Seite 1 von 2

Film scanner vs digital camera - biological/geological thin sections

Verfasst: Freitag 23. Oktober 2020, 09:58
von Tzontonel
Hi there,

I want to scan some thin sections/slices with geological or biological features. Currently, I took the photos with a digital camera (5D mark-iii) and a macro lens (100 mm 1:1 magnification). The output of my photos is at 5760 × 3840 pixels at an effective resolution of 22.3 mega pixels.

My question is, can I achieve better results with a film scanner? For example with a scanner from the range with effective resolution greater than 4000.
The thin sections/slices are smaller (generally can fit in the film placeholders).

Any opinion? Can I get better results in terms of quality/resolution/focus of the entire area (now with my camera, some areas are out of focus).

--
Andrei

Re: Film scanner vs digital camera - biological/geological thin sections

Verfasst: Freitag 23. Oktober 2020, 11:05
von Jossie
Hello Andrei,
Tzontonel hat geschrieben: Freitag 23. Oktober 2020, 09:58 Can I get better results in terms of quality/resolution/focus of the entire area (now with my camera, some areas are out of focus).
There are two issues:
  • dedicated filmscanners have a very small depth of field, only about the thickness of the film (0.1mm ... 0.3mm). So this will not help you to get sharper images. Flatbed scanners are different but they do not deliver the resolution you want.
  • a digital camera with 22.3 mega pixels (this is not a resolution but a number!) has only 11.15 megapixels in the G channel, which determines the resolution (Bayer matrix). So you only have 4070x2715 pixels determining the resolution, other pixel values are interpolated! A dedicated film scanner does much better and delivers all three colour channels for every pixel. So here you gain in resolution.
I am afraid you have to decide between these.

I would recommend to ask a scan service for some test scans to see what is possible. Alternatively, you could borrow a scanner for a short time to test yourself before buying one.

Hermann-Josef

Re: Film scanner vs digital camera - biological/geological thin sections

Verfasst: Freitag 23. Oktober 2020, 14:28
von Tzontonel
Dear Jossie,

Thank you so much for your response. The thickness of the rock is ~ 0.03 mm/30 microns. The slide itself is ~1-2 mm, but if the film scanners can focus on the thin slice of the rock is great! I have enough depth-of-field.

Regarding the digital cameras, thank you so much for the info. So if I'm right, a mid-range film scanner (e.g. with an effective resolution greater than 3000-4000 ppi) can increase a lot my resolution and quality?

Re: Film scanner vs digital camera - biological/geological thin sections

Verfasst: Freitag 23. Oktober 2020, 16:17
von Jossie
Hello Andrei,

take a look at the scanner test reports here at ScanDig.

Many scanners do not deliver the resolution specified in the data sheet. Furthermore, the measured resolution quoted in the test strictly applies only for the machine which was actually tested, not for the whole product series. This is due to the fact that the resolution critically depends on the alignment of the optics. It is quite common that scanners are not perfectly aligned (e.g. my scanner was delivered with a resolution of about 2300 ppi, although the specification was 5000 ppi. I complained and it was immediately replaced by the manufacturer). If everything is fine, the resolution should be at least 80% to 90% of that quoted in the data sheet.

Hermann-Josef

Re: Film scanner vs digital camera - biological/geological thin sections

Verfasst: Freitag 23. Oktober 2020, 17:20
von Tzontonel
Dear Hermann-Josef,

I read the whole Film scanners tests (except low-end film scanners). Also, I looked at different websites where the film scanners were used as WSI (Whole Slide Imaging) system.
I know that this comparison is a little bit forced, but I'm wondering if I can get better image resolution and quality with an OpticFilm Scanner 8200i Ai.
I mention again, that now I took photos with a 5d-mark iii DSLR camera with a 1:1 macro L lens.

So you say that every device can have a little bit up or down - resolution, compared with the datasheet? There is no quality control? I mean the effective resolution, not nominal or optical resolution.

Thank you once again for your support. I think I will try a mid-range film scanner (in terms of price and quality).

