Newb advice on scanners

[algarzai]

Hi,
I have spent the past week or so searching through the internet for a good drum scanner to scan my good 4x5 b&w negatives. it drove me nuts
- I intended to print not more that 44X55 inches. viewing distance will be about 20 - 40 inches.
- I am concerned if a company is out of business or if a certain software doesn't run on reasonably dated OS.
- I appreciate options that are below USD 10k. if you feel that this is shy from producing good quality then please propose something with higher price.
- I have questions like is 5000 DPI going to be enough for my needs
- flatbed scanners are quite good but I would rather get a drum scanner unless price difference is huge.

could you help me in suggesting brands and models of suitable scanners.

[Joanna Carter]

Your output requirements are 11x the original size of the neg. Thus, take the output resolution for a typical inkjet printer (240ppi) and multiply by 11, giving you 2640 ppi for scanning resolution, without having to use interpolation in the processing.

However, for the viewing distances you are talking about, 2400ppi should be plenty and a little interpolation will hardly be noticed.

You can get a very adequate print from a prosumer flatbed like the Epson V700 for a lot less than $10,000.

Going to resolutions like 5000ppi is really overkill and will only make for excessively large file sizes. It is doubtful that, if you could achieve such resolutions, anyone would ever notice the difference.

The jump in price from flatbed to drum scanners is significant, as is the extra room required to accommodate a drum scanner, along with all the fluids, mounts and cleaning requirements.

[Peter De Smidt]

The other option would be a refurbished Howtek (or Aztek) drum scanner from Aztek, but this will cost many times what V700 would cost. You might try getting a V700, learn to use it, and then make a print at your final output size. Is it acceptable? You could also have a good drum scanner operator, there are many mentioned in various threads, make a drum scan of the same negative. Now do your editing and compare prints. That would put you in the best position to decide if a drum scanner would be a good purchase for you. Even if you decide that it is, the V700 could be useful to have around, or you could sell it.

[Ben Syverson]

People are basically giving away drum scanners... Join the high end scanning group on Yahoo. You can get a drum scanner for $1000.

The big difference is not price, but workflow. It takes some skill and practice to use a drum scanner correctly, and even when you do, it's very laborious. Wet mounting, scanning and cleaning the negatives really takes some time.

If I were you, I would get a professional drum scan of a favorite 4x5, then do a 4990/V700/V750 scan of the same negative. Print both at your desired size, then make the decision. No need to torture yourself with drum scanning unless it's really necessary.

[Bruce Watson]

I'm a drum scanner owner/operator. I use my scanners to scan my own 5x4 film (both color negative and B&W). I've printed a number of image in the size range you state. My biggest so far is 148.5 cm long. About 58.5 inches for the Americans who haven't seen the metric light ;-) IOW, I know what you are trying to do.



- I intended to print not more that 44X55 inches. viewing distance will be about 20 - 40 inches.

Easily done from 5x4 film.


- I am concerned if a company is out of business or if a certain software doesn't run on reasonably dated OS.
There are no "new" drum scanners (OK, you can by a newly manufactured drum scanner (maybe) but the design for said scanner will be a decade old or more). The amount of drum scanner R&D this century has been little to none. Really. No, I'm not kidding. The number of new drum scanners sold in the world every year would be extremely lucky to break into the triple digits. This has of course driven most of the drum scanner makers out of the market. Even Heidelberg is out. There are only three left. Aztek in the USA, ICG in England, and Screen in Japan. If you really have to have current manufacturers, those are your only choices.

But you'll be passing up some excellent deals. Some of the "orphaned" drum scanners are excellent machines.

Then again, the software to run these older machines isn't being updated, so you'll need an ancient Mac running OS 9.x or older to run it. This isn't a problem either as old Macs are easy to find. Just dedicate the ancient Mac to run the scanner, and have it store the scan file across your LAN to the machine you want to edit with. Easy and nearly painless.

- I appreciate options that are below USD 10k. if you feel that this is shy from producing good quality then please propose something with higher price.
The problem here is that there are plenty of good quality used drum scanners available in the 1-2K USD range. But not usually from the remaining "big three" manufacturers. That's because guys like you put a premium on still being able to get parts and service from the manufacturer, which of course bids the prices up. And if you want to buy new or factory refurb, you'll be at or over your limit quickly.

