Sharpness vs. Resolution vs. MTF vs. Diffraction

[georgl]

I just wanted to post a reply on the "f64"-thread which seemed to raise very basic questions - so I've decided to give it it's own topic.
I hope it's well understandable, and helps to clear up certain missconcepts:

When we speak about "sharpness" we usually don't mean extinct resolution at high contrast from the lens, the film OR the scanner.

What's important is MTF - which basically means contrast at a certain resolution/ frequency. As most of us know, lenses have a MTF, but film and scanners also have MTF - they all "render" detail with a certain contrast.

Of course there is no "absolute" MTF-value for 4x5" or 8x10" - but let's play with a few values that are common:

- 20lp/mm frequency, which results in about 5000 "pixels" resolution @ 4x5", a frequency which is barely visible at 4x magnification (5lp/mm)

- 40lp/mm frequency to see what happens when we become more demanding - that is the Nyquist-limit (0% contrast, we have to scan with higher resolution to render detail at 40lp/mm) at 2000spi scanning resolution.

The lens:
Sironar-S 150mm - pretty much the best you can get for 4x5". It reaches 80% center contrast @20lp/mm at f11 and about 65% at f22 due to diffraction
There are no 40lp/mm-values so we can only guess given other MTFs:
60% @ 40lp/mm at f11 and 30% at f22?

The Film:
Ektar 100 or Portra 400 resolves 20lp/mm nearly perfect with >90% contrast and 40lp/mm with about 70% contrast (varies with the color spectrum)

The scanner:
I have no MTF-value except for the 1M$ 6000ppi ARRISCAN for the movie-industry (it's custom-made Zeiss lens propably costs as much as a drum scanner) - the MTF is worse than you would expect, because a scanner is a MTF-chain for itself (sensor, lens...).
Let's say 80% contrast @ 20lp/mm and 60% contrast @ 40lp/mm.
A very good drum scanner might be a little better, the very best Epsons are a lot worse. Let's say 50% contrast @20lp/mm and 20% contrast @ 40lp/mm

Now starts the fun:
We can calculate the final MTF of the file by multiplying the MTF-values.

Drum-scan from 4x5" at f11: 80% x 90% x 80% = ~57%
Very good rendition of detail at 20lp/mm.

Drum-scan from 4x5" at f22: 65% x 90% x 80% = ~47%
Still good rendition of detail, difference perhaps not visible below 5x magnification

Epson-scan from 4x5" at f11: 80% x 90% x 50% = 36%
The scanner resolves way more than 20lp/mm, the lens has higher contrast than at f22 but still, the file will appear "softer" than the drum-scan from the f22-image!

How far can we push resolution? So let's calculate with 40lp/mm (under ideal conditions):

Drum-scan 4x5" at f11: 60% x 70% x 60% = 25%
Well defined detail at 40lp/mm - that are 10000 x 8000 pixels of actual resolution rendered with 25% contrast - beyond a 80MP-MFDB!

But why are so many MFDB-samples so good in comparison to 4x5"?

Epson-scan 4x5" at f22: 30% x 70% x 20% = 4%
Each component renders detail @ 40lp/mm but the final image barely contains any information at this frequency! The 80MP-MFDB doesn't have contrast there, either but it holds contrast better below this frequency and therefore appears sharper without actually rendering more detail.

Finally a 8x10" negative with a different lens (let's say f22-MTF is the same) with the same file size as the 4x5" (so 20lp/mm at 8x10" equals 40lp/mm at 4x5" in the file).
Drum-scan 8x10" at f22: 65% x 90% x 80% = ~47%

47% vs. 25% - the lens resolves less contrast but due to the large negative, the 8x10" appears sharper, even below extinct resolution.

What have we learned today? ;-)


Extinct resolution (taking pictures of contrasty resolution charts) doesn't really matter in reality
Contrast over detail (MTF) is the key to sharpness - not megapixels or file sizes
Final MTF is the product of the whole image chain - from the lens to the scanner. One weak link can ruin it all while seeming sufficient on it's own


I hope this is helpful to some!?

[Ken Lee]

MFDB = Medium Format Digital Back

[Ken Lee]

I was able to follow your presentation. It is rather "compressed" :)

Could you please expand upon your conclusions: What have we learned ?

[Michael Graves]

I was able to follow your presentation. It is rather "compressed" :)

Could you please expand upon your conclusions: What have we learned ?

That a darkroom print from a 4x5 FP4 negative onto Oriental Fiberbase paper still renders a nicer print than a computer?

