Quality comparable to Hasselblad Flextight?

[Jarin Blaschke]

Hello:

I have a library of negatives that go back 28 years, and I’d finally like to look at all of it - my sporadic printing and contact printing has never allowed me to fully see it all. Certainly not all in one place. Formats vary between 35mm and 8x10, although 4x5 counts for the smallest portion and 8x10 is the most recent work, and easiest to keep up with. Lots of 35mm, 6x6 and 6x7.

I am ready to invest in a proper scanner to scan negatives. To scan in the past, I’ve rented a Flextight scanner at the local darkroom by the hour. As I have 16 binders bulging with negatives, obviously I can’t archive everything this way. As the Flextight is discontinued, and difficult to find for sale used, what is a currently available scanner that best approaches its quality? The preference is to scan once and be done, rather than pay for higher quality scans of select images later.

At the same time, when I make prints, I do so in the darkroom and not from digital files. These files are for reference first, but also for sending to galleries and publishers later as master digital versions.

As a separate question, if there is no current scanner comparable to the above on the market, I see from my first cursory research that Epson has two popular scanners that straddle the prosumer/pro market: the V850 and the 12000XL. Is there much quality difference between these, or is the difference in price primarily from heavy-use robustness and flatbed capacity?

Thanks in advance for guiding this ignorant analog photographer.

Best,
Jarin

 

[Helge]

Google camera scanning.
You might want to macro and stitch for print quality.
But very fast workflow combined with far better quality than almost any consumer scanner, can be a achieved with one shot from a pretty good, but not necessarily super new digital camera.

 

[Alan9940]

I don't know of any brand new scanner for up to 4x5 that can match the output from a Flextight. 8x10? The only flatbed scanners that might be a match are the Creo/Kodak IQsmart line or Scitex/Creo line. If you're interested in one of these scanners, I'd probably start at: https://www.scansolutionsonline.com/products/refurbished-flatbed-scanners/

That said, I would second Helge recommendation for camera scanning. I don't know of any market-ready setup for 8x10, but there might be something out there. From 4x5 on down you're good. I use a Canon 5D II with a Sigma macro lens and I'm very happy with results up to about 13x19". If you want to go larger, you need more sensor resolution or stitch multiple shots of one frame.

 

[Helge]

You can get reasonable results to excellent when camera scanning LF depending on technique.
It all depends on your use case.

A Flextight high quality scan of a 4x5 takes a hideously long time.

You can macro and stitch anything but an 8x10 at full quality will take a long time and a lot of tweaking and retries.

 

[Jarin Blaschke]

Thanks. Ok, forgetting 8x10 for just a moment, any suggestions for 35mm and 120, 4x5 as a bonus?

 

[madNbad]

Thanks. Ok, forgetting 8x10 for just a moment, any suggestions for 35mm and 120, 4x5 as a bonus?

https://www.negative.supply/

This is a good resource to learn about camera scanning.

 

[Bronson Dugnutt]

I started with an Epson V600 which has a transparency unit wide enough to handle up to 120 film. If you shoot panoramic aspect ratios you can scan the whole strip in one go. When digitizing with a camera, the more closely your images match the ratio of the sensor the better. I also used it to scan 4x5, but you have to perform 2 scans and stitch them together if you want the entire frame width.

For best image quality I opt for camera scanning small frame sizes (8mm, 16mm) and usually choose for the flatbed for everything else (half-35mm - 8X10) out of convenience, especially color films. Digital ICE is essential for scanning color film IMO plus the software does a pretty good job automatically reversing the negative. Epson's film holders are legendarily bad but if I really need a baller scan I can use the fluid mount or do a gigapixel mosaic via DSLR.

Take time to align and fine-tune everything regardless of the method and you can't really go wrong.

 

[flavio81]

Epson has two popular scanners that straddle the prosumer/pro market: the V850 and the 12000XL. Is there much quality difference between these, or is the difference in price primarily from heavy-use robustness and flatbed capacity?

If you're using an Epson flatbed, you're discarding information. A big amount of information. So, don't use it when you need to fully capture what is on the negative. For web usage they're just fine.

The optics on those scanners introduce smearing and chromatic aberration, that then is compensated for (by the users) at the digital stage by increased sharpening. The end result is a characteristic look that can be easily spotted (when pixel peeping).

They say the larger the format, the smaller the problem, yet even on 6x6 format you're discarding quite a bit of information compared to a DSLR scan or a good drum scan. There is plenty of information about this on the web.

