35mm Film - How many "equivalent pixels"?
Dave Leeper
When comparing digital cameras to good old 35mm film, is there an
approximate pixel equivalent for film?
I think I've heard estimates as high as 1 billion (yes billion) pixels
for 35mm film, but that sounds very high. Are there any experts out
there that know the answer?
Part of the answer may depend on how "pixel equivalent" is defined. I
know there is a point in enlarging traditional photographs where the
optics of the camera and enlarger may become the limiting factor, and
there are other cases where the "grain" of the film is the limiting
factor.
I guess I'm really asking: Approximately how many pixels would a
digital camera need to have so that its picture-enlargement limitations
would be comparable to those of a good quality 35mm camera & enlarger
combination?.
Thanks in advance for any responses!
Dave Leeper
Scottsdale, AZ
The Lonely Boy
I have no idea on your question!
However, the pixels of a digital image have a uniform size, says 10um X
10um. The pixels equivalent - the grains in a traditional images have
variable sizes even in the same picture area. It is because nowaday the
film has multilayers of grains which are of different size.
This makes your comparison more difficult to define.
--
Have a nice day and eat as much as you can.
Lonely Boy
http://www.geocities.com/Tokyo/Towers/7939/
ICQ : 1176912
Martin Tom Brown
In article <69ub3o$b...@bgtnsc02.worldnet.att.net>
dle...@mail.com "Dave Leeper" writes:
> When comparing digital cameras to good old 35mm film, is there an
> approximate pixel equivalent for film?
There is another thread on scanning slides to PCD which pretty much
answers this question. PCD base*16 or 3072x2048 is just good enough.
Rough comparison with common slide films at 100-200 ASA.
base*16 is good enough to see most off axis lens abberations clearly.
> I think I've heard estimates as high as 1 billion (yes billion) pixels
> for 35mm film, but that sounds very high. Are there any experts out
> there that know the answer?
It's far too high unless you are talking exotic specialist emulsions.
(most of which are never made into 35mm film anyway)
> Part of the answer may depend on how "pixel equivalent" is defined.
Whichever way you define it there is no point in having more pixels
than are needed to cope with diffraction limited optics so it is
a simple matter of physics to put a bound on it.
A 50mm lens at f8 has a diffraction limited point spread is 10um
(very few lenses are diffraction limited since getting rid of
other abberations and having flat focal plane are more important)
> would be comparable to those of a good quality 35mm camera & enlarger
> combination?.
On the diffraction limited optics at f8 assumption 3500 x 2500.
Regards,
--
Martin Brown <mar...@nezumi.demon.co.uk> __ CIS: 71651,470
Scientific Software Consultancy /^,,)__/
Tony Sleep
In article <69ub3o$b...@bgtnsc02.worldnet.att.net>,
lee...@worldnet.att.net (Dave Leeper) wrote:
> When comparing digital cameras to good old 35mm film, is there an
> approximate pixel equivalent for film?
I've seen two published estimates from Kodak, for the minimum bitmap size
required to capture all the information present in an ISO100 colour neg.
The first was 40Mb, quoted in the British Journal of Photography c.1995.
Since then, they seem to have revised the number downwards, to 35Mb.
Tony Sleep
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
http://www.cix.co.uk/~tsphoto/
Tony Sleep - online portfolio & exhibit
Stephan Piel
The Wall Street Journal had an article last week on digital cameras, in which
they mentioned that the equivalent resolution of 35mm film is about 6 million
pixels.
-sp
ampfax
Alex Nichol wrote:
>
> Dave Leeper wrote:
>
> >When comparing digital cameras to good old 35mm film, is there an
> >approximate pixel equivalent for film?
> used for resolution tests, gave upwards of 100 lines/mm - equals 200
> pixels/mm - so about 35million pixels per frame, or 4800x7200. A
> recent post here, measuring I guess with a more regular film, was
> getting about 40 lines - 80 pixels/mm, much the same as from some top
> end digital cameras he checked. Its getting to the point where lens
> quality is going to be an important consideration
> --
> Alex Nichol
> Bournemouth, U.K.
> al...@dial.pipex.com
Absolutely the right answer,
except that the resolution as a bit higher, up to 150-175 lines/mm.
From all manufactor are data sheets available who inform you about the
reolution of a specific film.
Digital imaging is coming strong but still far to go ...
W.S.
Blake Leyh
> When comparing digital cameras to good old 35mm film, is there an
> approximate pixel equivalent for film?
