Question for 1d owners?

Ok, leave aside number of pixels for the moment, because they are NOT IMPORTANT.

Look at it this way.

Put a 300mm lens on a D60, on a 1D and on a 1V, and take a shot with say a bird in the centre. Print all the shots the SAME SIZE (regardless of resolution, print them the SAME size).

Now, the picture from the D60 will have the largest bird, and the picture from the 1V will have the smallest bird. In order to get the bird from the film image the 1V recorded to the same size as the image the D60 recorded, you have to crop down the original 35mm negative to EXACTLY the same size as the D60's sensor, and THEN enlarge it to the same size as the other prints.

This crop-and-enlarge is done by the smaller sensor in the D60. There is no interpolation.

It's obvious you're a fairly intelligent guy, don't need any of this explanation, and are merely using it as a mental exercise. You still seem to have some confusion about the term interpolation, but that's nothing fatal.

Member said:

If there is no "standard size" then let's use the image "size" as
it's seen on the 35mm negative or transparency at a given optical
focal length. The D60 with its 1.6x crop factor produces an image
reduced in field of view and increased in relative size, as does
the D30.

Click to expand...

You also seem to want a conversion factor between subjective resolution of film and digital sensors, and since that's pretty subjective I'll avoid giving you mine.

Obviously, the orginal "1.6x" comes from the fact that you can take a picture with a 1V, then take the same shot (distance, lens) with a D60, then take your CF and film in and get them exposed at say Ritz, and you'll see smaller subject details on your 4x6s with the D60. Sure, the paper is limiting, you should crop the film, blah, blah, blah. Regardless of all that, that's the initial basis for such a number.

There's nothing wrong with being picky and wanting to go beyond this, and talk about actual resolving power (such as with bigger prints) to form a different definition of "X" factor. You'd need to assign a number to what film resolves. This has been done before in music for instance. The numbers 8 bits and 32.1kbps have been assigned as the "equivalent" to FM stereo. It's a subjective number, but almost universally agreed upon. In some ways FM analog broadcast is superior, in some ways it can be inferior. It drastically depends on implementation (such as with film).

Anyway, I think some similar number would have to be assigned to a standard 35mm frame. What's this number? Just like before it would be a bitrate (resolution) and the number of bits (signal to noise ratio). Now, you could go ahead and assign the number 20MP to the resolution due to some number of grains listed in a spec of some film somewhere, but the number of bits is going to suffer, due to all that grain in the image. If a mythical 20MP Canon 1DSV has a 20MP image that looks as bad as Provia 100 at 20MP, customers will complain for sure. So, you're back to numbering it to something lower. As I said, that number is subjective, but you'd have to pick something. If you read:

http://www.luminous-landscape.com/reviews/film/iso100-400.shtml

you'll see some think that number is rather low. Some think it's astronomically high. YMMV.

Isaac Sibson said:

Ok, leave aside number of pixels for the moment, because they are
NOT IMPORTANT.

Click to expand...

The pixel count doesn't appear to to be important here - whether or not is actually is, is part of the question.....

Isaac Sibson said:

Look at it this way.

Put a 300mm lens on a D60, on a 1D and on a 1V, and take a shot
with say a bird in the centre. Print all the shots the SAME SIZE
(regardless of resolution, print them the SAME size).

Now, the picture from the D60 will have the largest bird, and the
picture from the 1V will have the smallest bird. In order to get
the bird from the film image the 1V recorded to the same size as
the image the D60 recorded, you have to crop down the original 35mm
negative to EXACTLY the same size as the D60's sensor, and THEN
enlarge it to the same size as the other prints.

This crop-and-enlarge is done by the smaller sensor in the D60.
There is no interpolation.

Click to expand...

Yes, the crop and enlarge is obviously done by the sensor, or by the sensor and other electronics in the camera. The crop part is very simple to explain by optical physics - but explaining the "enlarge" part doesn't seem to be quite as easy, does it?

Lin
--
http://208.56.82.71

Pixels and croping and magnification make no difference if you cannot capture the image. When I switched from my 1V to my D30, I absolutely stopped taking wildlife and bird images. Two reasons. AF and shutter lag. If you shoot mostly wildlife you will be very disappointed with the D60. Your percent of hits will be substandard and you will be photographing the back of a lot of birds.

Ray Amos
--

Canon EOS digital & film cameras and lenses from 17-500mm. Waiting patiently for my new Canon 24-105 f/2.8 IS lens.