 

[albireo]

They say the larger the format, the smaller the problem, yet even on 6x6 format you're discarding quite a bit of information compared to a DSLR scan

A DSLR is discarding a lot of information even compared to a flatbed scan. To be more precise (though this has been discussed extensively before on these boards) a DSLR scan using an interpolating Bayer or X-trans sensor will discard 2/3rd of the colour information captured on the same surface area by a non-interpolating CMOS line sensor, such as the one in a V550, V700, V750, V800.

A good 6x6 scan, even with an Epson V550 or above, can be better than a poor DSRL scan according to many metrics (resolution is only one of them and for many people not even the most important). And there are many ways to get a poor DSLR scan as the platform's optimisation is completely down to the operator.

As another option, there are dedicated film scanners. A good dedicated film scanner scan will be very close to Flextight scan (it would be extremely difficult to pick them apart in an ABX blind test I wager) and completely obliterate the vast majority of DSLR scans out there. This with an investment ranging from 1/5th to 1/20th of the one needed for a beginner's DSLR scanning setup (if one does not own a DSLR).

It is true though that a DSLR scanning capture sequence will be faster and quieter than with a dedicated film scanner, so if speed and silence are a priority DSLR scans are probably a good option.

 

[Helge]

A DSLR is discarding a lot of information even compared to a flatbed scan. To be more precise (though this has been discussed extensively before on these boards) a DSRL scan using an interpolating Bayer or X-trans sensor will discard 2/3rd of the colour information captured on the same surface area by a non-interpolating CMOS line sensor, such as the one in a V550, V700, V750, V800.

A good 6x6 scan, even with an Epson V550 or above, can be better than a poor DSRL scan according to many metrics (resolution is only one of them and for many people not even the most important). And there are many ways to get a poor DSLR scan as the platform's optimisation is completely down to the operator.

As another option, there are dedicated film scanners. A good dedicated film scanner scan will be very close to Flextight scan (it would be extremely difficult to pick it up in an ABX blind test I wager) and completely obliterate the vast majority of DSLR scans out there. This with an investment ranging from 1/5th to 1/20th of the one needed for a beginner's DSLR scanning setup (if one does not own a DSLR).

It is true though that a DSLR scanning capture sequence will be faster and quieter than with a dedicated film scanner, so if speed and silence are a priority DSLR scans are probably a good option.

The optics and optical path in any flatbed is terrible. From keeping film flat, to sensor noise.
A camera scan takes a bit of care, but not unreasonably much.
A one shot scan with carefully controlled focus and flat film will beat any Epson scan by lengths.
You get rid of almost all Bayer filter artifacts by macro and stitch, or alternatively rotating the film a bit and shooting many shots of the same bit and combining.
You can also handle dense and high dynamic film that way.
Adjust the shutter or aperture.

What doesn’t go away is the weak filter strength of the Bayer filter. The filters are intentionally weakened in saturation to make interpolation and sharpening easier.
That’s why a B&W sensor camera module with “adaptall” mount would be very welcome. Then it would be a matter of filtering the backlight. Sharp/steep filters would make orange mask removal a cinch.

The colour results are still a lot better than any flatbeds.

 

[Bronson Dugnutt]

A DSLR is discarding a lot of information even compared to a flatbed scan. To be more precise (though this has been discussed extensively before on these boards) a DSRL(sic) scan using an interpolating Bayer or X-trans sensor will discard 2/3rd of the colour information captured on the same surface area by a non-interpolating CMOS line sensor, such as the one in a V550, V700, V750, V800.

I'd be interested in seeing digitization attempts on something like a Sigma DP Merrill w/ a Foveon sensor or even a Lecia Monochrom. A bayer-less b&w sensor is probably a gimmick to most, but it would be excellent for process work.

 

[McDiesel]

@Jarin Blaschke If you want an X1 equivalency, then you have no choice, frankly. Camera scanning is the "new drum". I agree with everything @Helge has said above. The state of imaging sensor technology has progressed by leaps and bounds since the last X1 was manufactured. There's no more need for active CCD cooling and signal to noise ratio of a modern BSI sensor is a dream to work with.

In terms of resolution, assuming a high-quality flat-field macro lens and a pixel-shifting 50-60MP sensor, you will easily exceed the X1 for 35mm (pixel shift is not even needed) and match it for 6x4.5, 6x6, and 6x7 with pixel shift but without stitching. For LF you'll have to stitch, which BTW is painless.

Camera scanning gets a bad rep, and has its share of detractors, because it depends hugely on the equipment (camera, lens, light source, copy stand, etc, which will be expensive if you want quality. But also - this is absolutely critical - the output hugely depends on the operator skill. The internet is full of botched camera scans made by beginners with shitty gear.

