PDA

View Full Version : Epson V700 scanner



Linhof
10-Feb-2013, 22:52
I acquired epson v700 scanner for 4 years and had tried many times, still could not have a good scan. sharp focus scan is already very difficult to achieve. And it is more difficult to deal with high contrast positive. The dark area of positive is almost block in the scanned file.

oysteroid
10-Feb-2013, 23:49
What resolution are you scanning at? You cannot possibly get sharp pictures at the resolutions advertised by the manufacturers. They are liars. The highest resolutions where you can get reasonable sharpness range from 1800 to 2400ppi. But it is still useful to scan at 4800 or 6400 ppi, as you are oversampling. You can downsample afterward and you then have an image with far less scan noise. Even then, you must sharpen liberally. You might also want to try different height settings on the film holder to try to achieve optimal focus. Further, put your scanner on a very solid surface, maybe the floor, as vibrations in the scanner also reduce resolution. My scans with my V750 are much worse if the scanner is on a wobbly, light table.

As far as the dark areas of slide film go, yes, this is a problem with these scanners. Velvia 50 is especially problematic. You can try using the multisampling and multi-exposure features of the scanning software. But they don't help much. With these scanners, you can only get about 4 stops of DR out of Velvia 50. A drum scan will get you more. With negative films, your scanner should easily capture the full range. I haven't tried Provia, Astia, or Velvia 100. Supposedly they are a little better than Velvia 50.

StoneNYC
11-Feb-2013, 01:39
What resolution are you scanning at? You cannot possibly get sharp pictures at the resolutions advertised by the manufacturers. They are liars. The highest resolutions where you can get reasonable sharpness range from 1800 to 2400ppi. But it is still useful to scan at 4800 or 6400 ppi, as you are oversampling. You can downsample afterward and you then have an image with far less scan noise. Even then, you must sharpen liberally. You might also want to try different height settings on the film holder to try to achieve optimal focus. Further, put your scanner on a very solid surface, maybe the floor, as vibrations in the scanner also reduce resolution. My scans with my V750 are much worse if the scanner is on a wobbly, light table.

As far as the dark areas of slide film go, yes, this is a problem with these scanners. Velvia 50 is especially problematic. You can try using the multisampling and multi-exposure features of the scanning software. But they don't help much. With these scanners, you can only get about 4 stops of DR out of Velvia 50. A drum scan will get you more. With negative films, your scanner should easily capture the full range. I haven't tried Provia, Astia, or Velvia 100. Supposedly they are a little better than Velvia 50.

I've gotten sharp image at 3200 and discover its mostly impossible without the aid of the betterscanning adjustable height holders AND the ANR glass.

With the Epson holders, 1 they don't hold the film flat so the focus on the edges is better but the center is off, and 2 the holders themselves don't hold the film at the proper height and are actually at a hyperfocal distance so it's impossible to get a clear scan film them.

My only suggestion is, I only own the betterscanning holders for 120, so for 35mm, I take the film and place it on the Epson wet scanning glass that comes with the scanner then I take the 120 ANR glass and lay it on top, the height of the wet scanning glass seems to be about right, sometimes with curly 35mm I need two pieces of ANR on top of each other.

Anyway hope that helps, it's not that the scanner is incapable, just that the scanner's lens distance isn't really at the sweet spot so it takes some adjusting of the height of the film and making sure to keep it flat to get it just right.

oysteroid
11-Feb-2013, 16:38
StoneNYC,

I have adjusted the holder in very small increments using paper shims, looking at only one tiny spot, and moved through a range, making sure I went from unfocused to ideally focused to unfocused again, then returning to the most focused height to make sure I was at the ideal height. Then I tested various resolutions. Realistically, on my scanner, I don't have good sharpness above 2400ppi, and that's pushing it. More realistically, it is like 2100 or so. And tests by many others seem similar to my results. If you are able to get 3200, you must have a better example of the scanner than I have. But I would be very interested in seeing a full resolution crop of what you consider to be sharp output at 3200ppi.

