Velocity Reviews - Computer Hardware Reviews

Velocity Reviews > Newsgroups > Computing > Digital Photography > Analogue Lenses on Digital Body

Reply
Thread Tools

Analogue Lenses on Digital Body

 
 
Philip Homburg
Guest
Posts: n/a
 
      12-24-2005
In article <dojadn$csr$(E-Mail Removed)>,
David J. Littleboy <(E-Mail Removed)> wrote:
>"Philip Homburg" <(E-Mail Removed)> wrote:
>> David J. Littleboy <(E-Mail Removed)> wrote:
>>>In actual comparisons, the 5D + 17-40 looks a lot
>>>better than the D200 + 12-24.

>>
>> Do you have any examples that clearly show the differences?

>
>There's a thread on the dpreview Canon forum with examples that shows the
>differences clearly. I think the examples were taken from a Japanese site.


I assume you are refering to the site
<http://digitalcamera.impress.co.jp/06_01/auth/toku1/index.htm>

I found two images taken with the 12-24.

A proper comparison we be the 12-24 at 12/4.0 and the 16-35 at 18/5.6.
Or the 12-24 at 12/5.6 and the 16-35 at 18/8.0.
Unfortunately, the shots with the 5D were taken at 16mm.

Anyhow comparing 12/4.0 and 12/5.6 shots with the 16/5.6 shot doesn't show
anything that suggests that 'a lot better' is the right summery of the
differences.

The 12-24 does seem to need f/5.6 and the D200 images seem to be less sharpened
compared to the 5D images.


--
That was it. Done. The faulty Monk was turned out into the desert where it
could believe what it liked, including the idea that it had been hard done
by. It was allowed to keep its horse, since horses were so cheap to make.
-- Douglas Adams in Dirk Gently's Holistic Detective Agency
 
Reply With Quote
 
 
 
 
no_name
Guest
Posts: n/a
 
      12-24-2005
David J. Littleboy wrote:

> "Philip Homburg" <(E-Mail Removed)> wrote:
>
>>David J. Littleboy <(E-Mail Removed)> wrote:
>>
>>>In actual comparisons, the 5D + 17-40 looks a lot
>>>better than the D200 + 12-24.

>>
>>Do you have any examples that clearly show the differences?

>
>
> There's a thread on the dpreview Canon forum with examples that shows the
> differences clearly. I think the examples were taken from a Japanese site.


I expect if you could find a similar thread in a Nikon forum, the
comparison would probably favor the other direction.
 
Reply With Quote
 
 
 
 
David J. Littleboy
Guest
Posts: n/a
 
      12-24-2005

"Rich" <(E-Mail Removed)> wrote:
> On Sat, 24 Dec 2005 13:17:30 +0900, "David J. Littleboy"
>>
>>Another problem is that all lenses are soft, even in the center, wide
>>open.

>
> Residual spherical aberration.
>
>>APS-C requires 50 or 60% more magnification to get to the same print size,
>>so for wide open work (with a center subject and OOF background), you are
>>going to be unhappy with FF lenses on a cropped sensor (a lens with a
>>limited image circle can have better resolution (in lp/mm terms but not
>>lines per height terms), but then you get the funky corners problem back).
>>The cropped sensor fans argue that it's only the pixel count that matters,
>>but that assumes you are using the lens stopped down to the point its
>>resolution exceeds that of the sensor. For available light work, were
>>lenses
>>are noticeably softer, the larger pixels on the larger format means
>>sharper
>>images at the same print size.

>
> Sharper? Maybe, possibly. But we should differentiate sharpness from
> resolution, they are not the same. I've seen shots taken with poor
> sharpness that had superior resolution.


