you are putting words in my mouth which I didn't say.


So do we agree that current DSLRs don't have 'sufficiently
high pixel density'?


> It appears that
> for today's combinations of those variables, with 24 MP DX cameras and
> 36 MP full-frame cameras we are either approaching or have reached
> "sufficiently high" for many people for much of the time.


But here you are saying that current DSLR pixelcounts are
close to a sufficiently high pixel density (at least for
many situations) --- which is what I 'put on your mouth'!


> Test cases, super-expensive lenses, and particular subjects excepting.


Good macro lenses are not super-expensive, for example.

-Wolfgang
 
Reply With Quote
 
Wolfgang Weisselberg
Guest
Posts: n/a
 
      03-12-2013
David Taylor <(E-Mail Removed)> wrote:
> On 11/03/2013 23:23, Floyd L. Davidson wrote:


>> Of particular significance is that last line, "you can
>> still get moire with the D800". That is with all these
>> other effects that David Taylor believes will remove it
>> for cameras without an AA filter *and* with an AA filter
>> in place.


> The optical anti-alias filters used with DSLRs are not the same as the
> analogue or digital brick-wall filters used in audio, and do not have a
> sharp cut-off just below the Nyquist frequency.


True but irrelevant unless the D800E will have less moire
(or less likely moire) in the identical conditions where the
D800 non-E has moire.

> As you have now resorted to personal attacks, don't expect any further
> response from me.


Uh --- where *is* the personal attack?

All I can see is the claim that you believe that stuff like
non-perfect lenses, stopping down a bit, imperfect focus
etc. will prevent moire for high pixel count cameras. That may
be a misrepresentation you can correct (e.g. "I agree that the
D800 pixeldensity is too low for preventing moire generally,
but by now I am talking about ..."), but a personal attack?

Unless you read that as "hehe, David is soo stupid, he actually
believes that the D800E can't have moire, the idiot". Which is
not what was written!

-Wolfgang
 
Reply With Quote
 
Wolfgang Weisselberg
Guest
Posts: n/a
 
      03-12-2013
David Taylor <(E-Mail Removed)> wrote:
> On 12/03/2013 02:25, Me wrote:


>> A worrying aspect to this, is that at present rate of progress, these
>> arguments are likely to to continue for a very long time indeed:
>> http://oi50.tinypic.com/10qgtjn.jpg
>> <g>


> Nice one! There really should not be an argument, as the physics and
> signal-processing theory are well understood. The discussion is really
> about "when are aliasing artefacts sufficiently weak that they can be
> ignored?". Different people will have different requirements.


Identical people have different requirements at different times.
And many won't have the money to have both an AA and a non-AA
camera, nor would enjoy carrying both.

If you however know your future requirements to not need an
AA filter *and* to need the (assumed) extra resolution, feel
free to buy such a camera, the matching lenses and the sturdy
tripod and cable release ...

As for my needs: I'll stay with AA, but then I'd be happy
enough with 8 MPix + (good) ISO 12,800 or more. But that's
not available today, so my camera has more MPix. (My old
camera had 8 MPix, and that was nearly never a concern, much
less a problem. Quite unlike high ISO.)


-Wolfgang
 
Reply With Quote
 
David Taylor
Guest
Posts: n/a
 
      03-12-2013
On 11/03/2013 17:19, Wolfgang Weisselberg wrote:
> David Taylor <(E-Mail Removed)> wrote:
>
>> The discussion is not about a 36 MP sensor, but one which has a pixel
>> density equivalent to a 56 MB full-frame one.

>
> That's just 25% more linearly and doesn't matter much.
>
>> To reach the resolving
>> limit of the lens requires (most likely) using a tripod,

>
> A short enough time or flash suffices with usable techniques.
>
>> being in
>> perfect focus,

>
> The assumption being that
> a) the AF isn't good enough
> (Canon's AF in the newest cameras, using their newest
> lenses is as good as LifeView AF and manual focussing
> using LifeView + 10x zoom!)
> b) you or your subject move too much for the focus to stay
> true to the closing shutter
> c) only the part you want to focus on can have moire
> Imagine photographing a clothed person. Can you guarantee
> that your plane of focus will never go through some clothing?
>
>> having a subject which doesn't move significantly during
>> the exposure time,

>
> Again: short enough time or flash.
> Manmade objects often don't move during photography and are
> common sources for close parallel lines ...
>
>> and having an atmospheric MTF which is perfect.

>
> There's a wide range between a macro shot with a couple cm
> air in a cleanroom and a photographing something from 10 km
> away in dense fog.
>
>> Most
>> practical lenses will not reach the diffraction-limited resolution when
>> fully open.

>
> Usable ones often will at f/4 or f/5.6.
>
> f/5.6 means airy disks of ~3.8µm (green light).
> If airy disk size ~= pixel size: ~61 MPix.
> But you want at least 3 pixels being hit with one airy disk:
> red, green and blue. With the standard Bayer pattern, this
> means 4 pixels, and that means ~245 MPix for f/5.6!
>
>
>> What is happening is that the sensor pixel density in the D7100 is now
>> at the levels where for the majority of users there is no need for an
>> anti-alias filter, or that's what Nikon appears to believe.

