How big do you scan a 35 mm original?

On 2007-09-22 18:57:34 -0700, Scott W said:

sheepdog 2007 wrote:
On 2007-09-22 13:56:32 -0700, Scott W said:

sheepdog 2007 wrote:
The thread where I started out to explain why pixel dimensions are
meaningful to me while dpi is not made me question my (amateur)
scanning technique. Have I been missing some quality by not scanning at
a higher resolution on my inexpensive desktop scanner?

I decided to do a test today. The original was a mediocre image,
literally the first 35 mm slide out of the first box I came to. I
scanned it once at about 3600 px wide, then my usual 1800.

The quality of the raw scans is shabby compared to what a high-end
drum scanner can do, and there were NO Photoshop tweaks applied, other
than saving both as JPEGs in the size you see, down-sampled to 1080
pixels wide, with the same amount of compression.

See if you can guess which is which before I say any mo

http://web.mac.com/olddognewtrick/iW...st_070922.html

I am not sure what the point is here, both images are the same size and
both are far to small to tell much of anything.

Scott

The point is I will stick to scanning @ 300 px/in (1800 wide for a
full-frame 35 mm neg or slide). Here's a crop from each:

http://web.mac.com/olddognewtrick/iW...st_detail.html

With this enlargement you can see a subtle difference where (to my eye
at least) the grain looks slightly better in the smaller scan, not what
I would have expected. I printed both and I am seeing about the same
amount of difference on paper.

Maybe using even higher res would be better for some originals. When I
first got the scanner I did tests on good Kodachrome 25 transparencies
and got results similar to this (Fuji 200 transparency). This 'chrome
was scanned in the early days of my website...

http://web.mac.com/olddognewtrick/iW...pinkrose-2.jpg

What still puzzles me is that you are two crops from a 3600 pixels wide
and a 1800 pixel wide crop, any yet the two seem to be the same size. I
would guess you have resized one or both of the scans.

The test was to see if scanning bigger and throwing a bit more
information information away in image editing would result in a
better-looking final product. I would normally do several steps to
improve the image along the way, and size down to my final dimensions
just before converting to RGB and saving as a JPEG. Yes, they're both
down-sampled to a final width of 1080. I don;t know too many folks who
post 3550 px wide on their web pages, do you?

For what is worth I would tend to agree with you that for many 35mm
photos there is little to be gained for scanning to higher resolution
then 1800 x 1200. However I have seen scans that show a fair bit of
detail past that amount, a lot depend on both the film and the scanner
used. A good scanner and film should get a fair bit of detail out to
300x 2000 pixels. Some film buff claim much more but I have yet to see
it.

The pink rose looks very good as a 6/c inkjet 8.5" x ii" print. It
could probably stand up to even more enlargement (K25, good lens, good
light). If I were setting out to shoot a two-page magazine spread, I'd
go with a larger format to feel well-prepared. I'm just shooting for
enjoyment these days, using a D80 (about 90% of my pictures), a Coolpix
L5 (5%), and 35 mm film (5%).

Scott


--
Cease then to grieve for your private afflictions, and address
yourselves instead to the safety of the republic