Re: Film scanner vs digital camera - biological/geological thin sections

Verfasst: Freitag 23. Oktober 2020, 17:38
von Jossie
Hello Andrei,
Tzontonel hat geschrieben: Freitag 23. Oktober 2020, 17:20 So you say that every device can have a little bit up or down - resolution, compared with the datasheet?
Only down, unfortunately. :( Up is not possible, because the data sheet specifies the resolution obtainable only in principle due to the optics layout, CCD-pixelsize etc.
Tzontonel hat geschrieben: Freitag 23. Oktober 2020, 17:20 There is no quality control?
Apparently not, since then my first scanner would not have been delivered with optics improperly aligned. To estimate the resolution obtained by any device you should use an USAF1951 target.

Before buying, I would strongly suggest to ask for test scans e.g. by ScanDig, or renting a scanner for a short period to see, if you will get what you want.

Hermann-Josef

Re: Film scanner vs digital camera - biological/geological thin sections

Verfasst: Freitag 23. Oktober 2020, 21:29
von micha12345
Hello!

Don't buy the Plustek. Decent quality, but sloooooooooooow, even without iSRD image correction (what you don't want do use in your case anyway).

Pacific Image offers a scanner for scanning probes. It's the Pacific Image PrimeHisto XE. I think (don't know for sure) it has the same hardware like the PrimeFilm XE (in Germany Reflecta Proscan 10T). So it's a very 'basic low end' scanner like the Plustek. But it is faster, offers a scanning software interface for scientific analysis purposes ( if it's meriting I don't know ). Best think is, with an optional film holder http://www.scanace.com.tw/science_pd_1.php?id=35 dhe PrimeHisto XE can scan rocks and other none flat small objects.

How did you plan to fixate the probes between the film holder? Is a probe smaller than the scan area of a 35mm negative film? Most film holders don't offer anti newton glass to hold the probe in place. A few high end scanners offer this option, optional or out of the box. But I would have concers the probes would scatch the glass over time.

Would the probe fit between a slide? If yes and you have a lot of probes to process regulary, I would consider buying a professional slide scanner like the Pacific Image PowerSlide X (Reflecta Digidia 7000) and frame every probe within it's own slide (still asuming the rock probes are transparent). It would be also good idea for archiving the probes in old slide boxes.

Re: Film scanner vs digital camera - biological/geological thin sections

Verfasst: Samstag 24. Oktober 2020, 09:36
von Jossie
micha12345 hat geschrieben: Freitag 23. Oktober 2020, 21:29 Most film holders don't offer anti newton glass to hold the probe in place.
If you use anti_newton glass plates you will need a scanner with auto focus, since the additional glass will shift the focus. With fixed focus scanners you will get unsharp images. reflecta ProScan 10T has a fixed focus.

USAF1951-Target scanned with my DD6000 which has a fixed focus. Left without glass, right target between glass plates.
USAF1951-Target scanned with my DD6000 which has a fixed focus. Left without glass, right target between glass plates.
Vergleich_ohne_mit_Glas.jpg (56.85 KiB) 8408 mal betrachtet

Hermann-Josef

Re: Film scanner vs digital camera - biological/geological thin sections

Verfasst: Samstag 24. Oktober 2020, 16:02
von micha12345
Because of the fixed focus and it's very high price for fixed focus I was wondering about the PrimeHisto XE, as well. But if PIE can sell it as a affordable laboratory scanner without complains from it's customers, it may work with fixed focus.
USAF1951-Target scanned with my DD6000 which has a fixed focus. Left without glass, right target between glass plates.
Really? That's a pity. I thought all recent Digidias (maybe all Digidias) are autofocus scanners. I thought it is mandatory for compensation different thickness of slide types and focusing at the stripe layer and not the glasses around.

So we still need to know whether or not his probes are width enough to fit within a 35mm film holder or slide or maybe the third party APS film holder for MA adapters of some Nikon Coolscans.

Of course and as always every road leads to Nikon Coolscan scanners (and maybe Minolta Multi Pro). But for a recommondation there we need to know if Andrei/his employer is willed to buy old, but excellent, 2nd hand scanners from shops with warranty and invoice.

