# How many equiv pixel?

Tags:
Last response: in Digital Camera
Share
Anonymous
February 5, 2005 5:52:59 AM

Using Fuji 400 Press film with a Nikon and scanning them into Photoshop
using an Epson Flatbed Scanner @3200 dpi @24bit color; how can you
equate the pixel count to a digital SLR and to a digital Point and Shoot?

Anonymous
February 5, 2005 9:42:45 AM

"measekite" <measekite@yahoo.com> wrote in message
news:fyWMd.622\$lz5.490@newssvr24.news.prodigy.net...
> Using Fuji 400 Press film with a Nikon and scanning them into Photoshop
> using an Epson Flatbed Scanner @3200 dpi @24bit color; how can you equate
> the pixel count to a digital SLR and to a digital Point and Shoot?

Try looking at this site.....
http://www.scantips.com/

Tim
Anonymous
February 5, 2005 1:21:05 PM

measekite wrote:
> Using Fuji 400 Press film with a Nikon and scanning them into
> Photoshop using an Epson Flatbed Scanner @3200 dpi @24bit color; how
> can you equate the pixel count to a digital SLR and to a digital
> Point and Shoot?

Frankly I believe there are so many other factors like color depth
optical issues etc. that you are talking apples and oranges. Try both and
see for yourself what works for the specific type of work you do and your
personal appreciation of the results. The two methods are close enough from
a technical point of view to make the only questions the artistic results
the deciding factor. That is likely to be different for any two people.

--
Joseph Meehan

26 + 6 = 1 It's Irish Math
Related resources
Anonymous
February 5, 2005 2:26:04 PM

"measekite" <measekite@yahoo.com> wrote in message
news:fyWMd.622\$lz5.490@newssvr24.news.prodigy.net...
> Using Fuji 400 Press film with a Nikon and scanning them into Photoshop
> using an Epson Flatbed Scanner @3200 dpi @24bit color; how can you
> equate the pixel count to a digital SLR and to a digital Point and Shoot?

The fact is, you can't get a clear scan with a flat bed film scanner like
yours so any digital image shot with a 2 Mp or greater camera and half
decent lens will look sharper and have more colour definition than your scan
will.

According to Kodak, the native resolution of the average colour film is in
the region of 3200 lpi to 4000 lpi (lines per inch) so in theory, if you
scan at this dpi, you should capture all the film has to yield - including
dust, fibres, grain and the texture of the film itself.

Digital images have none of this to contend with and you can also resize
them with little loss of sharpness and detail so you film scan is not going
to compete very well with a digital image. You might get better results by
having the film printed at 11" x 8" with an optical/chemical process and
scan it on the Epson. I used to do this with my perfection 4870 to get large
(digital) prints before I got a Nikon coolscan dedicated film scanner. Grain
is still a problem but colour and sharpness are not.

Doug
Anonymous
February 5, 2005 2:42:25 PM

> "measekite" <measekite@yahoo.com> wrote in message
> news:fyWMd.622\$lz5.490@newssvr24.news.prodigy.net...
>> Using Fuji 400 Press film with a Nikon and scanning them into
>> Photoshop using an Epson Flatbed Scanner @3200 dpi @24bit color; how
>> can you equate the pixel count to a digital SLR and to a digital
>> Point and Shoot?
>
> The fact is, you can't get a clear scan with a flat bed film scanner
> like yours

Maybe that is overstating the case a bit.

> so any digital image shot with a 2 Mp or greater camera
> and half decent lens will look sharper and have more colour
> definition than your scan will.

Sorry I can't accept that.

>
> According to Kodak, the native resolution of the average colour film
> is in the region of 3200 lpi to 4000 lpi (lines per inch) so in
> theory, if you scan at this dpi, you should capture all the film has
> to yield - including dust, fibres, grain and the texture of the film
> itself.
>
> Digital images have none of this to contend with

True, but they have their own problems like dust on the sensor.

> and you can also
> resize them with little loss of sharpness and detail

How is that different than a scanned image?

> so you film scan
> is not going to compete very well with a digital image. You might get
> better results by having the film printed at 11" x 8" with an
> optical/chemical process

> and scan it on the Epson. I used to do this
> with my perfection 4870 to get large (digital) prints before I got a
> Nikon coolscan dedicated film scanner.

No argument that a good dedicated scanner should outperform a flatbed.

> Grain is still a problem but
> colour and sharpness are not.
>
> Doug

--
Joseph Meehan

26 + 6 = 1 It's Irish Math
Anonymous
February 5, 2005 3:42:28 PM

"measekite" <measekite@yahoo.com> wrote:

> Using Fuji 400 Press film with a Nikon and scanning them into Photoshop
> using an Epson Flatbed Scanner @3200 dpi @24bit color; how can you
> equate the pixel count to a digital SLR and to a digital Point and Shoot?

