Any real world comparisons between Epson 4990 and v700(/v750)?

[jvuokko]

Hi,

It is finally time to say good bye to my old Epson 3200 flatbed scanner.

But which scanner to buy now? Epson 4990 is quite cheap but old. Epson v700 is expensive but really good. v750 is out of budget.

I have done quick test between my 3200 and v700 (http://jukkavuokko.com/public/epson3200vsV700/) that I had access.

The result is clear. v700 is way better than my old scanner. In both, sharpness and dense areas.

My main need for scanner is to get MF and LF negatives to web, which can be done perfectly with old 3200. The 3200 has dmax around 2.9 and typical negative has densest highlight somewhere around 1.8

I also have to scan all my MF color work for printing. This does not happen very often but this is where my 3200 really sucks. Actually I think that reason why I haven't done much color MF photography for couple of years is simple because I cannot get good scans from them.
For LF, I don't even think color slides before I have decent scanner.

Now I am asking have anyone done similar test between 4990 and v700/v750 that I have done between 3200 and v700.

If such test exist, it would show clearly how much difference there is between 4990 and v700.

I have google'd a lot. There are comparisons, but I have found a single one where dense areas are compared and shown. Usually the sample pictures are also so tiny that it's hard to judge anything.

Thanks!

Jukka

[Walter Calahan]

The v700 is not a repackaged 4990. There are quality differences between the two. I have sense side by side tests, but for the life of me now I can't remember where. The bottom line is the jump from the 4990 to the v700 is not huge. Many here still use their 4990 with great results for printing as long as they are aware of the scanner's limits.

As you said, for the web, your 3200 is fine.

Stay tuned for others to answer your question for this community is a great resource. I never owned a 4990, so I cannot reply from personal knowledge. I own a v750, and enjoy the results even though it is not a perfect machine.

[Brian Ellis]

I seem to recall that the late Ted Harris reviewed the 700/750 scanner in View Camera magazine about 3-4 years ago and that his review included some comparisons with the 4990. If you want to pursue this you probably could obtain a back issue from the magazine. FWIW, I decided not replace my 4990 with a 700/750 after reading the review. There didn't seem to be enough differences to make it worthwhile.

[Ken Lee]

If you perform a Google search for "v700 4990 (http://www.google.com/search?&rls=en&q=v700%204990&ie=UTF-8&oe=UTF-8)", or other similar arguments, you can find some comparisons and discussions - some of them on this very forum.

Like Brian, I didn't see a big enough jump in quality to merit an upgrade from my 4990. Instead, I got the Variable Height Mounting Station from betterscanning.com (http://www.betterscanning.com/scanning/vb_advantage.html) If I were to shoot a lot of Medium Format, I would buy their dedicated Medium Format holder as well.

However, I stay away from Medium Format whenever possible, and try to shoot 5x7 instead of 4x5. That's one way to solve the "scanner problem" once and for all. You just can't beat the image quality of big negatives - and even the most humble of scanners, can handle a 2x or 3x enlargement :-)

[venchka]

Ken Lee,

Thank you. Please stop it! I don't need a 5x7 camera. I really don't.

[ASRafferty]

I seem to recall that the late Ted Harris reviewed the 700/750 scanner in View Camera magazine about 3-4 years ago and that his review included some comparisons with the 4990. If you want to pursue this you probably could obtain a back issue from the magazine. FWIW, I decided not replace my 4990 with a 700/750 after reading the review. There didn't seem to be enough differences to make it worthwhile.

I'm not at liberty to actually share Ted's 2006 article. However, here are some notes from his email that may help.

He shared this website back in '06 in the context of a comparison that had just been added, between the 4990 and the 700. Although he seemed to have some substantial reservations about the website itself, he used it for the comparison data, and commented at the time that this supported his prediction that he could not see much difference, if any:

http://www.terrapinphoto.com/jmdavis/

He also had occasion to write this in an email in June:

"I just finished putting the V700 through its paces (few differences between it and the V750). You can read my quick comments on the Large Format Forum. Bottom line, I don't think these scanners give you much gain over the 4990. To get the full advantage of the new optical system you need to scan at 6400 spi which gives you very alrge files from a 4x5 and impossibly large from a 5x7 or larger. There is also a time penalty with file sizes that large. At a more reasonable (for most) 3200 spi the scanner doesn't perform quite as well as the 4990 and is still a good bit slower than the 1800f which it doesn't outperform either."