[jb7]

I've learned a few things-

Math and testing is not my thing; there are others who can do it better, and I have several opinions to choose from.
Resolution and scanning produces the most abrasive, defensive, and personal comments, of all topics posted on this forum.
Anyone who sticks his head above the parapet will likely have it shot off.
All the numbers, be they lp/mm, MP, %, they all come down to one thing-
a measure of how a print will look at a given (usually maximum) size-


A member here, bglick, who hasn't posted here in a long time, produced this graphic to show comparable print sizes for multiple dof scenarios, on multiple formats, based on a common 5 lp/mm print size.

I find it as good a guide as any- http://www.largeformatphotography.info/forum/showthread.php?t=45186

There are times when you need to make a print with less lp/mm-
and nothing is stopping you doing that.

However, if you want to claim that you can achieve many times the resolution than is considered physically possible, you might have a debate on your hands, and chances are, it will descend into personal abuse, and name calling.

I notice from your numbers, that you claim 80MP at 47% MTF for a drum scanned 8x10 negative at f/22-
although some of your parameters are estimated, I'm not going to argue with you...
I wonder what your numbers are for f/64?

[georgl]

Resolution and scanning produces the most abrasive, defensive, and personal comments, of all topics posted on this forum.

Which surprised me the most, when I stumbled over the f64-thread. This is the peaceful world of large-format-photography ;-) not a marketing platform for photographic gimmicks.
Some of the parameters I used are estimated, but realistic and the basic science behind it is still true and hopefully helps to calm down the discussion a bit.


What have we learned ?

Mostly "sharpness" ("resolution") is evaluated by shooting test charts, Whether it be the film, the lens or the scanner that is tested.
Several times (not just here) I stumbled over the missconception that the evaluated/ measured sharpness/ resolution is just one fix value, whether it be dpi, ppi, spi or lp/mm.
You use a not too-sharp lens at diffraction limit -> you don't need a sharp film
You have diffraction-limited, grainy slide at 400ASA -> you don't need anything but a Epson!
Wrong!
Just because an Epson-flatbed can resolve 2000 or 2500 spi, doesn't mean it affects "sharpness/ resolution" below these values not as well (double negative? ;-)).

You want a really "sharp" image? You have to consider the whole image chain!

[Brian Ellis]

That a darkroom print from a 4x5 FP4 negative onto Oriental Fiberbase paper still renders a nicer print than a computer?

Please tell me how I can make my computer render a print. Mine won't do that, it just sits there doing nothing until I make it do something.

[Bob Salomon]

"20 lp/mm....40lp/mm"

Just in case there is any confusion:

Those are 20 line pairs per mm and 40 line pairs per mm. They are not 20 lines or 40 lines per mm.

[Ken Lee]

You want a really "sharp" image? You have to consider the whole image chain!

That makes sense of course, and I'd be surprised if anyone would disagree.

I think you are also showing - which people do not always consider - is that a single system has different responses, depending on the level of detail that we send through it. That's why MTF charts contain more than one line.

By analogy, an automobile may get 100 kilo/liter at moderate speed, but when moving very fast or very slow, it will be less efficient. We can't say that a car has one rating. We might calculate an average rating, but that would be only a synthetic number.

Is that what you're saying ?

[Brian K]

Sharpness and resolution is a chain as others have stated.

For those of use using view cameras, this is Large format forum after all, that chain is extensive.

Film, the combination of films and various developers affects both the resolution and sharpness (acutance) of the recorded image.

Optics, you need to use the right lens i.e. a macro in a macro setting, and a non macro in a scenic setting, and the aperture that yields the required DOF while limiting the affects of diffusion and giving sufficient edge to edge resolution.

GG placement and accurate focus. You need to be certain that your ground glass is actually on the same plain of focus as the film and that your holders hold the film flat and consistently in that exact plane of focus. Sheet film will sag and move in holders. You then need to be able to focus on that point, most often relying on a loupe for assistance. And you also need to know WHERE in the scene to focus for optimum DOF as well as what camera movements can optimize that focus. You also need to be certain that your rear and front standards are plano parallel to each other and to the scene (this is assuming that no movements are required).

Vibration elimination. You need to have sufficient camera support to keep the camera steady and vibration free. If on vibration prone floors you need to be aware of footfalls and even heavy exterior traffic. When doing exposures in the 2 second to 1/15th second range, especially when using longer focal lengths or doing close up work, you need to be aware of shutter vibration, any vibrations still resonating in the camera or camera support from having inserted a film holder and any external sources of vibration to your camera or the subject.