Ray Amos said:

Pixels and croping and magnification make no difference if you
cannot capture the image. When I switched from my 1V to my D30, I
absolutely stopped taking wildlife and bird images. Two reasons.
AF and shutter lag. If you shoot mostly wildlife you will be very
disappointed with the D60. Your percent of hits will be
substandard and you will be photographing the back of a lot of
birds.

Ray Amos
--
Canon EOS digital & film cameras and lenses from 17-500mm. Waiting
patiently for my new Canon 24-105 f/2.8 IS lens.

Click to expand...

Hi Ray,

I'm sort of surprised that you have that much trouble shooting wildlife and birds with the D30. Though the autofocus is decidedly inferior to your 1V, there isn't a significant shutter lag and by using the center focus point and AI servo, many people get excellent in-flight bird and wildlife shots with the D30 and D60. I use the D30, 1D and Kodak DCS-760 and though it's "easier" to get in-flight shots with the better autofocus of my 1D and DCS-760, I have been able to get excellent results with the D30 as well. I suspect practice with differing techniques might help. Actually, if you check around the wildlife forums you will find some really great in-flight shots of birds taken with the D60 and D30...

Best regards,

Lin
--
http://208.56.82.71

there is no enlarge. There is only a crop. A 300 mm lens is still a
300 mm lens on a 1D. It is not 400 mm.

So one big advantage of using a full frame is that, if you subject is in the center 60%, you will get much sharper images with lower quality glass.

Steven

Lin Evans said:

Isaac Sibson said:

Ok, leave aside number of pixels for the moment, because they are
NOT IMPORTANT.

Click to expand...

The pixel count doesn't appear to to be important here - whether or
not is actually is, is part of the question.....

Member said:

Look at it this way.

Put a 300mm lens on a D60, on a 1D and on a 1V, and take a shot
with say a bird in the centre. Print all the shots the SAME SIZE
(regardless of resolution, print them the SAME size).

Now, the picture from the D60 will have the largest bird, and the
picture from the 1V will have the smallest bird. In order to get
the bird from the film image the 1V recorded to the same size as
the image the D60 recorded, you have to crop down the original 35mm
negative to EXACTLY the same size as the D60's sensor, and THEN
enlarge it to the same size as the other prints.

This crop-and-enlarge is done by the smaller sensor in the D60.
There is no interpolation.

Click to expand...

Yes, the crop and enlarge is obviously done by the sensor, or by
the sensor and other electronics in the camera. The crop part is
very simple to explain by optical physics - but explaining the
"enlarge" part doesn't seem to be quite as easy, does it?

Lin
--
http://208.56.82.71

Click to expand...

--
---
My really bad summer pictures:
http://www.pbase.com/snoyes/summer_travels

Steven Noyes said:

there is no enlarge. There is only a crop. A 300 mm lens is still a
300 mm lens on a 1D. It is not 400 mm.

So one big advantage of using a full frame is that, if you subject
is in the center 60%, you will get much sharper images with lower
quality glass.

Steven

Click to expand...

Hi Steven,

Yes, of course the lenses native focal length can't change (effectively it would be 390mm with the 1D) but the "effective" focal length does indeed change and the image is definitely larger and of the same apparent size as if it had been shot with a 390mm focal length lens. The true DOF and other factors are identical to what they would be on a 35mm film camera, but there is, I assure you, an enlargement.

Lin

http://208.56.82.71

Steven,

Here's a sample of the differences in a shot taken from a tripod with the 100-400IS at 400mm with my D30 and 1D. The D30 shot was interpolated to the same dimensions as the 1D, then both were reduced to about 800x533 for display.

The crop factor differences are fairly easy to see since the D30 subjects are larger and the FOV diminished accordingly.

If I had time, I would shoot a comparison slide and scan to identical frame dimensions, but I've done it in the past for examples several years ago when the D30 was first released.

There is obviously a crop from the 1D and D30, but the enlargement of the D30 image shows the differences. The crop factor is a prime reason why sports photographers really love the digitals like the Nikon D1H and EOS-1D. They get the greater DOF of the shorter focal length with the magnification of a longer lens.

Regardless of how the image eventually gets to the enlargement phase (software interpolation, digital crop, etc.) the end result is a closeup from a shorter focal length lens of less weight and cost. What I've been trying to show is that there is no clear understanding of how this actually occurs.

I do thoroughly understand interpolation, having written many "tomes" about the process. What I don't understand, and what I believe this is true of most who wrestle with this issue, is where the enlargement actually comes from. Obviously, it's not from pixel count since the D30 and D60 share a common sized sensor with vastly different pixel counts and with identical crop factors. I've always suspected that Canon and other digicam manufacturers use a similar firmware algorithm that which they use for digital zooms in the consumer cameras to get the original cropped image back to full frame size. This would explain the 1.5x for Nikons (corresponding to the frame size difference from 35mm in each case), the 1.6x for D60 and D30's, the 1.3x for the 1D and my DCS-760.

