Film Scanners: Bogus dpi specs?

[papo]

I keep reading these reviews from websites that do scanner tests and reviews and as it turns out, all of the "more affordable" scanners do not reach anywhere neard the dpi that the manufacturers claim.

Heres a review of the Epson V700 that i wanted to get (skip to the image quality paragraph). It states that the actual dpi the scanner was able to reach was 2300dpi, compared to the 6400 dpi that is indicated by the manufacturer.

Whats your take on this? And while we at it, i do need a affordable scanner 😀

Thanks.

 

[Ko.Fe.]

Yes, at least with Epson flatbed.

I'm using V550 1800 dpi (maximum optical resolution) files of 135 film scans to print on 8x10 without problems. And I could print much larger from MF scans.

Do you have computer capable of f6400 dpi files handling and print service available to print those huge files?

 

[papo]

Thanks for the info. Can i see some scans you did with the 550?

As far as if my computer can handle f6400 dpi files is somethign i dont know since all of these scanners who claim they can scan these resolutions have very vague requirements infos.

Other than that, i hear that the scan software itself is super important for scanning. How so? I can understand for post production but for scanning itself?

 

[BLKRCAT]

depends on how you want to scan. I used to use silverfast with my V700. It was a great one step scan that applied profiles and with some tweaks you can get consistent results roll to roll. When I switched to my Coolscan the game changed a bit and I resorted to scanning flat then applying my gamma in post to tone the image.

I always scoffed at vuescan but its worth the money. Good software and brought my unsupported nikon back into mac-land in 2016.

Once you decide the method you'll use to scan I think then you'll find strong points and weak points between software.

 

[majid]

The limiting factor is usually the optics. The V700 has a stated 4800/6400 x 9600 resolution. The second number can be dismissed out of hand, as it is simply the number of steps per inch the stepper motor can make. The second number is simply the number of pixels per inch of the linear CCD or CIS sensor.

As for a lens that would resolve 4800 dpi (100 lp/mm) over a 8.5x11in area, even the best enlarger lenses have at best 30% MTF contrast at 100 lp/mm, so the inexpensive probably $10-20 in the scanner Bill of Materials) lens in the V700 isn't going to get anywhere near that.

 

[Ko.Fe.]

Epson have best scanning software I ever seen. It gives tiff files which are easily processible in LR.
6400 files are going to be huge for computer, graphic card and LR to deal with.
Here is no point to look at v550 full size scans. On computer screen they look horrible. But printed on 8x10, it shows every hair on the head, no artifacts.

 

[flavio81]

I keep reading these reviews from websites that do scanner tests and reviews and as it turns out, all of the "more affordable" scanners do not reach anywhere neard the dpi that the manufacturers claim.

Heres a review of the Epson V700 that i wanted to get (skip to the image quality paragraph). It states that the actual dpi the scanner was able to reach was 2300dpi, compared to the 6400 dpi that is indicated by the manufacturer.

Whats your take on this? And while we at it, i do need a affordable scanner 😀

Thanks.

From all sample scans I've seen, plus the scans i've received from two labs that used the EPSON VXXX scanners, those scanners are discarding (read "wasting") information contained on the negative.

Not only they give less resolution than what's on the actual negative, but they also give "fuzzy" definition, introducing chromatic abberations, loss of contrast, etc.

Yes, you can use such Epson scanners to produce an acceptable 8x10 enlargement of a 35mm frame, but not a really good enlargement.

The best is to use a dedicated film scanner, or to use a DSLR camera for scanning.

 

[Prest_400]

I have a V550 that I purchased for scanning 6x9, as lab scans do get cumulatively costly and high resolution also means high prices and I have less control over the output vs a given JPEG.
Long story short, the ~1800-2000 real dpi is adequate for such format and have done decent 12x18" from it.

35mm you can have good results 2000x3000px for 8x12" but it is a bit of a PITA. I find it too little too slow (resolution and speed). In the pakon thread there is around another member recommended a Primefilm scanner which seems decent.

My point is, for larger formats the lower resolution of these scanners are not as critically felt as in 35mm and for fewer frames I also can put up with slower speeds. A 6x9 frame non-ICE took about 5 mins, up to 10-15 with ICE, which my unit I don't know why seems to not do well now.

