Weird Film Scan Situation

Hi,

I am mystified by the CD's I just got back from my photolab.

I asked them to scan a roll of slide film at "medium resolution"
(3000x2000 according to their literature) and also at "high
resolution" (6000x4000) so I could see the difference.

I got back 2 CD's. The one marked medium has two folders each with the
same set of .jpg files about 3000x2000 pixels. Each is about 1Mbyte in
size. This seems very small to me.

The one marked high has one set of the same images, again 3000x2000
jpg files. Each is about 5Mbytes. This seems more reasonable for a
high quality jpg with that number of pixels.

The image pixel sizes are reported by Win Explorer and match those
reported by Photoshop and Irfanview.

So, first, I don't know what happened to my 6000x4000 scans. I'll have
to ask the lab.

But the weirdest part is that these were pictures of flowers (lupines)
from several meters away, with lots of fine detail in the images. When
I blow up the pics to 100% in PS, I see no difference except that the
"hi res" ones look slightly less smooth. Both images look quite good
at the fine detail level. I would expect that the "medium res" ones
would have lots of artifacts due to the high compression that must
have happened.

How can 2 jpg files with the same pixel dimensions look the same if
one is 5x bigger than the other on the disk?

And is there some wonderful jpg compression that can squash a
3000x2000 raw scanned color image (18Mbytes or so I would think) into
1 Mbyte with no artifacts?

And I thought I had a good handle on all this stuff...

TIA
Duncan

" wrote in message . ..
Hi,

I am mystified by the CD's I just got back from my photolab.

I asked them to scan a roll of slide film at "medium resolution"
(3000x2000 according to their literature) and also at "high
resolution" (6000x4000) so I could see the difference.

I got back 2 CD's. The one marked medium has two folders each with the
same set of .jpg files about 3000x2000 pixels. Each is about 1Mbyte in
size. This seems very small to me.

The one marked high has one set of the same images, again 3000x2000
jpg files. Each is about 5Mbytes. This seems more reasonable for a
high quality jpg with that number of pixels.

The image pixel sizes are reported by Win Explorer and match those
reported by Photoshop and Irfanview.

So, first, I don't know what happened to my 6000x4000 scans. I'll have
to ask the lab.

But the weirdest part is that these were pictures of flowers (lupines)
from several meters away, with lots of fine detail in the images. When
I blow up the pics to 100% in PS, I see no difference except that the
"hi res" ones look slightly less smooth. Both images look quite good
at the fine detail level. I would expect that the "medium res" ones
would have lots of artifacts due to the high compression that must
have happened.

How can 2 jpg files with the same pixel dimensions look the same if
one is 5x bigger than the other on the disk?

And is there some wonderful jpg compression that can squash a
3000x2000 raw scanned color image (18Mbytes or so I would think) into
1 Mbyte with no artifacts?

And I thought I had a good handle on all this stuff...

TIA
Duncan

Why are you asking that ****ing question in this forum? This is not
the right place. Why don't you take your old fashioned film and ram it
deep inside the depths of your honkey tonk culo of yours. You gringo.
You have some really big elephant balls to post that **** here. And
don't **** with Mexican Sigma owners. NORTE! NORTENO XIV!!!

On Wed, 30 Jun 2004 13:46:29 +0200, Gisle Hannemyr
wrote:

[Snip]

Blow up the scans to 400 % and look for halos of off-colour around the
edges of objects. There should be a lot more of them them in the 1
Mbyte scan unless they've cheated. If you have Photoshop, use may
also try the difference blending mode to combine frames from 1 Mbyte
and 5 Mbyte scan to pinpoint the digfferences.

Btw.: If I paid for scanning, I would have insisted in getting the
results back in 16 bit/colour aTIFF - but I probably would not
bothered about getting a high res. (re. your order for a 6000 x 4000),
unless the quality of the originals was superb.

G,

You are right. That was extremely helpful. Thank you very much.

Duncan

Why are you asking that ****ing question in this forum? This is not
the right place. Why don't you take your old fashioned film and ram it
deep inside the depths of your honkey tonk culo of yours. You gringo.
You have some really big elephant balls to post that **** here. And
don't **** with Mexican Sigma owners. NORTE! NORTENO XIV!!!

Take your ****ing Sigma and put it deep into your mexican arse!! Nikon is
the best brand!!

Zorx wrote:
Why are you asking that ****ing question in this forum? This is not
the right place. Why don't you take your old fashioned film and ram it
deep inside the depths of your honkey tonk culo of yours. You gringo.
You have some really big elephant balls to post that **** here. And
don't **** with Mexican Sigma owners. NORTE! NORTENO XIV!!!

Take your ****ing Sigma and put it deep into your mexican arse!! Nikon is
the best brand!!

If only they weren't so expensive. Alas, I've turned into yet another Canon
dude.... ah well.

Cam

" writes:

The one marked high has one set of the same images, again 3000x2000
jpg files. Each is about 5Mbytes. This seems more reasonable for a
high quality jpg with that number of pixels.

Yes - this is the size I use for my 6 Mpx JPEGs.

So, first, I don't know what happened to my 6000x4000 scans. I'll
have to ask the lab.

Sounds like they've confused "high resolution" with "high quality
JPEG".

How can 2 jpg files with the same pixel dimensions look the same if
one is 5x bigger than the other on the disk?

They shouldn't. Blow up the detail and look for JPEG artifacts.
If they're identical they've scanned at "medium" quality and
just re-saved for "high" quality. That's useless.

And is there some wonderful jpg compression that can squash a
3000x2000 raw scanned color image (18Mbytes or so I would think)
into 1 Mbyte with no artifacts?

No.

You are right about 3000x2000 (6 Mpx, RGB, 8 bit/colour) will require
18 Mbytes if stored uncompressed in a non-lossy format (i.e.. TIFF).

Take that TIFF file and turn it to a "high" quality (or whatever
designation the software uses for best JPEG, and you should get
a file of about 5 Mbyte with almost no artifacts.

Now if you compress the same TIFF into a "medium" quality (or
whatever) JPEG - you should get a file of 1 Mbyte and *some*
artifacts.

But "medium" gualtity JPEG is actually quite good. It is when you
select "very low" quality (compressing an 18 Mbyte TIFF into only
200 Kbyte) that the artifacts tend to jump up and meet you.

And I thought I had a good handle on all this stuff...

:-(

Blow up the scans to 400 % and look for halos of off-colour around the
edges of objects. There should be a lot more of them them in the 1
Mbyte scan unless they've cheated. If you have Photoshop, use may
also try the difference blending mode to combine frames from 1 Mbyte
and 5 Mbyte scan to pinpoint the digfferences.

Btw.: If I paid for scanning, I would have insisted in getting the
results back in 16 bit/colour aTIFF - but I probably would not
bothered about getting a high res. (re. your order for a 6000 x 4000),
unless the quality of the originals was superb.
--
- gisle hannemyr [ gisle{at}hannemyr.no - http://folk.uio.no/gisle/ ]
================================================== ======================
«To live outside the law, you must be honest.» (Bob Dylan)