Perhaps I'm way off base with this reasoning - and I'm open to a clear, concise explanation of where the enlargement happens and how. But it's not possible to deny that it occurs.

Here's the images:





Lin

http://208.56.82.71

Lin:

Your notion of how a Bayer sensor works is way off... The only interpolation that goes on is the RGBG Bayer pattern to generate 4 RGB pixels per photosite. That's all! This yields an output file with the same number of photosites (and yes, I 'm not talking about the unused surrounding frame).

John

Lin Evans said:

Bob Jarman said:

I like most of you who have a 1d am in love with the camera. I
get great images, can usually get large prints, and it is simply a
joy to handle.

I shoot mainly nature and wildlife and now that the 1ds has been
announced, I am interested in your thoughts about upgrading. The
main reason we would want to upgrade is the additional resolution
which would alllow for more cropping. (albiet that's based on
sketchy heresay information). That would be a big plus in
wildlife shooting. I don't do a lot of Landscapes, so the full
frame isn't really that important to me.

Click to expand...

Hi Bob,
I really wouldn't consider the D1s an "upgrade," but rather a
supplementary camera. I definitely intend to purchase one as soon
as they become available, but the "s" is actually more of a
"studio" moniker than anything else.

As for the resolution issues, the 1Ds is nearly double the D60's
true electronic resolution. The crop factor with the D60 (1.6) is
actually a combination of "crop" plus electronic interpolation to
full frame size and original pixel count, though it's not commonly
thought of in those terms. Though it may seem that since the D60
has its full complement of pixels after the crop and enlarge, the
losses in absolute resolution (as opposed to electronic resolution)
are as if the image were taken at full frame, then cropped and
interpolated back to the original pixel count. This does take some
toll on the sharpness and though the D60, D30, 1D, etc., images are
extremely good, you can bet that the full frame 1Ds images will be
that much better.

What you gain from the D60 in terms of telephoto is that you don't
need to actually crop and enlarge in software to get to the
effective focal length advantage of the 60 percent sized sensor.
It's simply done for you and you have no choice. It's a big hit
with sports shooters since they get the advantage of a longer focal
length lens in terms of magnification, but with the depth of field
of the shorter lens. Of course you will loose that with the 1Ds,
but you will be starting with a full frame of a much greater pixel
count and the results of crop and enlarge should be that much
better than with the D60. I guess the bottom line here is that
there really is no "free lunch" with the crop factor, even though
it's a great time saver in terms of workflow.

So in answer to the original question, the quality of enlargements
from the 1Ds should considerably exceed that of the D60 or similar
instruments (D100, DCS-760, etc.).

Best regards,

Lin
--
http://208.56.82.71

Click to expand...

John Davis said:

Lin:

Your notion of how a Bayer sensor works is way off... The only
interpolation that goes on is the RGBG Bayer pattern to generate 4
RGB pixels per photosite. That's all! This yields an output file
with the same number of photosites (and yes, I 'm not talking about
the unused surrounding frame).

John

Click to expand...

John

I know how the sensor works - read the response more carefully...

Lin
http://208.56.82.71

you are talking rubbish ........
you don't understand resolution
Their is no "crop factor"
The pixels are eirther there or not there ...forget the sensor size.

Lin Evans said:

Member said:

The D60 wins the d60vs1ds argument (not considering costs) only if
you are constantly working with the highest zooms. If you plan to
purchase a 400 2.8 for the D60 when really needing a 600+ zoom for
full frame/film body, then you save a lot of money. After all,
with the 1ds, you'd be cropping most shots when using the 400/2.8.

BTW, if you take the predicted 1ds resolution ( 4100x2700) and
apply the 1.6 crop factor, you end up with a resolution that is a
bit under what the d60 has (4100/1.6= 2500, 2700/1.6= 1700). So in
the case that you do a lot of cropping, the d60 wins both in cost
and ultimate resolution.

Click to expand...

But you're forgetting how the D60 got this 1.6x boost in the first
place. As I said in the other post, there is no free lunch here.
Digital crops are a crop and enlarge to frame - not just a crop.
Because of this, interpolation was necessarily applied to get the
magnification factor from the telephoto lenses true focal length .
Some image quality is always sacrificed with interpolation. Not a
lot in this case, but when dealing with serious enlargements, there
will be a significant difference between the full frame sensor and
the D60's 1.6 crop factor. This difference will be apparent in
enlargements and sharpness.