I have owned almost every scanner commonly mentioned here: Coolscan 5000 ED, Plustek 120 Pro, Epson V600 and V850, and ordered plenty of 16-bit TIFFs from my local lab's X1. Nothing touches camera scanning. Basically any scanner you can get your hands on, and it's definitely true for scanners I had owned, is Flintstone tech in 2022. They work fine as a budget solution to get usable results, but see the 1st sentence in my comment - if you're looking for X1-like output, camera scanning is your only option.

BTW, the best-thing-money-can-buy is also a camera-scanning rig by Phase One, that's what museums use. You can get 80% of what it does for 10% of the cost.

@Bronson Dugnutt Bayer's effect on detail is often exaggerated. Pixel shifting allows digitizing without demosaicing. I've posted results here before, the only noticeable difference was slightly improved grain quality. Basically, Bayer sensors are fine and modern demosaicing algos are great, as long as you have enough resolution. I found that sharpening settings have far, far greater effect on the end result. People sharpen too much.

 

[albireo]

I'd be interested in seeing digitization attempts on something like a Sigma DP Merrill w/ a Foveon sensor or even a Lecia Monochrom. A bayer-less b&w sensor is probably a gimmick to most, but it would be excellent for process work.

That's a great idea. I'd love to see the above as well.

 

[JerseyDoug]

A lot depends on the negative size.

A digital camera scan is optimal for 35mm negatives. With the 4000 x 6000 pixel sensor of my Fiji X-T20 I get the equivalent of a 4000 dpi scan which beats the 2400 dpi (effective) scan of my V600 flatbed both objectively and subjectively.

Using the same camera for a single-shot of a (56x56) Hasselblad negative I get the equivalent of a 1714 dpi scan, and a comparison with a V600 scan of the same negative is pretty much a tossup subjectively. But the camera scan still beats the V600 hands down with regard to speed and dust spots.

I can produce a much better scan of a Hasselblad negative with six digital camera shots of the 56x56 negative stitched to produce a single file but this is time consuming and sometimes produces noticeable artifacts where elements of the image are misaligned.

 

[flavio81]

A DSLR is discarding a lot of information even compared to a flatbed scan. To be more precise (though this has been discussed extensively before on these boards) a DSLR scan using an interpolating Bayer or X-trans sensor will discard 2/3rd of the colour information captured on the same surface area by a non-interpolating CMOS line sensor, such as the one in a V550, V700, V750, V800.

A good 6x6 scan, even with an Epson V550 or above, can be better than a poor DSRL scan according to many metrics (resolution is only one of them and for many people not even the most important). And there are many ways to get a poor DSLR scan as the platform's optimisation is completely down to the operator.

As another option, there are dedicated film scanners. A good dedicated film scanner scan will be very close to Flextight scan (it would be extremely difficult to pick them apart in an ABX blind test I wager) and completely obliterate the vast majority of DSLR scans out there. This with an investment ranging from 1/5th to 1/20th of the one needed for a beginner's DSLR scanning setup (if one does not own a DSLR).

It is true though that a DSLR scanning capture sequence will be faster and quieter than with a dedicated film scanner, so if speed and silence are a priority DSLR scans are probably a good option.

I don't know really which point you're trying to make.

It has almost nothing to do with sensors. The optics and precision on those Epson flatbeds, ALL of them, are incapable of capturing what's on the film. That is, reproducing the image down to the grain level without adding artifacts.

A proper digital scan using a camera or a high-end device, can do it.

 

[flavio81]

A lot depends on the negative size.

A digital camera scan is optimal for 35mm negatives. With the 4000 x 6000 pixel sensor of my Fiji X-T20 I get the equivalent of a 4000 dpi scan which beats the 2400 dpi (effective) scan of my V600 flatbed both objectively and subjectively.

Using the same camera for a single-shot of a (56x56) Hasselblad negative I get the equivalent of a 1714 dpi scan, and a comparison with a V600 scan of the same negative is pretty much a tossup subjectively. But the camera scan still beats the V600 hands down with regard to speed and dust spots.

I can produce a much better scan of a Hasselblad negative with six digital camera shots of the 56x56 negative stitched to produce a single file but this is time consuming and sometimes produces noticeable artifacts where elements of the image are misaligned.

However there IS a difference. The V600 will add artifacts like smearing of high-contrast edges and some chromatic aberration too. The former can be compensated for by increased sharpening: this will also increase the presence of noise, dirt, etc.

Garbage in --> Garbage out.

These artifacts are even visible on reduced size images (i.e. 3MP), and in the worst cases they scream "Epson flatbed!" to discriminating observers.

 

[BCM]

I don't know really which point you're trying to make.

