jamesdfloyd
Film is cheap therapy!
First of all, my disclaimer that I am new to the party about scanning and just like “what camera, what film & what developer do you like” posting on this forum, I need to retread the scanner question again. This time however it is; “Can you tell me the truth about scanning”?
I have jumped back into film like a “born-again zealot” and I am teetering on the verge of developing my own film and scanning out of frustration with so called professional labs, but the more I ask about scanning the more confused and frustrated I become.
Here are my frustrations:
1) Every lab I talk with has a completely different approach to scanning technology – most treat scanning as merely a contact sheet without chemicals and when you ask them about the settings they use or the true resolution of their output, I get the sitcom response of the sound of crickets.
2) When I find a lab that seems to know what they are doing with scanning, I get the response that they use a scanner that I’ve never heard of before, that only scans b&w less than 6x7 and at that only a low/medium resolution.
3) When I find a lab that has a “reputation” for great technology, their costs are prohibitive.
4) But the common issue between all three issues is that each of the three claims that their type of scanner & methodology is significantly better than the other.
Even on this forum, the back-n-forth makes the topic even more confusing for me.
1) Why would you want to use a drum scanner?
2) I’ve use XXXXXXXXX for several years now and the results are great…well except for shadow details.
3) Why can’t XXXXXX’s new scanners work better than their old scanners?
4) <My personal favorite> XXXXXXX is a great scanner…for the web that is, but you don’t want to use it for a print.
“'And therein, as the Bard would tell us, lies the rub.”
Scanning has been around long enough, that without regard to personal preferences, truisms exist.
What is the truth about scanning? Can anything short of a $10,000 drum scanner produce a scan “worthy of fine-art printing”? Can you actually get good results from a color negative? Can a Canoscan 9000f really produce excellent output for $225? If the Nikon Coolscan 9000 is so good, why has Nikon stopped making it? And the most important question of all; Can anyone really tell the finite difference between a very meticulous scan workflow on a non-drum scanner vs. a drum scanner?
Thanks,
J.D.
I have jumped back into film like a “born-again zealot” and I am teetering on the verge of developing my own film and scanning out of frustration with so called professional labs, but the more I ask about scanning the more confused and frustrated I become.
Here are my frustrations:
1) Every lab I talk with has a completely different approach to scanning technology – most treat scanning as merely a contact sheet without chemicals and when you ask them about the settings they use or the true resolution of their output, I get the sitcom response of the sound of crickets.
2) When I find a lab that seems to know what they are doing with scanning, I get the response that they use a scanner that I’ve never heard of before, that only scans b&w less than 6x7 and at that only a low/medium resolution.
3) When I find a lab that has a “reputation” for great technology, their costs are prohibitive.
4) But the common issue between all three issues is that each of the three claims that their type of scanner & methodology is significantly better than the other.
Even on this forum, the back-n-forth makes the topic even more confusing for me.
1) Why would you want to use a drum scanner?
2) I’ve use XXXXXXXXX for several years now and the results are great…well except for shadow details.
3) Why can’t XXXXXX’s new scanners work better than their old scanners?
4) <My personal favorite> XXXXXXX is a great scanner…for the web that is, but you don’t want to use it for a print.
“'And therein, as the Bard would tell us, lies the rub.”
Scanning has been around long enough, that without regard to personal preferences, truisms exist.
What is the truth about scanning? Can anything short of a $10,000 drum scanner produce a scan “worthy of fine-art printing”? Can you actually get good results from a color negative? Can a Canoscan 9000f really produce excellent output for $225? If the Nikon Coolscan 9000 is so good, why has Nikon stopped making it? And the most important question of all; Can anyone really tell the finite difference between a very meticulous scan workflow on a non-drum scanner vs. a drum scanner?
Thanks,
J.D.
tammons
Established
Will try to keep it short.
In general a museum quality print is 4lp/mm (line pairs per mm) in print and good for a 20" view distance.
Let say your 35mm image resolves 80lp/mm on the negative.
That means you can enlarge 20x and still maintain 4lp/mm
So now to scanning. It roughly takes a 4000 dpi drum scan to resolve 80 lp/mm.
So half the first two digits of your film rez and add 2 zeros.