I do tiff with Epson scan itself and then load into Rawtherapee to do a LR esque treatment.

Mostly Portra and usually it inverts the colors OK but not perfectly, slightly cold cast specially in shadows... For slide films, there's a bit of red instead. If it is an issue, I do some curves or levels and that gets removed.


<Beamed through Tapatalk relay>

 

[michaelwj]

The quoted numbers at the theoretical best case, usually disregarding optics entirely. i.e., just how fine the motor can theoretically step (which is probably not reached in practice either). On top of this there is only so much resolution in the negative, so even if you could scan at 1,000,000 dpi, there's no point.

The actual dpi that you will be able achieve is effectively limited by the step size of the motor convolved with the point spread function of the illuminating beam. So with a 3mm diameter beam and a 1mm step, the resolution will be something between the two, I'd guess about 2mm (assuming that exists in the negative - a clear negative will have no resolution). The microscopy community do some work devolving the two, but to varied levels of success, and never getting down to the 1mm of the step - we did this a lot at the Synchrotron where I used to work.

As a rule, the less the motor has to move, the more accurate it will be, so ideally you match the scanner to the negative size, i.e. use the Epson flat bed for large format not 35mm, a dedicated 35mm scanner will be better (all being equal - which it's not).

For what it's worth I use a Plustek 8100 and scan at 3600dpi, where the max quoted is 7200dpi I think. At 7200dpi the motors are at there resolution limit and prone to jumping lines from what I can gather - I get lots of artefacts that I don't get at lower dpi.

 

[charjohncarter]

Does anyone see a difference between types of film (35mm). I can get lovely scans with TMX film, less lovely with TRI-X and other old style films. My C-41 scans are also good, but not as good a TMAX. I also find that slide film is almost always soft. I think this old slide was focus well but the scan is soft.

Ektachrome 64 by John Carter, on Flickr

Like someone said above it is hard to tell on a monitor, but my Costco print is soft.

I used a DSLR device I made and it was really better for 35mm. I'm wondering if anyone has tried wet scanning on the Vxxx?

 

[farlymac]

From what I've been told by the guys down at the lab, John, some films do not scan as well as others. It has to do with the grain structure, and since there is no grain in slide film, they do tend to be softer.

PF

 

[x-ray]

Epson fudges their number as do most other scanner makers until you get into the upper end machines. I have a V750 and its capable of ok scans within limits. The big issue is optimizing focus and flatness of the film. About the only thing I scan on the v750 are 8x10 negatives.

My primary scanner is a Hasselblad / Imacon 848 and there's a world of difference in dynamic range, noise, resolution and sharpness but there's also a word of difference in price.

Prior to the 848 and v750 I owned a Fuji Lanovia Quattro. The true optical resolution was 5000dpi. The dynamic range was equally as impressive. Unlike the Epson and most other flat beds in that price range the Fuji has 4 top quality Apo process lenses. Depending on the area scanner the scanner would automatically select the proper lens and auto focus up to 10mm in depth of the subject. Every spot on the platen isles an optimum scan where as epson and most others only give good resolution along a band in the center of the platen. The entire area of the platen on an epson is not optimized. The Creo / Kodak Eversmart does it by scanning strips of the film and automatically stitching them together.

Unfortunately all of the topend flatbeds were discontinued years ago. The Fuji was FireWire cut most are SCSI. Parts are no longer available and even on the Eversmart service is getting harder. Most if not all run on old OS and proprietary software. No one but the manufacture made software for these machines. I've not owned a Kodak / Creo but the Fuji requires a specific dongle to run. Not only does it require a dongle, the do gle is specific to the OS and you must have two sets of activation codes to get the dongle to function. Dongle but no codes, you're out of luck. You also must have special calibration negatives and prints. The cost of this scanner when new was around $50,000.

The only really good scanners currently made to my knowledge are the Aztek drum scanners and the Hasselblad. The Aztek I believe at in the $70,000 range new. I think the Hasselblad are in the $16,000 range.

There are a lot of high end used scanners but you need to make sure you have or can get everything needed to make it run or youll have an expensive boat anchor. Also you e got to be someone that understands circuitry and willing to work on it. From experience and owning several Fuji scanners they require a good deal of maintenence.

 

[x-ray]

From what I've been told by the guys down at the lab, John, some films do not scan as well as others. It has to do with the grain structure, and since there is no grain in slide film, they do tend to be softer.