A practical, useful rule of thumb is:
300dpi of true resolution when printed on a high-quality printer will have
"photographic" resolution to the naked eye.
I have printed a lot of my work on an Iris inkjet at 300dpi and the prints
are indistinguishable from photos. These were high-quality scans from
35mm negatives, not digital camera JPEG output. Applying the 300dpi rule
to the output resolution of an Oly D-500L or equivalent, you see that you
can print a 1024x786 resolution image at a size of about 3.5 x 2.5 inches
to achieve a "photo-quality" output.
Many people in this group are very happy with their D-500L images printed
at 4x6 inches or even 8x10, but I have found though much trial and error
that 300dpi is a very practical rule of thumb for "serious" work.
Blake Leyh
George Ou
I wish a the photo magazines would show more actual samples of blown up
targets.
All these numbers are highly questionalble no matter the source. Digital
cameras are not mature yet but I still think a digital pixel is more pure
than some grains randomly placed. You know how much error you are dealing
with in digital where film (mature) quality is harder to pin down.
Randy Anderson
I posted an article last week where I compared film to several mid-range
digital cameras as tested with an Air Force resolution test chart. Do a
search for "Pentax 645" or "resolution test" and you should be able to
find it. If it has dropped from the newgroup, then search Dejanews.com.
Randy Anderson
John G Dann
Dave Leeper <lee...@worldnet.att.net> wrote:
>When comparing digital cameras to good old 35mm film, is there an
>approximate pixel equivalent for film?
>
I despair of ever getting a good perspective on this question.
People so rarely distinguish between what a 35mm neg is capable of
under reasonably ideal conditions, (ie slow film optimally exposed and
processed, top quality lens and camera, utmost care taken to minimise
camera shake etc.) and what the system actually delivers for most
people most of the time.
I would guess that base*16 or around 2000dpi on the neg must be
getting on for the maximum resolution attainable. But I'd also guess
that even for reasonably serious amateurs, one or other of the
limiting factors (probably usually a combination of several factors)
in the system intervenes and pushes the real resolution down towards
1000-1500 dpi.
For those of us who can't afford top quality professional equipment,
processing etc (which is most people), this means that a digicam
providing say 1.5Mpixels of genuine (ie not interpolated) imaging is
going to give pictures every bit as good as traditional photography.
But I still would love to see some 'real-world' data on resolution in
everyday negatives from good amateurs, rather than ideal/maximum data.
Regards, John Dann
Mac McDougald
In article <34c4d123...@news.u-net.com>, jo...@prodata.u-net.com
says...
> But I still would love to see some 'real-world' data on resolution in
> everyday negatives from good amateurs, rather than ideal/maximum data.
Well, anything beyond the "facts" is art! The "facts" as I explain
them, since I do scanning and output 35mm slides/negs (at a very
reasonable price, I might add) :-) :
100 ISO 35mm film has equiv resolving power of 4096x2731 pixels, which
is also capability of a *true* 4K film recorder...makes about a 32Meg
Tiff file; also, of course, even at the full 4096, it's debatable whether
a digital printer will "equal" photo enlargement; dye sub etc might be
close enough, but certainly not inkjet (although, if it's "good enough",
then it's certainly good enough)...
--
Mac McDougald -- DOOGLE DIGITAL -- Knoxville, TN
doo...@icx.net 423-540-1308 http://www.doogle.com
[film recording|scanning|copystand photography|photo
restoration|presentation graphics|html|wordsmithing]
William Kaukler
There are and have been for decades, standard tests for resolution and
fidelity of films. This is also true of cameras. Film quality IS easy to
precisely pin down. However, the details are too technical for average
users to understand. This goes for ccd's and other digital technologies
too. You may be too young to know about the film wars a couple of decades
back where arguments abounded about which film was better.
Resolution is not everything. Look at the recent 'boot' magazine survey of
digital cameras. Particularly look at the low light images. This shows
just how bad digital cameras are and have a long way to go. This also
illustrates the stupidity of just buying resolution. Fidelity is in the
firmware of the camera.
I haven't heard anyone here talk about the ludicrous attempt at equating ccd
sensors with mos ones. Cheap, high res cameras are being sold with mos
sensors now.
Mos sensors have been around longer than ccds but they were not much good
even then. The ccd is far more sensitive and will be able to produce the
best pictures, and especaily in low light. I would never, ever consider
buying a mos sensor camera until a lot of add-on technology is added to
compensate for the inherent sensitivity of the ccd.