And this the following month:

"Still have not gotten a V750 but I have put the V700 through its paces. The V700 is close enough to the 750 that I don't expect anything but very small improvements from the V750 as the only hardware change is the addition of anti reflection coatings on th mirrored surfaces and the ccd array .... enough for a small improvement but that is all. Bottom line first ... not really an improvement over the 4990 for the LF shooter and some modest gains for the smaller format shooter.

--you need to scan at the highest setting of 6400 to activate the higher resolution lens and at that setting you get a real spi of 2400 - 2800 v the 2200 - 2300 for the 4870/4990. Big drawback for anything larger than 35mm is that you have a huge file size.

--at lower settings the performance is not as good as that of the 4870/4990

--the Dmax is the same as that of the 4870/4990 ..... 2.25

--the machine is slow slow slow .... much slower than the 4870/4990 ... a 2400 spi scan can take as much as 25 minutes ... at 6400 and scanning a 4x5 expect 3+ hours .... "

I hope this is of some help or interest... as he wrote then, you may be able to find his comments here on the forum via the Member List, as the mods here have left it there so as not to eradicate his posts. Nonetheless, it's good for me to get these things out of his email and someplace where they can help others, so... FWIW.

[venchka]

Thanks Amy! That information is priceless! It should also keep the price of 4990 scanners up.

[Kirk Gittings]

Yes Amy that reflects my memory of Ted's opinion too. I was talking to him almost daily during that period and at best he thought the 700 was an incremental improvement over the 4990 for LF use. That is my opinion too having owned both the 4990 and 750. The one other improvement that is never mentioned FME is the lack of out gassing on the 750, which is a problem on the 4990, coating the inside of the glass with a film which requires disassembly to clean periodically.

[ASRafferty]

Not my intention, Wayne... nor did he have any interest in doing so! :)

Michael Mutmansky may chime in when he sees all this, since he and Ted spent so much time discussing these matters at the time.

[venchka]

Sorry, Amy. Not my intention either.

I just meant that the Epson 4990 has a loyal following and a well deserved reputation in the market place. I have used a friend's 4990 with good success. In fact, I was on the lookout for a nice used Epson 4990 when I found a nice used Epson 1680 right here. I am pleased with the 1680 as well.

[Michael Mutmansky]

That hits the nail as I recall as well.

The large improvement in the Epson scanners happened between the 2450 and 3200, and then from the 3200 to the 4870 if I recall correctly. After that, the improvements have been somewhat small, but measureable for pixel peepers.

That said, I have a 4870 scan and a drum scanner scan of the same 4x5 piece of film, both printed at 16x20, and there are very subtle differences between the two. Take them apart, and you won't feel that there is any defficiency in the Epson scan, but put next to each other (and looked at VERY carefully) and the differences are visible. However, I still don't think there is any defficiency in the Epson scan. They just render slightly differently, and it is possible to choose one over the other based on certain preferences, but certainly not based on a real failing in one or the other.

However, are the differences meaningful at that size? I think that's a tough question, but I do think that for most people the answer is 'no'. That represents slightly larger than a 4x enlargement (full frame, with only a slight crop to remove the film border).

As the magnification increases from that point, I think the differences will become more apparent, until the difference is enough to be meaningful. As magnification increases, the difference will move into the 'undesirable' realm, and then finally into the 'unacceptable' realm.

I believe that all of the Epson scanners up to this point essentially (since the 4870) perform at the same level; about 4x is the limit for no meaningful difference from a top-of-the-line scanner, and above that the difference start to become greater as the magnification increases.

That's all for sharpness with B&W film, not DMAX, which is a different story, but I don't believe there has been any improvement in this aspect since the 4870 as well. However, none of the Epson scanners match the DMAX of a decent drum or high-end flatbed scanner.


---Michael

[aphexafx]

As an owner of the v750 I have just two notes:

1) The hi-resolution lens on the v700/v750 is NOT "activated" by scanning at 6400 spi. The lens in question is "acitvated" by scanning above the platen using the trancparency w/ film holder options and has a reduced coverage area over the normal platen-level lens. The lens is employed for any resolution when scanning with a holder or mounting tray.

2) Scanning a 4x5 at 6400 spi with the v700/v750 takes nowhere near 3+ hours on my system. I just did a single pass scan at 24-bits, 6400 spi of a 4x5 chrome producing a file that was just over 2GB. It took exactly 13 minutes, including the time for the software to save the file.

3+ hours vs. 13 minutes. Meh, no biggie, right? I fail to see how this information is helping others.

If this "test" was made with specific options turned on, it is not mentioned - but it should be - hardly a conclusive comparison and a worse statement to make.


