My idea for this little test, was to see which software would give me the best result, at the shortest time.
You can of course tweak and adjust for perfection for each scan, and that will obviously give you the best result for each scan – in either or – but like I said that wasn’t my goal here.

First off I want to make clear that I’m certainly no expert, there are a lot of technical stuff I know jack about, This is made purely from a personal perspective. But feel free to add a comment below, with your own experience or advice on what I could do better.

I started out spending some time finding the right settings in both pieces of software, I haven’t included that here. And this part is of course a matter of individual taste, you may prefer other settings then what I used.
That shouldn’t prolong the time, unless you add more features, like noise reduction or multi scan, which I don’t use.

I did this because I need to scan a lot of old negatives, and I for one, think that the scanning process is the only less joyful thing to do in the whole analog photography process, and with these old negatives – mainly holiday photos – I just wanted a quick way to get through them and not lose too much in quality. And secondly I simply was curious to know which one was fastest.

OK a bit about my procedure, I use (NLP) – a Lightroom (LR) plug-in – for my conversions from negative to positive, because it’s gives me the best result wise. So I just needed to scan my negatives as negatives, for this to work.
But just for fun I also tried to scan them, where I used the two programs to convert the negative, using their internal film profiles. And remember I want to do as few adjustments for each scan as possible, which means I’m going to have as much set to auto as possible.

I used a Plustek 8200i for this test, and it has the infrared feature, which I’m a huge fan of, so the only criteria I had for this test, was that IR had to be in action. Because that saves me a lot of work/time in the post process, with dust and scratch removal, and that is a huge factor in the “do as little as possible process”.

The film I used was a HDC 200 film.

My normal setup is with Vuescan, 3600dpi, 48bit, media set to image, color balance set to none, and infrared cleaner set to “heavy”, save as a DNG file. So Vuescan was pretty quickly done, setup wise.

Silverfast, was another thing, I only used it a couple of times before, and since I started using NLP, I stopped using it at all. I quickly found out that in order to use the infrared cleaning feature, I could only save the image file in Tiff format, if I wanted to save as an DNG, I wasn’t able to use the infrared feature, don’t ask me why, it doesn’t make any sense to me.
After a bit more testing I found out that the settings I preferred were Preset 300 dpi (photo quality), resolution 3600ppi, iSRD (IR) on automatic and then save as a Tiff file.

NLP pre-conversion settings were color model set to “Basic” and pre-saturation to “3 (default)”. Settings on the main panel were, color to “Autocolor 2.0 Neutral” and Tones to “Linear deep”.

Scanning times was as follows – these times may vary, depending on what kind of computer you are using – I used my laptop, which isn’t the fastest in the world, but OK – but the test here is to see the difference in scan/conversion times, not the actual speed. My guess is that the distance between them will be the same, on a faster machine, just short times.

  • Silverfast scan: 23,5 sec. (prescan) & 3 min. 20 sec. (with IR)
  • Vuescan scan: 15 sec (preview) & 2 min. 4 sec. (with IR)
  • NLP conversion from Silverfast Tiff: 9,5 sec.
  • NLP conversion from Vuescan DNG: 10,5 sec.
  • ——————————————————————————–
  • Tiff gamma conversion: 16 sec. (then a 9,5 sec. NLP conversion)
  • Tiff to DNG: 3,5 sec. (then a 9 sec. NLP conversion)

NB: These times are for one photo.

Well let’s look at some results.

First we have the direct conversion from the DNG file (Vuescan) and the Tiff file (Silverfast), I’m actually amazed by how much they look alike, there really isn’t that much of a difference. There is a bit more saturation in the Tiff file, and it’s a bit more magenta, but that’s really it.

But there are actually two more ways to handle the Tiff file from Silverfast, when using NLP. I can convert it to a DNG file in LR, or I can gamma correct it with NLP – not sure what that does, but it’s what recommend on the NLP website – but of course both of these actions add to the time it takes me to get a finished photo (see timetable above). But it doesn’t really matter because I’m not crazy with any of the results I get.