IMHO (and all the sensible comparison sites on the web), 35mm ISO 100 film,
6MP dSLRs, and 8MP dSLRs are pretty close. 5400 dpi scans of Provia 100F
probably edge out the 20D. Maybe.

But that's not what you asked.

My experience with the ISO 400 color negative films is that they are really
really bad compared to Provia 100F, and my experience with flatbed scanners
is that they are really really bad compared to real film scanners.

That's a lot of really really bad steps you have there between the subject
and the print.

In contrast, dSLRs are really really good at ISO 400. Slam dunk for the
dSLR. No comparison whatsoever.

On the other hand, current digital P&S cameras are really really bad at ISO
400. So you've got a more even dogfight* on your hand there.

If you shoot your P&S digicam at ISO 50, the images will be worlds better
than your Epson scanned ISO 400 film.

*: Fight between real dogs.

FWIW, my experience scanning medium format film with an Epson flatbed
scanner is that 6MP dSLR images look better than 645 at A4 (ouch!), but that
if you switch to a real film scanner, 6 and 8 MP dSLRs aren't even close to
MF.

But that's only for Provia 100F and other quality ISO 100 films. Film
quality goes all to hell at ISO 400 and above.

David J. Littleboy
Tokyo, Japan
Anonymous
February 5, 2005 6:38:18 PM

On Sat, 05 Feb 2005 03:42:28 GMT, you, "David J. Littleboy"
<davidjl@gol.com>, wrote in news:cu1fbg\$bge\$1@nnrp.gol.com:

> On the other hand, current digital P&S cameras are really really bad
> at ISO 400.

Not all that bad really. Some are very good, in fact. If you compare the
luma and chroma noise graphs on dpreview from dSLRs (7D, D70, 20D...) to
those of the 5mpxl digicams (V1, G5, 5400...), you'll see that it's only
one stop better in noise for dSLRs up to ISO800 compared to compact
digicams up to ISO400. And I've seen people boasting noise performance from
dSLRs at ISO1600, so ISO400 from digicams must not be that bad. In fact,
I've seen pictures from the Nikon D70 at ISO 1600, even at resized
resolutions, it's still no way comparable to ISO400 on my Sony W1 full res.
If dpreview is to be believed, the D70's noise at its best ISO200 is
actually worse than that of the Sony V1 at ISO100 and just slightly better
than V1's ISO200. Besides, the newer 1/1.8" sensors for 5 and 7mpxls are
getting even better these days.

As to film ISO400, I've never used print film at ISO400. I don't even like
those at ISO200 and have only used them when there was nothing else around.
Anonymous
February 5, 2005 8:36:42 PM

I scanned some Kodak High Definition ASA400 on a Nikon CoolScan V at 4000
dpi with normal ICE and no GEM and then downsampled it to 4MP to compare it
with my CanonPowershot S45 set at ASA200 and I would say the noise/grain is
about the same as viewed on the computer screen at 100%. I then did a
quarter frame crop of the original 35mm scan and did a 4x6 print on my Canon
i960 and I can cannot really see much grain. This should be the same quality
if I did a 8x10 of the full frame. If grain in an image is affected by the
quality of the lens then this may not be an accurate comparison. This was
not a tripod shot either. Also I looked at comparison images on
www.imaging-resource.com of the S45 at ASA200 to the 300D at ASA800 and I
thought the noise was about the same. From all this I think at 8x10 ASA400
CN film is better than 4MP Digital P&S and Digital SLR is better than the
film. I think this is consistent withthis graph
http://www.clarkvision.com/imagedetail/film.vs.digital....

Chris

"measekite" <measekite@yahoo.com> wrote in message
news:fyWMd.622\$lz5.490@newssvr24.news.prodigy.net...
> Using Fuji 400 Press film with a Nikon and scanning them into Photoshop
> using an Epson Flatbed Scanner @3200 dpi @24bit color; how can you equate
> the pixel count to a digital SLR and to a digital Point and Shoot?
February 5, 2005 9:28:06 PM

Joseph Meehan wrote:

>>
>> The fact is, you can't get a clear scan with a flat bed film scanner
>> like yours
>
> Maybe that is overstating the case a bit.
>

I don't think so, flatbed scans of 35mm film (especially asa 400 ones) are
pretty pathetic.
--

Stacey
Anonymous
February 5, 2005 10:11:43 PM

On Sat, 05 Feb 2005 18:28:06 -0500, Stacey <fotocord@yahoo.com> wrote:

>Joseph Meehan wrote:
>
>>>
>>> The fact is, you can't get a clear scan with a flat bed film scanner
>>> like yours
>>
>> Maybe that is overstating the case a bit.
>>
>
>
>I don't think so, flatbed scans of 35mm film (especially asa 400 ones) are
>pretty pathetic.