That's a film thing, which has a long "tail" of extremely low contrast
"resolution" out to the right of the MTF curve. Digital doesn't have that.
Usable resolution is pretty much gone around 2/3 of the Nyquist frequency
(the sensor's AA filter cuts in at roughly 2/3 of Nyquist (or should cut in,
sigh), and (correct) response is gone at Nyquist, so there can't be a tail).
So the <resolution and apparent sharpness> are largely determined by the MTF
of the lens at 2/3 of the sensor's Nyquist frequency.

If you make claims for the superiority of APS-C based on using the same
lenses, then you have to accept that you are going to have softer prints
wide open (and stopped down, which means that there isn't any extra DOF). If
you use specially designed lenses for APS-C, then the "use the good part" of
the lens argument doesn't hold. (I've seen one site that found that the
Sigma 30/1.4 has somewhat better resolution wide open than Canon the 50/1.4.
Exactly what you'd expect from a smaller image circle lens. But it has to
use the full image circle.)

> As for resolution, I've heard the reasons why a smaller sensor with
> smaller pixels aren't necessarily sharper, but isn't there something
> called, "oversampling" where a pixel smaller than the circle of
> confusion actually allows for better resolution?


No. You just see the lower MTF. Digital cameras are suprisingly sensitive to
lens MTF; I first noticed this in the Luminous landscape review of the older
17-35/2.8 vs. the new 16-35/2.8. Even with the D30's relatively large
pixels, the difference was noticeable. (Note that this somewhat contradicts
my "FF cameras can use cheap glass theory"<g>.)

> It must, because the
> best pictures of (for eg) planets where people use long telescope
> lenses are taken by Webcams with very tiny pixels. Some of these cams
> have 2um pixels and they kill DSLRs when it comes to these kinds of
> shots. The lenses being used (telescopes) typically have focal ratios
> of between f20 and f100. So the "spot size" argument is out the
> window.


With a good lens stopped down, the APS-C cameras, even the D2x, produce very
nice sharp images. The only difference is that the D2x is going to be more
sensitive to lens infelicities, since the magnification required to make a
print is 50% larger.

>>Basically, as long as photography consists of capturing light projected by
>>a
>>lens, a larger format is always going to have quality and/or speed
>>advantages over a smaller sensor, and the smaller sensor is going to have
>>convenience advantages. If you think you've found a quality advantage for
>>a
>>smaller sensor, you've made a mistake in your reasoning somewhere.

>
> I still think that in some situations, where more pixels can be put on
> a subject because of a sensor with smaller pixels, the image will have
> more detail. The solution of the FF users is simple; Use a lens that
> allows complete coverage of the sensor by the subject in question, but
> we know that does not happen all the time.


Sure it does: slap on a teleconverter. (Teleconverters degrade the angular
resolution of the lens only very slightly, so you can apply all your fat FF
pixels to the image.)

> Imagine a wildlife or
> sports photographer; They have to be able to predict exactly where
> an animal will appear in order to correctly frame them with a fixed
> telephoto. Virtually none of them use a zoom lens.


Yes. That's because most of the long zooms have lousy performance at the
long end.

> I think where FF sensors with large pixels show their stuff is
> 1. Where you can use the whole sensor.


Doh!

> 2. Where the lens can be stopped down enough to produce good
> aberration control.


This is only an issue in the wide angle, a range where cropped sensors have
problems of their own (such as a lack of availability of lenses). The rest
of the time, FF is more forgiving of funky lenses than APS-C, since you
don't enlarge the image as much.

> 3. Tonality and colour, apart from resolution. Like using a 4x5
> with medium resolution film versus a medium format with high
> resolution film. Of course, the enlargment process also included
> lenses so they'd have to be factored in too.


Yes, actually. But just tonality. I use the term "dynamic range" for this,
since the FF advantage is in lower noise, which can be used to get either
more shadow detail, better highlight rendition, or some of both. Of course,
high dynamic range images are low-contrast images, so they require a lot
more postprocessing to produce prints with the same pop. The small-pixel
cameras produce very contrasty images, since they map a smaller range of
tones onto the whole scale. The D5 in-camera tone curves just throw away
half the information captured to produce similar images. That's why people
think that the small-pixel cameras are "fine at ISO 100".