It has almost nothing to do with sensors. The optics and precision on those Epson flatbeds, ALL of them, are incapable of capturing what's on the film. That is, reproducing the image down to the grain level without adding artifacts.

A proper digital scan using a camera or a high-end device, can do it.

I've been researching this as well and am amazed at the level of confusion in the industry including those who don't address light source and lens issues on camera scanning and properly mounting of a negative (fluid mounted) on the V850. There are lots of flatbed comparisons (to very high-end professional machines) but little that compare a camera capture versus a V850 (as an example)? I'd love to see some actual raw scans instead of speculation and supposition.

 

[Helge]

A lot depends on the negative size.

A digital camera scan is optimal for 35mm negatives. With the 4000 x 6000 pixel sensor of my Fiji X-T20 I get the equivalent of a 4000 dpi scan which beats the 2400 dpi (effective) scan of my V600 flatbed both objectively and subjectively.

Using the same camera for a single-shot of a (56x56) Hasselblad negative I get the equivalent of a 1714 dpi scan, and a comparison with a V600 scan of the same negative is pretty much a tossup subjectively. But the camera scan still beats the V600 hands down with regard to speed and dust spots.

I can produce a much better scan of a Hasselblad negative with six digital camera shots of the 56x56 negative stitched to produce a single file but this is time consuming and sometimes produces noticeable artifacts where elements of the image are misaligned.

Either you are not keeping your film flat, or something else is awry. The emulsion is the same, between film types. So the advantage should be the same, all else being the same.

Normal film (ISO 50 - 400) only begins to max out at 8000 dpi (higher speed film still gets nice grain structure from higher res scanning), so it's worth going through the bother to get the stitch shots perfect, for the special photo you want to present as well as possible.

 

[flavio81]

I've been researching this as well and am amazed at the level of confusion in the industry including those who don't address light source and lens issues on camera scanning and properly mounting of a negative (fluid mounted) on the V850. There are lots of flatbed comparisons (to very high-end professional machines) but little that compare a camera capture versus a V850 (as an example)? I'd love to see some actual raw scans instead of speculation and supposition.

There are plenty of comparisons on the internet out there, this is old news, nothing new here.

Yeah you can fluid mount on your flatbed etc etc.This won't compensate for its mediocre optics. V850 isn't a good deal better than the lower models, when compared to other means of scanning like scanning with a digital camera, or with a high end scanner. Sorry.

I'm providing some images just as a teaser of the TON of information out there. As an example:

Versus DSLR scanning:

Versus enlargement: V600 scan versus the scan of an optical print done through enlargmement.

In other words, the real information vs what's captured by the flatbed.

Versus high end scanner (heidelberg tango at the left, which also shows the great recovery of tones done by the Heidelberg). This test is even more revealing since the image is not pixel-peeping but small, web-size.

Just to provide additional reading material:

Good tests by @Les Sarile here:

Best 120 scanning under $1200

Thanks everyone! Looks like the V600 is the best bet to hold me over until I can get my hands on a Coolscan 8000, which by all accounts should take care of both my 35mm and 120 format needs. Looks like it also has autofocus, which is a plus.

www.photrio.com

Interesting thread on this topic at depth:

Film vs. Scanning resolution

So last night, I got to thinking about how much resolution is actually needed to scan film. On the LF forum, the statement was made that the scanning resolution needs to be high enough to see the film grain, and I'm not sure of the "why" for that statement. I went looking, and Fuji very...

www.photrio.com

 

[BCM]

Thank you for the comments. I've read most of that and cannot believe that, based on direct experience, the examples of any V800 series were done after adjusting for focus (BetterScanning, etc) and fluid mount. Laying a negative on the glass or using the stock holders with out those adjustments isn't a fair comparison. TO BE CLEAR, I'm looking at FlexTight, and other high-end scanners so I get that the Epson isn't the be-all but the difference between un-focussed and non-fluid mount negatives seems unfair.

 

[bags27]

I'd be interested in seeing digitization attempts on something like a Sigma DP Merrill w/ a Foveon sensor or even a Lecia Monochrom. A bayer-less b&w sensor is probably a gimmick to most, but it would be excellent for process work.

Good thought. I just tried, but it's a lot harder to get the dp3-M focused down that close than the two cameras I use for scanning: Leica CL + Sigma 70 macro and esp. Fuji GFX-R with Mamiya 645 120 macro. I may keep trying, but it's a lot of work.