Re: Film scanner vs digital camera - biological/geological thin sections

Verfasst: Samstag 24. Oktober 2020, 16:22
von Jossie
micha12345 hat geschrieben: Samstag 24. Oktober 2020, 16:02 I thought it is mandatory for compensation different thickness of slide types and focusing at the stripe layer and not the glasses around.
The plane of the slide is in the middle of the frame. Obviously the mechanics ensures, that the film plane always comes to be placed in the focal plane, regardless of the thickness of the frame (which varies in my case between 1.5mm and 3mm). The problem is warped film. As Andrei wrote, his probes are very thin (30µ) and I understood his description that they do fit into a slide frame.

Hermann-Josef

Re: Film scanner vs digital camera - biological/geological thin sections

Verfasst: Sonntag 25. Oktober 2020, 08:40
von Tzontonel
Sorry for being late with my responses,

I attached a schematic overview of my sample. As you can see my area of interest is ~ 30 x 20 mm. The sample glass can be shortened or I can trim the plastic slide mounts one normally uses for 35mm film. Somehow I can make them so that the thin section would slide in through a slot on one end and be securely held. Once I had the thin sections (samples) in the slide mounts they fit in the holder that came with the scanner.

In terms of the focus point, my interests are on the rock surface plane, which is very thin (30 microns) (marked with red in the schematic photo, rock slice).

I know that Plustek OpticFilm 8200i AI has a fixed focal plane and no fine focus adjustment. The presence or absence of a cover glass, and the glass the rock is mounted on, shifts the rock section forward or backward enough to move the rock in or out of the focal plane. So I think that I need to experiment with the slide orientation to find the best way to get the clearest image.

The reason why I cannot buy the PrimeHisto XE is that I don't find a seller in my country (Romania). And, if I buy from online stores outside the EU, the final value fees are too much! But the PrimeHisto XE it's exactly what I'm looking for. So, I need to improvise with a low-cost film scanner.

Andrei

Re: Film scanner vs digital camera - biological/geological thin sections

Verfasst: Donnerstag 29. Oktober 2020, 23:34
von micha12345
I thought the last days about your problem, but cannot think about a good and cheap solution. So it's a standard tissue slide, what else :-) Yes, you are right. Looks like the PrimeHisto's default slide holder whould be perfect for your application. The pricing in German shops looks like rolled with the dice. The US price of 900 USD looks more promising. I would contact the (shop)-owner of this board. Maybe he can offer advice, roll the dice in a more affordable way.
Once I had the thin sections (samples) in the slide mounts they fit in the holder that came with the scanner.
You had a scanner for slides before? Which one?

I was interested in how my low entry fixedfocus/autofocus scanners would handle a tissue slide. Therefore, I took a slide with a feather glued to (thickness ~1.5mm). Results were a little bit sobering. I thought scanners with autofocus would perform far better, could even out small differences in thickness with ease.

In short:
- My Plustek 7500i performed worst (very similar to the 8200i, slide in the negative holder, strange colour artefacts).
- Surprisingly, my Reflecta CrystalScan 7200 performed best (maybe not enough 'real' resolution, no holder necessary, most easy handling).
- My Nikon Coolscan 2000 with MA-20 worked well (lack of resolution very well recognizable, slide fixed in the MA-20).
- Scanning software: Cyberview or Vuescan Demo (For this few scans I didn't want to set up my old XP PC with manufacture software.).

If the PrimeHisto is not an option, I would have a look at the Nikon Coolscan V/LS50 5000 and the MA-21 film holder. High price, but you would reach your high-resolution target of almost 4000dpi and could easily create/3D print a tissue holder for the MA-21.

If you still want to go with the Plustek 8200i as a low budget solution, I would have a look at the Reflecta Crystalscan 7200, too. Easy handling, target "maybe" in the right focus level (The tissue slide scanner Pacific Image NatureScan was sold this way.). According the filmscanner.info tests both scanners archive almost the same 'real' resolution (remembering results may vary between scanners of the same line) and both densities are decent, not more.

I am still wondering why you have problems with your macro lens and areas out of focus. Is your tripod/copy stand/set up in alignment? Did you test the alignment with a mirror? Has the aperture an influence on your focus problem? Are you using mirror lock-up for avoiding shakes?