Technically you have to over scan about 10% to get the true resolution, but scanning 80lp/mm film at 4000 dpi on a drum scanner will resolve 95% of the info so thats good enough for me.
Not all scanners are so efficient. The Nikon scanners are close like maybe a true 3800 dpi, Minolta 5400 is close to a true 5400 dpi if you can keep the film flat.
With a minolta in a typical scan scenario, lets say I am doing an 8% overscan, I would be able to scan and resolve 100 lp/mm film and print at 25x.
An Epson V700 is more like a true 2200-2400 dpi.
With an Epson V700 you can enlarge roughly 11x or 12X and stay within the
4lp/mm standard, but you may have to scan at 4800 dpi and reduce to get it.
The ultimate printer is a lightjet. At 204 dpi it prints at 4lp/mm or museum standard.
So from there it is simple math.
Yes. A nikon 5000, 9000, V700/750. minolta 5400
Yes
Skip it. Mediocre. Better off with a real film scanner like a Nikon 5000 or 9000 or if that is out of the budget and you need to scan big film then a V700
Film is fading fast and Digital cameras are taking over. Not as much profit in a 9000.
If you are never going to drum scan then it is not worth going into. They are complicated machine but can produce superb results.
If you want the best from a scanner, the basics are you need to make sure the film is flat and the scanner is focused properly and when you make adjustments make sure you are not clipping any info. You can check that with a histogram. Adjust color etc. Scan in either 48 or 16 bit tiff, archive then work in PS. If you have a good image and a scanner and make the correct adjustments, and you monitor gamma and balance is correct your raw scans should be almost good enough to print as-is.
I have a lot of drum scan and V750 etc etc comparisons on my pbase site if you want to look around.
http://www.pbase.com/tammons
In general a museum quality print is 4lp/mm (line pairs per mm) in print and good for a 20" view distance.
Let say your 35mm image resolves 80lp/mm on the negative.
That means you can enlarge 20x and still maintain 4lp/mm
So now to scanning. It roughly takes a 4000 dpi drum scan to resolve 80 lp/mm.
So half the first two digits of your film rez and add 2 zeros.
Technically you have to over scan about 10% to get the true resolution, but scanning 80lp/mm film at 4000 dpi on a drum scanner will resolve 95% of the info so thats good enough for me.
Not all scanners are so efficient. The Nikon scanners are close like maybe a true 3800 dpi, Minolta 5400 is close to a true 5400 dpi if you can keep the film flat.
With a minolta in a typical scan scenario, lets say I am doing an 8% overscan, I would be able to scan and resolve 100 lp/mm film and print at 25x.
An Epson V700 is more like a true 2200-2400 dpi.
With an Epson V700 you can enlarge roughly 11x or 12X and stay within the
4lp/mm standard, but you may have to scan at 4800 dpi and reduce to get it.
The ultimate printer is a lightjet. At 204 dpi it prints at 4lp/mm or museum standard.
So from there it is simple math.
Yes. A nikon 5000, 9000, V700/750. minolta 5400
Yes
Skip it. Mediocre. Better off with a real film scanner like a Nikon 5000 or 9000 or if that is out of the budget and you need to scan big film then a V700
Film is fading fast and Digital cameras are taking over. Not as much profit in a 9000.
If you are never going to drum scan then it is not worth going into. They are complicated machine but can produce superb results.
If you want the best from a scanner, the basics are you need to make sure the film is flat and the scanner is focused properly and when you make adjustments make sure you are not clipping any info. You can check that with a histogram. Adjust color etc. Scan in either 48 or 16 bit tiff, archive then work in PS. If you have a good image and a scanner and make the correct adjustments, and you monitor gamma and balance is correct your raw scans should be almost good enough to print as-is.
I have a lot of drum scan and V750 etc etc comparisons on my pbase site if you want to look around.
http://www.pbase.com/tammons
Last edited:
tlitody
Well-known
nikon stopped making them because photographers are cheapskates. They won't pay what it costs and the pros who would have, have all gone digital. That leaves a small number of semi pro or entusiasts who don't buy enough to make it worthwhile for anyone to make them.
