4000dpi 35mm vs. 1200dpi LF

Some guy is claiming that a 4000dpi 35mm with a dedicated film scanner
will beat a 1200dpi 4x5" scan with an Epson 4990. I don't believe it.
Anybody tried this?

Talk about comparing apples and oranges.

The 4000dpi is likely from a dedicated film scanner. You aren't saying if
you're comparing your Epson 4x5 scans (of transparencies) to his film
scanner scans of negative film.

It's somewhat likely, though not guaranteed, that the dedicated film scanner
will have a higher Dmax, meaning it will dig more details out of shadows
than the flatbed scanner. Also the dedicated scanner is not going to have
any Newton ring issues that you may or may not have when one scans 4x5 on a
flatbed scanner. Lastly the many dedicated scanners have features like
digital ICE to more easily deal with dust and scratches. Another feature
that my film scanner has is automated multiple pass scanning of the same
image to allow for averaging to deal with sensor noise is the inky dark
portions of the slides or blown out portions of negatives.

So one could rightly argue that for things like image quality and
flexibility a 4000 DPI 35mm scan is better, but in reality its not the 4000
DPI that makes it better; this the bells and whistles of a dedicated film
scanner that make the difference.

But what about raw pixels?

My Nikon Coolscan 5000 captures 3946x5782 pixels or 22.8Mpixel. With slides
the edges of this area is cardboard. If I scale my 4x5 scans to 1200 dpi I
get a 4534x5639 or 25.6Mpixel. Here too there is a portion around the out
side that is all black, or has clip marks and will need to be cropped out.
But start with the raw scan dimensions and I'd be hard pressed to say 22.8
is much better than 25.6. Now if the 35mm scan is either a negative or a
35mm slide removed from its mount, then almost all the 22.8Mp will be usual
where the 4x5 will need to be cropped. Still I highly doubt that the even a
22.8Mp would win hands down over a cropped 4x5 which may only have 21Mp or
so.

However there are more ways this is an apples-oranges comparison. A
dedicated film scanner may come with a holder the will hold the 35mm frame
nearly flat. A 4x5 sitting of a flat bed will first need to be suspended
above the glass to avoid Newton rings. The holder that do this are
suspending a rather large sheet of film, and there is ample chance that the
4x5 film will not be perfectly flat when it is scanned. So *perhaps* the
dedicated film scanner will not loose as much sharpness from the original
because it holds the film flatter. However the flip side of this argument
is that flatbed scanners have a truly remarkable depth or field and the
ability to sharply scan images of objects well off the glass.

So there is that perception thing. At 4000dpi the dedicated film scanner is
starting to capture the grain structure of the film, while at 1200 the
flatbed is not. Since sharp grain gives the appearance of a sharp image,
one could say that the 4000dpi scan of the 35mm *looks* crisper. One the
other side one might argue that the smooth textures of a 1200dpi 4x5 is
better as no grain appears.

Lastly if one is comparing 4x5 transparencies to 35mm film, then the 35mm
film has superior latitude for capturing shadow and highlight detail. One
could level this particular advantage or even reverse it. My point is if
you and your friend are comparing scans, the original film image is also
factor.

Personally I'm going to take a 2048dpi of 4x5 over 4000dpi of 35mm. Also
some flatbeds are up to 4800dpi optical. Even still the features of the
dedicated film scanner could make the final desired result easier to obtain.

Todd
--
See fine art photography at: www.konabear.com
"Mike" wrote in message
news
Some guy is claiming that a 4000dpi 35mm with a dedicated film scanner
will beat a 1200dpi 4x5" scan with an Epson 4990. I don't believe it.
Anybody tried this?

In article ,
says...
Some guy is claiming that a 4000dpi 35mm with a dedicated film scanner
will beat a 1200dpi 4x5" scan with an Epson 4990. I don't believe it.
Anybody tried this?

One factor to add to the confusion is that generally 35mm lenses
have a higher resolving power than large format lens when the overall
process (including the film) is considered.
So the degree of magnification desired needs to be considered.
I find my 6x7 scans on an Epson 4870 at 8x are about as "sharp"
as my 35mm scans on a Minolta 5400 at 16x. Part of this is due
to the resolution of the Minolta being about twice as high.
The grain of the larger originals is still lower, however.

--
Robert D Feinman
Landscapes, Cityscapes and Panoramic Photographs
http://robertdfeinman.com
mail:

Mike wrote:

Some guy is claiming that a 4000dpi 35mm with a dedicated film scanner
will beat a 1200dpi 4x5" scan with an Epson 4990. I don't believe it.
Anybody tried this?