Also, on the page linked above, the re-focused v700 blows the 4990 scan out of the water to my eyes, as far as pulling detail goes.

[jvuokko]

It really seems that 4990 is right scanner for me :)

[venchka]

If it isn't, there will be plenty of interest if you decide to sell it.

[Arthur Nichols]

Anyone using the v750 for medium format? Would the 4x enlargement ratio still hold?

[venchka]

I draw your attention to posts #42 and #51 in the Large Format Landscapes thread.

http://www.largeformatphotography.info/forum/showthread.php?t=44249&page=5


Scanned on Epson V750, printed on Epson 7880 on Red River UltraPro Gloss paper.

I would be pleased to have this photo hanging on my wall.

I would be ecstatic if I had also made the photograph.

The quality of scanning and enlarging and printing is all relative folks.

[aphexafx]

Anyone using the v750 for medium format? Would the 4x enlargement ratio still hold?


I scan 6x7 (cm) chromes shot with Mamiya RZ lenses using my v750 and I print them at 12" x 14" on various Epson printers - they are sharp and gorgeous with only very slight USM'ing, and they exhibit only very shallow grain in the smooth tones (skies), and I like that. The best prints that I have produced are from my prosumer R2880 -> fantastic quality on exhibition matte papers, if slightly more expensive per print over the larger Epson printers.

[ASRafferty]

I fail to see how this information is helping others.


Thanks... I'll tell him when I see him.

[aphexafx]

Thanks... I'll tell him when I see him.

*** comment removed; unknowingly made in bad taste ***

Seriously, though, I just think stating that 4x5's take 3+ hours to scan on a v750 is inaccurate...is all. Again, I just pulled one off in less than 15 minutes, albeit at 24-bits. (I honestly don’t think my system would handle a 48-bit scan of this scale…) However I do scan 6x7 (cm) frames at 48-bits and it does not add much time to similar 24-bit scans - nor should it, nothing changes at the CCD pickup level between the two color resolutions.

[Michael Mutmansky]

Matt,

Boy, you might want to consider your tone here. You don't know the history of the people in this discussion obviously, or you would probably see that comment as being in really bad taste.

My recollection is that there were issues with the driver when it first came out and the scan time were very long as a result (I think Ted had a pre-release model scanner operating on a Mac Pro). Those things got straightened out, but the actual performance of the scanner did not improve, so those aspects of the scan tests he did certainly are applicable.


---Michael

[Michael Mutmansky]

Anyone using the v750 for medium format? Would the 4x enlargement ratio still hold?

Arthur,

Yes, the magnification ratio does hold for MF. You might be able to get to 4.5x before you start to compromise on the print. These scanners really need a glass carrier to perorm at their best. Noticeably better on glass than in the air.


---Michael

[ASRafferty]

Or perhaps you could ask him... (haha)

Don't I wish.

Brian referred at the beginning of this thread to an article Ted had written, and I happened to have some of the material that had appeared in it, so I posted that, since I can't post the article. I can't defend it or carry it forward, so perhaps I shouldn't have shared it. Or perhaps you just have an agenda where my husband was concerned, and the ridicule has nothing to do with the subject at hand. Whatever it is, glad to have been of service.

[aphexafx]

Well I was just PM'd regarding this thread, and of course I did not know the history of the person whose claim I was arguing, nor was I aware of his relationship to any members here. So, in that respect, I am sorry if I've offended anyone.

However, and more importantly to this topic, I am not intending to come off in any offensive manner - I am merely arguing the 3+ hour scan time for a 4x5 at 6400 dpi. Thank you, Michael, for your insight regarding the driver. Obviously, and as was just explained to me, the author of that statement was highly respected and extremely knowledgeable about scanners, and I would not argue this.

I just don't want people who are interested in a new scanner to come in here and dismiss what I think is a fantastic machine because they read that it will take as long to scan their 4x5 transparencies as it takes to drive across a mid-western state.

That is my only intention, and I certainly do not have any agenda towards anyone here, nor am I intending to ridicule anyone. Again, my apologies; and I've inadvertently treaded onto a sensitive subject that I will fairly step out now. Amy, I did not mean to offend you or ridicule anyone.



Matt,

Boy, you might want to consider your tone here. You don't know the history of the people in this discussion obviously, or you would probably see that camment as being in really bad taste.

My recollection is that there were issues with the driver when it first came out and the scan time were very long as a result (I think Ted had a pre-release model scanner operating on a Mac Pro). Those things got straightened out, but the actual performance of the scanner did not improve, so those aspects of the scan tests he did certainly are applicable.


