Scan paper to see texture

fotofreek

Printer Master
Platinum Printer Member
Joined
Nov 5, 2004
Messages
1,811
Reaction score
434
Points
253
Location
San Francisco
High quality slide/film scanners are not only better able to provide better resolution, they also use ICE or a comparable surface scan to eliminate dust from the digitized image. High res Scans of B/W negatives tend to show the silver halide grains, and there is software available to counter that as well. My scanner is not new - it is an Epson 3170 with 3200x6400 ppi. In my estimation, 35 mm scans were a little better than mediocre. For the few slides I periodically scan, I use a digital camera with closeup lens attachements, an xray viewer light box, and a copy stand to standardize distance and placement. By filling the camera frame with the slide content I get the resolution of the camera and the speed of a shutter click rather than the slow scan at highest scanning resolution. If your camera has a suitable macro setting that will work well, but many of the macro settings on consumer grade cameras function at the wide angle setting which can create barrel distortion. Using some old darkroom tools - a small squeeze bulb air blower and an anti-static brush I've used for cleaning negatives - i eliminate nearly all dust and foreign material prior to digitizing the image.

I've done medical/dental photography and know that direct, shadowless light sources (like a scanner light source) eliminate texture. I originally used ringlight flash units but found that a single flash off the axis of the lens provided texture detail that was lost with the shadowless ringlight. I also used electricians tape over sections of the ringlight to pick up the shadows that show texture. Try crosslighting a textured paper and photographing it to see if the result is to your satisfaction.
 

Smile

Printer Master
Joined
Aug 23, 2006
Messages
1,914
Reaction score
417
Points
253
Location
Europe EU
Printer Model
Canon, Brother, HP, Ricoh etc.
"I would also comment that the other scanner appears to show "banding" in several areas of the photo papers."

This is a known defect of lide80, they seem to be related to early CIS technology compared to normal CCD. I don't know anyone with more recent scanner like lide200 to say that this problem is solved in new lide scanners.

Your scans seem to be the same, I can't see more detail 9600x9600 vs. 2400x2400.
Why you scanned in color? In my configuration for VueScan there seems to "B/W Photo" scan type?
Also, if you did not used VueScan then the images can't be compared side by side because there is an option that dramatically increases scan quality in Vuescan called "Number of samples"

Also your scans seem to be normal 24-Bit, when I try to shift levels in photoshop once more I got visible pixels, this is because there are not enough information for even gradations. The scan I posted is 48-Bit the fore you can do levels again.

"From what I can see, there is no visible texture on the photo paper, but maybe the Epson scanner rejects this effect better than the other scanner."

There is texture your scanned simply makes "too bright" scans, I tried to shift levels to the right once more and there is texture on photo paper. It's better than my friends scan, but I begin to see pixels that distorts the image. You can try to do levels 2 times yourself on 48-Bit Grayscale scan.

It would be interesting to see the results.

I would also note that there seems to be emerging "paper fingerprint" technology to identify paper documents by looking at paper fiber composition.
 

Grandad35

Printer Master
Platinum Printer Member
Joined
Feb 24, 2005
Messages
1,669
Reaction score
182
Points
223
Location
North of Boston, USA
Printer Model
Canon i9900 (plus 5 spares)
Smile said:
Your scans seem to be the same, I can't see more detail 9600x9600 vs. 2400x2400.
I agree - I don't know if that's all the detail that exists in the paper or if the scanner claims more detail than it can deliver.
Smile said:
Why you scanned in color?
Your posted scans were in color, so I assumed....
Smile said:
Also, if you did not used VueScan then the images can't be compared side by side because there is an option that dramatically increases scan quality in Vuescan called "Number of samples"
I don't use (or even have) Vuescan.
Smile said:
Also your scans seem to be normal 24-Bit, when I try to shift levels in photoshop once more I got visible pixels, this is because there are not enough information for even gradations. The scan I posted is 48-Bit the fore you can do levels again.
I did the levels shifting in the scanner's software (200/1/255), so there should be up to 255 tonal values in the resulting image. Even if you do another shift in PS, the resulting posterization shouldn't be too bad.