So using a 300 ppi file for printing, the 35 mm would be just over 12" by
18" while the 4" by 5" would give 13" by 20". Not exactly even. This is
still slightly over 3x for the 4" by 5", and more than 12x for the 35 mm.
While there are some nice dedicated film scanners, like the Imacons, I
still think about 10x is a more reasonable limit for 35 mm.

However, an Epson is not a high end scanner, and could be much more
limited. That 1200 dpi figure should easily within reach of the optics
for a true resolution from the 4990, even using that test from May/June
View Camera as a guide.

I think where the dedicated film scanner (some, but not all) would likely
beat out the Epson would be in the Dmax realm. An image with more shadow
details might turn out to be a better image file from a dedicated 35 mm
film scanner.

Now the other question is what does he mean by "beat"? If it was colour
or shadow detail, then he might be correct. Also, some dedicated film
scanners capable of 4000 dpi cost more than an Epson 4990, but you get
what you pay for . . . . .

Ciao!

Gordon Moat
A G Studio
http://www.allgstudio.com

"Mike" wrote in message
news
Some guy is claiming that a 4000dpi 35mm with a dedicated film scanner
will beat a 1200dpi 4x5" scan with an Epson 4990. I don't believe it.
Anybody tried this?


I recently did something similar. I shot an old family marriage certificate
using my Nikon D-70 and F-100 with a 60mm Nikon macro, my 4x5 with a
Rodenstock 150mm Sironar-S, and had a friend of mine who owns a Lab shoot
the item with his 4x5 and scan it with his Creo Scitex flatbed. I used Kodak
E100G and a Nikon Coolscan 4000 and an Epson 4990 for 35mm and 4x5
respectively, all in 48 bit mode. When the 35mm film scan was reduced to the
same pixel size as the digital "35", the digital won, hands down.

Only using my monitor as a reference, I liked the 4990 scan at 1200dpi
better than the 35mm at 4000dpi. - pretty subjective. Film grain was the
main issue. For a cheap desktop flatbed, the 4990 does a terrific job on 4x5
transparencies, if you take the trouble to properly profile it. The Creo
Scitex blew away my Epson, I was crushed. He even had the nerve to shoot a
better shot than mine. A Heidelberg Tango makes My Nikon Coolscan 4000 look
pretty limp, too. Both scanners are really useful, but not major league
equipment. They make sorting the wheat from the chaff easy.

Gordon Moat wrote:

Mike wrote:


Some guy is claiming that a 4000dpi 35mm with a dedicated film scanner
will beat a 1200dpi 4x5" scan with an Epson 4990. I don't believe it.
Anybody tried this?


So using a 300 ppi file for printing, the 35 mm would be just over 12" by
18" while the 4" by 5" would give 13" by 20". Not exactly even. This is
still slightly over 3x for the 4" by 5", and more than 12x for the 35 mm.
While there are some nice dedicated film scanners, like the Imacons, I
still think about 10x is a more reasonable limit for 35 mm.

However, an Epson is not a high end scanner, and could be much more
limited. That 1200 dpi figure should easily within reach of the optics
for a true resolution from the 4990, even using that test from May/June
View Camera as a guide.

I think where the dedicated film scanner (some, but not all) would likely
beat out the Epson would be in the Dmax realm. An image with more shadow
details might turn out to be a better image file from a dedicated 35 mm
film scanner.

Now the other question is what does he mean by "beat"? If it was colour
or shadow detail, then he might be correct. Also, some dedicated film
scanners capable of 4000 dpi cost more than an Epson 4990, but you get
what you pay for . . . . .

Ciao!

Gordon Moat
A G Studio
http://www.allgstudio.com



Another thing, and in my opinion, the major factor in this whole argument
is that the MTF of all scanners but drum scanners are rather poor at their
stated optical resolution, about 2/3 that of equivalent DPI drum scans in
my experience. This would mean the 4000 dpi scanner is more like
2700 dpi at high MTF. The Epson at 1200 DPI would be working at
the high region of its MTF response, so I would say there would be
no comparison, as the 4x5+ Epson scan would be high lens MTF, high scanner
MTF. The 35mm at 4000 dpi would be at its lower lens MTF and
low scanner MTF. I predict the 4x5+Epson on would win hands down.