A flatbed is useless for 35mm scans. It's OK for MF and works fine for 4x5. Consensus is that you won't get more than 2400spi from a flat bed in true resolution. You do the maths to see how big a print you can get from 35mm or 6x7 from that. i.e. if you want a 20x16 print area you need 4x5 if you are using a flatbed. Anyone who tells you otherwise has low quality standards.
Do you need drum scans? No you need access to a quality flatbed such as a Kodak IQsmart. Or a Nikon 5000 or Nikon 9000. But they are scanners and are the weak link in the process of getting to a print from film via digital.
Minilabs use inbuilt scanners which are great for small prints. Useless for big prints.
Labs with old Kodak 500 scanners were OK. Don't know what most labs use for rollfilm scanning these days but they are not good for big prints. So it depends on size of print you want.
Labs with Imacon/hassy flex scanners require time consuming operator hours and so do drum scans which is why they cost a lot to get scanned.
Suggest you either go digital or setup your own darkroom and print yourself. That way you remove the scanning element completely. Scanning was never a good solution and never will be. Thats one of the reasons so many pros went digtal. You get a better quality print from a digi camera than you do by scanning film unless you pay a lot for highend scanning.
A flatbed is useless for 35mm scans. It's OK for MF and works fine for 4x5. Consensus is that you won't get more than 2400spi from a flat bed in true resolution. You do the maths to see how big a print you can get from 35mm or 6x7 from that. i.e. if you want a 20x16 print area you need 4x5 if you are using a flatbed. Anyone who tells you otherwise has low quality standards.
Do you need drum scans? No you need access to a quality flatbed such as a Kodak IQsmart. Or a Nikon 5000 or Nikon 9000. But they are scanners and are the weak link in the process of getting to a print from film via digital.
Minilabs use inbuilt scanners which are great for small prints. Useless for big prints.
Labs with old Kodak 500 scanners were OK. Don't know what most labs use for rollfilm scanning these days but they are not good for big prints. So it depends on size of print you want.
Labs with Imacon/hassy flex scanners require time consuming operator hours and so do drum scans which is why they cost a lot to get scanned.
Suggest you either go digital or setup your own darkroom and print yourself. That way you remove the scanning element completely. Scanning was never a good solution and never will be. Thats one of the reasons so many pros went digtal. You get a better quality print from a digi camera than you do by scanning film unless you pay a lot for highend scanning.
Particular
a.k.a. CNNY, disassembler
Scanning (and photography in general) is a bit like cooking. If the meal looks and tastes good, nobody cares what went on in the kitchen. Every cook on the other hand has their own tricks that they claim made all the difference.
I have lots of scans done on high end scanners that are practically useless, and much worse than I managed to do on my own Epson flatbed, and Nikon LS5000. Doing your own tests, and establishing a coherent repeatable workflow can help you get the most out of any system. Sloppy scanner operators, and crappy software settings can ruin a scan on the best machine. That is why I prefer to scan myself as opposed paying someone to do it, even if their equipment is better.
ps. the white text in the OP is invisible to anyone using a white background.
I have lots of scans done on high end scanners that are practically useless, and much worse than I managed to do on my own Epson flatbed, and Nikon LS5000. Doing your own tests, and establishing a coherent repeatable workflow can help you get the most out of any system. Sloppy scanner operators, and crappy software settings can ruin a scan on the best machine. That is why I prefer to scan myself as opposed paying someone to do it, even if their equipment is better.
ps. the white text in the OP is invisible to anyone using a white background.
Sparrow
Veteran
What he said, it's mostly the Indian ...
jamesdfloyd
Film is cheap therapy!
Thanks for the replies, this is helping.
Let me make an update to this inquiry; I am most interested in how my questions relate to medium format film and possibly 4x5. Does any of this change because of the larger film area? I shoot 645, 6x9 and I might get back into 4x5.
Let me make an update to this inquiry; I am most interested in how my questions relate to medium format film and possibly 4x5. Does any of this change because of the larger film area? I shoot 645, 6x9 and I might get back into 4x5.
thegman
Veteran
I can't give the in-depth answers above, but I will say that I have recently started using a V700 for 6x7 negs, and I'm happy with it. Maybe I could squeeze out more with a betterscanning neg holder, but the point of weakness in my photography is my photography, not scanning. I think it's a personal thing whether you feel flatbed results are good enough for you. 99% of people are happy with Facebook size shots from a tiny-sensor compact, I think 99% of people would be happy with the results from a cheap scanner. But there is always 1% who are not.