Well, there's one flatbed scanner
that I wouldn't mind owning --
a Creo Scitex Eversmart or IQSmart.

Trouble is, they cost a lot more
than most dedicated filmscanners,
even my trusty LS-8000. Even used,
on eBay, I don't think I've seen
one sell for under \$3K or so.

These babies resolve around 5000
dpi, and they use stitching
technology to extend that resolution
over a considerable width.

rafe b.
http://www.terrapinphoto.com
Anonymous
February 5, 2005 11:39:16 PM

"Joseph Meehan" <sligojoe_Spamno@hotmail.com> wrote in message
news:Bi2Nd.5785\$4x5.5244@fe2.columbus.rr.com...
> > "measekite" <measekite@yahoo.com> wrote in message
> > news:fyWMd.622\$lz5.490@newssvr24.news.prodigy.net...
> >> Using Fuji 400 Press film with a Nikon and scanning them into
> >> Photoshop using an Epson Flatbed Scanner @3200 dpi @24bit color; how
> >> can you equate the pixel count to a digital SLR and to a digital
> >> Point and Shoot?
> >
> > The fact is, you can't get a clear scan with a flat bed film scanner
> > like yours
>
> Maybe that is overstating the case a bit.
>
> > so any digital image shot with a 2 Mp or greater camera
> > and half decent lens will look sharper and have more colour
> > definition than your scan will.
>
> Sorry I can't accept that.
>
As is usual with these forums... Everyone, including those armchair experts
who never actually do anything but read specifications disagree on anything
and everything someone else with experience with to say on the subject.

It is simpler than Irish math, Joseph.

1. *Flat bed scanners cannot produce a scan of a 35 mm film as good as that
from a dedicated film scanner*
2. *Digital images shot with decent digicams of over 2 Megapixels will be
sharper, clearer and have more defined colour than an image scanned from a
35mm film on an Epson flat bed scanner*
3. There is no 3. I put that there for the Irish reader so his stout doesn't
tip off the armchair when he reads it!

Doug
Anonymous
February 6, 2005 12:47:15 AM

> "Joseph Meehan" <sligojoe_Spamno@hotmail.com> wrote in message
> news:Bi2Nd.5785\$4x5.5244@fe2.columbus.rr.com...
>>> "measekite" <measekite@yahoo.com> wrote in message
>>> news:fyWMd.622\$lz5.490@newssvr24.news.prodigy.net...
>>>> Using Fuji 400 Press film with a Nikon and scanning them into
>>>> Photoshop using an Epson Flatbed Scanner @3200 dpi @24bit color;
>>>> how can you equate the pixel count to a digital SLR and to a
>>>> digital Point and Shoot?
>>>
>>> The fact is, you can't get a clear scan with a flat bed film scanner
>>> like yours
>>
>> Maybe that is overstating the case a bit.
>>
>>> so any digital image shot with a 2 Mp or greater camera
>>> and half decent lens will look sharper and have more colour
>>> definition than your scan will.
>>
>> Sorry I can't accept that.
>>
> As is usual with these forums... Everyone, including those armchair
> experts who never actually do anything but read specifications
> disagree on anything and everything someone else with experience with
> to say on the subject.
>
> It is simpler than Irish math, Joseph.
>
> 1. *Flat bed scanners cannot produce a scan of a 35 mm film as good
> as that from a dedicated film scanner*

While there are exceptions, I will generally agree that dedicated
scanners are better, but that was not the question was it?

> 2. *Digital images shot with decent digicams of over 2 Megapixels
> will be sharper, clearer and have more defined colour than an image
> scanned from a 35mm film on an Epson flat bed scanner*

I don't know what Epson has to offer in flat bed scanners, but I would
not otherwise agree with that statement. I also would not suggest that one
or the other was always better

> 3. There is no 3. I put that there for the Irish reader so his stout
> doesn't tip off the armchair when he reads it!

I don't care for stout, or ale. About the only alcohol I drink is a
little wine.