- I have questions like is 5000 DPI going to be enough for my needs
There are very few films that need 5000 spi. Of those, it takes an excellent optical system, outstanding conditions, the perfect subject, and excellent capture technique to put enough optical information on the film to justify scanner resolutions that high. IOW, scanning above around 4000 spi is nearly always overkill. For LF, I'll go out on a limb and say it's my opinion that scanning any LF film over 4000 spi is always overkill. LF is nearly always lens limited due to the lens design (think coverage) and the small apertures (think diffraction) used to met photographers' DOF requirements.

- flatbed scanners are quite good but I would rather get a drum scanner unless price difference is huge.
Again, depends on your preference. Older orphaned used drum scanners can be had for much less than used flatbeds. But non-orphaned drum scanners can cost as much or more than used flatbeds.

Finally, I second another poster's suggestion that you join the Yahoo group "ScanHi-End" Lots of good resources there, and you can mine the group's history and find answers to a lot of your questions. It's almost a certainty that if you have a drum scanner question, it's already been asked and answered (many times) in that group.

[Nathan Potter]

I like the comments above by Joanna. I need to scan 4X5 chromes and negatives for inkjet printing. I'm a digital duffer; have always done analogue work. So last week I went out and bought an Epson V750 Pro scanner thinking that I could at least get a bunch of my stuff on paper and of good enough quality for medium sized prints. I'm a systematic kind of guy so I was curious just what kind of resolution could be obtained from the noisy little machine. At first cut it appeared sort of like the usual consumer junk. But see my results below.

The following data was obtained using a chrome on glass resolution plate from Toppan Printing Co. Japan. Linewidths on plate range from 20um down to 2 um.
These are line space pairs running in X and Y so resolution difference as a function of direction can be determined. This is a lot like the USAF target available from Edmund Scientific.

Where above the glass platen is the point of best focus?
Turns out I get 2.9 mm. for my machine, but this moves slightly over a 4X5 frame - (about 0.5 mm). I was still within 10um (2540 DPI) resolution from 2.2 to 3.7 mm. above platen. At my measurement extremities - on the platen >20um; 5.1 mm above the platen >20um. 20um is equivalent to 1270 SPI. Don't put your image on the platen if you want the intrinsic resolution of the machine.

What is the best resolution at point of best focus?
Point of best focus yields about 7.0 um lines and spaces - (3607 SPI), or in another metric 71 lp/mm. This of course is a judgement call so I printed the target at 50X using PS. Used a levels adjustment to get a black bar with a completely defined white space between. No sharpening in scan or PS. Bicubic sampling for the printout at 300DPI. For me it is the printed image that matters. BTW with some PS sharpening I could squeeze the res. down to 5um lines and spaces (5080 SPI). Predictably the line pairs running in the direction of the scan were a couple of um better than those orthogonal to the scan.

The Dmax was a bit more disappointing. Using a Stouffer step wedge and a single scan and using a levels adj. (input 0-1.00-255; output 24-227) I could squeeze step 1 to 15 (0.05 to 2.1) onto paper using a crappy HP Deskjet D4260 with Vivera ink.

I think Joanna has a good point above, although I believe 11X is larger than most here would advise. But there again the detail you are after is a subjective thing.

Nate Potter, Austin TX.

[algarzai]

Joanna
Peter
Ben
Bruce
Nathan

thanks guys for your valuable input. It seems I was underestimating the Epson flatbeds. I guess I will try this for a while and concentrate on finding the sweet spot of configs for the film, scanner and printer. I will evaluate the need for a drum scanner then.

[Frank Petronio]

Most people seem to think that if you get a drum scanner then you can't own an Epson too. The truth is that I bet most drum scan owners have an Epson (or similar consumer flatbed) kicking around because they will need to scan different size films and prints for editing or quick+dirties, etc.

So start with an Epson to learn the ropes and do a little tire-kicking with the big stuff.

In the end it may make more sense to scan your own medium-quality stuff on your $500 Epson and then send out the big stuff to people like Bruce and Lenny for the occasional drum scan. How many giant prints can you afford and is your work even good enough to justify it?

[sanking]

Most people seem to think that if you get a drum scanner then you can't own an Epson too. The truth is that I bet most drum scan owners have an Epson (or similar consumer flatbed) kicking around because they will need to scan different size films and prints for editing or quick+dirties, etc.

So start with an Epson to learn the ropes and do a little tire-kicking with the big stuff.