David J. Littleboy
Tokyo, Japan


 
Reply With Quote
 
Rich
Guest
Posts: n/a
 
      12-25-2005
On Sat, 24 Dec 2005 17:42:24 +0100, http://www.velocityreviews.com/forums/(E-Mail Removed) (Philip
Homburg) wrote:

>In article <dojadn$csr$(E-Mail Removed)>,
>David J. Littleboy <(E-Mail Removed)> wrote:
>>"Philip Homburg" <(E-Mail Removed)> wrote:
>>> David J. Littleboy <(E-Mail Removed)> wrote:
>>>>In actual comparisons, the 5D + 17-40 looks a lot
>>>>better than the D200 + 12-24.
>>>
>>> Do you have any examples that clearly show the differences?

>>
>>There's a thread on the dpreview Canon forum with examples that shows the
>>differences clearly. I think the examples were taken from a Japanese site.

>
>I assume you are refering to the site
><http://digitalcamera.impress.co.jp/06_01/auth/toku1/index.htm>
>
>I found two images taken with the 12-24.
>
>A proper comparison we be the 12-24 at 12/4.0 and the 16-35 at 18/5.6.
>Or the 12-24 at 12/5.6 and the 16-35 at 18/8.0.
>Unfortunately, the shots with the 5D were taken at 16mm.
>
>Anyhow comparing 12/4.0 and 12/5.6 shots with the 16/5.6 shot doesn't show
>anything that suggests that 'a lot better' is the right summery of the
>differences.
>
>The 12-24 does seem to need f/5.6 and the D200 images seem to be less sharpened
>compared to the 5D images.


The Nikon photo suffers from lack of sharpness (which IMO means the
lens has some kind of residual spherical aberration), look at the
peeling paint on the centre beam, and noticeable
chromatic aberration, look at the edges of the tree trunk and
branches.
IMO, the Nikon lens is inferior and the image is not a good comparison
image compared to the Canon's.
-Rich.
 
Reply With Quote
 
Rich
Guest
Posts: n/a
 
      12-25-2005
On Sun, 25 Dec 2005 07:56:57 +0900, "David J. Littleboy"
<(E-Mail Removed)> wrote:

>
>> 2. Where the lens can be stopped down enough to produce good
>> aberration control.

>
>This is only an issue in the wide angle, a range where cropped sensors have
>problems of their own (such as a lack of availability of lenses). The rest
>of the time, FF is more forgiving of funky lenses than APS-C, since you
>don't enlarge the image as much.


I'm not sure I understand this point. A lens of 100mm focal length is
going to produce the same "size" image on a sensor, no matter how
large the sensor is. The only addition you see from a FF sensor over
an APS-C sensor is an increased field of view. So for a given area of
the image, the size is going to be the same. If the size did change,
then the myth about smaller sensors producing "longer" focal length
lenses out of shorter ones would be true.

But to the point; It's my understanding that long lenses are
"self-fulfilling" high resolution "prophesies" because when they have
larger f-ratios (f4-f8, whatever)
they are easier to make to achieve excellent aberration control
and the lenses made at f2-f2.8 (high speed units that cost a fortune)
simply have better glass, better (more labour intensive) designs so
they can work well wide open. In both cases, long lenses are better
optically than either WA or medium focal length lenses. Because of
that, they should work equally well with both sensor sizes and neither
should be at any kind of disadvantage. In fact, I can attest to that
as a $300 500mm lens will produces as much resolution and contrast
as a $3000 lens provided it's focal length is long enough and it's
basic design is a good one. I once shot images though a pair of
telephoto lenses, one costing $2000, the other $250 (100 ISO film)
and asked people if they could tell them apart. They couldn't.
This isn't to say all cheap telephotos are good, on the contrary, but
some can be. A longer focal ratio is a great equalizer.