---Michael

[venchka]

Would somebody who owns a V700 or V750 scanner please attempt a 6400 SPI 48 bit scan of a 4x5 original and report back?

I will do the same on my ancient Epson 1680.

[Ken Lee]

Thank you. Please stop it! I don't need a 5x7 camera. I really don't.

Oops - I thought I could just sneak that one in there un-noticed. :rolleyes:

[Kirk Gittings]

Wayne what is the point you are trying to get at? Without some controls like scanning the same negative or a test target what is the point?

[PenGun]

I can do a 16 bit greyscale, got no colour 4x5. Is Vuescan be alright? I don't like spending time in windose and I can do the Vuescan in Linux.

[Ken Lee]

PenGun: Speaking of Linux: I have been able to get CinePaint to run on Ubuntu running on VMWare on OS X - but I don't know how to print from CinePaint. Is there an easy way to install the Gutenprint package ? With Cinepaint and a way to print, there would be little need for Photo$hop.

[PenGun]

I think I remember that the Ted tests were all on the glass with the low res lens. For what it's worth.

[Michael Mutmansky]

Would somebody who owns a V700 or V750 scanner please attempt a 6400 SPI 48 bit scan of a 4x5 original and report back?

I will do the same on my ancient Epson 1680.

Wayne,

No need to test to know that the V series will beat the 1680 in sharpness. If all you are doing is scanning B&W negatives, you will get a sharper scan out of a V-series scanner. Probably out of a 4870 as well. Maybe about equal out of the 3200, and probably a little worse out of the 2450.

I think the DMAX will be about equal or a little better out of the 1680. If better, not by much, though.

Regardless, it's the tool you have, so use it and don't worry about it toom much, unless you have a desire to spend money on a new scanner.



---Michael

[sanking]

I looked on the Epson web site a few days ago and they had some refurbished 1680 scanners for about $400. That seems a lot to me given the fact that their performance is so much below that of the 4870 and 4990.

BTW, I used a 4870 for a while and sold it to buy a 4990 because I needed a full 8X10 area for scanning contact sheets. But my 4870 was sharper than the 4990.

I agree with Ted Harris and Michael Mutmansky in that the V700 and V750 provide only an incremental improvement over the 4990. And with the observation that about 4X is about the limit in terms of magnification.

In fact, I think Ted was a bit generous if he actually said that one could get 2400-2800 spi effective resolution from the V700/V750. I have tested both and could not get higher than about 2200 spi effective resolution.

Sandy King

[venchka]

I was just trying to get a handle on the 3 hour time.

Never mind.

[PenGun]

PenGun: Speaking of Linux: I have been able to get CinePaint to run on Ubuntu running on VMWare on OS X - but I don't know how to print from CinePaint. Is there an easy way to install the Gutenprint package ? With Cinepaint and a way to print, there would be little need for Photo$hop.

I'll be looking at that pretty soon. I have not had much time lately and I have not beat on the OSs for a while now. I think the CUPS setup can use the PS part of the drivers but I have not got to it yet.

[Keith S. Walklet]

Scan times for a 4x5 on the V750 with a G5 using single pass Epsonscan, no USM, no ICE, with the Film Holder Setting are:

3200spi @ 48bits 7 minutes 1.14gb file
4800spi @ 48bits 8 minutes 2.25gb file
6400spi @ 48bits 16 minutes (8 min. x 2 passes, top & bottom half)* Each half is 2.2gbs.

The differences between the 4870 and 4990 series were marginal, in essence an expanded transparency scan area which went from 6x9 inches to 8x10 inches, with a small boost in the advertised dmax of 3.8 to 4.0. A larger buffer in the 4990 permits scanning a 4x5 in a single pass. With the 4870, it needs to be done in two passes when using the highest claimed optical resolution (4800spi)

And, yes, while the 4870/4990 are a value, there is a significant jump in the quality with the V750 as viewed on a monitor. As Michael has mentioned, whether that translates to the print has to do with the size at which it is printed. With the V750, the scans I achieve with my current workflow are much closer to those made of the same image on a drum scanner. I'll try to post side-by-sides scans tomorrow if I have the time.

And, yes I am now using the Epson for medium format as well, wet-mounted with wonderful results.

*Post processing of the highest native resolution scans (6400spi) includes downsampling the two files to 3200dpi and merging them in Photoshop, which essentially doubles the time involved with getting to a workable file.

Almost forgot...The V750 is much faster than the 4870. Again, I try and post some times by tomorrow.