StoneNYC
11-Feb-2013, 16:41
StoneNYC,

I have adjusted the holder in very small increments using paper shims, looking at only one tiny spot, and moved through a range, making sure I went from unfocused to ideally focused to unfocused again, then returning to the most focused height to make sure I was at the ideal height. Then I tested various resolutions. Realistically, on my scanner, I don't have good sharpness above 2400ppi, and that's pushing it. More realistically, it is like 2100 or so. And tests by many others seem similar to my results. If you are able to get 3200, you must have a better example of the scanner than I have. But I would be very interested in seeing a full resolution crop of what you consider to be sharp output at 3200ppi.

Sure ill try and crop it out in the next few days, I had a great example but deleted the fuzzy one, wish I kept it.

It's taken me a long time of perseverance and each type of film is different. If I don't post in a few days send me a reminder.

Nathan Potter
11-Feb-2013, 19:54
I did some tests using a standard chrome on glass resolution mask a while back with my V750.
The target I used was chrome on glass with a 16 X 16 arrayed set of 2540 X 2540 µm sized targets. The plate was sloped (wedged) from 0 to about 4mm in order to capture the point of best focus. I did three scans at 1800 spi, 2400 spi and 4800 spi.

The scans were measured at three cell line pairs for resolution and contrast using the difference divided by the sum method with photoshop K values converted to an 8 bit scale. A bit tedious to do this so I picked line pairs that would give reasonable data at the low and high bit ends. Summary below after checking this older data again.

SCAN / CELL ID / lp/mm / CONTRAST vertical / CONTRAST horizontal

1800 spi / 1-1 / 20 / 39% / 39%

2400 spi / 1-6 / 40 / 28% / 24%

4800 spi / 2-1 / 42 / 52% / 45%

The 1800 spi data is considerably inferior to the 2400 and 4800 data while there is a gain in contrast at 4800 spi over 2400. This is not inconsistent with what people have seen previously, although I was a bit surprised at the better contrast at close to the same lp/mm using 4800 over 2400.

Nate Potter Austin TX.

StoneNYC
11-Feb-2013, 21:39
I did some tests using a standard chrome on glass resolution mask a while back with my V750.
The target I used was chrome on glass with a 16 X 16 arrayed set of 2540 X 2540 µm sized targets. The plate was sloped (wedged) from 0 to about 4mm in order to capture the point of best focus. I did three scans at 1800 spi, 2400 spi and 4800 spi.

The scans were measured at three cell line pairs for resolution and contrast using the difference divided by the sum method with photoshop K values converted to an 8 bit scale. A bit tedious to do this so I picked line pairs that would give reasonable data at the low and high bit ends. Summary below after checking this older data again.

SCAN / CELL ID / lp/mm / CONTRAST vertical / CONTRAST horizontal

1800 spi / 1-1 / 20 / 39% / 39%

2400 spi / 1-6 / 40 / 28% / 24%

4800 spi / 2-1 / 42 / 52% / 45%

The 1800 spi data is considerably inferior to the 2400 and 4800 data while there is a gain in contrast at 4800 spi over 2400. This is not inconsistent with what people have seen previously, although I was a bit surprised at the better contrast at close to the same lp/mm using 4800 over 2400.

Nate Potter Austin TX.

I don't understand most of what you said honestly, but I believe that you discovered the V750 is capable if 4800 with good contrast/sharpness? And that the 1200 was actually somehow worse?

Can you explain this in "for dummies" terms?

SergeiR
12-Feb-2013, 15:13
Nathan, thanks for the test, its interesting, however i dont get this - if contrasts on 2400 are worse than on 1800/4500 how is 1800 is inferior? Also, one more thing - which lens array you used? One for 8x10 or smaller one?

Nathan Potter
12-Feb-2013, 22:43
I don't understand most of what you said honestly, but I believe that you discovered the V750 is capable if 4800 with good contrast/sharpness? And that the 1200 was actually somehow worse?

Can you explain this in "for dummies" terms?

