Which DPI To Use When Scanning Film

Rafromak wrote in post #11650555
I have had similar results scanning 35mm slides on my Epson V7000 scanner.

gjl711 wrote in post #11651450
Epson V500

Tony-S wrote in post #11651525
The Epson Perfection V500 tops out around 1800 dpi (rated by Epson at 6400 optical) and the V700 tops out around 1900 dpi.

So would that mean my v600 is somewhere in that range?

Generic question to all.

I have read that when scanning film, especially slides, that using the epson software is better than vuescan (which I have) because it finds the position of the slides better* and focuses on the film plane better.

True or not?

Image enhancement features in the apps is irrelevant to me. I want the best possible original tiff scan first, I'll do other stuff if PS.

(* - I'm not sure what that means - I haven't scanned slides yet, but retrieved hundreds this weekend from a relative and am about to start.)

Picture North Carolina wrote in post #11653936
So would that mean my v600 is somewhere in that range?

Yes. The v500 and v600 have the exact same scanning mechanism. The only difference is that the v600 can do 6x12cm film while the v500 can only do 6x9cm film.

I have never heard this. Where did you see it? If you can point me to a web site that describes it, I'll test it on my v500.

Focusing on the film plane sounds like a fixed hardware characteristic that would not be influenced by software.

I have never heard that about VueScan, and I prefer this one to all others because of how easy it's to use and control all of the settings, from dpi to bit.

When I made the post I wanted the reference but could not find the link. I just looked again and (finally!) found it.

To put it in context, it's in a review of the V600 and at the time of the review, neither VueScan or SilverFast supported the (new) model.

Here's part of the quote:

The focus question is actually talking about the film carriers:

It's a pain in the neck to stop larger films from curling on the V700. But it's easier if you wear gloves to load the tray. I have been scanning 35mm slides with my V7000, and it does a wonderful job since it can scan 12 slides at a time. But there is one slide of my son that I can't get to look sharp (I still have to figure that one).

And yes, one has to select each slide when these are shown in the preview menu. But this only takes a few seconds to do, and to save time one can set VueScan to scan from the photos from the preview feature instead of rescanning them once selected.

One last experiment before I melt back into the forum. I don't have the funds to blow on a $250 airforce resolution chart so I tried scanning something much smaller that has a better focus then a negative or print. I settled on a CD. (pic #1) It has tiny writing and the backslash is actually a series of dots. They can not be seen with the nakid eye.

gjl711 wrote in post #11659408
One last experiment before I melt back into the forum. I don't have the funds to blow on a $250 airforce resolution chart so I tried scanning something much smaller that has a better focus then a negative or print. I settled on a CD. (pic #1) It has tiny writing and the backslash is actually a series of dots. They can not be seen with the nakid eye.


Anyway, scanned at 600, 1200, 1800, 2400, 4800, 6400, 12800 and again normalized in screen so that the same patch was being viewed. At 600 and 1200 the resolution is not enough to separate the lines into distinct dots. It is just a blocky slash. If your looking for detail, these two resolution will not get you the maximum detail.

At 1800 dpi, the optical limit if tested with the ASAF resolution method as posted in this thread, it is still very difficult to resolve the slash as separate dots, but they can be seen.


At 2400 dpi, the dots become clearly visible and the image is sharper and more detailed than at 1800.


At 4800 the dots seem to resolve much better but the image smoothness is much better. If your looking for larger prints, this looks to be the best resolution to scan at.


At 6400 it is no sharper or smoother than at 4800 and at 12800 it looks even less sharp. So from my experimenting, I seem to be getting the best image quality and most detail at 4800dpi. If I wanted to save space scanning at 2400 dpi gets a much smaller file without loosing too much detail. Lines do start getting the jaggies though. Anything less than 2400 and your just not getting the most from the scanner. This could be different depending on the scanner but I'm guessing that most modern scanners perform similarly.

Thanks for taking the time to run the tests and post the results. Very informative.

Well, I'm the one who opened this can of worms and I just scanned a negative on my Canon MP990 with built in film carriers for scanning, and I scanned at 1200dpi and printed out a 4x6. At that picture size, it looked pretty damn good to my old eyes.
Once I get some 8x10 paper, I'll try that.

Picture North Carolina wrote in post #11661219
Thanks for taking the time to run the tests and post the results. Very informative.