The equivalent pixels of film is difficult to pin down, but there are books
that describe the technical capabilities of film. Others have pointed out
the old rules of thumb here in this newsgroup. One can go by 'looks' and
say 4K pixels X 4K pixels will do it. I side with the old calculations that
hover around 45 million pixels. You won't find anyone selling a digital
camera face that fact. Many claim digital cameras will replace film cameras
now. This is not anything but hype to sell cameras. You will aways have
someone wanting to
crop and enlarge the image. This you can do with film, what do you do with
digital? You go higher in resolution! How high?
Again, we don't know.
I take photographs and scan them in, sometimes at 600 dpi on a 3.5 x 5
picture ( 3000 pixels accross). Then, I can keep just a portion of the
image I wanted. The scanner still isn't limited by the resolution of the
photograph.
Enough for now.
George Ou wrote in message <6a1p6f$i...@sjx-ixn6.ix.netcom.com>...
Martin Tom Brown
In article <6a6i88$2ve$1...@hammer.msfc.nasa.gov>
william...@msfc.nasa.gov "William Kaukler" writes:
> There are and have been for decades, standard tests for resolution and
> fidelity of films. This is also true of cameras. Film quality IS easy to
> precisely pin down. However, the details are too technical
I don't know about US magazines, but some of the better UK ones
do routinely publish MTF data for lenses and films under test.
> Resolution is not everything. Look at the recent 'boot' magazine survey of
> digital cameras.
Resolution, dynamic range and colour balance will do quite nicely...
Most systems have some colour balance "features" - even films.
> Particularly look at the low light images. This shows
> just how bad digital cameras are and have a long way to go.
No it shows that *uncooled* CCD's have thermal noise.
If you want a specifically low light camera you buy one!
You should be aware that conventional film emulsions suffer
terrible problems of reciprocity failure on long exposures too.
Cooling will also help here (some films are better than others).
> Mos sensors have been around longer than ccds but they were not much good
> even then. The ccd is far more sensitive and will be able to produce the
> best pictures, and especaily in low light.
But a lot of snap and point is in good light and very cost sensitive.
> The equivalent pixels of film is difficult to pin down, but there are books
> that describe the technical capabilities of film. Others have pointed out
> the old rules of thumb here in this newsgroup. One can go by 'looks' and
> say 4K pixels X 4K pixels will do it. I side with the old calculations that
> hover around 45 million pixels.
It depends whether you insist on sampling every detail of the grain
structure of the film as well as the image. The gain in information
content for quadrupling the file size is minimal for all but the
very slowest fine grain emulsions used with the best lenses.
PCD 4kx3k is good enough for all but the most demanding uses.
(it generally beats most commercial lenses)
terrapin
>Alex Nichol wrote:
>>
>> Dave Leeper wrote:
>>
>> >When comparing digital cameras to good old 35mm film, is there an
>> >approximate pixel equivalent for film?
>> It depends very much on the type of film used. Really fine grain b&w
>> used for resolution tests, gave upwards of 100 lines/mm - equals 200
>> pixels/mm - so about 35million pixels per frame, or 4800x7200. A
>> recent post here, measuring I guess with a more regular film, was
>> getting about 40 lines - 80 pixels/mm, much the same as from some top
>> end digital cameras he checked. Its getting to the point where lens
>> quality is going to be an important consideration
>> --
>> Alex Nichol
>> Bournemouth, U.K.
>> al...@dial.pipex.com
>
>Absolutely the right answer,
>except that the resolution as a bit higher, up to 150-175 lines/mm.
>From all manufactor are data sheets available who inform you about the
>reolution of a specific film.
>Digital imaging is coming strong but still far to go ...
>
>W.S.
As I recall... and my info is dated, for sure... I remember years ago
data for the very best Nikon lenses was in the neighborhood of
50-60 lines per mm. And that was only in the center of the frame;
at the edges, the figures were 20-30% lower.
rafe b.
terrapin
>When comparing digital cameras to good old 35mm film, is there an
>approximate pixel equivalent for film?
>
>I think I've heard estimates as high as 1 billion (yes billion) pixels
>for 35mm film, but that sounds very high. Are there any experts out
>there that know the answer?