Lin

http://208.56.82.71

Click to expand...

Damm straight he got it right ....

glowluzid said:

Jason Hutchinson said:

I assure you packing more photosites onto an imager the way the D60
does is not "interpolation." It's absolutely positively not
interpolation.

Click to expand...

Click to expand...

Paul Pope said:

you are talking rubbish ........
you don't understand resolution
Their is no "crop factor"
The pixels are eirther there or not there ...forget the sensor size.

Click to expand...

I see Paul,

And you have it all figured out, so why don't you enlighten me? I suspect I understand resolution as well as you, but I'm willing to listen - inform me...

If there is no "crop factor" then why does my D30 image taken with a 400mm prime produce a larger image than my 1D at the same focal length?

Lin

--
http://208.56.82.71

Because the D30 is 3.1 mp and the 1D is 4.15 mp ....doh!

Lin Evans said:

Paul Pope said:

you are talking rubbish ........
you don't understand resolution
Their is no "crop factor"
The pixels are eirther there or not there ...forget the sensor size.

Click to expand...

I see Paul,
And you have it all figured out, so why don't you enlighten me? I
suspect I understand resolution as well as you, but I'm willing to
listen - inform me...

If there is no "crop factor" then why does my D30 image taken with
a 400mm prime produce a larger image than my 1D at the same focal
length?

Lin

--
http://208.56.82.71

Click to expand...

Paul Pope said:

Because the D30 is 3.1 mp and the 1D is 4.15 mp ....doh!

Click to expand...

So to follow your logic, the lower the pixel count on a sensor the larger the image it produces? Then maybe you can explain how cameras with equal pixel counts (a DCS-760 and a Canon D60) produce different effective focal lengths with the same lens?

Are you sure we are even talking about the same world? You've jumped into the middle of a discussion aparently without reading what came before. The issue is how the 1.6x factor (you can choose to call it whatever you wish) produces the same image frame with a 400mm telephoto lens as a 35mm camera attached to a lens with a 640mm focal length. What is the precise mechanism which causes this?

Lin

http://208.56.82.71

You have got it all backwards.

There are the same amount of pixels (D60 and 760), but on the D60 they are smaller (to fit the smaller sensor). This is why they have the same resolution.

Put a 200mm on both of the cameras. Now imagine youself looking through the sensor (or a hole as big as the sensor). When looking through the D60, you will only be able to see a small part of the buildings, but you will be able to see them with all 6MP.

When looking through the 760, you will be able to see more thank't to the "bigger hole", but still with the same 6MP.

Therefore the D60 can creat a "bigger" (i.e. closer to the buildings) image with the same resolution and NO interpolation, than the 760.

I hope I cleared this dilemma out for ya!
Mathias

Lin Evans said:

Jason Hutchinson said:

I assure you packing more photosites onto an imager the way the D60
does is not "interpolation." It's absolutely positively not
interpolation.

Interpolation is the process of defining data at an unknown data
point between two other known data points. The D60 does not do
this. As long as your lens is reasonable, the D60 takes an ACTUAL
reading at the pixel.

Yes, there is a compromise in doing this, and it's precisely this:

If the D60 had a "1.4x crop", the effect would be the image
degredation of a 1.4x telephoto, minus the actual degredation of
the glass in the adapter.

Jason

Click to expand...

Let me see if I understand you correctly. You are saying that
placing more sampling sites within the same sized sensor results in
electronic "magnification" or a larger image than the true optical
focal length of the lens, with the downside being image
degredation to the extent of the crop factor? If that's correct,
then why is the crop factor identical with the D30 at half the
photosite count with the identical sensor size as the D60?

Lin

Member said:

Lin Evans said:

No Stephen,
That's not the way it works. First, it doesn't "interpolate down
from a full pixel count to effective pixels." The sensor has more
photosites than it uses to frame the image. These additional
photosites are used for other purposes.

The notion that an electronic "crop" is simply "cutting off or not
recording what is beyond the crop on what would be a full frame
sensor" is not accurate. If this were the case, then there would be
no advantage at all for telephoto lenses when using a camera with a
crop factor. You would simply have an image which would be exactly
like what you would get with a 35mm film frame with the center 60
percent cropped out. What you get, is the 35mm film frame with the
center 60 percent cropped out and enlarged back to the full frame
which is where the 60 percent magnification comes from.

