Scanning Percentages

[kemal_mumcu]

Hi all, I have been learning how to scan by b&w negs for a couple months now and am still very much near the bottom of the learning curve. I notice my scanner software gives me an option of what percentage to scan my negs. Lately I have been working with 300 dpi at 400%.

What about scanning at only %100 percent? Which setting will ultimately give me the better result? I have been using the 400% at 300 dpi only because the manual recommended it. What do you recommend?

By the way, I notice many of you scan at super high dpi levels. 4000dpi and such. Doing 300dpi at 400% takes my computer 2 minutes per frame. I hate to think how long each scan would be at 4000dpi.

Thanks in advance.

 

[JTK]

Nikon V, scanning full 35mm frame @ 4000ppi, no matter what film, with Ice takes 1.5 minutes...wristwatch second hand measured...many times....certainly within a few seconds of that, either way.

There's no reason to scan at lower resolution unless you have a huge number of visually non-critical images (eg for website or small catalog purpose only) or have an **extremely** slow computer. My 1.5 sec was measured with an XP laptop and an XP desktop, each with 1G Ram and (as I remember) middle of the road system (Sony and Compaq, off the shelf) , 4&6 years ago.

Scan time can get VERY long if you use all the scanner's bells and whistles, attempting to adjust the scan (virtually always a bad idea) rather than modifying the scanned file in post-processing (Elements, Photoshop).

The old Minolta 5400 was almost 9X slower than the Nikon (ie scans with ice took 12 minutes) but it did a comparably good job except for Minolta's inferior focus. The later 5400II, which was mechanical junk, was Nikon-fast but still had poor focus.

 

[kemal_mumcu]

Im using a bargain Canon flatbed with a Macbook with 512mb. Still though, my main question was about the differences between 100% and 300%.

Still though I'll try some high resolution scans.

 

[photophorous]

I don't understand what the percentages stand for. You may need to read your instruction manual to figure out what that percentage means so you can explain it to the rest of us. It sounds like something that is specific to your scanner software.

I use a Coolscan V and have used a Canoscan 8400f in the past, neither of which had anything measured as a percentage. Typically, you scan at a certain DPI, and 300 DPI is very low for negatives. That's a typical setting for scanning prints. I always scan at 4000 DPI and I wouldn't scan anything less than about 1200 DPI, even if it was just for web display.

Paul

 

[Pherdinand]

short side of 35mm frame is about 1 inch. 300dpi scan will result in 300 pixels on the short side of the digital image.
If you get more than that, you are either scanning at a true 300dpi x 4, that is 1200, the factor 4 being from 400%, or your scanner software blows up the image four times after scanning. The latter would mean a lot of quality loss.

 

[JPSuisse]

Hi there

This is a tough question to answer because you have a process with so many variables: scanner type, scan software, quality of negatives, camera type, etc. Also I don't understand what the 100% thing is.


Now, to my experience on hand from 2 examples usuing Vuescan, Sanning to DNGs (i.e. raw), using a Coolscan 5000 and Summilux ASPH 50mm, making a 16bit grey scale or 8bit RGB, respectively.:

1.) T-Max 400, makes no difference if scanned at 2000dpi or 4000dpi. At least when viewed on a screen.
2.) Kodak E100G, there's a significant difference at 2000dpi and 400dpi. At least when viewed on a screen.

If you're interested I can post the pics. But as you see, the variables are many and the experience is key.

Best of luck in scanning!!!

JP

 

[wray]

Here's an example of a Kodak Portra 160 neg scanned at: 3200 dpi at 100% neg size and 360 dpi at 1000% neg size. Both files were processed identically in PS to give same size images. Contrary to common wisdom, I can see no discernable difference in image detail. Before anyone says that you can't do comparisions at web sizes I have done comparisons with final print images and can still see no differences! It would appear that there is little to no difference between letting PS or my scanner (Epson V500) software do the up sizing.

 

[rogue_designer]

the percentages, it's mathematically blowing up the final print size to 300% of the original. This is meaningless from a pixel dimension except to say that at 3X of a normal frame the new file will be 300dpi.

(or the same as scanning at 100% at 900dpi, then converting to 300dpi in the image size options - no resampling).

It probably takes more math from the scanning software to do anything other than 100%. Which could account for the increased time.

 

[kemal_mumcu]

Thanks for the replies. Scanning has been a mystery and is still a mystery to me. So many variables, so many different opinions. I'm slowly trying to make some sense of it all.

My scanner is a CanoScan 4400 and the software provided with it is something called PhotoStudio. i'm more than disappointed with the software but I have learned how to work through some things.

This morning I tried some test scans at 1200 dpi @ 100% and was pleased to find that the scan time was reduced significantly as rogue_designer suggested. Scanning at this setting gave me a scan 1656 by 1183, as we would expect. I checked my older scans done at 300dpi @ 400% and the file size and the pixel dimensions were for the most part identical. Maybe the % thing was only a multiplication factor as Pherdinand suggested.

Unfortunately, when I tried the new 1200dpi @ 100 setting I also tried scanning the TriX negs as a positive and then inverted them later in PS. The scans were certainly better than I had previously done but I don't know if they were from the positive scan or from the different dpi settings.

I'll do another test.

Thanks for the help.

 

[djon]

There's no way to do ultimate evaluation of scans on the monitor if you intend to print. Print resolution is more demanding than monitor resolution in practice.
With my Nikon V and all film there's a distinct advantage to scan @ 4000ppi rather than 2000ppi when printing as small as letter size. My old Epson 3200 flatbed with adjustable focus carrier and 645 format resolves more detail at maximum nominal (not real) 3200ppi
than it does at 1600ppi although as we all know, 3200 is really more like 1400ppi in Nikon terms...I think 1600 is more like 800ppi in Nikon terms. 645 with that flatbed almost equals 35mm with Nikon, and does seem to equal optical enlargement.

Incidentally, if you get better results scanning the neg as a positive you're fooling yourself somehow..IMO. The only situation in which that contributes is with Nikon scanners and Nikon's application on B&W silver film only (ie not on color film or c41 B&W).

I suggest following and sticking with the scanner's simplest instructions until you get great results. Only then should one start exploring the extra variables, most of which are entirely useless...you'll probably find.