Better to scan at a lower resolution than the optical resolution of the scanner?

[Colin O]

<p>I have a Nikon Coolscan V film scanner and an archive of somewhere in the region of 8000 negatives and slides

that I want to scan. Before starting this job, I'd like to determine the optimal workflow, so I just worked my

way through <a href="http://www.rockynook.com/book/91/scanning-negatives-and-slides%2C-2nd-edition.html">Scanning

Negatives and Slides</a> by Sascha Steinhoff, and I'd like to ask your opinion on something he recommends which I

would never have done myself. I'm sure you understand that I don't want to do this job once and then realise that

I need to start all over again in order to do it properly.</p>

 

<p>Sascha Steinhoff says in the book that film grain will often look more pleasing if the scanner scales the

image, in particular in the case of Nikon scanners. He says, "Here is a very simple trick to improve image

quality: set the scan resolution to the optical resolution and then scale to 50%. This effect cannot be

reproduced afterwards in an image editor; it has to be done during scanning. It is due to the effect that

scanning in full resolution produces more noise than scanning with a lower resolution. If you just scale down the

scan afterwards, the noise is still there; at least in subjective matter, it gets even more visible through

scaling. The best practice is to scan the slide at the same resolution you want to use later. If you scan

primarily for archiving purposes, 2000 dpi is a good compromise."</p>

 

<p>However, at two different points later in the book, he contradicts the previous statement by saying that

images should be scanned at the optical resolution of the scanner.</p>

 

<p>Personally, I don't think 2000 dpi is enough for scanning 135 film for archival purposes. I'd be interested to

hear what others think of his suggestion, or if anyone has experience with this.</p>

 

<p>Thanks.</p>

[q.g._de_bakker]

Scan at the optical resolution.<br>Noise isn't less when you scan at a lower resolution. And there is no way that throwing away every n-th pixel at scanning time is better than doing that in post.<br>Get all out of your negs that you can, even if for the purposes you can think of now it wouldn't be necessary. You may need it later, and scanning is't that much fun that you would want to rescan even a single negative.

[AlanKlein]

<p>Scanner adjustments are basically post processing done by the scanner. The scan is the same. Might as well do them with a true post processing program.</p>

[peterbcarter]

<p>I have a V700 and I do all my scans at 64k. Vuescan has a feature to reduce on save, which I use / 2 for 35mm and /3 for 120. Reducing more than that cripples post as the more pixels you have minimises the ability to notice the repairs you do. By the time it hits flickr, it's down to a 1024x768 box.</p>

<p>If the image was for show / print, I don't reduce it at all.</p>

[david_henderson]

<p>Peter I don't understand your 64k. I have a V700 too and I'm used to using numbers like 4000ppi ( lots of bloat) or 2000ppi,approaching the real resolution capability of the scanner I believe. Do you mean that you scan at maximum quoted resolution?</p>

[peterbcarter]

<p>DH: Sorry, 6400</p>

 

[digitaldog]

<p>Scan at the highest optical resolution (to add/clarify what was said). Do that high bit (what the scanner may call 16-bit even if it's not). Scan into the scanner RGB color space as defined by the software if possible, convert into the largetst appropriate RGB working space when editing is necessary (try to do as much global color and tone correction at the scan stage assuming the software is up to the task). That will save you time. The native RGB scanner space is or should be defined by an input RGB scanner profile. As other's have suggested, I'd worry about grain and noise after the scan on a case by case basis. </p>

[q.g._de_bakker]

And do also consider the scanner's capabilities. A V700's specs and scanning software may suggest you can scan at 6400 ppi, but if the scanner doesn't actually deliver that high a resolution, it would just be a waste of time and storage space to scan at that resolution setting. So check what tests say about the true capabilities of the scanner you are using.

[peterbcarter]

<p>A v700 spec (and most flat beds) list different resolution for x & y axis. Depending on the lines of your image, it does make a difference.</p>

[steve m smith]

<blockquote>

<p>And do also consider the scanner's capabilities.</p>

</blockquote>

<p>And your own capabilities. Scanning 8,000 slides/negatives is going to take a long time and you would likely not take as much care over it towards the end as you would when you start.<br>

<br />I would do a bit of editing and reduce the number.</p>

<p>(Actually I would only scan as and when I needed to rather than embark on doing the whole lot).</p>

[AlanKlein]

<p>I was going to scan all the 35mm chromes residing in my dozen or so slide trays. After two trays of 80 slides each with the associate post processing, I had enough. I figured I finish the rest when I retired. Well I retired 8 months ago. Guess what?</p>

[peterbcarter]

<p>There is a lot to be said for getting someone else to do them. You can always redo the ones that are worth the effort.</p>

[sebastianmoran]

<p>Scan for archival purposes on a Coolscan V at 4000 dpi for your good images on 35mm. For the keepers, who knows what you'll want to do with them years down the road.</p>

[john_a._mozzer]

<p>I read <em>Scanning Negatives and Slides</em> by Sascha Steinhoff, and the section quoted by <strong>Colin O</strong> in the original post above, in 2009. Scanning for archival purposes also means future-proofing, and the quoted section didn't make sense to me. Therefore, I asked Sascha for clarification by email. Sascha wrote in an email dated Oct 19, 2009, " . . . the described effect depends very much on the scanner. my example was nikon. as every scanner has its own characteristique the effect my not be given with other scanners."<br>

<br /> Also, in 2009, I had two different services scan the same negative film strips as a test. One scanning service, ScanDigital (now in Torrance, CA), used a Nikon Coolscan, at 4000 pixels per inch per my instructions. The other scanning service, A&I (now in Burbank, CA), used Nortisu equipment at its Ultra High resolution setting, over 4000 pixels per inch, per my instructions. When comparing scans displayed at 100 percent on my MacBook Pro monitor, I saw digital noise in the Nikon Coolscan scans, and film grain (pleasing to me) in the Noritsu scans. For that reason, I chose to have about 3000 negative film images scanned on the Noritsu by A&I, rather than on the Nikon Coolscan.<br>

<br /> I agree with you, Colin O, that 2000 pixels per inch is not enough for scanning 35mm film for archival purposes. However, I get the impression, the Nikon Coolscan has some sort of sweet spot when it comes to resolution, apparently lower than its its optical resolution.</p>