>
>Part of the answer may depend on how "pixel equivalent" is defined. I
>know there is a point in enlarging traditional photographs where the
>optics of the camera and enlarger may become the limiting factor, and
>there are other cases where the "grain" of the film is the limiting
>factor.
>
>I guess I'm really asking: Approximately how many pixels would a
>digital camera need to have so that its picture-enlargement limitations
>would be comparable to those of a good quality 35mm camera & enlarger
>combination?.
>
>Thanks in advance for any responses!
>
>Dave Leeper
>Scottsdale, AZ
>
Here's my 2 bits.
20+ yrs ago I used to spend lots of time in a darkroom, making 8x10
BW prints from 35 mm negs, taken with a Miranda FE. Cheapo, yes.
Later, (1980) I graduated to a Nikon FE.
I have no experience with digital cameras, but recently purchased
a $750 slide/negative scanner (Microtek) and a $270 Epson Printer
(Stylus Color 600.)
I am amazed at the quality of the results.
The scanner is 1950 dpi, tops. I resample/resize in Photoshop to
get a PSD file of 12-18 MB, by reducing the output dpi to 220-250,
and resizing to the final print output size, usually around 10" x 7"
With a few minor qualifications, the prints, at any distance over 6"
or so, are every bit as good as the best darkroom work that I was able
to do... with an Omega B22 and El-Nikkor lens... 20 years ago.
No more Dektol, stop bath and fixer for me....
rafe b.
Bert Janssen
Stephan Piel wrote:
> The Wall Street Journal had an article last week on digital cameras, in which
> they mentioned that the equivalent resolution of 35mm film is about 6 million
> pixels.
>
> -sp
I agree,
when I scan a 100ASA film at 2400 DPI I can (just) see some film-grain. Of coarse
slower film would require more pixels to be the same quality.
Remember though that 6 million pixels at 24bits/colour would produce a filesize
of 18Mb uncompressed.
Regards,
Bert Janssen
===========================================================
Living on Earth may be expensive, but it includes an annual
free trip around the Sun.
===========================================================
bjan...@tref.nl (home)
Homepage: http://leden.tref.nl/bjanssen
Fotoclub Oce: http://www.dsvenlo.nl/~fcoce
Badminton Club Olympia '56: http://www.dsvenlo.nl/~olympia
######################################################################
# #
# This note does not necessarily represent the position of Oce- #
# Technologies B.V. Therefore no liability or responsibility for #
# whatever will be accepted. #
# #
######################################################################
Gene Alan Townsend
Bert Janssen wrote:
>
> Stephan Piel wrote:
>
> > The Wall Street Journal had an article last week on digital cameras, in which
> > they mentioned that the equivalent resolution of 35mm film is about 6 million
> > pixels.
> >
> > -sp
>
> I agree,
>
> when I scan a 100ASA film at 2400 DPI I can (just) see some film-grain. Of coarse
> slower film would require more pixels to be the same quality.
> Remember though that 6 million pixels at 24bits/colour would produce a filesize
> of 18Mb uncompressed.
Bert and Stephen:
The "equivalent resolution" of 35 mm film is indeterminate. Film
records image information in silver halide grains, or their dye images,
which have variable size and position, unlike pixelated detectors that
have fixed pixel sizes and grid arrangements.
In order to fully resolve the information in a 35 mm negative, the
grains need to be fully resolved, since that's where the information
is. This can take extremely high pixelation values, like a 4 by 4 pixel
grid for the very smallest grains.
I strongly disagree with the 6 million pixel equivalence statement in
the Wall Street Journal. 6 million pixels would result in a 6 MB file
NOT 24 MB. The digital cameras use a color mask over the CCD array and
interpolate colors between pixels, where 35 mm film scanners use three
separate CCD arrays. Tri-color CCD arrays have three separate pixels
per output space element.
35 mm film requires on the order of 30-120 MB file size for nearly full
quality scanning with fine grained films and careful photographic
technique when large print sizes (like 16 by 24 inch or so) are desired.
6 MB files would be adequate for amateur snapshots, however, when 4 by 6
prints are anticipated.
Regards,
Gene A. Townsend
David Swager
Most experts claim there is a maximum of between 2000 and 3000 pixels per
inch required to fully resolve the data from film. At 35mm this is
12,268,800 pixels at 3,000 pixels per inch and .96" x 1.42" approximate size
of a 35mm frame. Unfortunately, it is all dependent on the film itself
(grain size and structure), light qualities, exposure duration, apeture used
for expsoure, developing qualities etc.