>
>> 3. Tonality and colour, apart from resolution. Like using a 4x5
>> with medium resolution film versus a medium format with high
>> resolution film. Of course, the enlargment process also included
>> lenses so they'd have to be factored in too.

>
>Yes, actually. But just tonality. I use the term "dynamic range" for this,
>since the FF advantage is in lower noise, which can be used to get either
>more shadow detail, better highlight rendition, or some of both. Of course,
>high dynamic range images are low-contrast images, so they require a lot
>more postprocessing to produce prints with the same pop. The small-pixel
>cameras produce very contrasty images, since they map a smaller range of
>tones onto the whole scale. The D5 in-camera tone curves just throw away
>half the information captured to produce similar images. That's why people
>think that the small-pixel cameras are "fine at ISO 100".


Until you see the noise in the shadowed areas. But it's also pretty
easy to see the dynamic range compression of smaller sensors.
My concern was only about resolution and even a lithographic image can
have high resolution and basically no intermediate dynamic range to
speak of.

But what I would like to know is this; If you take the Canon 5D and
the Nikon D200 and shoot a picture through both, at what point
(pixel coverage) would the Nikon and the Canon produce equally
detailed shots? Lets assume that the lenses are equally good.
Lets assume also that you need more magnification on a subject with
the Nikon to glean as much detail from any given area of the subject
as the Canon shot.
Would the Nikon produce as much detail with a 125mm lens as say the
Canon would with a 100mm lens?
-Rich
 
Reply With Quote
 
David J. Littleboy
Guest
Posts: n/a
 
      12-25-2005
"Rich" <(E-Mail Removed)> wrote:
> "David J. Littleboy"<(E-Mail Removed)> wrote:
>>
>>> 2. Where the lens can be stopped down enough to produce good
>>> aberration control.

>>
>>This is only an issue in the wide angle, a range where cropped sensors
>>have
>>problems of their own (such as a lack of availability of lenses). The rest
>>of the time, FF is more forgiving of funky lenses than APS-C, since you
>>don't enlarge the image as much.

>
> I'm not sure I understand this point. A lens of 100mm focal length is
> going to produce the same "size" image on a sensor, no matter how
> large the sensor is. The only addition you see from a FF sensor over
> an APS-C sensor is an increased field of view. So for a given area of
> the image, the size is going to be the same. If the size did change,
> then the myth about smaller sensors producing "longer" focal length
> lenses out of shorter ones would be true.


Suppose we want to take the same photogtaph. So I put my 50/1.4 on my 5D and
you put your Sigma 30/1.4 on your D2x. When I make my 12x16 print, it's a
13x enlargement, and when you make your 12x16 print it's a 19.5x
enlargement. If we agree that 3 lp/mm at 50% MTF is "sharp" (just a random
criteria, pick any that you please), then I have to stop down my 50/1.4
until it coughs up (3 x 13 = 39) lp/mm at 50% MTF, but you have to stop down
your 30/1.4 until it coughs up (3 x 19.5 = 58.5) lp/mm at 50% MTF.

Whatever criteria we pick for sharpness in the final print, the smaller
format lens must provide that MTF at 1.5x the frequency.

If these were telephoto lenses, there might not be a problem. (Maybe) And if
they were superwides, then most of the superwides for FF require a lot of
stopping down to sharpen up the corners.

But while 39 lp/mm at 50% MTF is within reason for better 35mm SLR lenses of
focal lengths of 50mm or shorter, 58.7 lp/mm at 50% MTF is completely out of
the question. (Actually, 39 lp/mm at 50% MTF is fairly hard at those focal
lengths, but don't tell anyone.) So the idea of using FF lenses on cropped
sensors is problematic; they won't have the resolution, especially wide
open.

Not only is the noise a couple of stops worse, your lenses are going to be
painfully soft near wide open. So high-res APS-C sensors are a really bad
idea for low light.