Re: Film scanner vs digital camera - biological/geological thin sections

Verfasst: Freitag 30. Oktober 2020, 10:37
von Tzontonel
Regarding the PrimeHisto scanner, I've contacted different customers, and I cannot find it in Romania. Even in Europe, just one website from Germany, but they don't respond to my quote/request. I think the price and what it's offer is a good option! Thank you for telling me about the PrimeHisto. I will hunt that scanner, and I hope that I will buy/try in the future.
You had a scanner for slides before? Which one?
No, I have tried on different professional flatbed scanners. If you look at this link: https://prd-wret.s3.us-west-2.amazonaws ... gabbro.jpg, you will find an image of a thin section of a rock with the FOV ~20 x 13 mm. The Exif info says that it's made with Plustek OpticFilm 7600i. In my opinion, it's a very good result, compared with what I've got with my DSLR + macro lens.

One more option that I want to explore is infrared. My sections have a lot of dust and speckles, and I think the infrared option will solve this problem.

The biggest problem of my current workflow (with DSRL camera and macro 1:1 lens attached, on the tripod, mirror lock, etc.) it is lightning. I cannot perform a good illumination on the slide - considering that I need to use polarizer thin films.

This weekend I have time to 'play' with a Plustek 8200i AI, and I will come back with my results.

Can you say more about the scale of your feather tissue? And the thickness...

Re: Film scanner vs digital camera - biological/geological thin sections

Verfasst: Freitag 30. Oktober 2020, 16:08
von micha12345
My feather slide is just a very basic plastic slide. I think you will find it in every kid's microscope set. The plastic has a thickness of around 1,5mm maybe little bit less. The feather glued to... very thin. The diameter of the cycle is around 0,8cm, the width x of the feather is around 0,3cm, the heigth y of the feather around 0,4cm (You can measure it by yourself,too. Just change resolution in your image editing software to the value of the filename (3600 or 2700dpi) without resizing an you can measure the precise values.).

Strange. If I am searching for PrimeHisto there are several vendors displayed. Cheapest PrimeHisto for 469€, but 2-4 weeks delivery time. Severall in the range of 500-800€, others over 1000€. I even found a "PrimeHisto" for "5.355,00 RON". The Romanien shop offers the NatureScan, too (I won't name the shops, because it is against board rules.

The scanner is not named as Pacific Image PrimeHisto XE in every case. Sometimes it's the Euromex Pathoscanner or BioImager BioScan 360 (The XE? I am not sure because of the optical 3600dpi, interpolated 10000dpi resolution shop comment), BioImager BioScan 1000 (The XE or XEs? I am not sure. No picture.) or PathScan Enabler 5 / V.

And there are scientific variants of the Plustek OpticFilm models, too. E. g. BioImager BioScan 720 or PathScan Enabler 4 / IV. But why go with the Plustek, if the PrimeHisto offers the much higher real resolution?

One UK shop listet the PrimeHisto XE as outlet. Maybe there will be a PrimeHisto XEs soon? I would still write to the board owner for advice.

Re: Film scanner vs digital camera - biological/geological thin sections

Verfasst: Freitag 30. Oktober 2020, 16:20
von Jossie
Tzontonel hat geschrieben: Freitag 30. Oktober 2020, 10:37 One more option that I want to explore is infrared. My sections have a lot of dust and speckles, and I think the infrared option will solve this problem.
This will only work if your sections are transparent to IR-radiation.

Hermann-Josef

Re: Film scanner vs digital camera - biological/geological thin sections

Verfasst: Samstag 31. Oktober 2020, 08:00
von Tzontonel
This will only work if your sections are transparent to IR-radiation.
Indeed. Many scans are made on transparent thin sections.

--

I've finally tested the OpticFilm 8200i with SilverFast AI. I've got very good results. See attachment for a small piece of microscale glass (the smallest unit is 0.1 mm). The .dng (RAW) format it's a big plus! For me, the ~2-3 minute scan-time @ 48bit HDR - 3600ppi, it's a good time-scan. If I using the HDRi the time increases a little bit. According to filmscanner.info, there is no point to scan at 7200ppi to double the pixels without any increase in resolution (effective resolution).

I have plans to make a custom holder that fits my needs (in terms of size and height).

