High resolution photos from a digital camera.

Philip Bailey writes:

Bruce Chang wrote:
"Scott W" wrote in message

One of the things I want to try sometime it to take a lot of photos
of a very busy road and by combining the right photos together
remove all the cars but leave the people on the sidewalks, I think
it might make for an interesting photo. This is how one of the
shots for the highway scene in Matrix Reloaded was taken. They
taped the highway and stitched parts together to make it look
vacant.

I've seen this sort of thing done with a stack of neutral density
filters... SERIOUSLY long exposure. Corrected for reciprosity failure
by trying multiple exposures. Basically, nothing remained in the field
of view long enough to be exposed except "landmarks". It was a picture
of a California freeway. NO cars.

Couldn't the same be achieved by averaging many short exposures?

--
Måns Rullgård

Måns Rullgård wrote:

Philip Bailey writes:

Bruce Chang wrote:

"Scott W" wrote in message

One of the things I want to try sometime it to take a lot of photos
of a very busy road and by combining the right photos together
remove all the cars but leave the people on the sidewalks, I think
it might make for an interesting photo. This is how one of the
shots for the highway scene in Matrix Reloaded was taken. They
taped the highway and stitched parts together to make it look
vacant.

I've seen this sort of thing done with a stack of neutral density
filters... SERIOUSLY long exposure. Corrected for reciprosity failure
by trying multiple exposures. Basically, nothing remained in the field
of view long enough to be exposed except "landmarks". It was a picture
of a California freeway. NO cars.

Couldn't the same be achieved by averaging many short exposures?


I'm kind of a beginner at all this... I'm probably not qualified to
answer that question. The shot I saw was made with film. I'm not
sure if many short exposures would lend itself to EITHER film or
digital. I don't know how the noise adds (or doesn't) for the
various techniques. I'd certainly like to hear some opinions!

I guess my point was, lots of ND filters SEEMS to me to be easier
than trying to "stitch" frames together "around" the moving
objects...

On 2005-11-06 11:08:56 -0500, "Scott W" said:


Dave Cohen wrote:

I must be losing it in my old age. So I'm standing alongside this guy who is
carefully composing an image of this beautiful old church and is using the
swing and tilt feature of his 4x5 to include the steeple. Now using the
technique described in this post, what exactly do I do, get close to the
subject and take a shot of a few bricks (or stones at a time), climb up a
ladder to shoot the steeple, then stitch the whole thing together.
Since I'm using dial-up, I can't view the sample. I'm confident it's very
good and I have stitched landscape views myself, so I'm both aware of and
certainly not opposed to stitching as a useful technique, I just think the
rational of this post is missing something.
Dave Cohen

You would set up your camera at the same spot the guy shooting the 4 x
5 view camera would. With the 4 x 5 camera he can get the whole photo
in one shoot, with the digital it would take a number of shoot, the
camera stays in the same spot but is aimed at different parts of the
church. The software can stitch the photo as if a view camera was
being used, at least the shift part which is what corrects for the
perspective.

Most people do not understand how a shifting lens works, basically the
camera lens that is used with a view camera has a much larger field of
view then the film, if you want to shoot something like a church you
point the camera straight at the horizon and then shift the lens up or
the film down. You could get the same effect by using a 8 x 10 sheet
of film in the 4 x 5 camera, not shifting the lens and cropping the
photo.

I think using a view camera is a great way to get a fantastic photo and
am not arguing against it. What I am trying to say is that there is a
lot more that you can do with a digital camera then many people are
aware of.

Scott

True! But, every time I see a well made 24x from a 4x5, it blows me
away... far surpasses anything digital I have seen printed. Find a
gallery of accomplished photographers using silver media... I wish I
was that good. A 24x is only 6 times enlargement from the 4x5
negative. Enough said. On the computer monitor there is no
difference. On a wall, well there is still a difference.