Filters, the use of resin or other non glass filters, with the exception of kodak wratten filters behind the lens, will cause a loss of sharpness with longer focal length lenses and are to be avoided.

While all of this may sound too "perfectionist" to some of you, I can assure you that myself and every professional still life photographer I have ever known has taken all of these steps as part of their normal course of business. This is how a professional does their job.

As for scanning, the use of a prosumer scanner may be acceptable to some, but I doubt if it's output quality is considered sufficient for serious editorial or advertising work. And when one compares the image quality of an Epson 750 to that of a drum scanner or pre press flatbed scanner, one will see a world of difference.

[Dan Fromm]

Um, there's a book on the subject. Image Clarity, by John Williams. Short posts on a bulletin board are poor substitutes for it. Those who want to know more about the topic should buy a copy and read it.

[engl]

The problem with the mathematical approach is that it relies on simplifications, disregarding a lot of factors that influence the MTF chain, as well as ignoring practical limitations.

The most obvious one is deviation from design specifications. Lenses perform below their simulated ideal. One does not have to look further than Perez/Thalmann testing to find that there are WF Ektars from the 60's performing better than samples of the APO-Sironar-S. Another spec deviation issue is GG/film register distance and flatness. Having the best MTF ever at a relatively wide aperture is not helpful if focused beyond infinity in the final picture.

Also consider sharpening. This will improve MTF for any frequency at which there still is response. So how much can you use? "It depends". Taste, sharpening algorithm, film choice, development choice, mounting method, exposure etc. There is much more to be gained in terms of high frequency MTF for an Epson flatbed compared to a drum scanner, since the MTF drop is not abrupt as the frequency increases with a flatbed.

Light level also matters. Photon noise limits resolution in dark areas. MTF varies with light intensity for film/sensors/scanners.

The basic idea that "all parts of the chain matter" is still intact, but in practice, estimating the advantage one system might have over another through math and theoretical performance is impossible. Using them and comparing directly is the only way to get meaningful information about real consistently achievable performance (and the impact, if any, of the difference). Theory might be a guide for how to use them, such as remembering to sharpen flatbed scans due to low contrast still being present at relatively high frequencies. Diffraction theory is a hint that you probably do not want to shoot 4x5 at F64 unless absolutely needed for DOF reasons.

[Ken Lee]

What engl said.

[Henry Ambrose]

Then there's the part where you can only see what you can see. At a given print size its all a waste after you make a print with all the resolution you can see.

(and I'm not contradicting the previous posts or trying to over simplify this - I agree that using a bigger camera is a wonderful thing)

[Ed Kelsey]

Drum-scan 4x5" at f11: 60% x 70% x 60% = 25%
Well defined detail at 40lp/mm - that are 10000 x 8000 pixels of actual resolution rendered with 25% contrast - beyond a 80MP-MFDB!


This is an often used comparison to digital but in fact is false. You are comparing apples to oranges. The film scan is not the same as digital because of grain for one thing. If you have 10000x8000 pixels from a digital back be assured its all good clean data; with film not so.

[Ben Syverson]

Now starts the fun:
We can calculate the final MTF of the file by multiplying the MTF-values.
You can't obtain a meaningful number by individually testing each component of the imaging chain and multiplying the MTF numbers. It just doesn't work like that.

If you want to know the MTF of the system, the only real way to do it is to measure it at the end of the chain—in this case, you'd shoot a resolution chart with a LF camera, scan the negative, remove the orange mask and reintroduce contrast to get a "normal" looking image. THEN you would use that file to take your MTF measurements.

However, I agree with your overall point that in LF, diffraction is not the bogeyman some people think it is.

[onnect17]

This is an often used comparison to digital but in fact is false. You are comparing apples to oranges. The film scan is not the same as digital because of grain for one thing. If you have 10000x8000 pixels from a digital back be assured its all good clean data; with film not so.

Very good filtered and interpolated 20 MB ( 5000x4000 ) of data.

[georgl]

We're talking about ideal situations here, that's for sure. But neither value was really unrealistically high (picking the very best MTF of any 4x5" lens I could find is the most extreme one). We're not talking about an average image dealing with DoF or a missaligned holder - this is the best-case scenario, for all components.

Sharpening enhances MTF - but it's not recovering information - this is not about subjective image impression, dealing with grain or tonal rendition - just contrast at a certain frequency.