Gene Alan Townsend wrote in message <34D2F7...@primenet.com>...
Wayne D Johnson
Kodak agrees with David and I tend to believe them on stuff like this.
Wayne
David Swager wrote in message <34d37...@news.siscom.net>...
Gene Alan Townsend
David Swager wrote:
>
> Most experts claim there is a maximum of between 2000 and 3000 pixels per
> inch required to fully resolve the data from film. At 35mm this is
> 12,268,800 pixels at 3,000 pixels per inch and .96" x 1.42" approximate size
> of a 35mm frame.
David:
A 3000 PPI scan of COLOR film would produce a file 3X larger than your
calculation, or about 36 MB. You are correct if you refer to
monochrome.
GAT
Jim Kajpust
Somebody wrote....
>>
>> > The Wall Street Journal had an article last week on digital cameras, in which
>> > they mentioned that the equivalent resolution of 35mm film is about 6 million
>> > pixels.
Just talking "pixels" doesn't really mean anything. You have to relate
it to some size. Either pixels per inch, per square inch, (or as my
graphics prof would say, pixels per pica).
After all, a pixel can really be any size. If you have a mountain to
paint, a pixel 12 inches square could still be "hi-res".
Jim Kajpust - Personal Freedoms - Michigan
http://www.cris.com/~jkajpust
MTCO GROUPS
I read in a photo trade magazine that a 35mm neg was equivalent to about
25megs of data
opm...@aol.com
Warren Young
"MTCO GROUPS" <opm...@aol.com> wrote:
(Required preparation: find salt shaker -- you'll need a few grains.)
I've heard that standard film has about 2000dpi of info in it. I just
did a little eyeballing with a film negative in one hand and a Vernier
caliper in another and found that the frame is about 35mm by 24mm.
Converting those measurements to inches and multiplying, I come up
with about 5.2 megapixels. Now assuming about 36 bits per pixel, I
end up with about 22.4 megabytes.
So, 25 megs sounds reasonable to me.
= Warren -- http://www.cyberport.com/~tangent
=
= Remove the SPAMCATCHER to email. -- Finger me!
Y. Samuel Arai
That also all depends on the film as well...I bet Kodak Royal Gold 25
will hold about 200 megs of information...probably the sharpest film in
production now.
Frank van der Pol
Y. Samuel Arai wrote:
> Warren Young wrote:
> > with about 5.2 megapixels. Now assuming about 36 bits per pixel, I
> > end up with about 22.4 megabytes.
Might be a wrong assumption. Depending on the film used, the 36 bits
might not be enough.
> > So, 25 megs sounds reasonable to me.
....which is possibly the reason a non-compressed PhotoCD scan of a 35
mm photo is around 18 Mbyte (3072 x 2048 x 24 bits)
> That also all depends on the film as well...I bet Kodak Royal Gold 25
> will hold about 200 megs of information...probably the sharpest film in
> production now.
200 Mb at 36 bits means 44.44 megapixels resolution. This means a good
8300 x 5400 pix resolution which results in 237 pixels per mm = 118
lines per mm
On paper this might be theoretically possible in the center of your
image using a rocksteady mount and the right shutterspeed combined with
optimal lense conditions. In real life even your heartbeat gets these
results down.
Now lets bet again...
Frank van der Pol
fr...@mediacenter.demon.nl
remove NOSPAM from reply address
Steve Ballantyne
In article <34E0FB6B...@NOSPAMmediacenter.demon.nl>, Frank van der
Pol <fr...@NOSPAMmediacenter.demon.nl> wrote:
>Y. Samuel Arai wrote:
>
>200 Mb at 36 bits means 44.44 megapixels resolution. This means a good
>8300 x 5400 pix resolution which results in 237 pixels per mm = 118
>lines per mm
>
>On paper this might be theoretically possible in the center of your
>image using a rocksteady mount and the right shutterspeed combined with
>optimal lense conditions. In real life even your heartbeat gets these
>results down.
>
>Now lets bet again...
Also, the film may be capable of a lot, but the resolution of 35mm lenses
can be amazingly low -- like, as low as 0.1mm! Under these circumstances,
dynamic range of the medium is going to have a big effect on perceived
print quality.
By the way, standard 35mm still frames are 36mm x 24mm.