--
Jim

On $DATE , Jim wrote:

On 2005-11-06 11:08:56 -0500, "Scott W"
said:


Dave Cohen wrote:

I must be losing it in my old age. So I'm standing alongside
this guy who is carefully composing an image of this beautiful
old church and is using the swing and tilt feature of his 4x5
to include the steeple. Now using the technique described in
this post, what exactly do I do, get close to the subject and
take a shot of a few bricks (or stones at a time), climb up a
ladder to shoot the steeple, then stitch the whole thing
together. Since I'm using dial-up, I can't view the sample. I'm
confident it's very good and I have stitched landscape views
myself, so I'm both aware of and certainly not opposed to
stitching as a useful technique, I just think the rational of
this post is missing something. Dave Cohen

You would set up your camera at the same spot the guy shooting
the 4 x
5 view camera would. With the 4 x 5 camera he can get the
whole photo in one shoot, with the digital it would take a
number of shoot, the camera stays in the same spot but is aimed
at different parts of the
church. The software can stitch the photo as if a view camera
was being used, at least the shift part which is what corrects
for the perspective.

Most people do not understand how a shifting lens works,
basically the camera lens that is used with a view camera has a
much larger field of view then the film, if you want to shoot
something like a church you point the camera straight at the
horizon and then shift the lens up or
the film down. You could get the same effect by using a 8 x 10
sheet of film in the 4 x 5 camera, not shifting the lens and
cropping the photo.

I think using a view camera is a great way to get a fantastic
photo and
am not arguing against it. What I am trying to say is that
there is a lot more that you can do with a digital camera then
many people are aware of.

Scott

True! But, every time I see a well made 24x from a 4x5, it
blows me
away... far surpasses anything digital I have seen printed. Find
a gallery of accomplished photographers using silver media... I
wish I
was that good. A 24x is only 6 times enlargement from the 4x5
negative. Enough said. On the computer monitor there is no
difference. On a wall, well there is still a difference.

I'm sorry that I missed the beginning of this thread, and may be
recovering old ground, but you can obtain digital backs for some
quality MF cameras, but those megapixels come at a very high price
and you'd better have a use for them if you expect to recoup your
investment.

--
Regards,
Fred.
(Please remove FFFf from my email address to reply, if by email)

On Sun, 06 Nov 2005 08:41:11 -0700, "Richard H." wrote:

Scott W wrote:
The photo is of course stitched, it is a way to get a lot of pixels
using a digital camera. This photos does not even come close to what
some others have done, I have seen a 2.5 GP photo. But the high
resolution stitched photos that I have seen to date have been of pretty
static scenes, I wanted something with a bit of a dynamic feel to it,
something where people are doing things in the photo.

Interesting test - what did you use for the stitching? How much overlap
was there between the shots? Did you use a rigging to take the
photos, or was it handheld?

Buried on my list of to-dos, I'd like to experiment with very
large-scale stitching, with a goal in the 1000MP range (wall-sized
high-res print).

I expected to do a static scene, and probably make a rig to pan & scan
the ~400 images. This could even fit on one memory card, but flash
recording time will be the limiting factor for a live scene - capturing
a single scene could easily take 2 minutes. Using a bank of several
cameras might be an (expensive) idea, if the colors / exposures can be
balanced.

Your example is encouraging; maybe a live scene is even viable if the
images can be captured quickly enough. Perhaps by rapid-firing the live
areas and methodically collecting the static portions, then compiling
the result - what was your technique?.

Cheers,
Richard


Having only followed this half heartedly, this may be totally
redundant and/or unsuited, but here's my own 2 bits worth.

I've used my digital cameras for some panoramas, nothing spectacular.
The software wasn't anything special either, but just wanted to assure
I've tried it.

The real point of this subject is using a modest digital camera (like
3 mp or so) to take multiple shots and blend them into something with
100s of MP - but the OBJECT is to increase the depth of detail. For
this to be done, the entire image captured with the modest camera must
be of a very tiny portion of the overall subject matter. This seems
to demand some special lens that covers only a limited field of the
total image. Either I've missed how this is proposed to be
accomplished, or I've slept thru them. So, I propose a method that
makes sense to me.