> But to the point; It's my understanding that long lenses are
> "self-fulfilling" high resolution "prophesies" because when they have
> larger f-ratios (f4-f8, whatever)
> they are easier to make to achieve excellent aberration control
> and the lenses made at f2-f2.8 (high speed units that cost a fortune)
> simply have better glass, better (more labour intensive) designs so
> they can work well wide open. In both cases, long lenses are better
> optically than either WA or medium focal length lenses. Because of
> that, they should work equally well with both sensor sizes and neither
> should be at any kind of disadvantage. In fact, I can attest to that
> as a $300 500mm lens will produces as much resolution and contrast
> as a $3000 lens provided it's focal length is long enough and it's
> basic design is a good one. I once shot images though a pair of
> telephoto lenses, one costing $2000, the other $250 (100 ISO film)
> and asked people if they could tell them apart. They couldn't.
> This isn't to say all cheap telephotos are good, on the contrary, but
> some can be. A longer focal ratio is a great equalizer.


Yes. To the best of my knowledge, that's quite right. Longer lenses are
(probably) fine for high-res cropped cameras. (I focus on the wide to normal
range, because that's where most of my photography happens.) Still, the D2x
really does require 60 lp/mm with decent contrast at the sensor, so I'd
guess that people with tele zooms (which tend to be funky at the long end)
would be unhappy.

> But what I would like to know is this; If you take the Canon 5D and
> the Nikon D200 and shoot a picture through both, at what point
> (pixel coverage) would the Nikon and the Canon produce equally
> detailed shots? Lets assume that the lenses are equally good.
> Lets assume also that you need more magnification on a subject with
> the Nikon to glean as much detail from any given area of the subject
> as the Canon shot.
> Would the Nikon produce as much detail with a 125mm lens as say the
> Canon would with a 100mm lens?


I looked _very_ closely at some 5D vs. 1Dsmk2 comparisons, and couldn't
_honestly_ see any difference in detail rendition, despite the 15% extra
linear resolution. So for telephoto work, I'd be seriously surprised if
anyone could see any difference between 5D and D200.

I bet you would see a difference between the D200 and the 1Ds2, which
represents a 25% increase in linear resolution, but the 10% difference
between the D200 and the 5D really shouldn't make a difference in prints up
to 11x14 or so.

(I think I'm going to get excommunicated from the Canonistas for saying
that.)

David J. Littleboy
Tokyo, Japan


 
Reply With Quote
 
Philip Homburg
Guest
Posts: n/a
 
      12-25-2005
In article <(E-Mail Removed)>, Rich <dfs> wrote:
>The Nikon photo suffers from lack of sharpness (which IMO means the
>lens has some kind of residual spherical aberration), look at the
>peeling paint on the centre beam, and noticeable
>chromatic aberration, look at the edges of the tree trunk and
>branches.
>IMO, the Nikon lens is inferior and the image is not a good comparison
>image compared to the Canon's.


Which images are you talking about?

I was looking at the series that starts with
http://digitalcamera.impress.co.jp/0...1/60108001.jpg

I can't find anything that can be described as 'the centre beam'.
And 'the three trunk' is not really there either.


--
That was it. Done. The faulty Monk was turned out into the desert where it
could believe what it liked, including the idea that it had been hard done
by. It was allowed to keep its horse, since horses were so cheap to make.
-- Douglas Adams in Dirk Gently's Holistic Detective Agency
 
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
Difference between lenses for film and lenses for digital? One4All Digital Photography 22 01-14-2007 05:00 AM
Digital/Analogue Comparisons Steve Hodgson Digital Photography 47 10-28-2006 02:18 PM
Gamepads digital v analogue John Computer Information 2 12-09-2005 03:54 PM
Analogue or digital camcorder? Kenny Computer Support 1 11-24-2004 09:17 PM
Re: Transfer from analogue to digital Trick Computer Information 2 06-24-2003 06:43 AM



Advertisments