Matus
Well-known
Your best bet is to find a lab like digitalcopy24.de (in Germany) which uses the Nikon Coolscan 9000 and will deliver you 35mm scans well bellow $1 a a frame. That is what I do with 35mm. 6x6 and 4x5" I mostly scan myself with Microtek F1 (could have as well get the Epson V750, seems to give less problems) and send only those images out for scanning that I plan to elarge beyond 5x (4 lines per mm is maybe fine in large prints you view from a distance, but you need more for really sharp print if it is smaller and gets viewed from 10" instead 20").
Truth to be told - If I would not shoot also 4x5" I would probably bite the bullet and get a new Coolscan 9000 while there are some left.
I did a 20x24" print from a 4x5" scan I did with mine Microtek. It looks very nice although close inspection shows noise and bit of missing resolution (it is 360 dpi print). That is the advantage of large format. Prints beyond A3 are usually viewed from larger distance so even scan of medium quality may yield satisfactory results. Smaller photos you often hold in your hand and expect them to be sharp. So an 8x10 print from a 645 neg may not be satisfactory even if the enlargement factor is the same.
Truth to be told - If I would not shoot also 4x5" I would probably bite the bullet and get a new Coolscan 9000 while there are some left.
I did a 20x24" print from a 4x5" scan I did with mine Microtek. It looks very nice although close inspection shows noise and bit of missing resolution (it is 360 dpi print). That is the advantage of large format. Prints beyond A3 are usually viewed from larger distance so even scan of medium quality may yield satisfactory results. Smaller photos you often hold in your hand and expect them to be sharp. So an 8x10 print from a 645 neg may not be satisfactory even if the enlargement factor is the same.
Frankie
Speaking Frankly
My 2+ decades of experience is in serious digital photogrammetry. Image format size 23 x 23cm (9" x 9"...yes inches):
1. Good scanning starts with holding the film flat...a vacuum system is used.
2. The scanner lens must be able to re-photograph a flat [film] scene into a flat image plane.
3. The scanner might have a stationery glass bed with the optics moving, or have stationery optics with the bed moving. In either case, the speed must be steady. No drum scanners ever pass the steady-speed test...too many variables.
4. Such scanners have native resolution of 5u...5/1000mm.
5. The light source must be even and cool...or unknown distortion creeps in during the scan...and measurable.
6. Such scanners starts at $100k, and must be installed in a temperature controlled clean-room...a grain of dust looks like a boulder, a human hair is ~60u.
Now consider the small format scanners we are dealing with here:
A. Almost all of them cannot hold a film strip nor a cardboard mount slide flat.
B. The lens is of unknown quality.
C. 600dpi scanners is only 42u...far from the best in resolution. Many 1200dpi scanner achieved higher dpi numbers via interpolation...cheating. Agfa used to offer a 1200dpi optical resolution one...price $25k.
D. Temperature, light source, scanner motion are not great issues because of short scan time and small scan travels. BUT, dust still is.
E. Almost all film strip holder place some distance from the glass bed surface...out of focus to some degree.
F. No one bother to re-mount slides with glass mounts...the best way is actually half-glass --emulsion side exposed backed with anti-newton glass.
I bypassed most of the issues by:
Or:
Order a CD with each film processed...cheap, and untouched by human hands yet. The scanner such labs used are high-priced industrial grade units...manufactured by the film processing machine maker. Many offers 2000 x 3000 or 12u scans.
1. Good scanning starts with holding the film flat...a vacuum system is used.
2. The scanner lens must be able to re-photograph a flat [film] scene into a flat image plane.
3. The scanner might have a stationery glass bed with the optics moving, or have stationery optics with the bed moving. In either case, the speed must be steady. No drum scanners ever pass the steady-speed test...too many variables.
4. Such scanners have native resolution of 5u...5/1000mm.
5. The light source must be even and cool...or unknown distortion creeps in during the scan...and measurable.
6. Such scanners starts at $100k, and must be installed in a temperature controlled clean-room...a grain of dust looks like a boulder, a human hair is ~60u.