How did you post a 16 bit .jpg image? AFAIK, you have to use a different file format to get additional resolution.

Yes, my scanner supports "48 bit" (16/channel), but that's just sales hype, IMHO. I once played with it, but the penalty of having to store the much larger .tiff images just wasn't worth any minor gain that I could see.

Smile said:
There is texture your scanned simply makes "too bright" scans, I tried to shift levels to the right once more and there is texture on photo paper. It's better than my friends scan, but I begin to see pixels that distorts the image. You can try to do levels 2 times yourself on 48-Bit Grayscale scan.
As was stated previously, the levels values in the scanner were 200/1/255, so if there are blown pixels at the high end, that's the best that this scanner will deliver.

As fotofreek stated, why not use cross lighting and a camera? That's an option with good microscopes the sample can be illuminated with a shallow angle light coming from within the lens. A scanner isn't designed for this type of illumination.
 

Smile

Printer Master
Joined
Aug 23, 2006
Messages
1,914
Reaction score
417
Points
253
Location
Europe EU
Printer Model
Canon, Brother, HP, Ricoh etc.
fotofreek said:
In my estimation, 35 mm scans were a little better than mediocre.
Well, epson V700 or V750 are more than hard core for that price as aposed by Nikon. And Epson's scanners can be used after you scan all your slides :D

fotofreek said:
In my estimation, 35 mm scans were a little better than mediocre. For the few slides I periodically scan, I use a digital camera with close-up lens attachments, an xray viewer light box, and a copy stand to standardize distance and placement.
Since I need to compare the scans with each other I think this will not work. I do have 100mm canon macro lens, but don't have the body to srew it on :rolleyes:

I have G10 that has more megapixels than I need and a custom close-up lens that I put together for my 8mm film digitization project. That makes shots where I can see films grain. But for this task the paper texture changes very much with slight camera angle changes and lighting changes.

You can even think that same paper shot is actually two different papers. That is not good. I do use tripod and try to make paper stable but it can work for slides but not my task.
 

Smile

Printer Master
Joined
Aug 23, 2006
Messages
1,914
Reaction score
417
Points
253
Location
Europe EU
Printer Model
Canon, Brother, HP, Ricoh etc.
Grandad35 said:
Your posted scans were in color, so I assumed....
Yes, my scans were color but we tried to make test scant to compare with my friend scans, and they were BW.

Grandad35 said:
I don't use (or even have) Vuescan.
Thats a shame, you should try it http://www.hamrick.com/vsm.html
Don't remember if trial version puts any watermarks or just time limited trial. Maybe VueScan will unlock full resolution of your scanner?

Grandad35 said:
Even if you do another shift in PS, the resulting posterization shouldn't be too bad.
See for yourself

Grandad35 said:
How did you post a 16 bit .jpg image? AFAIK, you have to use a different file format to get additional resolution.
Yes my file is jpg 8-bit (scanned as 48-bit) but for some reason it does not have gradation problems compare it in photoshop yourself.

Grandad35 said:
As was stated previously, the levels values in the scanner were 200/1/255, so if there are blown pixels at the high end, that's the best that this scanner will deliver.
Your scanner is CCD with CCFL bulb, that means it gives more light than my firend scanned with CIS system. That's why you can scan circuit boards etc on "normal scanner" but not LED scanners.

So it seems to reach same level of detail for CCD with CCFL you have to use levels two times.

Grandad35 said:
As fotofreek stated, why not use cross lighting and a camera? That's an option with good microscopes the sample can be illuminated with a shallow angle light coming from within the lens. A scanner isn't designed for this type of illumination.
As I posted above I need to compare the images between each other and my concern is stability, scanned seems to provide the stability I need. I did make a custom made power source for one of my projects to power LED, I do know how lighting stability is important. And how expensive it is to make one cause the parts are not ordinary resistors and capacitors they must pre all precision type ones.
 

Latest posts

Top