MF scanning dpi

I am assuming you mean on a flatbed. I don't have to scan very many but I set my Epson to scan at 300dpi but more importantly set the output print size to anything larger than the intended print. So when the scan is to be printed I will have a lot of leeway in the final resolution. It can make for large files but you will have a lot of latitude to play with when setting up the image to be printed. It also means I can largely ignore the 'unsharp mask' and therefor cut down the more obvious grain.
 
Last edited by a moderator:
I am assuming you mean on a flatbed. I don't have to scan very many but I set my Epson to scan at 300dpi but more importantly set the output print size to anything larger than the intended print. So when the scan is to be printed I will have a lot of leeway in the final resolution. It can make for large files but you will have a lot of latitude to play with when setting up the image to be printed. It also means I can largely ignore the 'unsharp mask' and therefor cut down the more obvious grain.
I presume you mean 300dpi as a target output resolution of a specified size of print, rather than the actual res the negative is scanned at?
 
Optical resolution on flatbed scanners like the V550 is no where near 3200 DPI, so it's likely better to stick to 1800 to 2400 and then upscale in something far better at the task, like photoshop. I personally scan LF frames at 2400 for B&W and 1200-1800 for colour frames - though now I have a bigger HDD, so will probably choose 2400 for everything.
 
Optical resolution on flatbed scanners like the V550 is no where near 3200 DPI, so it's likely better to stick to 1800 to 2400 and then upscale in something far better at the task, like photoshop. I personally scan LF frames at 2400 for B&W and 1200-1800 for colour frames - though now I have a bigger HDD, so will probably choose 2400 for everything.

To be honest I've never bothered testing it, files are always resized down for printing/web but I suppose I might be adding to my scan times/file sizes a bit!
 
Well you can't go wrong scanning on an Epson flatbed e.g. V550 at 2400 (or higher) dpi for 35mm or 120 as although the true optical resolution is about 1800 dpi (for V550)..the software crams in more pixels per sq centimetre.....so the scan looks better esp for crops.
Anyway scan at say 500, 1200, 1800, 2400 and 3200 dpi and see what results you like.
 
Last edited:
I did some comparison tests when I first got my Epson V600 scanner and found 3200 dpi was the sweet-spot, anything above or below this seemed to have progressively slightly less detail/sharpness. The difference may be more apparent on 35mm but I scan 120 at 3200 dpi to get the best out of it. @Stegosaurus Perhaps do some tests with your scanner and do some pixel peeping to work out what's best for you. (y)
 
Last edited:
Thanks guys. I have an Epson Perfection 3170. I think the optical resolution is only about 1200 or so.

The software will make up for it so you should still get good results scanning higher dpi, after all the adverts for scanners going up to 6400 dpi must be true (cough porky).
 
Epson V700, 3200 dpi for both MF and 35mm. Resolutions above seem only produce bigger files and no real improvement in detail. Also 120 scanned at a higher resolution is harder to navigate when I'm spotting, which I do zoomed in to full resolution in Photoshop. I drop the resolution to 2400 for LF otherwise the files become unmanagably large.
 
Epson V700, 3200 dpi for both MF and 35mm. Resolutions above seem only produce bigger files and no real improvement in detail. Also 120 scanned at a higher resolution is harder to navigate when I'm spotting, which I do zoomed in to full resolution in Photoshop. I drop the resolution to 2400 for LF otherwise the files become unmanagably large.

The optical resolution for a scanner is fixed, even the Fuji frontier can only scan at about true 2600 dpi, well film has more detail than this and no one has bothered to produce a joe public easy to use scanner to match what you can get out of film and guessing maybe one reason is:- joe public printers can never produce the details that you can get from film so why bother and of course it's all digi now.
I wonder what Nasa used to scan the film shots (assuming they did)
 
So is the consensus to allow the scanner to do the interpolation, or to do it in Photoshop or some such afterwards?

Well if you have an old cheap scanner that's all you can do..I got VG results from an Epson 2480 and you will only see the difference if you compared the results from say a Fuji frontier.....or look at it this way:- what you don't know you won't miss ;)
 
I scan at maximum resolution with my V800 flatbed. My reasoning is to scan once and never again. As I use VueScan I can save the raw scan and reprocess it as many times as I want with different settings to get the best scan. I find this very useful as some changes to settings can give a big difference. Yes, my scans may be 1.3 gb (I think it's gb...) but so what? My negatives are big as well. :)

I do sometimes use VueScan's interpolate down function to average adjacent pixels to reduce noise, but this can be done post scanning. I never have to interpolate upwards to make a print, but I don't use 35mm, and any scanned 35mm don't go above A4 except in "emergencies".