 

[Richard Man]

I scanned a lot of negs for David Wing, from 35mm to 4x5, then I convinced him to get his own X1, and he has been very happy. Here are some samples

The Negative Is The Score…

“The negative is the equivalent of the composer’s score, and the print the performance.” – Ansel Adams When I shoot film (which is most of the time), I generally use the film’s “b…

richardman.photo

 

[flavio81]

I've been researching this as well and am amazed at the level of confusion in the industry including those who don't address light source and lens issues on camera scanning and properly mounting of a negative (fluid mounted) on the V850. There are lots of flatbed comparisons (to very high-end professional machines) but little that compare a camera capture versus a V850 (as an example)? I'd love to see some actual raw scans instead of speculation and supposition.

Yes you can use a fluid mount (and adjust its height), however the improvement for the V850 isn't too big compared to correctly adjusted film holder height. This doesn't surprise me at all, coming from the enlarging generation i find fluid mounting (and glass holders) to be a bit overrated, and i'm not alone in this assessment.

Epson Prefection V850 Pro the review - fluid mount

Wet mounting film on the Epson Perfection V850 Pro with the optional fluid mount accessory.

www.stockholmviews.com

Anyways, the tests done by filmscannerinfo show that the max real resolution you can get off the V850 pro is 2600ppi, this is pulling out all stops using better software, adjusting height, using best settings, etc.

Test report film-flatbed-scanner Epson Perfection V850 Pro transparency unit: experiences report; image quality, scanning slides, medium formats

www.filmscanner.info

Compare to dedicated desktop scanners (i.e. Nikon) that easily achieve 4000ppi. And the 4000ppi is easy to get with DSLR/Digital camera scanning. The examples abound, just google. You can even surpass that and theoretically achieve as high a resolution you want, with a digital camera and macro lens, by scanning portions of the film and stitch (it is routinely done by meticulous users too).

Decades ago people said "2000ppi is sufficient, there is not much detail on a negative" when that was the only thing available for little money and people were used to 6MP digital cameras, then Minolta and Nikon brought out the true 4000ppi scanners and the difference was remarkable. One would need a 8000ppi scanning system to really get what is inside most (not all) the best negatives (i.e. b/w fuji acros) -- that would be equivalent to being able to capture about 160 lp/mm.

So, back to the 2600ppi optical scanner -- you're discarding a lot of information. And introducing artifacts too.

 

[NortheastPhotographic]

"Pro" Flatbed (IQSmart/Eversmart etc) > Flextight > Coolscan > Camera scan is a good guide.

We used to do a lot of camera scanning with the best rig we could assemble, Panasonic S1R using pixel shift, 99cri light source, heavy duty copy stand, Sigma 70mm ART macro. The results were often very good, but just as often very problematic. Once we started doing a lot of sheet film the problems multiplied. Flare, geometric distortion, evenness of lighting...at high res there were weird color artifacts.

I'd still probably recommend a camera scan rig for a home user who has time to tinker. The speed can't be argued with.

If you'd like to go high end, a Flextight or pro flatbed will make your life EASY. We acquired an Eversmart Supreme II from Micheal Streeter and they're the best scans I've ever seen. We even still use Negative Lab Pro for conversions. It's as close to one click conversions as they come.

 

[Patrick Robert James]

If you want a scan once and be done then a graphic arts scanner is the way to go. Get one of the good ones. There is enough space on the bed to do multiple larger negs at a time and quite a lot of smaller ones. Setup is going to burn a lot of time though. Once setup, you can go eat dinner or whatever until you need to setup again.

Frankly though if you have that much to do, you are going to be doing it a looooong time, especially if you want the highest quality out of everything.

I'm not convinced on camera scanning yet. With the technical challenges of it, it can be a pain. You really need a laser to align everything. Special lenses, lightsources (if you are doing color). And you have to deal with every neg as an individual entity. Setting it all up is a pain. Plus to get the highest quality out of it, you are going to have to spend some bucks. And don't believe the whole "stitching is a breeze" spiel for larger negs because it isn't. It is the most labor intensive way to scan as well. For one shot 35mm it is pretty damn good though. If the majority of your negs are not 35mm then take a pass.

If you have the scratch you should hire someone to do it after you buy the equipment. Maybe get a college student to come in and do it for you. In 2012 I scanned everything (five stuffed binders) I made before I started scanning everything as I shot it in 2007. It took me more than a month working on it every day, all day, nothing else. And that was just for catalog quality scans just so I could see the images. 16 binders? That is going to be a massive amount of work. It is worth it in the end though. I found all kinds of stuff languishing in those binders.

You should also come up with a system to catalog everything before you start, if you don't already have one, so you can go back later and find a neg easily. Lightroom is by far the best program to keep track of everything. Year shot-roll #-neg # is the easiest most logical way. so 2022-001-36.

Hope that helps you.