Resolution of the Plustek 8100

[Alan Johnson]

I photographed a test chart using a Canon 50mm f1.8 at f8 lit by off camera flash, at the correct distance (50x focal length).
Film was Ilford Delta 100, developed in Rodinal 1:50 . The chart was at the center of the frame.
A negative of the chart was scanned at 7200 dpi in the Plustek 8100 and as normal, black and white levels set and sharpening applied.
IMO the resolution read off the scan (attachment) is ~80 lppm
This corresponds to 80x2x25.4= 4064 dpi.
This is higher than the more reliable result obtained at scandig.com of 3600 dpi, probably because my reading of 80 is a bit optimistic.

 

[Diapositivo]

I agree with you that 80 lppm is adequately resolved, in this high contrast target.
Overall this seems to me an excellent performance.
I was worried about the health of my expensive Nikon LS-5000, which has a 4000 dpi resolution in real terms.
I am glad to hear that a replacement would be not very expensive.

Thanks for sharing.

 

[Les Sarile]

Your results seems to far exceed scandig's. You're using a DIY test frame on Delta 100 while scandig uses a commercially bought test target. Have you tried scanning your target with any other scanner covered by scandig to corroborate your results?

 

[Alan Johnson]

As stated, my results are slightly higher than Scandigs, they do not "far exceed" them.
Also I stated the Scandig result more likely to be accurate.
I obtained a similar slightly higher result with the Epson V700:
https://www.photrio.com/forum/threads/resolution-of-the-epson-v700.150649/
Here too I regard the Scandig result as more reliable.

 

[Patrick Robert James]

The only problem i see here is that you are filtering your results through a lens, so your scanner may in fact be able to resolve more than your result. Or your lens may resolve more than the scanner. Hard to tell. At any rate, it is good enough!

 

[Alan Johnson]

Under my old Lomo microscope I examined the negative of the very same test chart as shown in the original post above.
By this means it could be seen that 100 lppm (but not 125 lppm) is visible on the negative, though only somewhat less than 80 lppm was resolved by the scanner.

 

[Les Sarile]

Also I stated the Scandig result more likely to be accurate.
. . .
Here too I regard the Scandig result as more reliable.

Why would you say that scandig's results are more accurate and reliable? If so this obviously makes your results less accurate and less reliable.

It seems to me you are using different standards and therefore to correlate your results with scandig's a mathematical factor may have to be applied. This works well since both of your results are higher then scandig's. The more scanners you test with the same higher results will validate your standards and procedures.

 

[Alan Johnson]

Les,
I don't have shares in Scandig but they seem to have tested most of the 55 scanners they offer and I take my 2 results as merely verifying that theirs is very likely a good resource for reliable data on scanner performance.
https://scandig.com/filmscanner/plustek/index.html

 

[faberryman]

filmscanner.info has the PlusTek 8100's resolution at 3800, so the OP's findings are in the ball park.

 

[Les Sarile]

filmscanner.info has the PlusTek 8100's resolution at 3800, so the OP's findings are in the ball park.

About a 13% discrepancy can be considered in the ballpark since no tolerances were represented.

 

[Alan Johnson]

I obtained a higher resolution from this scanner using Adox CMS document film developed in TDLC-103 (Film Developing Cookbook formula).
With this 100 lppm is nearly reached.

 

[albireo]

OP - interesting results.

As an aside - did you notice any issues with performance at or nearby frame borders using this scanner? I purchased a unit some time ago, and noticed the constant presence of a darker band running in parallel to the smaller dimension of the frame, and mainly visible against evenly coloured background (eg. sky). At the time I tentatively attributed this to internal reflections of light against the negative scanning holder, and returned the scanner. Would be interested to know if this is a known issue with this model.

 

[Alan Johnson]

I have made about 20 prints each 40cm wide ( they fit in the standard 50x 40cm mounts and frames) and I never noticed anything but I usually crop off the extreme edges, maybe 1mm of negative.
It's not an issue for me.

 

[albireo]

Thanks. I've just re-purchased a Plustek negative scanner (a used 7500i, older than the 8100, but seemingly identical sensor-wise) and do not notice anymore the problems I had with the first unit. Must have been a faulty unit or faulty negative holder.

As an aside, the results I'm getting with this simple, relatively inexpensive machine are extremely interesting. Will post examples in a separate thread later.

 

[Alan Johnson]

Here is a youtube review from Dec 2020 comparing the Plustek 8100 to the Nikon 9000.
It seems to agree with my conclusion that for 35mm the Plustek is a notable improvement on flatbed scanners.
But I only use it for B/W .
(44) $300 vs $3000 film scanner comparison | Plustek 8100 - YouTube

 

[Helge]

I obtained a higher resolution from this scanner using Adox CMS document film developed in TDLC-103 (Film Developing Cookbook formula).
With this 100 lppm is nearly reached.

View attachment 224804

I’d say 100 is clearly reached.
Is there any kind of even partially agreed upon measure of what constitutes a line pair?

It’s worth considering that you’d need at least a 90MP camera to out resolve the film here, even with an affordable scanner like this.

 

[Alan Johnson]

Before the digital era things were well defined as indicated here:
https://www.image-engineering.de/library/technotes/761-resolution-measurement-and-its-units
Now I believe various different methods are used , some mentioned in the conversion table but my knowledge of this is limited.
I did need to increase the contrast to demonstrate 100 lppm on Adox CMS20 so this may be on the optimistic side for pictorial photography.

 

[Helge]

Before the digital era things were well defined as indicated here:
https://www.image-engineering.de/library/technotes/761-resolution-measurement-and-its-units
Now I believe various different methods are used , some mentioned in the conversion table but my knowledge of this is limited.
I did need to increase the contrast to demonstrate 100 lppm on Adox CMS20 so this may be on the optimistic side for pictorial photography.

If the line pair is there it’s there.
Increasing contrast and sharpening is not cheating, it’s part of image processing.

It’s easier on digital because it’s been possible and necessary and obligatory since day one, as part of demosaicing and pre-RAW processing and is almost always done without the user being aware of it.

Filtering grain out has never really been a priority in the same way. Partly because grain is not as objectionable and partly because it never became a big field of research.

Of course reducing any noise makes increasing micro contrast easier.