--
Steve Ballantyne
Martin Tom Brown
In article <critic-ya02408000...@news.ihug.co.nz>
cri...@NoChoppedPorkShoulder.ihug.co.nz "Steve Ballantyne" writes:
Even quite modest 35mm lenses usually manage 30 lpm. I can't believe
there are any brands sold that are as bad as you seem to be claiming.
A fair proportion are good for something like 60-100 lpm on axis
at the right aperture setting.
Steve Ballantyne
In article <critic-ya02408000...@news.ihug.co.nz>,
cri...@NoChoppedPorkShoulder.ihug.co.nz (Steve Ballantyne) wrote:
>In article <34E0FB6B...@NOSPAMmediacenter.demon.nl>, Frank van der
>Pol <fr...@NOSPAMmediacenter.demon.nl> wrote:
>
>>Y. Samuel Arai wrote:
>>
>>200 Mb at 36 bits means 44.44 megapixels resolution. This means a good
>>8300 x 5400 pix resolution which results in 237 pixels per mm = 118
>>lines per mm
>>
>>On paper this might be theoretically possible in the center of your
>>image using a rocksteady mount and the right shutterspeed combined with
>>optimal lense conditions. In real life even your heartbeat gets these
>>results down.
>>
>>Now lets bet again...
>
>Also, the film may be capable of a lot, but the resolution of 35mm lenses
>can be amazingly low -- like, as low as 0.1mm!
Clarification: that is a circle of confusion with a diameter of 0.1mm.
Check out the Focal Cyclopaedia of Photography, Cox's Optics etc for more
details.
>Under these circumstances,
>dynamic range of the medium is going to have a big effect on perceived
>print quality.
>
Jim Kajpust
tan...@SPAMCATCHER.cyberport.com (Warren Young) wrote:
>"MTCO GROUPS" <opm...@aol.com> wrote:
>
>>I read in a photo trade magazine that a 35mm neg was equivalent to about
>>25megs of data
As poorly as I take pictures, about 14K oughta do it.
Bruno Barsella
On Fri, 13 Feb 1998, Jim Kajpust wrote:
> tan...@SPAMCATCHER.cyberport.com (Warren Young) wrote:
>
> >"MTCO GROUPS" <opm...@aol.com> wrote:
> >
> >>I read in a photo trade magazine that a 35mm neg was equivalent to about
> >>25megs of data
>
> As poorly as I take pictures, about 14K oughta do it.
Do the following calculation:
a 35 mm slide is about 24x36 mm, so the surface is about 864 mm^2
Suppose (it is reasonable) that your slide has a resolution of 100
lines/mm and multiply by two (theorem of Shannon) You obtain that you must
scan at 5x5 microns = 25 microns^2 per pixel.
Divide the surface by the pixel surface and obtain
33.84 million pixels
Do a truecolor image (3 bytes/pixel) and you have
101.52 MegaBytes/image!!!!
6 images per CD ....... :)
If you use B&W negs divide by three, but remember that good B&W negs have
300-500 lines/mm resolution :) and it is possible to scan at 16 bits per
pixel :)))))
Bruno Barsella - br...@astr17pi.difi.unipi.it
http://astrpi.difi.unipi.it
David Lawton
My God, that's a lot of math. I think they contain around 30-32 mb of data,
according to folks at a local telemedicine program that scans them in at a local
hospital.
Dave
Michael McGuire
0]>
: > tan...@SPAMCATCHER.cyberport.com (Warren Young) wrote:
: >
: > >"MTCO GROUPS" <opm...@aol.com> wrote:
: > >
: > >>I read in a photo trade magazine that a 35mm neg was equivalent to about
: > >>25megs of data
: >
: > As poorly as I take pictures, about 14K oughta do it.
: Do the following calculation:
: a 35 mm slide is about 24x36 mm, so the surface is about 864 mm^2
: Suppose (it is reasonable) that your slide has a resolution of 100
: lines/mm and multiply by two (theorem of Shannon) You obtain that you must
: scan at 5x5 microns = 25 microns^2 per pixel.
: Divide the surface by the pixel surface and obtain
: 33.84 million pixels
: Do a truecolor image (3 bytes/pixel) and you have
: 101.52 MegaBytes/image!!!!