PF

Not sure where they got this. I've scanned considerable film on high end scanners, primarily transparency film up to 11x14, and it's the easiest to scan. Kodachrome is the exception but that's due to a unique property of the Reds formed in the film. Color net is much more difficult to scan. Grain has nothing to do with it. Even transparencies have a grain structure.

 

[ColSebastianMoran]

Scanner DPI numbers are completely bogus, theoretical in some universe.

I tested a V500 and got 2000 ppi when scanning lines parallel to the scan track and only 1270 for lines perpendicular to the scan track. See this page.

This produces good 12x18 images from 6x9 film, that is, about 6-8x the linear dimension of the film. Examples and a sample image on this page

 

[charjohncarter]

Thanks, Col. Moran, that is a lot to read but really interesting. I'm going to go over it again. It also somewhat answered my question about 'wet' scanning.

 

[flavio81]

Scanner DPI numbers are completely bogus, theoretical in some universe.

I tested a V500 and got 2000 ppi when scanning lines parallel to the scan track and only 1270 for lines perpendicular to the scan track. See this page.

This produces good 12x18 images from 6x9 film, that is, about 6-8x the linear dimension of the film. Examples and a sample image on this page

Thanks for the tests!!

Yes, 6x sounds reasonable. But for 35mm film that means a 14x21cm print, or 5.5x8.4" print.

I agree that for medium format like 6x9 or 6x7 a flatbed is good enough for a 8x10" and even a bit bigger... but then if you took the increased burden of shooting medium format (higher film cost, sort of heavier and slower equipment, less shots per roll) then I think you expect it (as I do) to blow up to huge sizes with fantastic definition.

I recall when my local lab scanner my 35mm film using a Fuji Frontier film scanner. That thing would output 13MP files of pure down-to-the-pixel resolution with excellent color!! Prints using the Frontier system rivalled the prints done optically using an enlarger... at least for color (b&w was better done in the optical domain.)

So 10 years later, the first time I received a scan done using a V750, i was not dissapointed but actually angry !!

 

[Prest_400]

I recall when my local lab scanner my 35mm film using a Fuji Frontier film scanner. That thing would output 13MP files of pure down-to-the-pixel resolution with excellent color!! Prints using the Frontier system rivalled the prints done optically using an enlarger... at least for color (b&w was better done in the optical domain.)

So 10 years later, the first time I received a scan done using a V750, i was not dissapointed but actually angry !!

The frontiers are all the rage, specially when with color neg density is adjusted and the images tend to be airly.

I got back a couple of rolls at 5400x3500px (19MP) from 35mm and was surprised to see how wonderful Portra 400 is; and how rather decent but non super resolving is the Nikkor 28-80 kit zoom in. I used to get Dev+Scan at the normal 3600x2400 (8MP). Heaps of labs do Dev+Scan with these scanners and the results are mostly really nice.

Also the MF rationale can happen here, you'd spend lots if you want Dev+Scan at high resolution... And often supplied JPEGS, TIFF files having a surcharge (let's keep it lossless). But there was someone's rationale which said "Having a roll dev+scanned and manually graded for $20 is very well priced".

 

[Ronald M]

Nikon D800 is 4900 DPI high. that is one inch

My computers have no issue with it so 6000 should be ok.

The whole problem is the scanner does not achieve 6400, probably 3200. It is non focusable and negs are not flat so that is more problems.

 

[user237428934]

Does anyone see a difference between types of film (35mm). I can get lovely scans with TMX film, less lovely with TRI-X and other old style films. My C-41 scans are also good, but not as good a TMAX. I also find that slide film is almost always soft. I think this old slide was focus well but the scan is soft.

I read more than once that modern T-grain film is better to scan. That is my experience too.

 

[charjohncarter]

I read all of the two threads the Sebastian posited in one of his post. He has certainly done some hard work on this subject. I thought I would do an experiment (nothing like the Col. has done). I scanned this an old 35mm Ektachrome slide and isolated a small portion. It was scanned on a V500 @ 2400 dpi, no adjustments. the first was in the cardboard holder from Kodak, the second without the holder but held down with the a Better Scanning ANG for 35mm. The glass on glass in better, but not by much (no Newton rings ???)