Re: Film scanner vs digital camera - biological/geological thin sections

Verfasst: Samstag 31. Oktober 2020, 12:00
von micha12345
According to filmscanner.info, there is no point to scan at 7200ppi to double the pixels without any increase in resolution (effective resolution).
No. You entierly misunderstood the filmscanner review. But that's were this forum is for. ;)
Naturally the high nominal resolution has the same disadvantage like the predecessor. One has to scan with 7200 dpi, to get the effective 3250 dpi. This leads to very long scan-durations and to swollen picture files, which afterwards again have to be edited/ compressed.
It means, your 7200 dpi scans obtains a real effectiv resolution of around 3250 dpi for this very one tested unit (results may vary). That's why it is advised to scan at 7200dpi and bin, not resize (A 1x1cm@7200dpi test image remains at 1x1cm@3600dpi after binning. It's O.K., not to bin. The filesize of your image is just much more bigger than needed for it's effectiv resolution.).

If you scan at 3600 dpi you will obtain a real effectiv resolution of around 2600 dbi according the Plutek 7600i test.

With IR activated the Plustek 8200@7200dpi is much slower than the Reflecta 10T@5000dpi which is slower than the Nikon LS50@4000dpi.

Also, the Nikon will grant you with the best results if it comes to shadowly areas. If you maybe need the Nikon scanner only for a thesis (for a company it's price shouldn't be a problem), prices of the Nikon scanner are increasing. If you didn't buy it way overprised you can sell it without loss later.

Nikon also offered the tissue slide holder FH-G1 (medical purpse, no polarisation options) for the Nikon V/LS50 LS5000. But I haven't seen a seller for this holder on ebay/amazon.de, today.

PS: Yes, you are right. If you need the IR channel/RAW files for processing later, you will need Silverfast AI Studio (The lesser Silverfast version can safe raw file, but cannot porcess them later. Please correct me, if I am wrong) or Vuescan.

Re: Film scanner vs digital camera - biological/geological thin sections

Verfasst: Samstag 31. Oktober 2020, 13:28
von Jossie
micha12345 hat geschrieben: Samstag 31. Oktober 2020, 12:00 That's why it is advised to scan at 7200dpi and bin, not resize (A 1x1cm@7200dpi test image remains at 1x1cm@3600dpi after binning.
The problem is, that you never get completely rid of the excessive pixels. If you scan at 7200ppi and get an effective resolution of only 3600ppi, binning with a factor of 2x2 will not end at a resolution of 3600ppi but with considerably less! This is due to the fact that binning in itself results in smoothing.

Hermann-Josef

Re: Film scanner vs digital camera - biological/geological thin sections

Verfasst: Sonntag 1. November 2020, 09:52
von Tzontonel
Dear Micha and Hermann-Josef,
Thank you for the technical corrections,

I have attached two scans, one @3600ppi and the other one @7200ppi, both of them with HDR settings without any post-processing in Camera RAW/Photoshop (e.g. sharpening, contrast, shadows, etc.). I've just saved them to jpg format (compression quality 8). After several tests, I will consider saving them in .tiff format.
Also, the Nikon will grant you with the best results if it comes to shadowly areas.
I know that Nikon film-scanners (e.g. LS5000) gives much better results (in terms of quality, resolution, etc.) but the price is too much for my budget. And I read about the price-increase of these scanners. I will wait to win the lottery :roll:.

Can you tell me where can I find more about binning? The workflow and more details?

Files @ http://macrockscopic.ro/scans_3600ppi_vs_7200ppi.zip

(zip file is too large for the attachments ~ 5 MB)

--
Andrei

Re: Film scanner vs digital camera - biological/geological thin sections

Verfasst: Sonntag 1. November 2020, 10:44
von Jossie
Hello Andrei,
Tzontonel hat geschrieben: Sonntag 1. November 2020, 09:52 Can you tell me where can I find more about binning? The workflow and more details?
There is only a short description in Wikipedia. Otherwise there is no magic about binning, if you bin with integer numbers, like 2x2. Binning 2x2 just means that you average blocks of 2x2 pixels, which reduces also the noise. However, binning is also possible by non-integer numbers, which requires interpolation. More details on this are provided e.g. in the textbook by Burger & Burge in chapter 22.

Binning should be applied before sharpening as the two last steps in image processing.

Hermann-Josef