Much of my nature photography is concentrated on what is often
referred to as "digiscoping". We use high quality spotting scopes
adapted to our digital cameras. Compared to the angle subtended by a
normal camera lens, set at equivalent to 50mm on a 35mm camera, which
is usually in the 70 degree range - these spotting scope equipped
cameras at max. optical zoom subtend very narrow angles. Like maybe a
fraction of a degree, up to maybe 2 or 3 degrees. Now if you took
this rig and carefully moved it across the desired subject in tiny
steps of only a degree or so, then stitched the resulting images, you
could get both the detail and magnitude of final image being
discussed. Of course the careful movement to cover a large
rectangular area must be in both the horizontal and vertical
directions.

I believe this is how many of the scientific and astronomical photos
are assembled, incidentally. But with far more sophisticated gear.

Sorry if this is irrelevant.

Olin McDaniel

Olin K. McDaniel wrote:


Having only followed this half heartedly, this may be totally
redundant and/or unsuited, but here's my own 2 bits worth.

I've used my digital cameras for some panoramas, nothing spectacular.
The software wasn't anything special either, but just wanted to assure
I've tried it.

The real point of this subject is using a modest digital camera (like
3 mp or so) to take multiple shots and blend them into something with
100s of MP - but the OBJECT is to increase the depth of detail. For
this to be done, the entire image captured with the modest camera must
be of a very tiny portion of the overall subject matter. This seems
to demand some special lens that covers only a limited field of the
total image.

This is simply a long lens, nothing really special about it.

The camera motion is done with a special panoramic tripod head.

Scott

http://www.tawbaware.com/maxlyons/gigapixel.htm
http://www.outbackphoto.com/workshop...icTutorial.pdf

"Scott W" wrote in message
ups.com...

Olin K. McDaniel wrote:
[...] For
this to be done, the entire image captured with the modest camera must
be of a very tiny portion of the overall subject matter. This seems
to demand some special lens that covers only a limited field of the
total image.

This is simply a long lens, nothing really special about it.

In fact, a very long lens presents a couple problems. First, depth-of-field
is remarkably shallow so that if the complete object in question is not at
infinity (or ideal hyperfocal), then it is refocused for some part(s) of the
image, the focal length, thus the image segment(s) change size. Second, for
a telephoto lens, the nodal point can be VERY far in front of the lens. Of
course, the later can be accomodated using an offset camera mount.

That said, somewhere in the literature (I may have it in the lab) is an
outstanding study of 'infinite depth of field' algorythms and real-world
applications made as part of a dissertation. It is obscure, and brilliant. I
am surprised it hasn't seen applications outside of research efforts. I will
look for it.

Lorem Ipsum wrote:
http://www.tawbaware.com/maxlyons/gigapixel.htm
http://www.outbackphoto.com/workshop...icTutorial.pdf

Both good referances and ones that I have used.

Lorem Ipsum wrote:
"Scott W" wrote in message
ups.com...

Olin K. McDaniel wrote:
[...] For
this to be done, the entire image captured with the modest camera must
be of a very tiny portion of the overall subject matter. This seems
to demand some special lens that covers only a limited field of the
total image.

This is simply a long lens, nothing really special about it.

In fact, a very long lens presents a couple problems. First, depth-of-field
is remarkably shallow so that if the complete object in question is not at
infinity (or ideal hyperfocal), then it is refocused for some part(s) of the
image, the focal length, thus the image segment(s) change size.
You shoot like you are shooting LF, a pretty large FNumber.

Second, for
a telephoto lens, the nodal point can be VERY far in front of the lens. Of
course, the later can be accomodated using an offset camera mount.

Well the nodal point can't be way in front of the lens, in fact for my
telephoto
the nodal point is way inside the lens. I have made a chart as to
where the
nodal point is on my zoom lens for different focal lengths.

To date I have been leaving the camera in manual focus mode to avoid
some
of the problem with stitching different images that are focused
differently.

I find that getting a photo in the range of 40 MP is like falling off a
log and 100 MP a whole lot of work.

I also have found that taking photos with people in them is not as hard
as you might think.
It takes a bit of adjusting the seams between photos but this is not
hard to do.
I took 16 wide angle panoramic photos at our canoe race this last
weekend, some of them had
so many people in them I was not sure how they would come out, turned
out not to be a problem

Here is one with a lot of people who were on the move, the line to get
food.
http://www.pbase.com/konascott/image/52303782/large
Hit original at the bottom of the photo for more detail, even then it
is half scale of the stitched photo.

Scott