In the end it may make more sense to scan your own medium-quality stuff on your $500 Epson and then send out the big stuff to people like Bruce and Lenny for the occasional drum scan. How many giant prints can you afford and is your work even good enough to justify it?

Excellent advice. Virtually everyone I know who owns a high end scanner also has an Epson close by do editing.

Plus, as Nate and others have observed, the Epson V700/V750 is a pretty decent scanner. At the very least you can learn a lot about as digital work flow without a huge expenditure of cash.

For another test of the Epson V700/V750 see http://www.filmscanner.info/en/FilmscannerTestberichte.html
Their tests show resolution of about 2300 spi, which is what I have observed in my own testing of the V700.

Sandy King



http://www.filmscanner.info/en/FilmscannerTestberichte.html

[stevebrot]

In regards to the real world scan resolution of the V700, my experience is the same as Sandy's and the German scanner review. Things seem to fall apart real fast above 2400 dpi. It sounds like Nathan has done quite a bit of work fine tuning his technique and is getting better results. His message read like he was going to include some photos, but I guess those got left out.

Here is a link to a side-by-side test I did for the Pentax Forums between the V700 and the Nikon 5000 ED. Sorry that it is 35mm film and not LF, but the principle should be the same when comparing full resolution crops from both scanners. Be sure to look at the follow-up tests as well as the initial images:


http://www.pentaxforums.com/forums/post-processing-printing-software-darkroom/103002-scanner-comparison-images.html

The Nikon is pretty much the gold standard for 35mm consumer grade film scanners and is one of the few that comes close to its advertised resolution. It is often the more expensive option for users considering the Epson V700 and V750 scanners for 35mm work, hence the rational for the comparison. For those not willing to wade through the thread, my conclusion was that film flatness and focus are critical to get the best from the V700. Even with close attention being paid to flatness, at settings above 2400 dpi, the Nikon simply pulled away from the Epson. That being said, I might have had better results with a focusable film holder and mounting to glass or wet mount. There are also a few posts from other users comparing 120 film negatives using the V700 and the Nikon 8000 showing similar results.

I did some additional inquiry that I did not post on the Pentax Forums thread. It was possible to increase apparent sharpness by adjusting both global and local contrast (clarity) and adding significant sharpening (USM) in PP. This was done at the expense of tonality and increased artifact. The results looked sort of nasty on-screen, but may have been fine when printed. YMMV.


Steve

[Keith S. Walklet]

Great info Nathan. I agree that the DMAX of the V750 isn't on par with other scanners (NIKON 8000 and TANGO in my comparisons), but I was able to squeeze some more out of mine by boosting the gamma to 3.0. So, if you're interested, give that approach a whirl. The resulting scans look washed out, but capture much greater detail in the shadows than the default gamma, and I can correct for density and contrast in post production.

[Nathan Potter]

Well, a bit more information. I wanted to check the difference between scanning at 1800, 2400 and 4800 SPI. This at the best focus point of my machine which is .110 inch (2.75 mm) above the glass platen. I scanned using the B&W Negative Film setting and Positive Film setting - not significant difference between the two. Used the Toppan glass mask again.

1800 SPI - 13um scan direction - 17.5 um orthogonal
2400 SPI - 6 um scan direction - 8 um orthogonal
4800 SPI - 4 um scan direction - 7 um orthogonal

The line pairs were read off a printed copy of the mask at about 50 X and the imaged adjusted using levels only in PS. No sharpening anywhere. The logic only being that I wanted to know how fine a resolution could be obtained on print paper with a criteria of a good black line and a clear white space.

As mentioned previously in a PM to a member, I calibrated the linewidths using a Leitz substage reticle on a Zeiss Axiomat microscope and reconfirmed using a Tencor profilometer matched to an NBS standard - just to be sure of the dimensions.

Yes, when I get my act together I'll try to post the 50X images - well at least one.

Nate Potter, Austin TX.

[Nathan Potter]

Great info Nathan. I agree that the DMAX of the V750 isn't on par with other scanners (NIKON 8000 and TANGO in my comparisons), but I was able to squeeze some more out of mine by boosting the gamma to 3.0. So, if you're interested, give that approach a whirl. The resulting scans look washed out, but capture much greater detail in the shadows than the default gamma, and I can correct for density and contrast in post production.

Indeed, photomultiplier detectors in Aztec machines cannot be beat for obtaining high Dmax values AND lower noise due to the extraordinarily high gain from electron multiplication. (Gains even of a million are possible).