No prob. As I mentioned, I'm doing a lot of archiving and from earlier testing the 4800dpi setting seemed to be best but I wanted to make sure. I admit it's far from a scientific test, but at least for my scanner, the best IQ and most resolution is 4800. 2400 gives an outstanding scan and for anything 8x12 or smaller is more than enough resolution.

gjl711 wrote in post #11659408
One last experiment before I melt back into the forum. I don't have the funds to blow on a $250 airforce resolution chart so I tried scanning something much smaller that has a better focus then a negative or print. I settled on a CD. (pic #1) It has tiny writing and the backslash is actually a series of dots. They can not be seen with the nakid eye.


Anyway, scanned at 600, 1200, 1800, 2400, 4800, 6400, 12800 and again normalized in screen so that the same patch was being viewed. At 600 and 1200 the resolution is not enough to separate the lines into distinct dots. It is just a blocky slash. If your looking for detail, these two resolution will not get you the maximum detail.


At 1800 dpi, the optical limit if tested with the ASAF resolution method as posted in this thread, it is still very difficult to resolve the slash as separate dots, but they can be seen.


At 2400 dpi, the dots become clearly visible and the image is sharper and more detailed than at 1800.


At 4800 the dots seem to resolve much better but the image smoothness is much better. If your looking for larger prints, this looks to be the best resolution to scan at.


At 6400 it is no sharper or smoother than at 4800 and at 12800 it looks even less sharp. So from my experimenting, I seem to be getting the best image quality and most detail at 4800dpi. If I wanted to save space scanning at 2400 dpi gets a much smaller file without loosing too much detail. Lines do start getting the jaggies though. Anything less than 2400 and your just not getting the most from the scanner. This could be different depending on the scanner but I'm guessing that most modern scanners perform similarly.

To me it look "best" at 2400 and softer (and smoother) at 4800. This confirms to me that the scanner is only capable of up to 2400 "real" resolution that that there must be some sort of interpolation that starts to kick in after 2400 to make the higher "pseudo-resolution" numbers.

Based on what you show I'd probably not scan past 2400 DPI with a flatbed and anything above 4800 is a definitely waste of scan time that will just end up needing more USM in post.

I think you're right, Bob. Here is the 2400 ppi scan uprezzed to 4800 ppi with Photozoom's S-spline interpolation, no sharpening, and compared with the 4800 ppi scan. I'd say Photozoom does a better resample than Epson.

HOSTED PHOTO
please log in to view hosted photos in full size.

tzalman wrote in post #11678424
I think you're right, Bob. Here is the 2400 ppi scan uprezzed to 4800 ppi with Photozoom's S-spline interpolation, no sharpening, and compared with the 4800 ppi scan. I'd say Photozoom does a better resample than Epson.

That won't work for the pics I posted. What you are seeing is not the 2400dpi scan itself, but an portion of the 2400dpi scan already enlarged so that detail can be seen. The 2400dpi scan was viewed at 400%, a screen shot taken and cropped. The 4800dpi scan was viewed at 200%, a screen shot taken and cropped. The area under the image is about equal but the magnification is not.

gjl711 wrote in post #11678548
That won't work for the pics I posted. What you are seeing is not the 2400dpi scan itself, but an portion of the 2400dpi scan already enlarged so that detail can be seen. The 2400dpi scan was viewed at 400%, a screen shot taken and cropped. The 4800dpi scan was viewed at 200%, a screen shot taken and cropped. The area under the image is about equal but the magnification is not.

Sure, but in my comparison screen shot your 4800 was displayed at 100% and my 4800 (2400x2) was displayed at 50%. As long as the image size is the same it all evens out.

Well... I've always wanted to be part of the Diplomatic Corp, so how about this? We negotiate and find a middle ground at 3600 dpi. Sound ok to you guys?

I'll tell you one thing, tho. Of all the things I've researched, scanning film and slides seems to create an inordinate amount of differing opinions and "debates." Lots of bad info from so-called experts, too.

Case in point:

I ran across a tutorial from a photographer/scanning guru. It was nicely done and covered everything (sorry, don't have the link right now). In it he scanned B/W negatives. At one point he said he did not like Digital ICE because it made images soft, and he recommended not using it. Sounded reasonable to me because de-noising an image can do the same thing, so why not removing scratches and dust?

But then, I found this. It's a nicely-done free ebook from a commercial scanning company on how to scan. More importantly, in it they say this:

Interesting. If the other guy was such a guru, why didn't he know about the problem with B/W film. Hmmmmmmm.

So I guess the debate continues and you need to be aware that differing opinions exist and not all info may be correct.