>
> Doug

--
Joseph Meehan

26 + 6 = 1 It's Irish Math
Anonymous
February 6, 2005 3:30:17 AM

In article <U9aNd.148753\$K7.37180@news-server.bigpond.net.au>,
>
>2. *Digital images shot with decent digicams of over 2 Megapixels will be
>sharper, clearer and have more defined colour than an image scanned from a
>35mm film on an Epson flat bed scanner*

You're wrong here. Whilst they are noisier, the scans of 35mm slides I get
from my Epson 4870 flatbed are easilly a match for my 10D in terms of
sharpness, and that's 6 megapixels, not 2.
Anonymous
February 6, 2005 3:30:18 AM

On Sun, 06 Feb 2005 00:30:17 GMT, Chris Brown

>In article <U9aNd.148753\$K7.37180@news-server.bigpond.net.au>,
>>
>>2. *Digital images shot with decent digicams of over 2 Megapixels will be
>>sharper, clearer and have more defined colour than an image scanned from a
>>35mm film on an Epson flat bed scanner*
>
>You're wrong here. Whilst they are noisier, the scans of 35mm slides I get
>from my Epson 4870 flatbed are easilly a match for my 10D in terms of
>sharpness, and that's 6 megapixels, not 2.

Care to post or send me a 1200 x 1200 pixel
full-res scan sample from your 4870?

To see where I'm getting at, have a look here:

http://www.terrapinphoto.com/jmdavis

rafe b.
http://www.terrapinphoto.com
Anonymous
February 6, 2005 7:35:16 AM

My own experience with ASA 400 film has been pretty dismal. I shot
film for close to 30 years before switching to digital. Having a
rather large collection of slides and negatives I invested in a film
scanner, A Minolta Dimage Scan Dual III, not a high end scanner but it
give way better results then I can get off my flat bed scanner.

I tended to shoot both 100 and 400 ASA film, at the time the largest
prints I would make would be 4 x 6 or there about. I really thought
that when I got the film scanner there would be all these great photos,
it turned out to be a mixed bag. For the lower ASA film I got good
scans and was delighted to see that many photo that looked bad on the
prints that I got back from the film processor look good when scan and
printed. On the down side the 400 ASA film it really not good for
prints larger then 4 x 6.

Others seem to be saying that they are getting good 8 x 10 prints from
400 ASA film, I would like to see a scan of one of these photos. Maybe
others are getting more out of a roll of 400 ASA film then I did but
when I went back over all the photos I took I really wish I had never
bought a roll of 400 film.

I know there are film scanners that can pull out more detail from film
then mine, but from what I have seen that detail is simply not there
when the speed of the film is 400 or higher.

Scott
Anonymous
February 6, 2005 2:29:41 PM

In article <k8ra01hi1unmu51g2lo24neof9nng97tij@4ax.com>,
rafe bustin <rafeb@speakeasy.net> wrote:
>
>Care to post or send me a 1200 x 1200 pixel
>full-res scan sample from your 4870?
>
>To see where I'm getting at, have a look here:
>
>http://www.terrapinphoto.com/jmdavis

Seen that before. I'll look into getting you a sample - have some scanning I
need to do today (for David Littleboy as it happens - long story), and I'll
see what I can do.
Anonymous
February 6, 2005 2:29:42 PM

On Sun, 06 Feb 2005 11:29:41 GMT, Chris Brown

>In article <k8ra01hi1unmu51g2lo24neof9nng97tij@4ax.com>,
>rafe bustin <rafeb@speakeasy.net> wrote:
>>
>>Care to post or send me a 1200 x 1200 pixel
>>full-res scan sample from your 4870?
>>
>>To see where I'm getting at, have a look here:
>>
>>http://www.terrapinphoto.com/jmdavis
>
>Seen that before. I'll look into getting you a sample - have some scanning I
>need to do today (for David Littleboy as it happens - long story), and I'll
>see what I can do.

Scans for Dave Littleboy? We can have some fun with this.

Here's what you do... there must be some Photoshop
filter that adds grain. Or maybe run NeatImage with
settings that enhance the grain. There's gotta be
a way.

Make it look like Tri-X at ISO 3200. I'd love to see
the look on Dave's face.

rafe b.
http://www.terrapinphoto.com
Anonymous
February 6, 2005 5:58:16 PM

"T.N.T." <tnt@localhost.ca> wrote in message
news:Xns95F46C37FBCC3gehatagubzrpbz@corporate.utopia.disorg...
> On Sat, 05 Feb 2005 03:42:28 GMT, you, "David J. Littleboy"
> <davidjl@gol.com>, wrote in news:cu1fbg\$bge\$1@nnrp.gol.com:
>
> > On the other hand, current digital P&S cameras are really really bad
> > at ISO 400.
>
> Not all that bad really.

Yes, they are. Notice that I said _current_. You are basing your comments on
the best of the previous generation. None of the _current_ cameras are even
close to the Sony V1. (The V3 is worse than the G6, and its noise reduction
messes up the images.)