>
> That's at least what Nikon appears to believe.
>
>> I don't see
>> significant aliasing issues in these images, do you?
>> http://imaging.nikon.com/lineup/dslr/d7100/sample.htm

>
> Even if the camera was only able to produce usable pictures
> after sacrificing several virgins and waving a black chicken
> over it, I'd expect near perfect samples from the manufacturer.
> (And a huge pile of ex-virgins and a whole zoo of black
> chicken in their backyard, as well as reams of photos they'll
> never show.)
>
> You don't, I take it?
>
> Not that these images had much stuff that'd tend to produce
> moire ... to bad that the air filter(?) on the bottom of the
> biplane's engine doesn't seem to resolve whatever pattern it
> has ...
>
> -Wolfgang


25% may not be a lot on its own, but it's the successive 25% (or
whatever) that count, the cumulative effect of going from 6 to 8 to 10
etc. MP.

Studio conditions will be different to outdoor photography (where I
don't use flash very often). No one set of options will suit everyone.

Agreed with the need to consider Bayer requirements, although I don't
think it's as simple doubling the spatial requirement, or quadrupling
the pixel requirement.

Do you mean:
http://chsvimg.nikon.com/lineup/dslr...e/img_04_l.jpg ? I
can see detail in the air filter. On closer examination, though, I can
also see aliasing on the wires between the wings, which I had not seen
before.

As I said, more sample photos are needed, and independent ones.
--
Cheers,
David
Web: http://www.satsignal.eu
 
Reply With Quote
 
David Taylor
Guest
Posts: n/a
 
      03-12-2013
On 11/03/2013 16:34, Wolfgang Weisselberg wrote:
> David Taylor <(E-Mail Removed)> wrote:
>> On 10/03/2013 10:49, Floyd L. Davidson wrote:

>
>>> 36MP is not even close, unless you stop the lens down to
>>> f/45 or something very small to get excessive
>>> diffraction.

>
>> Well, I was perhaps thinking more of 24 MB on DX format - as offered by
>> the Nikon D7100 - which is nearer 56 MB on full-frame.

>
> So an increase of less than 25% linearly is your difference
> between "not even close" and "more than needed"?


Not my words....

> And yes, you still get moire, AA filter or not, with these
> cameras.


Of course, as I already mentioned.

>> If 36 MP on full
>> frame were as bad as you seem to imply, wouldn't we be hearing more
>> reports of how bad the D800E was, how many were returning them, and what
>> a mistake it was?

>
> Why should they?
> The ones who know what they are doing, i.e. weighted the
> possible false detail against the increased resolution, are
> checking while shooting and reshoot if necessary.
>
> The ones who are blinded by the "more MP, more sharpness"
> drivel probably aren't seeing the moire --- but will once
> they are going to recheck their photos in a couple years,
> having become wiser. (I remember my first digital camera.
> It was decentered, producing a rather large spot of softness
> on the right hand side. Took me a while to spot that ---
> I wasn't shooting test charts --- but once you knew it it was
> visible in many shots.)
>
> The ones who shoot non-critical subjects will encounter it
> rarely.
>
> But for general photography moire might be the same as if your
> AF misfocusses every 20th, 50th or 100th frame and mars or ruins
> that picture. I understand that's not acceptable to many,
> wheras the slight decrease in sharpness isn't much of a
> problem; most of it can be recovered.
>
> In addition, colourful single pixel highlights --- while
> correctable in post --- aren't a joy to have either. With AA
> filters, you get at least the right colour.


Yes, but there comes a point where there isn't sufficient information at
the above Nyquist spatial frequency to produce the aliasing artefacts in
the first place. Getting that perfect sharpness isn't easy, as reports
on the D800 demonstrated. Different photographers will have different
needs.

>> Perhaps "Bruce" would tell us what the return rate
>> was, if he were still listening?

>
> Which, see above, won't tell you much about the camera's ability
> to perform in general photography without restrictions. *If*
> /you/ can live with the restrictions, fine! Just don't tell
> me /I/ and John Q. Photographer and Jane Q. Snapshooter have
> to live with them, too!
>
> -Wolfgang


Nikon would appear to be doing that with the D7100, though. It will be
interesting to see how well it's accepted.
--
Cheers,
David
Web: http://www.satsignal.eu
 
Reply With Quote
 
Wolfgang Weisselberg
Guest
Posts: n/a
 
      03-12-2013
David Taylor <(E-Mail Removed)> wrote:
> On 11/03/2013 17:19, Wolfgang Weisselberg wrote:
>> David Taylor <(E-Mail Removed)> wrote:


>>> The discussion is not about a 36 MP sensor, but one which has a pixel
>>> density equivalent to a 56 MB full-frame one.


>> That's just 25% more linearly and doesn't matter much.


>>> To reach the resolving
>>> limit of the lens requires (most likely) using a tripod,


>> A short enough time or flash suffices with usable techniques.