StoneNYC, sorry for too much tecno babble. I forget that many are just not used to resolution measurements. I did three scans on the Epson V750 using the glass resolution mask pictured below.
One each at 1800, 2400 and 4800 samples per inch using the high resolution lens.

http://farm9.staticflickr.com/8236/8469145179_a3d44d0d9b_z.jpg (http://www.flickr.com/photos/argiolus/8469145179/)
TOPPANmsk-2 (http://www.flickr.com/photos/argiolus/8469145179/) by hypolimnas (http://www.flickr.com/people/argiolus/), on Flickr

The arrayed mask of 16 X 16 identical cells allowed me to slope the mask on the Epson using shims. That common technique is used to capture the point of best focus somewhere along the mask so that these cells then can be used for focus accurate resolution measurements. Since the Epson is an all digital capture device the most convenient way to measure the contrast between light line areas and dark line areas is to use photoshop eyedropper K % values. I'm interested in doing this measurement over the full density region that can be captured by the Epson so I try not to use any contrast enhancement and no sharpening in PS. I scan the eye dropper over the pairs of lines available and average the light and dark readings. For example I may have a string of readings; 70-1-53-6-58-0-70-1-43; a pretty noisy set so I may take more than one set for each line pair both horizontal and vertical. Lastly I apply Korens' (and others) method to determine the absolute contrast ratio between dark line and light lines.

The idea here is that we want to find the maximum Line Pairs / mm. that also yield a decent contrast value. A sort of standard contrast used by many is 50%. In digital print quality as well as all analog technique the human eye is very sensitive to contrast in an image so a full disclosure of scanner quality really needs both linewidth and contrast data to be truly quantitative.

Notice for this exercise I tried to guess ahead of time which lwpairs would yield about 50% contrast but failed to get it quite right. The scan at 1800 spi produced a pretty inferior image at best focus so I had to choose the 20 lp/mm pair to obtain 39% contrast. As I mentioned above both the 2400 and 4800 scans showed a notable difference in contrast at nearly similar lp/mm of around 40.

Now this kind of exercise may not tell you much about the actual reproduction quality of a film (especially color film) on the Epson. This is simply a quantitative comparison technique. For instance it appears there is a slight advantage in resolution and contrast using 4800 spi vs 2400 spi.

Nate Potter, Austin TX.

Nathan Potter
12-Feb-2013, 22:51
Nathan, thanks for the test, its interesting, however i dont get this - if contrasts on 2400 are worse than on 1800/4500 how is 1800 is inferior? Also, one more thing - which lens array you used? One for 8x10 or smaller one?

SergeiR, for the 1800 scan I needed to go down to pairs 1-1 (20 lp/mm) in order to get a 39% contrast while 2400 and 4800 yielded reasonably useable contrast at around 40 lp/mm. Used higher resolution lens array.

Nate Potter, Austin TX.

StoneNYC
12-Feb-2013, 22:59
Nathan thanks for trying to educate me, I think it's like trying to explain to a Neanderthal how to fly a jet airplane... when he hasn't even seen a wheel yet... ok here's my examples of not data just image stuff... lol
OK, best I could do, not a good image either just about the time I got frustrated with fuzzyness and wanted to try some experimentation.

The image, I haven't even cropped out the edges of the can yet......
89327


The cropped area...
89328

The scan at 3200 with Epson holder cropped 1:1 ...
89329

The scan at 3200 with betterscanning holder and ANR glass cropped 1:1 ...
89330

It's certainly not super great, this was Provia 100F, I seem to do a lot better with B&W film than color, but this is SUCH a tiny part of the frame, you could blow this up really big and not see any blur. I don't have technical understanding of film and scanning and all the tech side, but this looks really sharp even larger than 20x24. It still has fuzzy edges, and I recognize that, I think I could probably hone it even better I just don't have the patience honestly.

Hope this is in some way helpful.

StoneNYC
12-Feb-2013, 23:01
Oh, and this is from a 6x7 Medium format image NOT a LF image... just clarifying...

oysteroid
13-Feb-2013, 02:29
Thanks for the examples. Here is a sample of my own, scanned on a V750 at 4800ppi, resampled to 3200ppi, and sharpened liberally. This is from a 4x5 sheet of Portra 160, shot with a Schneider Apo-Symmar 210mm. I scanned it in the plastic holder.