By the way, thanks for the heads up on the V1, though: it's ISO 400 really
is very nice.

> Some are very good, in fact. If you compare the
> luma and chroma noise graphs on dpreview from dSLRs (7D, D70, 20D...) to
> those of the 5mpxl digicams (V1, G5, 5400...),

Comparing the V1 to the 300D noise patches, I see you are right: the V1 is
exactly one stop worse than the 300D. (The V1 is by far the best of the
recent consumer dcams.)

But the g5 and 5400 are problematic in the noise area, as are the current
generation. The g5 and 5400 are major disasters at ISO 400.

http://www.dpreview.com/reviews/canong5/page12.asp
http://www.dpreview.com/reviews/canoneos300d/page13.asp

If you compare the ISO 800 patches on the 300D/10D to the ISO 400 patches
for the G5 and 5400, you'll see that the 300D/10D are significantly better
at 800 than the others are at 400. The experience here, and all the comments
I've read anywhere basically say exactly that: ISO 400 on those cameras
really isn't usable, whereas ISO 800 on the 300D/10D/20D is. That's a big
difference. Especially since you can put f/2.0 and f/1.4 lenses on the
dSLRs. (The V1's usable ISO 400 puts it in a class by itself amongst P&S
dcams.)

Grumble. I wish Sony would make a 2/3" 5MP sensor with that
technology and build a new model of the F717. The V1's slow lens makes it's
nice ISO 400 not all that useful. The F707 was really a pleasure to use, but
its ISO 400 was pretty funky.

> you'll see that it's only
> one stop better in noise for dSLRs up to ISO800 compared to compact
> digicams up to ISO400. And I've seen people boasting noise performance
from
> dSLRs at ISO1600, so ISO400 from digicams must not be that bad.

The problem, though, is that ISO 1600 with that much noise is giving us
something we've never had before: the 300D + 50/1.4 at ISO 1600 is seriously
amazing. ISO 400 with that level of noise (and a 2-stop slower lens) isn't
anything to write home about. (Again, I'm ignoring the V1<g>.)

> In fact,
> I've seen pictures from the Nikon D70 at ISO 1600, even at resized
> resolutions, it's still no way comparable to ISO400 on my Sony W1 full
res.

Yes. But they'd be equivalent to the G5/5400 at ISO 400.

> If dpreview is to be believed, the D70's noise at its best ISO200 is
> actually worse than that of the Sony V1 at ISO100 and just slightly better
> than V1's ISO200. Besides, the newer 1/1.8" sensors for 5 and 7mpxls are
> getting even better these days.

No, the new sensors are getting worse, since they are packing more pixels

http://www.dpreview.com/reviews/sonydscf828/page14.asp

> As to film ISO400, I've never used print film at ISO400. I don't even like
> those at ISO200 and have only used them when there was nothing else
around.

Really. ISO 100 films are quite nice, but ISO 400 is a disaster. The new
Fuji professional color negative films are all ISO 160 (120/220 only films),
but I haven't tried those yet.

David J. Littleboy
Tokyo, Japan
Anonymous
February 7, 2005 1:06:29 AM

In article <vv0d01pvmq6ql4g9fgtrq8m1qrfsmq0l6s@4ax.com>,
rafe bustin <rafeb@speakeasy.net> wrote:
>
>It is, in a way. No offense, but this is nowhere
>near 4800 dpi. Oh, Epson may have sampled the image
>at that resolution, but the sharpness isn't anywhere
>near enough to justify it.

I'm well aware that flatbed film scanners don't get the same detail as
dedicated film scanners at any given resolution, but you can't really beat
the 4870 for medium format without spending far more money. You already have
a few 4870 scans on your comparison, so I'm surprised that you seem to be
surprised. ;-)

>I'll bet if I downsampled my Nikon scans to
>1500 and up-res'ed them to 4800, they'd beat or
>match this Provia scan.

Maybe, maybe not. I know from direct experience that the 4870 just edges out
a 2700 dpi Nikon Coolscan III on 35mm.

>I think you will be very
>surprised at what you see from Dave Littleboy's

I do hope not - I'm expecting it to be quite a bit better than mine!
Anonymous
February 7, 2005 1:06:30 AM

Chris Brown wrote:

> In article <vv0d01pvmq6ql4g9fgtrq8m1qrfsmq0l6s@4ax.com>,
> rafe bustin <rafeb@speakeasy.net> wrote:
>
>>It is, in a way. No offense, but this is nowhere
>>near 4800 dpi. Oh, Epson may have sampled the image
>>at that resolution, but the sharpness isn't anywhere
>>near enough to justify it.