This only shows that even on larger formats, things like diffraction, lens-quality and scanner matter and are an important part of missleading comparisons.


This is an often used comparison to digital but in fact is false. You are comparing apples to oranges. The film scan is not the same as digital because of grain for one thing. If you have 10000x8000 pixels from a digital back be assured its all good clean data; with film not so.

Remember, I said 60% MTF at 40lp/mm (and translated it into MP worth of information) for the scanner - not 2000spi (40lp/mm = 0%) - in fact the MTF was measured at 4000 spi! So we are comparing a 80MP-MFDB-file to a 320MP-scan - and given these MTFs, it contains more information.

40lp/mm with reasonable contrast can be dealt with even in 8x10", IMHO. Rigid, precise camera, good holder, careful work, good tripod/head, some Ektar100, a nice modern lens design (I have never seen a 60s design that could hold up with modern high-end-lenses - but I have more experience with smaller formats?) and a drum/creo-scan and I. Would say this calculation is realistic.

[engl]

This only shows that even on larger formats, things like diffraction, lens-quality and scanner matter and are an important part of missleading comparisons.


I do not think that anyone disagrees with this, F128 (on 4x5), an old poor oscilloscope lens and an ancient budget flatbed scanner are all ways to ruin sharpness.

The problem is that you are trying to quantify the difference between two setups using theory. I would call your comparison misleading. I would also call your advice uninformed, since unless I'm mistaken, you do not even shoot large format.

My belief is that consistently getting close to the theoretical performance limit of 4x5 would require going much farther than right about anyone is interested in doing. You'd need a massive precision monorail for absolute control and parallelism, heavy duty tripod/head, flawless focusing, vacuum backs calibrated to the GG, shoot film with possibly undesirable characteristics besides resolution, lenses selected by trying out and measuring several samples shot at relatively wide apertures. I have not used a drum scanner so I do not know how much is required to reach peak performance, but I assume that you'd have to be fairly knowledgeable to get the most out of the scanner as well as confirm that the scanner is fully within specifications.

If you drop over to the image sharing forum you will notice that nobody shoots like this.

[rdenney]

The point, from my perspective, is that the image-forming chain is not really like a chain. A chain is only as strong as its weakest link, and unnecessarily strong links have no effect on the chain's performance other than to make it heavier.

The point here is that any step in the process that imposes an MTF less than 100% affects the image that passes through it, even if that step provides much better performance than other steps. The effect may not be important but it will certainly exist.

Thus, instead of the chain analogy, we should think of every step in the process like looking through another pane of glass. If one looks through five panes of glass, one will see substantial overall degradation of the scene, even if each pane by itself is clean and invisible. Yes, there are things we can do if we are considering that that we will be looking through five panes of glass. One of them might be to apply anti-reflective coatings to each pane. Maybe we wouldn't see the effects of doing so while just looking through one pane, but we would certainly see the effects looking through all five.

Degradation is cumulative. We might decide to use a better scanner than an Epson, but we also might mitigate the weaknesses of the Epson by improving the film and the lens, even when they are abundantly better than the Epson already. We often argue that it makes no sense to greatly exceed the quality of the worst step in the choices we make for the other steps, but maybe we are wrong.

Rick "getting the point" Denney

[georgl]

This is an excellent analogy! When speaking of a "chain" with it's weakest link defining it's overall strength - I actually made the same mistake I've tried to avoid by explaining multiplying contrast...

As a chain, you would just have to find the weakest link you couldn't change (the largest aperture with sufficient DoF, the best scanner you can afford...), "measure" it's performance and don't care about components that might be more powerful within the chain - because it doesn't matter.

You're right, it's the exact opposite!
You want a sharp, high-res image? Try to get the best quality of each component, even when you feel one of them is limiting (like a 400ASA film or f32).

[Bob McCarthy]

The chain may be less important than the threshhold of visability.

The paper used, the printing system, either optical or inkjet, and finally the ability of the human eyeball to resolve, are the limiting factor.

Inside this threshhold, money is wasted.

It's why an inferior scanner (Epson vs drum) can be equally effective.

It why a aperture limited by defraction can be effective.

Its why a big sheet of film with old, less capable lenses can be effective.

To me the only reason to optimize the chain, is when the prints are large enough to illustrate the weakness in the chain. That is a damn big print with an 8x10 negative.

bob

[Jack Dahlgren]

The chain may be less important than the threshhold of visability.