But a big caveat in all very high resolution scanning is what kind of detail does one have on the original image. Hell, I see under good microscopy, (a Wild M3 binocular microscope) that a lot of my chromes are not better than 15 lp/mm. These are certainly fine for the V 750.

Nate Potter, Austin TX.

[stevebrot]

Thanks Nathan. I understand better now what you have done. The 3600 dpi sounds pretty good to me. Am I correct in reading that the focus point varied over a range from 2.65mm to 3.15mm over a 4x5" scan and that the 3600 dpi is not reliably attainable across the full scan frame? The reason I ask is to help me determine whether it is worth purchasing a focusable negative holder such as that from betterscanning.com.

BTW...what were you using to support your target plate?


Steve

[stevebrot]

...But a big caveat in all very high resolution scanning is what kind of detail does one have on the original image. Hell, I see under good microscopy, (a Wild M3 binocular microscope) that a lot of my chromes are not better than 15 lp/mm. These are certainly fine for the V 750...

:D :D :D

So true! It isn't until we try to go big with our prints that we realize that our images are not really that great in the first place.


Steve

[urs0polar]

If you don't want to throw your V700/V750 out the window in frustration at fuzziness, get the betterscanning holder or something similar. My focus point happens to be 1.8mm above the bed; they are all different.

[SCHWARZZEIT]

For another test of the Epson V700/V750 see http://www.filmscanner.info/en/FilmscannerTestberichte.html
Their tests show resolution of about 2300 spi, which is what I have observed in my own testing of the V700.

Sandy King
One thing to consider when interpreting these resolution figures is contrast. A scanner target like the USAF1951 retains very high contrast up to the higher spatial frequencies, much higher than what can be achieved photographically through a lens on film. The overall appearance of sharpness in a scan depends on how well the contrast of the film is modulated by the scanner.

Even though the V700/750's top resolution for high contrast details might not be that bad at an effective 2300ppi the contrast modulation is extremely soft, at least from the samples I've seen so far. The low contrast details are blurred out from the image. It's very obvious in the samples Steve posted at the Pentax forum. It would be great to see if this scanner can do much better. It's much more significant to see how a scanner performs with a very low contrast target. Take a Tango on the other hand with it's very limited max resolution but being a drum scanner it's still very good at rendering low contrast detail.

When I tested the resolution of my ICG drum scanner with a scanner target it maxed out at 180 lp/mm which seems impressive. But when scanning photographed resolution test charts (contrast ratio of approximately 1:8) with 35mm film I wasn't able to capture all the detail from the film with my scanner. For example I was able to extract 80-90 lp/mm from the low contrast Fuji Pro 160S while the film holds a little over 130 lp/mm as can be seen under a microscope at 100x magnification.

What I'm trying to communicate is that low contrast resolution of a scanner is far more important for the overall image quality than scanner target results.

-Dominique

[Nathan Potter]

Dominique, yes very good points. I've not been able to make any conclusions about the contrast of my V750. Would not even know how to do that. I was at first only interested in the resolution obtainable at any contrast and secondarily how much damage I would do to the resolution by artificially increasing the contrast or employing other digital enhancements using software (Photoshop). As part of trying to figure out what the maximum resolution might be I also looked at the size of the capture pixels at 2400 SPI. These are about 5 um squares as I view them at about 5000X. Thus 2 to 4 pixels are required to characterize a line or space say 8 um wide. My criteria for resolution is that there be, more or less, at least a clear pixel and a black pixel to define a line or a space. Of course all this is quite arbitrary in that the actual pixels as viewed are generated by software so do not, I presume, represent the hardware reality.

Mark, for the money, this is a very nice little scanner, considerably better than I imagined. I'll keep it.

It seems to me that for doing wet mounting one will need an intermediate piece of glass in order to raise the film above the platen to the point of best focus. Then the film would need to be held flat using a top piece of glass on the emulsion side of the film. Is anyone doing this?

Nate Potter, Austin TX

[Peter De Smidt]

<snip>
It seems to me that for doing wet mounting one will need an intermediate piece of glass in order to raise the film above the platen to the point of best focus. Then the film would need to be held flat using a top piece of glass on the emulsion side of the film. Is anyone doing this?