I use an epson 4870 to scan large format at 3200 ppi.
It gets a little less than drum scans at 3300 ppi.
I seem to find scanners, whether flatbed or film scanners
about 2/3 to 1/2 the ppi rating of the manufacturer
compared to drum scans. I rate my 4870 at 4800 "optical"
at about 3000 ppi compared to a drum scanner.
One factor I've seen is focal point: sometimes it is
not on the glass. Everyone should run a test to see
where the optimum focus is if they want to maximize
resolution. Fortunately, mine is right on for
the 4x5 film (which sits just above the glass).

An example result, though not full res, but two mosaicked 4x5s:
http://www.clarkvision.com/galleries/gallery.landscape-...
(it makes a great 78 x 39 inch print at 300 ppi!)

Roger
Anonymous
February 7, 2005 1:06:30 AM

On Sun, 06 Feb 2005 22:06:29 GMT, Chris Brown

>In article <vv0d01pvmq6ql4g9fgtrq8m1qrfsmq0l6s@4ax.com>,
>rafe bustin <rafeb@speakeasy.net> wrote:
>>
>>It is, in a way. No offense, but this is nowhere
>>near 4800 dpi. Oh, Epson may have sampled the image
>>at that resolution, but the sharpness isn't anywhere
>>near enough to justify it.
>
>I'm well aware that flatbed film scanners don't get the same detail as
>dedicated film scanners at any given resolution, but you can't really beat
>the 4870 for medium format without spending far more money. You already have
>a few 4870 scans on your comparison, so I'm surprised that you seem to be
>surprised. ;-)

Yes and no. I thought maybe there was some variability,
(or maybe a lot) and that there might be better specimens
to be found.

It's not entirely an academic matter for me. My Microtek
2500 -- which I use for scanning 4x5 -- is reasonbly sharp,
but has some serious issues with banding on dense negatives.
By "reasonably sharp" I mean, judged as 2500 dpi scanner.
Plus, it has no dICE, and I really miss that. (Plus, it's
a huge honking machine that takes up a lot of space in my
study.)

>>I'll bet if I downsampled my Nikon scans to
>>1500 and up-res'ed them to 4800, they'd beat or
>>match this Provia scan.
>
>Maybe, maybe not. I know from direct experience that the 4870 just edges out
>a 2700 dpi Nikon Coolscan III on 35mm.

Think so? I worked for a long time with a
Polarid Sprintscan 35+, another 2700 dpi
scanner from the same generation as that Nikon.

I took a 675 x 675 pixel snippet from one
of those scans and upsampled it to 4800 dpi.
Here's the result:

<http://www.terrapinphoto.com/sprintscan_upsampled.jpg&g...;

It's plenty grainy (I can hear Dave L. groaning
as I write this) but then again it's from Reala.

I repeated this exercise with snippets from my
Microtek 2500... and have concluded that it's
roughly in the same ballpark as the Epson 4870.
Here's a couple of samples -- 625 x 625 snippets
upsampled to 1200 x 1200.

<http://www.terrapinphoto.com/microtek_2500_upsampled.jp...;
<http://www.terrapinphoto.com/microtek_2500_upsampled_2....;

That's saying quite a lot for the Epson, by the
way, given that the Microtek sells for about \$2500.

OK. Here's the last sample -- the hypothetical
I initially suggested. Initally scanned on the
LS-8000 at 4000 dpi. Downsampled to 1500 dpi
and then back up to 4800 dpi.

<http://www.terrapinphoto.com/coolscan_down_and_upsample...;

From all this, I conclude that the Epson 4870
is providing somewhere in the neighborhood of
2000-2500 dpi. Again, not too bad at all
considering its cost, and at least as good as
that honking Microtek 2500 that I'm using for
my LF scans.

rafe b.
http://www.terrapinphoto.com
Anonymous
February 7, 2005 1:48:47 AM

"rafe bustin" <rafeb@speakeasy.net> wrote:
> On Sun, 06 Feb 2005 11:29:41 GMT, Chris Brown
> >
> >Seen that before. I'll look into getting you a sample - have some
scanning I
> >need to do today (for David Littleboy as it happens - long story), and
I'll
> >see what I can do.
>
> Scans for Dave Littleboy? We can have some fun with this.
>
> Here's what you do... there must be some Photoshop
> filter that adds grain. Or maybe run NeatImage with
> settings that enhance the grain. There's gotta be
> a way.
>
> Make it look like Tri-X at ISO 3200. I'd love to see
> the look on Dave's face.

That'll make me happier. He's sending me the film to see if I can do a
better job scanning...