Here's a simple test for you. Set a 400mm lens on a tripod and
attach your D60. Frame a subject and take a shot. Mount a DCS-760
beside it on a 400mm lens and frame an identical shot. Compare the
images. The D60's image with a 1.6x crop factor and the DCS-760's
image with the 1.3x crop factor will have equivalent electronic
resolution (pixel count), but any subjects in the D60's frame will
appear larger and closer and there will be less "real-estate"
covered in the frame. That with for all practical purposes
identical lenses and identical pixel counts. How did the D60's
image get larger than the DCS-760's image? It was interpolated.
It's impossible for the optics to create a longer focal length for
one than for the other. The focal lengths have not changed, but the
image has. "Electronic" enlargement is done by interpolation -
therefore the D60's image was interpolated to the .6 crop factor
and the DCS-760's image was interpolated to the .3 crop factor.

Unfortunately, there is no "free lunch" with digital crop factors.
You can't get the magnification with a fixed focal length without
interpolation.

Lin

--
http://208.56.82.71

Click to expand...

Click to expand...

--
http://208.56.82.71

Click to expand...

Smaller physical sensor cause field of capture to be limited to the area available.
The smaller the sensor on a 35mm system the larger the "crop" effect.
There is NO interpolation.

I'll "jump" in when i see someone trying to pass off something so blantently false as fact.

And the DCS-760 and a D60 do not have equal pixel counts ...

Lin Evans said:

Paul Pope said:

Because the D30 is 3.1 mp and the 1D is 4.15 mp ....doh!

Click to expand...

So to follow your logic, the lower the pixel count on a sensor the
larger the image it produces? Then maybe you can explain how
cameras with equal pixel counts (a DCS-760 and a Canon D60) produce
different effective focal lengths with the same lens?

Are you sure we are even talking about the same world? You've
jumped into the middle of a discussion aparently without reading
what came before. The issue is how the 1.6x factor (you can choose
to call it whatever you wish) produces the same image frame with a
400mm telephoto lens as a 35mm camera attached to a lens with a
640mm focal length. What is the precise mechanism which causes this?

Lin

http://208.56.82.71

Click to expand...

you are absolutely correct. As also I said previously (look above in forum). I did however, for the sake of the argument say that the D60 and 760 has the same pixelcount.
Mathias

Paul Pope said:

And the DCS-760 and a D60 do not have equal pixel counts ...

Lin Evans said:

Paul Pope said:

Because the D30 is 3.1 mp and the 1D is 4.15 mp ....doh!

Click to expand...

So to follow your logic, the lower the pixel count on a sensor the
larger the image it produces? Then maybe you can explain how
cameras with equal pixel counts (a DCS-760 and a Canon D60) produce
different effective focal lengths with the same lens?

Are you sure we are even talking about the same world? You've
jumped into the middle of a discussion aparently without reading
what came before. The issue is how the 1.6x factor (you can choose
to call it whatever you wish) produces the same image frame with a
400mm telephoto lens as a 35mm camera attached to a lens with a
640mm focal length. What is the precise mechanism which causes this?

Lin

http://208.56.82.71

Click to expand...

--

Click to expand...

I must agree with Mathias! It can't be explained much simpler than this.

Patrick

mathias said:

You have got it all backwards.
There are the same amount of pixels (D60 and 760), but on the D60
they are smaller (to fit the smaller sensor). This is why they have
the same resolution.
Put a 200mm on both of the cameras. Now imagine youself looking
through the sensor (or a hole as big as the sensor). When looking
through the D60, you will only be able to see a small part of the
buildings, but you will be able to see them with all 6MP.
When looking through the 760, you will be able to see more thank't
to the "bigger hole", but still with the same 6MP.
Therefore the D60 can creat a "bigger" (i.e. closer to the
buildings) image with the same resolution and NO interpolation,
than the 760.

Click to expand...

Come on, Lin. There really is no interpolation involved in this 'crop' story. Really, really, there isn't

Patrick

mathias said:

Paul Pope said:

And the DCS-760 and a D60 do not have equal pixel counts ...

Lin Evans said:

Paul Pope said:

Because the D30 is 3.1 mp and the 1D is 4.15 mp ....doh!

Click to expand...

So to follow your logic, the lower the pixel count on a sensor the
larger the image it produces? Then maybe you can explain how
cameras with equal pixel counts (a DCS-760 and a Canon D60) produce
different effective focal lengths with the same lens?

Are you sure we are even talking about the same world? You've
jumped into the middle of a discussion aparently without reading
what came before. The issue is how the 1.6x factor (you can choose
to call it whatever you wish) produces the same image frame with a
400mm telephoto lens as a 35mm camera attached to a lens with a
640mm focal length. What is the precise mechanism which causes this?

Lin

http://208.56.82.71

Click to expand...

--

Click to expand...

Click to expand...