: 6 images per CD ....... :)
: If you use B&W negs divide by three, but remember that good B&W negs have
: 300-500 lines/mm resolution :) and it is possible to scan at 16 bits per
: pixel :)))))
: Bruno Barsella - br...@astr17pi.difi.unipi.it
: http://astrpi.difi.unipi.it
This is an overestimate. The Modulation Transfer Function (MTF) of film is
typically 10% at 100 lines/mm. The MTF of a good quality lens for a 35mm
camera is quite low at 100 lines/mm, probably quite a bit less than 10%. But
assuming it is 10%, the combined MTF of film and lens is the product of the
two--1%. This means that out at this high a frequency given a maximum contrast
input, say a black/white bar target that has a 255 level difference between
black and white, you actually record a 2.55 level difference on the film. The
target is barely distinguishable from a gray smudge of the same mean level. For
practical photography (meaning subjects more interesting than bar targets), 50
lines/mm is a reasonable place to cut off. Giving Shannon his due, you sample
at 100 pixels/mm which gives you 2400 x 3600 pixels x 3 color planes at 8 bits
per plane. This works out to 25.92 Mbytes.
The only b&w films that I know of that reach 300-500 lines/mm are designed for
microfilm use. They are very high in contrast, very slow, and very difficult to
use to make continuous tone images.
Mike
--
Michael McGuire Hewlett Packard Laboratories
email:xmcg...@xhpl.xhp.com P.0. Box 10490 (1501 Page Mill Rd.)
(remove x's from email if not Palo Alto, CA 94303-0971
a spammer)
Phone: (650)-857-5491
************BE SURE TO DOUBLE CLUTCH WHEN YOU PARADIGM SHIFT.**********
Carl Donath
Michael McGuire wrote:
> : Do the following calculation:
> : a 35 mm slide is about 24x36 mm,
> : 101.52 MegaBytes/image!!!!
> This is an overestimate.
Not really. The frame size is roughly equivalent to the largest image printed by
our Lightning II Digital Movie Film Recorder, which takes about 100MB. Smaller
frames typically used in movies (smaller than a normal 35mm still frame) take
20-40MB.
--
Carl Donath http://www.ei.kodak.com/~donath
http://www2.rpa.net/~ctdonath
"He who goes unarmed in paradise had better be sure that
that is where he is." - James Thurber
Michael McGuire
0]>
: > : Do the following calculation:
: > : a 35 mm slide is about 24x36 mm,
: <snip>
: > : 101.52 MegaBytes/image!!!!
: > This is an overestimate.
: Not really. The frame size is roughly equivalent to the largest image printed by
: our Lightning II Digital Movie Film Recorder, which takes about 100MB. Smaller
: frames typically used in movies (smaller than a normal 35mm still frame) take
: 20-40MB.
: --
: Carl Donath http://www.ei.kodak.com/~donath
: http://www2.rpa.net/~ctdonath
We are discussing _real_ films used in _real_ cameras with _real_ lenses by
_real_ people taking _real_ pictures to set a practical upper limit on the
pixel equivalent of a 35mm frame. The microlithography techniques used in
chipmaking now put down 0.1 micron lines which amounts to 5000 line(pairs)/mm.
A 24 bit scan of a 35mm frame of this would amount to 2.6 x 10^5 MB. However
the relationship of the device that writes these lines to conventional cameras
is pretty far fetched. Your film recorder is not quite as far fetched but it's
still far.
je...@tiny.net
:>: Not really. The frame size is roughly equivalent to the largest image printed by
:>: our Lightning II Digital Movie Film Recorder, which takes about 100MB. Smaller
:>: frames typically used in movies (smaller than a normal 35mm still frame) take
:>: 20-40MB.
:>
:>: --
:>: Carl Donath http://www.ei.kodak.com/~donath
:>: http://www2.rpa.net/~ctdonath
:>We are discussing _real_ films used in _real_ cameras with _real_ lenses by
:>_real_ people taking _real_ pictures to set a practical upper limit on the
:>pixel equivalent of a 35mm frame.:>
:>Michael McGuire Hewlett Packard Laboratories
:>email:xmcg...@xhpl.xhp.com P.0. Box 10490 (1501 Page Mill Rd.)
:> (remove x's from email if not Palo Alto, CA 94303-0971
:> a spammer)
Really.
Keep it at the lowest common denominator.
Any image over 2k pixels contains unnecessary information and is harmful
to Rec.
Rec says: Quality is elitist. Education is dangerous. Attempt neither.
fwiw. two nearby service bureaus have both recommended file sizes of
25-30megs for Real People wishing 35mm film recording. Others may
choose...
unreal
jehnk
.