David J. Littleboy
Tokyo, Japan
Anonymous
February 7, 2005 1:48:48 AM

In article <cu57dr\$d7r\$1@nnrp.gol.com>,
David J. Littleboy <davidjl@gol.com> wrote:
>
>"rafe bustin" <rafeb@speakeasy.net> wrote:
>> On Sun, 06 Feb 2005 11:29:41 GMT, Chris Brown
>>
>> Here's what you do... there must be some Photoshop
>> filter that adds grain. Or maybe run NeatImage with
>> settings that enhance the grain. There's gotta be
>> a way.
>>
>> Make it look like Tri-X at ISO 3200. I'd love to see
>> the look on Dave's face.

;->

>That'll make me happier. He's sending me the film to see if I can do a
>better job scanning...

I think it's less a case of *if* you can do better, and more a case of
*how much* better you can do, given our respective scanning hardware.
Anonymous
February 7, 2005 3:28:56 AM

In article <8t7d01tsfdsvnf0b753j1g6gq27gnn7k1b@4ax.com>,
rafe bustin <rafeb@speakeasy.net> wrote:
>On Sun, 06 Feb 2005 22:06:29 GMT, Chris Brown
>>
>>Maybe, maybe not. I know from direct experience that the 4870 just edges out
>>a 2700 dpi Nikon Coolscan III on 35mm.
>
>
>Think so? I worked for a long time with a
>Polarid Sprintscan 35+, another 2700 dpi
>scanner from the same generation as that Nikon.

A friend of mine has the Coolscan III, and we've done a few comparisons.
There's not a lot in it, but the 4870 does produce the slightly better
results.

>OK. Here's the last sample -- the hypothetical
>I initially suggested. Initally scanned on the
>LS-8000 at 4000 dpi. Downsampled to 1500 dpi
>and then back up to 4800 dpi.
>
><http://www.terrapinphoto.com/coolscan_down_and_upsample...;
>
>
>From all this, I conclude that the Epson 4870
>is providing somewhere in the neighborhood of
>2000-2500 dpi. Again, not too bad at all
>considering its cost, and at least as good as
>that honking Microtek 2500 that I'm using for
>my LF scans.

Sounds about right. I tried the downsampling to 1500 dpi, then back to 4800
dpi with the sample I posted, and some of the mortar lines which are visible
on the original are no-longer discernible afterwards, which suggests that it
is doing better than 1500.

Also, on the version of the full (6*7 cm) scan which I downsampled and
sharpened, there are lines which are obviously distinct on the original
(sharpened) scan, but which start to merge together when downsized to 2400
dpi/30 megapixels.

What I haven't yet determined to my satisfaction is whether scanning at 2400
dpi does any worse than scanning at 4800 and then downsizing in Photoshop
using Bicubic Sharper. I usually downsample to anywhere between 2000 and
3000 dpi for storage/printing, depending on the image.
Anonymous
February 7, 2005 4:41:10 PM

"Chris Brown" <cpbrown@ntlworld.no_uce_please.com> wrote in message
news roid2-9ri.ln1@narcissus.dyndns.org...
SNIP
> What I haven't yet determined to my satisfaction is whether
> scanning at 2400 dpi does any worse than scanning at 4800
> and then downsizing in Photoshop using Bicubic Sharper.

I've found the Photoshop down-sampling quality to be inferior
(Bi-cubic Sharper produces halo and aliasing artifacts) to some of the
alternatives.
See <http://www.xs4all.nl/~bvdwolf/main/foto/down_sample/exa...;
for examples.

Good quality down-sampling and separate sharpening produces better
results with lower graininess.

Bart
Anonymous
February 8, 2005 11:28:22 PM

In article <420762c7\$0\$28984\$e4fe514c@news.xs4all.nl>,
Bart van der Wolf <bvdwolf@no.spam> wrote:
>
>I've found the Photoshop down-sampling quality to be inferior
>(Bi-cubic Sharper produces halo and aliasing artifacts) to some of the
>alternatives.
>See <http://www.xs4all.nl/~bvdwolf/main/foto/down_sample/exa...;
>for examples.