89335

I don't consider this sharp. And even to get that level of sharpness, I had to sharpen a lot, which accentuates the grain quite a bit. If I visually compare this to a 2000ppi version, I really don't see that it contains any additional useful information. Even the 2000ppi version isn't "sharp" and doesn't really have pixel level image information, not like the output of a Foveon sensor. Here it is at 2000ppi, sharpened as much as I think it can tolerate without making the grain so strong that it ruins the image:

89336

oysteroid
13-Feb-2013, 02:39
StoneNYC, I suspect that your scan with the regular holder could be improved some by adjusting its height a bit. It looks rather out of focus to me. It has a lot more CAs than I normally see as well. Are you scanning as close to the center of the glass as possible? Performance is a little better in the center. Theoretically, a dry scan with the Betterscanning holder shouldn't be any sharper if the film is at the same height in both cases. I am talking locally of course. The problem with the regular holder is lack of film flatness, so the sharpness might not be consistent across the image.

Linhof
13-Feb-2013, 07:02
In addition to the issue of sharpness, contrast is another big issue, especially Velvia 50 film that I had a lot. How do you deal with this problem with V700?

Nathan Potter
13-Feb-2013, 12:05
StoneNYC, your Provia image scanned on the Epson with the betterscanning holder seems somewhat less than what I would expect in clarity. I assume this is 4X5 film and if so your small flower section is about 10mm (10,000 µm) wide referenced to the film. Your magnified view of the 10mm section shows detail (a reasonable line detail of about 1/200 of the 10 mm) which would be in the vicinity of a 50 µm line width. This is 20 lines/mm or 10 lp/mm. The limit may be on the film but should not be in the Epson. Scaling off a jpeg image on a computer screen is a poor way to assess image quality but one can get a rough estimate.

Nate Potter, Austin TX.

StoneNYC
13-Feb-2013, 12:32
Thanks for the examples. Here is a sample of my own, scanned on a V750 at 4800ppi, resampled to 3200ppi, and sharpened liberally. This is from a 4x5 sheet of Portra 160, shot with a Schneider Apo-Symmar 210mm. I scanned it in the plastic holder.

89335

I don't consider this sharp. And even to get that level of sharpness, I had to sharpen a lot, which accentuates the grain quite a bit. If I visually compare this to a 2000ppi version, I really don't see that it contains any additional useful information. Even the 2000ppi version isn't "sharp" and doesn't really have pixel level image information, not like the output of a Foveon sensor. Here it is at 2000ppi, sharpened as much as I think it can tolerate without making the grain so strong that it ruins the image:

89336

I wanted to show a straight scan, I could have easily re sampled it smaller and sharpened it further, I wanted the pure scan...

StoneNYC
13-Feb-2013, 12:35
StoneNYC, I suspect that your scan with the regular holder could be improved some by adjusting its height a bit. It looks rather out of focus to me. It has a lot more CAs than I normally see as well. Are you scanning as close to the center of the glass as possible? Performance is a little better in the center. Theoretically, a dry scan with the Betterscanning holder shouldn't be any sharper if the film is at the same height in both cases. I am talking locally of course. The problem with the regular holder is lack of film flatness, so the sharpness might not be consistent across the image.

Well the betterscanning does two things, one it keeps the plane of film almost entirely flat, were the Epson holders do not, so there's always a bend at the edges.

Secondly the Epson holders aren't adjustable and I can't get them any closer to the glass, so they are impossible to get any clearer, it's a really big fail. I suspect they planned for pros to get better holders and made these cheap holders for the masses.

StoneNYC
13-Feb-2013, 12:36
StoneNYC, your Provia image scanned on the Epson with the betterscanning holder seems somewhat less than what I would expect in clarity. I assume this is 4X5 film and if so your small flower section is about 10mm (10,000 µm) wide referenced to the film. Your magnified view of the 10mm section shows detail (a reasonable line detail of about 1/200 of the 10 mm) which would be in the vicinity of a 50 µm line width. This is 20 lines/mm or 10 lp/mm. The limit may be on the film but should not be in the Epson. Scaling off a jpeg image on a computer screen is a poor way to assess image quality but one can get a rough estimate.

Nate Potter, Austin TX.