Here is my comparison of the Epson 4990 with a drum scan:

http://www.clarkvision.com/imagedeta...atbed-scanners

It does very well in the shadows. I'm making beautiful Lightjet,
Fuji crystal archive 30x40 inch prints from Epson 4990 scans of
4x5 velvia transparencies. One key is to do all work in 16-bit,
including the scan, going to 8-bit at final output.

Roger
Large format gallery: http://www.clarkvision.com/galleries...y.large_format

Do you really think scanning at 16-bit makes a huge difference? I'll do
more testing later tonight, but my initial tests I didn't see a huge
difference between 8-bit and 16-bit.

Also, if my final goal is 1200dpi, would you recommend scanning at a
higher resolution and then down-sampling? Say a 2400dpi scan at 16-bit
and then downsample to 1200dpi 8-bit for final output?





Another thing, and in my opinion, the major factor in this whole argument
is that the MTF of all scanners but drum scanners are rather poor at their
stated optical resolution, about 2/3 that of equivalent DPI drum scans in
my experience. This would mean the 4000 dpi scanner is more like
2700 dpi at high MTF. The Epson at 1200 DPI would be working at
the high region of its MTF response, so I would say there would be
no comparison, as the 4x5+ Epson scan would be high lens MTF, high scanner
MTF. The 35mm at 4000 dpi would be at its lower lens MTF and
low scanner MTF. I predict the 4x5+Epson on would win hands down.

Here is my comparison of the Epson 4990 with a drum scan:

http://www.clarkvision.com/imagedeta...atbed-scanners

It does very well in the shadows. I'm making beautiful Lightjet,
Fuji crystal archive 30x40 inch prints from Epson 4990 scans of
4x5 velvia transparencies. One key is to do all work in 16-bit,
including the scan, going to 8-bit at final output.

Roger
Large format gallery: http://www.clarkvision.com/galleries...y.large_format

Mike wrote:

Do you really think scanning at 16-bit makes a huge difference? I'll do
more testing later tonight, but my initial tests I didn't see a huge
difference between 8-bit and 16-bit.

Also, if my final goal is 1200dpi, would you recommend scanning at a
higher resolution and then down-sampling? Say a 2400dpi scan at 16-bit
and then downsample to 1200dpi 8-bit for final output?

Mike,
If you are simply scanning and printing, then you will not gain much
by going to 16-bit. But if you are making corrections, dodging
and burning, then it is my opinion the 16-bits does make a difference.

Roger


Another thing, and in my opinion, the major factor in this whole argument
is that the MTF of all scanners but drum scanners are rather poor at their
stated optical resolution, about 2/3 that of equivalent DPI drum scans in
my experience. This would mean the 4000 dpi scanner is more like
2700 dpi at high MTF. The Epson at 1200 DPI would be working at
the high region of its MTF response, so I would say there would be
no comparison, as the 4x5+ Epson scan would be high lens MTF, high scanner
MTF. The 35mm at 4000 dpi would be at its lower lens MTF and
low scanner MTF. I predict the 4x5+Epson on would win hands down.

Here is my comparison of the Epson 4990 with a drum scan:

http://www.clarkvision.com/imagedeta...atbed-scanners

It does very well in the shadows. I'm making beautiful Lightjet,
Fuji crystal archive 30x40 inch prints from Epson 4990 scans of
4x5 velvia transparencies. One key is to do all work in 16-bit,
including the scan, going to 8-bit at final output.

Roger
Large format gallery: http://www.clarkvision.com/galleries...y.large_format

Mike wrote:

Do you really think scanning at 16-bit makes a huge difference? I'll do
more testing later tonight, but my initial tests I didn't see a huge
difference between 8-bit and 16-bit.

While the file sizes are larger, I did find that it makes a difference when any
adjustments are done to that image file. Considering that nearly all printing needs
8-bit files, you could save some time and file storage space just scanning an 8-bit
file. That would also depend on if your scanning software and set-up works well
enough that you are not further adjusting that 8-bit file.



Also, if my final goal is 1200dpi, would you recommend scanning at a
higher resolution and then down-sampling? Say a 2400dpi scan at 16-bit
and then downsample to 1200dpi 8-bit for final output?

It might not hurt to try it. If you find that letting PhotoShop convert to the
output specifications works better than the scanner software, then it is a good
choice. If you see little to no difference, then you are just adding extra time and
complication.



Ciao!

Gordon Moat
A G Studio
http://www.allgstudio.com