That's interesting, thanks.
Anonymous
February 9, 2005 7:49:50 PM

Bart van der Wolf wrote:
> "Chris Brown" <cpbrown@ntlworld.no_uce_please.com> wrote in message
> news:luhnd2-3pm.ln1@narcissus.dyndns.org...
> > In article <420762c7\$0\$28984\$e4fe514c@news.xs4all.nl>,
> > Bart van der Wolf <bvdwolf@no.spam> wrote:
> >>
> >>I've found the Photoshop down-sampling quality to be inferior
> >>(Bi-cubic Sharper produces halo and aliasing artifacts) to some of
> >>the
> >>alternatives.
> >>See
> >><http://www.xs4all.nl/~bvdwolf/main/foto/down_sample/exa...;
> >>for examples.
> >
> > That's interesting, thanks.
>
> You're welcome. Here is a more technical view on the risk of
> down-sampling artifacts:
> http://www.xs4all.nl/~bvdwolf/main/foto/down_sample/dow...
>
> Bart
Nice write up. The problems with Photoshop's down-sampling are
fairly well known, what a number of people do is down-sample in steps,
kind of a pain but it does work. Up-sampling has some of the same
issues and some people do this in steps as well. It is pretty rare
that I have a photo that the artifacts are noticeable from the
down-sampling in Photoshop, but I have seen it happen.
February 9, 2005 11:20:18 PM

biphoto@hotmail.com says...
>
> Bart van der Wolf wrote:
> > "Chris Brown" <cpbrown@ntlworld.no_uce_please.com> wrote in message
> > news:luhnd2-3pm.ln1@narcissus.dyndns.org...
> > > In article <420762c7\$0\$28984\$e4fe514c@news.xs4all.nl>,
> > > Bart van der Wolf <bvdwolf@no.spam> wrote:
> > >>
> > >>I've found the Photoshop down-sampling quality to be inferior
> > >>(Bi-cubic Sharper produces halo and aliasing artifacts) to some of
> > >>the
> > >>alternatives.
> > >>See
> > >><http://www.xs4all.nl/~bvdwolf/main/foto/down_sample/exa...;
> > >>for examples.
> > >
> > > That's interesting, thanks.
> >
> > You're welcome. Here is a more technical view on the risk of
> > down-sampling artifacts:
> > http://www.xs4all.nl/~bvdwolf/main/foto/down_sample/dow...
> >
> > Bart
> Nice write up. The problems with Photoshop's down-sampling are
> fairly well known, what a number of people do is down-sample in steps,
> kind of a pain but it does work. Up-sampling has some of the same
> issues and some people do this in steps as well. It is pretty rare
> that I have a photo that the artifacts are noticeable from the
> down-sampling in Photoshop, but I have seen it happen.
>
>

The downsampling in PhotoShop CS is set (by default) to bicubic smoothing,
which, it is claimed, accomplishes the same result as downsampling in steps.

I've used it as a test, and I must admit, the results SEEM the same as
"stepped" downsampling.

--
Larry Lynch
Mystic, Ct.
Anonymous
February 10, 2005 1:53:52 AM

"Chris Brown" <cpbrown@ntlworld.no_uce_please.com> wrote in message
news:luhnd2-3pm.ln1@narcissus.dyndns.org...
> In article <420762c7\$0\$28984\$e4fe514c@news.xs4all.nl>,
> Bart van der Wolf <bvdwolf@no.spam> wrote:
>>
>>I've found the Photoshop down-sampling quality to be inferior
>>(Bi-cubic Sharper produces halo and aliasing artifacts) to some of
>>the
>>alternatives.
>>See
>><http://www.xs4all.nl/~bvdwolf/main/foto/down_sample/exa...;
>>for examples.
>
> That's interesting, thanks.

You're welcome. Here is a more technical view on the risk of
down-sampling artifacts:
http://www.xs4all.nl/~bvdwolf/main/foto/down_sample/dow...

Bart
Anonymous
February 10, 2005 8:29:19 PM

"Scott W" <biphoto@hotmail.com> wrote in message
>
> Bart van der Wolf wrote:
SNIP
>> Here is a more technical view on the risk of down-sampling
>> artifacts:
>> http://www.xs4all.nl/~bvdwolf/main/foto/down_sample/dow...
>>
>> Bart
> Nice write up. The problems with Photoshop's down-sampling
> are fairly well known, what a number of people do is down-
> sample in steps, kind of a pain but it does work. Up-sampling
> has some of the same issues and some people do this in
> steps as well. It is pretty rare that I have a photo that the
> artifacts are noticeable from the down-sampling in Photoshop,
> but I have seen it happen.

Well, I've also tested that one, but overall it wasn't better than
some of the others (see
http://www.xs4all.nl/~bvdwolf/temp/Rings1_BCstep10.gif) with steps of
10% reduction (using Photoshop CS's Bicubic resampling).

> It is pretty rare that I have a photo that the artifacts are
> noticeable from the down-sampling in Photoshop, but I have
> seen it happen.

As predicted by the above sample. It doesn't look too bad at first
glance, but there are problems that really show as soon as you try to
sharpen (an RGB mode version of) it. It also is less sharp than some
of the other methods, which begs for sharpening, and that's when it
causes trouble.

Bart
!