For prints, I stick to the limits that I worked out when darkroom printing: that in a few cases 35mm would produce an acceptable 10x8, but was better not pushed beyond whole plate (6.5 x 8.5 in) and regard the maximum print size to be 5 times the negative size (and less is better). I know most people regard this as hopelessly conservative. I hate grain in a print, and any unsharpness that is not a result of depth of field.
 
Last edited:
I scan at maximum resolution with my V800 flatbed. My reasoning is to scan once and never again. As I use VueScan I can save the raw scan and reprocess it as many times as I want with different settings to get the best scan. I find this very useful as same changes to settings can give a big difference. Yes, my scans may be 1.3 gb (I think it's gb...) but so what? My negatives are big as well. :)

I do sometimes use VueScan's interpolate down function to average adjacent pixels to reduce noise, but this can be done post scanning. I never have to interpolate upwards to make a print, but I don't use 35mm, and any scanned 35mm don't go above A4 except in "emergencies".

For prints, I stick to the limits that I worked out when darkroom printing: that in a few cases 35mm would produce an acceptable 10x8, but was better not pushed beyond whole plate (6.5 x 8.5 in) and regard the maximum print size to be 5 times the negative size (and less is better). I know most people regard this as hopelessly conservative. I hate grain in a print, and any unsharpness that is not a result of depth of field.

The Epson 700, 750 and 800 are not far off from a Fuji Frontier (comparisons on the net)....it's probaly more down to getting the shadow detail off the film.
I suppose there are comparisons showing the difference between the Fuji frontier and a drum scan if anyone wants to look, but for me I don't want to pay for a drum scan and happy with my Epson V750....and again what I don't know I won't miss.
Anyway the problem is scanning 35mm and not 120 and larger film.
 
regard the maximum print size to be 5 times the negative size (and less is better).

Sounds good to me does that, however I have a question which relates more to wet prints rather than scans as of course one can't up the resolution of a neg sat in an enlarger.

If my assumption is correct then you base the idea on sq area of the neg?

Thus 5x4 = 20 inches sq

5 x 20 inches sq = 100 inches sq

that would comfortably make a 'good' 10x8 print ( 80 inches sq) or a 'decent' 9.5 x 12 print ( 114 inches sq.)


I typically print 12x16 prints ( 192 inches sq) so effectively I'm pushing your (possibly conservative) ideals well beyond their limits ( ie near to 100% more)

Does this now mean that I need to sell all my kit barring the 10x8 outfit? :runaway::runaway:

I think , in fact Im sure that you have much higher standards for your prints than I so the fact that I'm happy enough with most of my work is probably due to my slack attitude ( It simply follows on from when i was at school many moons ago :LOL: ), however I am now a little concerned about obtaing a good quality result as a wet print from my 5x4 exposures :(

Perhaps i have totally misunderstood and your '5 times the negative size' relates to external measurments , 5x4 inch neg equating to a 25 x 20 inch print. If so, it will be good news as my heartbeat will return to a slower pace and I'll sleep soundly this evening :ROFLMAO:
 
There's always the "linear or area" question, so I should have been specific. For me, in this context, 5 times means linear, so 12x16 from a 6x7 negative, 20x25 from 5x4 etc. With scanned negatives, I would be happy possibly pushing it further. I can produce an A3 (12x16 near enough) from 6x7 where I can see extra detail under a magnifying lens than I can't see unaided. For those unfamiliar with my physical infirmities :) I'm short sighted, and can read comfortably print a hand span away from my eyes.
 
Last edited:
There's always the "linear or area" question, so I should have been specific. For me, in this context, 5 times means linear, so 12x16 from a 6x7 negative, 20x25 from 5x4 etc. With scanned negatives, I would be happy possibly pushing it further. I can produce an A3 (12x16 near enough) from 6x7 where I can see extra detail under a magnifying lens than I can see unaided. For those unfamiliar with my physical infirmities :) I'm short sighted, and can read comfortably print a hand span away from my eyes.
Saved!(y)
Thank the lawd for that!:LOL:
 
I did some comparison tests when I first got my Epson V600 scanner and found 3200 dpi was the sweet-spot, anything above or below this seemed to have progressively slightly less detail/sharpness. The difference may be more apparent on 35mm but I scan 120 at 3200 dpi to get the best out of it. @Stegosaurus Perhaps do some tests with your scanner and do some pixel peeping to work out what's best for you. (y)

This really, it doesn't make a lot sense to me but on my V550 I find 3200 to be the "sweet spot". On some negs it seems not to make a lot of difference between 1600 and 3200 but on others it is very noticeable so that's what I tend to go for. It does make for large files though but then you have the negative so after finishing in LR/PS a sensible sized file can be saved and the original scan deleted.
 