Now consider the small format scanners we are dealing with here:
A. Almost all of them cannot hold a film strip nor a cardboard mount slide flat.
B. The lens is of unknown quality.
C. 600dpi scanners is only 42u...far from the best in resolution. Many 1200dpi scanner achieved higher dpi numbers via interpolation...cheating. Agfa used to offer a 1200dpi optical resolution one...price $25k.
D. Temperature, light source, scanner motion are not great issues because of short scan time and small scan travels. BUT, dust still is.
E. Almost all film strip holder place some distance from the glass bed surface...out of focus to some degree.
F. No one bother to re-mount slides with glass mounts...the best way is actually half-glass --emulsion side exposed backed with anti-newton glass.
I bypassed most of the issues by:
- Using a good high optical resolution scanner.
- Marking the frame/strip position on the glass bed using another piece of film slightly thicker the the negative.
- Holding the film strip down with an anti-newton glass plate...enlargers use such glass plates.
- Remove slides from mounts.
- Blow dust off all surfaces for each scan...I use a $5 Radio Shack solder suck-up delrin tipped rubber bulb, refills are free.
Or:
Order a CD with each film processed...cheap, and untouched by human hands yet. The scanner such labs used are high-priced industrial grade units...manufactured by the film processing machine maker. Many offers 2000 x 3000 or 12u scans.
mob81
Well-known
It's hard to keep the Film flat "35mm". I'm still trying!
I put book over it for hours with little improvement. I will try leave it for a day or so and see.
If anyone has better solution please provide it.
Regards.
I put book over it for hours with little improvement. I will try leave it for a day or so and see.
If anyone has better solution please provide it.
Regards.
tammons
Established
Medium format, IMO either get a Nikon 9000 or the high rez Minolta not in production any more. Both about 2G.
The V700/750 is an option for MF, but to me it just does not return a crisp enough resolution. For 645 vs 35mm you can scan 35mm film at 4000 dpi on a Nikon V or 5000 and 645 on a v750 and get close to the same rez.
For 4x5 the only reasonable options are to buy a used drum scanner or an Epson V700/750.
That said just the purchase of a drum scanning is only the beginning of money spent.
I think there was a polaroid sprintscan ultra 45 pro that was good, but I have never used one.
Maybe a few others here and there, but you want at least 2000 dpi for 4x5.
All that said most normal medium format film and large format film is limited in resolution to about 40-50 lp/mm so a 2000-2400 dpi scanner is fine for that but for 35mm you need a real film scanner up around 3600-4000 dpi.
Even though a V700 will barely resolve 44 lp/mm, IE a tru 2200 dpi, you have to overscan it at 4800 dpi to get it especially with medium and large format film.
Scanning a 4x5 piece of film at 4800 dpi in 48 bit tiff, then reducing it to 2000 or 2400 dpi is agonizing to me.
The V700/750 is an option for MF, but to me it just does not return a crisp enough resolution. For 645 vs 35mm you can scan 35mm film at 4000 dpi on a Nikon V or 5000 and 645 on a v750 and get close to the same rez.
For 4x5 the only reasonable options are to buy a used drum scanner or an Epson V700/750.
That said just the purchase of a drum scanning is only the beginning of money spent.
I think there was a polaroid sprintscan ultra 45 pro that was good, but I have never used one.
Maybe a few others here and there, but you want at least 2000 dpi for 4x5.
All that said most normal medium format film and large format film is limited in resolution to about 40-50 lp/mm so a 2000-2400 dpi scanner is fine for that but for 35mm you need a real film scanner up around 3600-4000 dpi.
Even though a V700 will barely resolve 44 lp/mm, IE a tru 2200 dpi, you have to overscan it at 4800 dpi to get it especially with medium and large format film.
Scanning a 4x5 piece of film at 4800 dpi in 48 bit tiff, then reducing it to 2000 or 2400 dpi is agonizing to me.
Frankie
Speaking Frankly
I am an old-timer in the aerial mapping industry and cut my teeth in very large format films before venturing into digital photogrammetry. One of the many techniques used then was counter-intuitive:
Films curl toward the emulsion side...roll the film strip emulsion side out and wrap it with a 35mm strip of paper. Put the works back into the canister and let it rest a while. Try not to scratch it of course. The film will flaten...but only long enough for an hour or so. Scan it quickly.