>>> being in
>>> perfect focus,


>> The assumption being that
>> a) the AF isn't good enough
>> (Canon's AF in the newest cameras, using their newest
>> lenses is as good as LifeView AF and manual focussing
>> using LifeView + 10x zoom!)
>> b) you or your subject move too much for the focus to stay
>> true to the closing shutter
>> c) only the part you want to focus on can have moire
>> Imagine photographing a clothed person. Can you guarantee
>> that your plane of focus will never go through some clothing?


>>> having a subject which doesn't move significantly during
>>> the exposure time,


>> Again: short enough time or flash.
>> Manmade objects often don't move during photography and are
>> common sources for close parallel lines ...


>>> and having an atmospheric MTF which is perfect.


>> There's a wide range between a macro shot with a couple cm
>> air in a cleanroom and a photographing something from 10 km
>> away in dense fog.


>>> Most
>>> practical lenses will not reach the diffraction-limited resolution when
>>> fully open.


>> Usable ones often will at f/4 or f/5.6.


>> f/5.6 means airy disks of ~3.8µm (green light).
>> If airy disk size ~= pixel size: ~61 MPix.
>> But you want at least 3 pixels being hit with one airy disk:
>> red, green and blue. With the standard Bayer pattern, this
>> means 4 pixels, and that means ~245 MPix for f/5.6!


>>> What is happening is that the sensor pixel density in the D7100 is now
>>> at the levels where for the majority of users there is no need for an
>>> anti-alias filter, or that's what Nikon appears to believe.


>> That's at least what Nikon appears to believe.


>>> I don't see
>>> significant aliasing issues in these images, do you?
>>> http://imaging.nikon.com/lineup/dslr/d7100/sample.htm


>> Even if the camera was only able to produce usable pictures
>> after sacrificing several virgins and waving a black chicken
>> over it, I'd expect near perfect samples from the manufacturer.
>> (And a huge pile of ex-virgins and a whole zoo of black
>> chicken in their backyard, as well as reams of photos they'll
>> never show.)


>> You don't, I take it?


>> Not that these images had much stuff that'd tend to produce
>> moire ... to bad that the air filter(?) on the bottom of the
>> biplane's engine doesn't seem to resolve whatever pattern it
>> has ...


>> -Wolfgang


> 25% may not be a lot on its own, but it's the successive 25% (or
> whatever) that count, the cumulative effect of going from 6 to 8 to 10
> etc. MP.


Correct. However, since we started with 6MP very far away
from not needing an AA filter[1] a *lot* of successive 25%
increases are necessary. It has been shown that current DSLR
sensor densities can have moire even with AA filters ...


> Studio conditions will be different to outdoor photography (where I
> don't use flash very often). No one set of options will suit everyone.


There are however sets of options which are able to support
an all-round camera --- they are for most usages the safe bet.

Of course, for some the chance to wring every last bit of
resolution out of their camera is much more important than
the risk of aliasing and moire and coloured highlights and
therefore a camera that's less viable for some situations.
It's a good thing for these people to be able to buy what they
need/want (as long as they understand the drawbacks).


> Agreed with the need to consider Bayer requirements, although I don't
> think it's as simple doubling the spatial requirement, or quadrupling
> the pixel requirement.


It's at least what I understand the AA filter does: spread
one pixel's light to e.g. the pixel to the right and both to
the pixels below.

> Do you mean:
> http://chsvimg.nikon.com/lineup/dslr...e/img_04_l.jpg ? I
> can see detail in the air filter.


Yes --- I see some noise-like pattern of gray with darker
random dots, and some (comparatively large) stains; I guess
there's a regular grid there but I cannot see it --- it's too
fine to be resolved. (Or maybe it's indeed a sort of grey
foam or cloth which is so low contrast that it's structure
wouldn't be visible even if it was large enough.) (Of couse
I can see the metal holder grid

> On closer examination, though, I can
> also see aliasing on the wires between the wings, which I had not seen
> before.


There's something strange there, but for all I know the wires
could be vibrating with a high frequency in the airstream and
we're seeing some static waveform: the more blurred parts is
where the wire has lots of amplitude and the sharper parts
where the vibration nodes are (and where the wire has almost
no amplitude and is static).

> As I said, more sample photos are needed, and independent ones.


Yep, indeed.

-Wolfgang

[1] 1995's Kodak DCS460 had 6 MPix, and unless I don't mix up
things it had no AA filter and had reportedly fun with
strangely coloured highlights, one of the things an AA
filter is designed to combat)
 
Reply With Quote
 
 
 
Reply

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Would Nikon release new telescopes? RichA Digital Photography 9 08-30-2010 09:43 AM
Nikon hints at "New Concept" camera release, possibly this year Bruce Digital Photography 30 08-16-2010 03:25 AM
J2ME: javac: target release 1.1. conflicts with default source release 1.5 John Goche Java 1 12-17-2005 08:12 PM
Nikon 8800 vs Nikon 990 vs Canon 8MP Rebel vs Nikon D70 fj40rockcrawler@gmail.com Digital Photography 10 06-07-2005 06:53 PM
NEWS RELEASE: Leone's "Nobody" Films 1st time mastered in HD for DVD release with 240 Minutes Extras Torsten Kaiser \(TLEFilms\) DVD Video 1 06-06-2005 03:17 PM



Advertisments