You can also get a piece of glass, wet mount the negative emulsion down under the glass, and use shims under each corner of the glass to get the emulsion at the ideal height. The piece of mylar used for wet mounting will need to be fairly large to generate enough tension to hold the negative flat. It's also a good idea to place a mask on the top side of the glass to block extraneous light. Using the preceding method with a Canon 9950F lead to a fairly substantial increase in quality. Doing it with my current scanner, a Screen Cezanne, did not.

[PenGun]

The V700/V750 are pretty good scanners. To the point that the drum scan crowd is threatened by them. Keep that in mind perhaps when reading the responses.

[Keith S. Walklet]

It seems to me that for doing wet mounting one will need an intermediate piece of glass in order to raise the film above the platen to the point of best focus. Then the film would need to be held flat using a top piece of glass on the emulsion side of the film. Is anyone doing this?


Nate, with my workflow, I wet mount a piece of 3/16" glass directly to the scanner platen, then wet mount the film on top of that. My glass is not adjustable like the custom film holders, but I found that it was at the optimum level for my particular scanner. I reasoned that by wet mounting both the glass and film, that I created a continuous fluid envelope and eliminated multiple surfaces that introduce some level of image degradation.

As I've stated in other threads, the best solution would be for the platen itself to be adjustable. Perhaps some brave soul will undertake that modification.

Something else you might want to investigate is the effect of a layer of diffuse mylar on top of the film. This idea was shared by another photographer and showed improvement in my own results, and I can only speculate why it did? Softening the light source? Introducing the appearance of grain? I'm not sure, but I could see improvement in the results.

[Bruce Watson]

The V700/V750 are pretty good scanners. To the point that the drum scan crowd is threatened by them.

Right. I fear for my masculinity anytime I get too near to one. Or is it more like light to a vampire? IDK, but it sure is threatening! Somehow. Really?

Back to the topic at hand, I find that the Epson consumer flatbeds do an acceptable job, as long as you don't print too large. Up to around 4x enlargement they are fine. I've got a 4x enlargement of one of my own photographs, framed and hanging on my dining room wall, scanned with an old Epson 2450. That one has actually sold a couple of copies, a rarity for me.

But for prints larger than around 4x enlargement, I do use a better scanner. If I didn't already own a drum scanner, I might be looking to use a professional flatbed like a Kodak/Creo/Scitex IQ Smart, or maybe a Fuji Lanovia, or Screen Cezanne, for enlargements up to around 8x. For 8x and up, I really do want a drum scan. Clearly it's worth it to me or I wouldn't own one. But as always, YMMV.

[PenGun]

Right. I fear for my masculinity anytime I get too near to one. Or is it more like light to a vampire? IDK, but it sure is threatening! Somehow. Really?



Yup. Say something good about A v700/v750 and watch what happens.

You can do searches to see if you care.

[stevebrot]

...I find that the Epson consumer flatbeds do an acceptable job, as long as you don't print too large. Up to around 4x enlargement they are fine. I've got a 4x enlargement of one of my own photographs, framed and hanging on my dining room wall, scanned with an old Epson 2450. That one has actually sold a couple of copies, a rarity for me.

But for prints larger than around 4x enlargement, I do use a better scanner. If I didn't already own a drum scanner, I might be looking to use a professional flatbed like a Kodak/Creo/Scitex IQ Smart, or maybe a Fuji Lanovia, or Screen Cezanne, for enlargements up to around 8x. For 8x and up, I really do want a drum scan. Clearly it's worth it to me or I wouldn't own one. But as always, YMMV.

I guess it comes down to the intersection of time, skill, money, computer resources, and needs. For myself:

I seldom need more than the V700 is capable of giving
I don't have the money for the next step up for 4x5 negatives (commercial flat bed or drum)
I don't the time to develop drum scanning skills or to devote to the drum scan work flow
I don't have the computer resources (hardware/software) to work with with 200+ Megapixel images

There it is in a nutshell. The first point is the most important. With the V700 I can easily support 16x20 prints from a 4x5 negative and, with care, 32x40. With 6x7cm (120 roll film), the situation is not quite as rosy, but I can still go as large as 18x22 with good technique. The day when I have a need to go bigger may come, but the number of images that will fall into that category...well, simply put, I am not that good a photographer.


Steve

[algarzai]

guys i am ever so thankful for this rich discussion!
I have concluded that Epsons will produce good quality with proper technique up to certain level which covers about 80% of my needs. It seems this is what I am going to do.