Experiment.
Also:
Most films are 0.006" thick. Use a 0.007" (7 thou') to make a film channel and covered it the whole works with anti-newton glass before scanning is like simulating the guide-rail and pressure plate inside the cameras. You can't get better than that.
tlitody
Well-known
How big do you intend print from each format?
Work backwards from your intended output. So if you want to make an a print which is 16x12 inches (actual image area) printed at 300ppi, then you would need a file which is (16*300) X (12*300) = 4800 x 3600.
Then take the long side (4800) and divide by neg size. So from a 6x4.5 neg you actuallly have 56mm (thereabouts and it does make a difference) of neg length. 56mm = 2.204inches. 4800 / 2.204 = 2177. That means you need to scan at 2177 spi to get a file just big enough. On an epson flatbed I would scan at 2400 spi or 4800 and then downsize. On an epson that just about maxes out its capabilities because you won't get more than 2400spi resolution from it.
What if you want a bigger print from a 6x4.5 neg? Well just scann at 4800 and downsize. You will lose a tad of quality unless you want very big prints in which case you had better to go to 6x9 or 4x5 film.
But 4x5 will give you 5*2400 = 12000 pixels on long side and 12000 / 300(print resolution) = 40 inch print. But that assumes you have very high quality neg and its flat and in the focus plane of scanner. In reality you'll only get maybe 2000 real resolution so (5*2000) / 300 = 33 inch print but knock off a bit because image area on a 4x5 neg is not 5 inches long and you start to see that you are entering the twenty something image size from a 4x5 neg.
At those dimensions you will get just as good if not better from a digital camera with plenty pixels because you don't have the problem of using a cheap scanner which is simply not upto getting the most out of any neg or pos.
Bobonli
Established
I have the same frustrations as the OP. Inconsistent scanning and developing from the local shops. Really good scans in NYC area cost a premium. I decided to learn to develop my own B&W and scan with a Plustek 7600. It's certainly not Nikon 9000 calibre but good enough for what I do. If I need better, I can pay to scan an individual negative here and there.
I haven't even developed my first roll yet (this weekend I hope!) and I'm already wondering how difficult could it be to rent time in a darkroom to print my own stuff. As one respondent said: scanning is an intermediate step that tends to diminish quality not improve it unless using the very best ($$$) scanner.
I haven't even developed my first roll yet (this weekend I hope!) and I'm already wondering how difficult could it be to rent time in a darkroom to print my own stuff. As one respondent said: scanning is an intermediate step that tends to diminish quality not improve it unless using the very best ($$$) scanner.
jamesdfloyd
Film is cheap therapy!
My apologies on the white text issue; problems with my work monitor made the text against background disappear - I thought I was helping by changing the text to white.
sojournerphoto
Veteran
Quick reply based on my experience:
- Epson V750 is better than I expected. I can make 6 by 4s for the album and family from 35mm with no issue - I print at 300 or 600dpi, so scan at 1200 or 2400dpi for this. It's also excellent for 'contact' sheets or proofing. With 6 by 7 (mamiya 7ii) I've made some nice 20 by 16s as well.
- Nikon 9000 is what you really need to make good bigger enlargements. It's better with monochrome than the 5000 (had one before) and is able to pull a lot of information from the film. Slow/medium speed 6 by 7 scanned on the 5000 makes lovely prints of any size I've tried (27.5 by 22 is about 400dpi to the printer).
I wouldn't want to try an make big enlargements from 25mm with the Epson.
MIke
- Epson V750 is better than I expected. I can make 6 by 4s for the album and family from 35mm with no issue - I print at 300 or 600dpi, so scan at 1200 or 2400dpi for this. It's also excellent for 'contact' sheets or proofing. With 6 by 7 (mamiya 7ii) I've made some nice 20 by 16s as well.
- Nikon 9000 is what you really need to make good bigger enlargements. It's better with monochrome than the 5000 (had one before) and is able to pull a lot of information from the film. Slow/medium speed 6 by 7 scanned on the 5000 makes lovely prints of any size I've tried (27.5 by 22 is about 400dpi to the printer).
I wouldn't want to try an make big enlargements from 25mm with the Epson.
MIke