The gamma corrected version is to blue/cold for me, and the Tiff to DNG is actually ok, it’s just a desaturated version of the direct from Tiff version above, so I see no reason to do the extra step.

Of course another fast way to scan a negative, would be to use the film profiles there are in the two software. But they are not doing a very good job if you ask me. Silverfast had a profile for the film – the AGFA HDC 200 – but Vuescan didn’t here I used a AGFA XRG/XRS 200 profile.

Silverfast is actually doing a pretty good job, with its film profile – at least with the AGFA HDC 200 film – it’s a little dark and a bit to magenta for my taste. Vuescan how come, is just way off, nothing to like here. I don’t know if it would do a better job with a film that it actually had the profile for, but that’s another test for another day 🙂

So which software would I choose?

I not really sure, if I look at speed as the only factor, then there is no doubt, Vuescan is the winner.
But if I take a look at some other factors as well, then I’m not so sure.

Vuescan makes 16 bit files, Silverfast only 8bit.
Vuescan files are 106MB per photo, Silverfast only 54MB
Vuescan makes DNG (RAW) files, Silverfast Tiff
By the look of it – I’m working on a little test with that as well – Silverfast is better at dust/scratch removal.
You can quickly adjust highlights and shadows – via the histogram – in Silverfast, it takes a bit more work to do the same in Vuescan.

In this case here, I like the end result from both pieces of software. Like I wrote earlier, they look almost the same.
But I have done a few more scans, to compare the results, and I might just like the Silverfast a bit better than Vuescan’s, they tend to be a little warmer (more green) than Vuescan’s scans, which are more cold (and magenta), but we are really talking minor details here, and again it all comes down to personal taste.

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Previous articleNovember 2020
Next articleDecember 2020

4 COMMENTS

  1. It appears the bundled editions of Silverfast (SE and SE Plus) are limited to basic user interfaces and export only as 24-bit (8-bit per channel).

    I did not see mention of which edition your were using, but AI Studio has the full “expert” interface and can export 16-bit per channel it seems.
    It’s expensive though.

    • Hi Paul

      I’m pretty sure I have the Studio edition, it came with my scanner, so haven’t really thought of it.
      Have to check next time I’m using it.

      Cheers
      Bo

  2. Thanks for the article, I am going through the same tests with the same set-up (8200i with Silverfast and VueScan). I am curious what you are using as your output profile in silverfast when exporting to TIFF? I have tried untagged, Prophoto, AdobeRGB all with various results. To my eye, I think the Silverfast 16 bit tiff iSRD with no embedded profile offers more pleasing results then the DNG from silverfast. I do not use the TIFF utility within NLP and do not use the white balance dropper off of the film base. Sometimes, I will crop very tight into the image for my conversion, perform the conversion then remove the crop in LR. I also use ME with both TIFF and DNG, not sure how much of a difference it truly makes, but I feel that when I pixel peep, I see better nose reduction and transitions within the image vs non ME scans. My issue though, in most cases is that my scanned image is usually very warm/orange. I am able to balance in NLP, but there still seems to be strange colors in skintones. This happens whether I WB or not.

    As for VueScan, that is somewhat new to me and I just purchased a license about a week ago. I like the idea of maintaining a DNG really prefer the iSRD application. I have been working on comparing DNG with IR scanned files to what I get from VueScan to what I get out of Silverfast TIFF. So far, the jury is still out and I have yet to dial in my settings that I feel offer optimal results.

    • Thanks! Glad you liked it.

      I use the AdobeRGB profile, must admit I haven’t tested the other profiles. Maybe because I haven’t felt the need for it because I’m quite happy with the Adobe profile.
      I would also love to have an all DNG workflow, but with my testing/comparing I ended up with the Silverfast Tiff setting.

      Are you publishing your results anywhere?
      Would like to see what you end up with.

      Cheers
      Bo

I would love to hear from you!

This site uses Akismet to reduce spam. Learn how your comment data is processed.