I you read the second post, I did say this is from a 6x7 piece of film NOT a 4x5.

Do your calculations based on that surface area if it helps you.

oysteroid
13-Feb-2013, 23:54
Well the betterscanning does two things, one it keeps the plane of film almost entirely flat, were the Epson holders do not, so there's always a bend at the edges.

Secondly the Epson holders aren't adjustable and I can't get them any closer to the glass, so they are impossible to get any clearer, it's a really big fail. I suspect they planned for pros to get better holders and made these cheap holders for the masses.

Yes, that holder with anti-newton-ring-glass would certainly be better for holding the film flat. I am surprised that you feel that you need to get the included plastic holder closer to the glass than it can possibly go in order to get optimal focus. Mine is optimally focused with the little adjustment things pointed toward the "O", which is basically the middle setting. Taking them off entirely is the lowest setting. Pointing them toward the "+" is the highest setting. But you can get levels in between with little paper shims if you like. Anyway, if yours truly has a focal point below the lowest setting, this points to rather poor consistency between different samples of these scanners.

But as far as flatness goes, I don't really have any problem. I just stick the negatives between some pages of a heavy textbook and stick it on a warm radiator if they are curved and they flatten right out and remain quite flat in my holder. I get consistent focus across the frame. This is for 4x5 sheet film though. I haven't tried that with roll film. For my purposes, it doesn't seem worth messing with fluid mounts or anything, especially since solvents are involved, and it is cold here in Colorado, and so good ventilation is out of the question. And while I don't really know how much of a problem it is, it seems to me that introducing more refractive surfaces, such as another sheet of glass, could reduce quality in some way or another. For one thing, light bounces back and forth between the two surfaces of a sheet of glass, just like you see in a silver-backed mirror. Ideally, there would be no glass at all. But dust getting inside would be a nightmare.

StoneNYC
14-Feb-2013, 01:36
Yes, that holder with anti-newton-ring-glass would certainly be better for holding the film flat. I am surprised that you feel that you need to get the included plastic holder closer to the glass than it can possibly go in order to get optimal focus. Mine is optimally focused with the little adjustment things pointed toward the "O", which is basically the middle setting. Taking them off entirely is the lowest setting. Pointing them toward the "+" is the highest setting. But you can get levels in between with little paper shims if you like. Anyway, if yours truly has a focal point below the lowest setting, this points to rather poor consistency between different samples of these scanners.

But as far as flatness goes, I don't really have any problem. I just stick the negatives between some pages of a heavy textbook and stick it on a warm radiator if they are curved and they flatten right out and remain quite flat in my holder. I get consistent focus across the frame. This is for 4x5 sheet film though. I haven't tried that with roll film. For my purposes, it doesn't seem worth messing with fluid mounts or anything, especially since solvents are involved, and it is cold here in Colorado, and so good ventilation is out of the question. And while I don't really know how much of a problem it is, it seems to me that introducing more refractive surfaces, such as another sheet of glass, could reduce quality in some way or another. For one thing, light bounces back and forth between the two surfaces of a sheet of glass, just like you see in a silver-backed mirror. Ideally, there would be no glass at all. But dust getting inside would be a nightmare.

I should clarify I don't actually use the wet moony as a wet mount, I use it dry, 35mm and 70mm film is so curly that I have to lay 2 120 size ANR glass pieces on top if the 35mm film to get it remotely thin, and even that isn't perfect. But it's not bad, if I had 4x5's I'm sure that wouldn't be as bad the larger the film the less total curl seems to happen so it's more the 35mm stuff. The film is so curly I don't have to worry about newton ring problems with the wet mount glass because with two pieces of glass on top the center of the film still doesn't touch the wet mount glass, just the sprocket edges.

I've been really lazy about adjusting the holders to be perfect so I'll try gettin the betterscanning holders even more well adjusted and see if I can get more detail out of 3200 on the 120 film.

I came here to the LF forum because I want to buy a 4x5 or 5x7 field camera and skip the rail stuff I just haven't found one in my budget that has all the movements I want, like a Toyo bit a Crown Graphic so that's why I don't have 4x5 to test with yet.