This really, it doesn't make a lot sense to me but on my V550 I find 3200 to be the "sweet spot". On some negs it seems not to make a lot of difference between 1600 and 3200 but on others it is very noticeable so that's what I tend to go for. It does make for large files though but then you have the negative so after finishing in LR/PS a sensible sized file can be saved and the original scan deleted.

Well one thing for sure for 35mm:- I've had 100s scanned by Asda, Tesco and filmdev all at low scan i.e. @ 1800dpi and if you do any cropping the results get worse the greater the crop.so you have to scan higher to cram the pixels in if cropping.
And just to add after recently doing about 30 A4 prints you can get quite a good print from a 35mm neg scanned at 1800 dpi (shot not cropped)...and what is a good print? A print that the family or friends see with no complaints about quality.
 
Unless photographing for other people, one that you as the tog are content with! ;)

AAMOI as the last time I was in the darkroom scanners weren't invented o_O....For 35mm, Is a darkroom print (using a VG enlarger lens) equivalent of a drum scan, Fuji Frontier scan or whatever and then using a pro printer for the same size print to compare to see if any difference OR another way of looking at it is:- can a drum scan get more detail off the neg than an enlarger lens can.
 
Having read Ctein's comments about the tests he made on enlarger lenses (Post Exposure), I'm almost inclined to go with a drum scanner, even though I don't know their claimed or actual resolution figures. The test results make for sorry reading.
 
AAMOI as the last time I was in the darkroom scanners weren't invented o_O....For 35mm, Is a darkroom print (using a VG enlarger lens) equivalent of a drum scan, Fuji Frontier scan or whatever and then using a pro printer for the same size print to compare to see if any difference OR another way of looking at it is:- can a drum scan get more detail off the neg than an enlarger lens can.

i personally don’t see that they are comparable.
One is made produced via millions of digitized pixel thingies which then in turn communicate with a printer which in turn vomits ink onto some specific paper.
The enlarger lens projects an image onto a piece of specific paper which in turn gets itself on a high whilst going through a 100% chemical process ( I just get the vapours lol)

The two processess are worlds apart and the resulting prints , although from the same negative are never going to look identical in tones etc.
I now have a number of prints duplicated out of my interest to see which result I actually prefer.... inkjet print or wet print.
There is no preference globally as it very much depends on the actual photograph ( subject matter, tones, exposure etc etc ) .... some look better ( imo) wetprinted, others look better scanned and inkjet printed.

Trying to compare the two like you are suggesting to see which is ‘better’, isnothing short of the digi/ analogue camera comparisons.
 
As a counter argument, conventional prints (and negatives) are no more continous tone than their digital equivalents. Analogue photography (at least in black and white) depends on grain thingies, that at a suitable viewing distance give the illusion of continous tone.

If you read the old books on photography, back in the dry plate era, the holy grail of analogue prints was to achieve permanence, and carbon (pigment, if you like) was seen as the best way of doing it. Carbon prints apart, this had to wait until inkjet printers with pigment inks came along. :exit:
 
Trying to compare the two like you are suggesting to see which is ‘better’, isnothing short of the digi/ analogue camera comparisons.

Well Asha I was interested in the ability to get detail off the neg rather than compare analogue to digital. ;)
 
I do sometimes use VueScan's interpolate down function to average adjacent pixels to reduce noise, but this can be done post scanning...

My Plustek 7500i (I know it's 135 only, but bear with me) has stopped doing multi-scans reliably (registration of second or subsequent passes can be off). I used to use multiscans quite a lot, 2 or 4 passes, and it seems like the results are very much smoother. Recently I've been scanning at twice the desired resolution and using the size reduction feature that Stephen's referring to (in Vuescan Pro's Output tab). I think the results are better than a straight scan.

My general rule for scanning 135 has been 2400 samples per inch for reasonable frames, 1200 for "placeholders", 3600 for the better ones. At the 3600 setting I'm probably actually achieving something reasonably close to that figure in effective resolution (filmscanner.info reports a real resolution of around 3500 spi).

Scanning at 2400 spi, I'm getting a frame a bit less than 2400 x 3600 pixels. Printing at 300 dpi, a 12x8" print seems reasonable, and I've had very nice-looking A4 prints, and A3 from the higher res scans. Any prints are for looking at, not inspecting!

I don't feel that I need to achieve many more pixels per frame with 120 than with 135. Others clearly differ, but again, huge detailed high resolution has never been a target for me. So if I have a "reasonable" 6x6 frame (which is roughly 2 1/4" square, of course), I'd aim for something like 3600 pixels each way, which works out at 1600 pixels per inch (I can't remember what the nearest nominal resolution is on my V500). Again, I might well scan at twice the resolution and then size reduce with Vuescan, though I've not scanned enough MF to be sure if it's worth it.
 
Back
Top