The paper used, the printing system, either optical or inkjet, and finally the ability of the human eyeball to resolve, are the limiting factor.

Inside this threshhold, money is wasted.

It's why an inferior scanner (Epson vs drum) can be equally effective.

It why a aperture limited by defraction can be effective.

Its why a big sheet of film with old, less capable lenses can be effective.

To me the only reason to optimize the chain, is when the prints are large enough to illustrate the weakness in the chain. That is a damn big print with an 8x10 negative.

bob

I agree, both the chain and the stack of glass metaphors are useful in describing the entire system. Observation is definitely of chain type, while other parts of the system are layered.

Don't forget subject lighting as well. It plays as big a part as anything else.

From my personal point of view, most normal equipment and practices are good enough for what I want out of the system so return on optimization of a specific part is low - if not negative (resulting in higher cost, more time spent, heavier and larger equipment). Moving from an uncoated Ilex Paragon to more modern lenses made a noticeable difference, but I don't think that the differences between lenses made in the past 30 years are worth chasing after. I'd improve my results more by getting out in the right sort of light or spending time perfecting printing.

If you have a limitation that is preventing you from achieving what you want, then by all means attack it. I think my limitations are not in the equipment anymore.

[Bob McCarthy]

VC, right on the mark.

Before I came to large format (for the second time), I was stitching a whole flock of digital shots into one large file. One shutter press vs. 12 is a very different thing. It works when it works but any single mistake or change in light over time wastes all the effort.

Eventually the return to large format solved most of the problems.

bob

[paulr]

The chain analogy works reasonably well in a digital workflow, where sharpness can be recovered. In an all-analog workflow, it's a poor description of what we see, because the threshold of visibility tells us nothing about how good the detail looks. Detail rendered in a print at 10% contrast looks like mush. At 80% contrast it looks tactile.

When you loose MTF at each stage without possibility for recovery, the concept of a weakest link becomes unimportant.

[onnect17]

Um, there's a book on the subject. Image Clarity, by John Williams. Short posts on a bulletin board are poor substitutes for it. Those who want to know more about the topic should buy a copy and read it.

Just got my copy today. Thanks for the recommendation!

[Acheron Photography]

This is an often used comparison to digital but in fact is false. You are comparing apples to oranges. The film scan is not the same as digital because of grain for one thing. If you have 10000x8000 pixels from a digital back be assured its all good clean data; with film not so.

Others have commented amply on other aspects, but I wanted to pick this one up. It's not true. Sensors have to contend with various form of shot noise and kT/C noise for a start, even before you get into response non-uniformity, the frequency dependence of dynamic range and any issues in getting the pixels out. Remember a 'full' site might only contain 50,000 electrons in a high pixel count sensor: capturing and reading that few accurately isn't easy.

This isn't to diss digital, simply to point out that the performance of any DB is depenent on a massive amount of post processing of the raw signal which attempts (often reasonably successfully) to deal with the inherent noise and non-linearities of the photo sites. This means that a resolution number for a DB is subject to a lot of caveats about how well that processing works, just as the theoretical performance of a film arrangement depends on film flatness, vibration etc etc.

[paulr]

Sensors have to contend with various form of shot noise and kT/C noise for a start, even before you get into response non-uniformity, the frequency dependence of dynamic range and any issues in getting the pixels out. Remember a 'full' site might only contain 50,000 electrons in a high pixel count sensor: capturing and reading that few accurately isn't easy.

You're right, but in practice the signal / noise ratio is much higher with digital capture than with scanned film. I'm doing my first ever project now with digital capture after working for years with scanned film (color 2-1/4 and bw 4x5) and the phenomenon is interesting. I'm working with fewer pixels than ever, but the quality of the individual pictures is higher and compensates a lot.

Obviously the DSLR images do not contain as much useful information as the 4x5 scans, but the difference isn't nearly what you'd expect based on pixel numbers alone.

[Bob McCarthy]

You're right, but in practice the signal / noise ratio is much higher with digital capture than with scanned film. I'm doing my first ever project now with digital capture after working for years with scanned film (color 2-1/4 and bw 4x5) and the phenomenon is interesting. I'm working with fewer pixels than ever, but the quality of the individual pictures is higher and compensates a lot.

Obviously the DSLR images do not contain as much useful information as the 4x5 scans, but the difference isn't nearly what you'd expect based on pixel numbers alone.

I think that's right and agrees with my years of using digital.

But in the end a big negative/positive wins by brute force.