[jvuokko]

I see that there are some confusion with v700/v750's dual lens system.

The sharper lens is always in use when scanning is performed "with holder" setting.
The selected dpi does not have anything to do with lens.

According what I have tested by myself (the link was in the first post) and seen from my friend's scans, the v700 is much sharper than (my) Epson 3200. I have measured resolving power of my 3200 to 1800dpi horizontally and 2200dpi vertically.
That's by scanning USAF 1951 resolution chart from the film.

Actually there is visible gain in resolution that can be achieved above that.

I checked where is the point when scanner does not resolve any more details compared to lower resolution scan and interpolated to match larger size.
The result was:

1600dpi -> 2200. 2200 scan will have much more fine details than 1600 scan which is interplated to 2200dpi.

2200dpi -> 3200. The 3200dpi scan have a little more details than image scanned at 2200 and interpolated to 3200dpi. But the amount of required USM is so enormous that it makes scanners exeggared grain patterns really visible (the Epson 3200 sees grain where it can't really see it cleanly, it's called grain aliasing effect).

So I prefer to use 2200 or 2400dpi setting, it is a good compromise.


My dedicated film scanner is bit old, it's Minolta scan dual IV which has 3200dpi resolution according to the specifications. The real, measured resolving power is 3000dpi.
When comparing v700 to scan dual IV, there is only a little difference. Actually when looking cropped portion of scans withouth knowing which one is from which scanner it is impossible to say whether image is from v700 or scan dual IV.
Perhaps scan dual IV sees grain of fine grained color films little sharper, but it is hard to judge.

One big difference is need of USM. The v700 requires lot of USM like all flatbet scanners will do. The dedicated film scanner like scan dual IV does not need any USM. Actually the image quality will only suffer when USM is applied.

The contrast response curve of flatbed scanners are always so poor that USM is really necessary. Without it, the most of scanners resolving power is thrown away.


In term of resolution, there are a lot of different results from similar scanners. One tells that v700 is not better than 3200, one tells that 4780 is way better than 3200 and perhaps one tells that 4780 is better than v700 etc.

The main reason for that is the lack of any focusing system in the flatbed scanners. The film scanner will give option for AF and MF, but flatbed... All that you can do is to try find perfect distance from the surface of the glass by adjusting the height of the holder. The Betterscan holders may have good distance for accuiring best focus, but it depends a lot of individual scanner.
These flatbed scanners really are not designed for professional purposes and thus the tolerances used during manufacturing are somewhat larger than in really specialized scanners.

The conclusion is that resolution of scanner is hard to judge without careful and documented testing.


The DMAX is totally different beast. It is easy to test. The scanning of step wedge is good for showing where scanner's ability to resolve density changes starts to drop and finally flat out.
The ultimate test is ofcourse color slide where is a lot of dark tones. It is easy to take a scan and check where practical limits are. But it requires densitometer for checking the density of areas where scanner's ability to separate dark tones drops.

My current scanner, Epson 3200 has DMAX 3.4 according to the specs. The real DMAX is way belove that. It is only 2.80. After that nothing can be seen.
Even with standard Stouffer step wedge, it will flat out before the densiest part of wedge.

The difference to Scan Dual IV is visible, but the DMAX of Scan Dual IV is also far from promised 4.8DMAX. The real DMAX is around 3.2 which is insufficient for some color slides.
As far as I have been able to access v700, the v700 seems to have better DMAX.
And no doubt that v700's DMAX is way above 3200's poor 2.8DMAX.

That leads again to the original question, is there any measured or side by side comparisons between 4990 and v700. Especially in term of DMAX.

The one that was pointed, was test at davis's site http://www.terrapinphoto.com/jmdavis/

There is same slide scanned with both scanners:

http://www.terrapinphoto.com/jmdavis/bridgewater_4990.jpg

http://www.terrapinphoto.com/jmdavis/v700_4_5mm_no_usm.jpg

What these will show is that there probably slight difference in the DMAX. The 4990 sample is dark, but when download image and open the shadows like has done to v700 image, there is no much difference.

The sharpness has clear difference, the v700 is better. But in the height of v700's holder is adjusted to getting better focus. The 4990 is used without any adjustments as it's holder does not allow adjustments by default. So the real resolving power of 4990 remains still unknown.


Best Regards,
Jukka

[venchka]

Thank you Keith! That was very kind of you.

I attempted a 6400 scan on my 1680, but my copy of Epson Scan won't allow saving a file that large. No worries. I wasn't planning to scan 4x5 negatives at 6400 spi anyway.