The lenses are not stressed with large reproduction ratios

A better scanner improves the film scanned pixel.

View camera movements really add to DOF.

Defraction is a monster issue in digital with optimum apertures in the 5.6 area. All digital camera pixels are so tiny today. Even the inherent short lens focal length doesn't come close to compensating.

Brute force with everything in the reproduction chain wins easily.

Of course birding with a view camera is .......

And shooting for publication or web is just not convenient with film and digital capture is so much more to the customers liking.

But for fine art, large sheets of film is the best capture medium, by far...

Bob

[paulr]

I don't know about fine art being a special case, but sure, any time you want a ton of detail in something bigger than a small print ... or if you need camera movements for any reason, LF still seems like a better choice than a dslr.

Smaller film formats have a much harder time competing. I've abandoned medium format and 35mm in favor of the dslr.

[C.T. Greene]

I don't know about fine art being a special case, but sure, any time you want a ton of detail in something bigger than a small print ... or if you need camera movements for any reason, LF still seems like a better choice than a dslr.

Smaller film formats have a much harder time competing. I've abandoned medium format and 35mm in favor of the dslr.

Hmm . . I still shoot 35mm, in fact took it out this morning, shooting some Ecktar 100. But the point is yes, I need to upgrade & the scanner is probably the most probiting factor so not having or wanting to spend an arm or a leg . . what flat beds scan 35mm & 4x5? Most seem only to scan medium formats & of course LF prohibits dedicated film scanning.

[engl]

The Epson V700 and V750 flatbeds give good results with 4x5. The results with 35mm film will not come close to what a dedicated scanner can offer, but it is good enough for me. You could get a V700 for 4x5 and a Plustek 7600i for 35mm.

If you want a single scanner capable of great results from 4x5 and 35mm you'd need one of the professional flatbeds, in which case you will be spending a ton more than what the V700+7600i cost together, and will have to fight compatibility issues.

[Leigh]

Diffraction?

Here's an Apo-Ronar 300mm/f9 with calibrated apertures down to f/250
http://www.mayadate.org/pix/Apo-Ronar.jpg

Given that this is a high-end process lens, used in an environment where the lighting is completely controllable, I doubt they would provide an f/stop that gave significantly degraded performance.

- Leigh

[rdenney]

Diffraction?

Here's an Apo-Ronar 300mm/f9 with calibrated apertures down to f/250
http://www.mayadate.org/pix/Apo-Ronar.jpg

Given that this is a high-end process lens, used in an environment where the lighting is completely controllable, I doubt they would provide an f/stop that gave significantly degraded performance.

- Leigh

Hmm. Diffraction is what it is. I suspect they provided those apertures because in the macro world, depth of field is particularly challenging. And then they leave it to us to determine if the resulting diffraction is acceptable or not.

Rick "not drawing too much conclusion from that aperture ring" Denney

[Drew Wiley]

Leigh - the aperture ring is like a speedometer on a car that goes up to 150mph, even
though it's a compact 4-cylinder vehicle. Just a standardized component which doesn't
imply recommended usage. Diffraction is optical fact, not a variable of labeling or
marketing. Anyone who actually uses lenses like these knows it from practical experience too.

[C.T. Greene]

Thank you for the info, a separate dedicated 35mm may be the best solution. It would be alot quicker & easier when doing rolls of 36 frames.
As for LF, I have looked (considered) the both V700 & V750. I'm still looking around, heard some about the 4990 but they seem to be out of stock with unknown delivery date which leaves the Hp G4050 aas the only option?

[engl]

The 4990 is since long discontinued. For some time they were available as refurbs from Epson, but I think those days are gone as well. They are sometimes available used, but I think I'd rather get a V700 with warranty, better holders and possibly better image quality (there is some debate about the last point).

As for scanning 35mm negatives quick and easy, the V700 is probably a lot better than the Plustek. You simply cut the negatives into strips of 6, load 24 frames at once and hit scan (the software finds the frames and balances exposure per frame). With the Plustek you are scanning one frame at a time, waiting 1-10 minute per frame, then manually advancing. Every 4 slides or 6 photos on negative strips, you have to reload the holder. Unless you do the 3 minute 7200DPI scans and downsample, you are not going to see any advantage over the V700. If you want ICE dust reduction, scans take 11 minutes at 7200DPI, per photo. You'd be finished with a 36 frame roll in about 7 hours.

From what I quickly skimmed about the HP G4050, it does not seem like a very good scanner.