8x10 with a V700, Vuescan can't scan above 1600 DPI

[urs0polar]

Hi John,

I don't know what everyone else is saying, but what I'm saying is that no matter what I set the resolution at above 1600 ppi, the resulting file is 1600 ppi if you divide the resolution by the number of inches it actually scanned (~8" in one case, ~10" in the other). Maybe it does some internal interpolating or something, especially since as I understand it, 2400 is an interpolated resolution anyhow... so, it either interpolates up form 1600, or interpolates down from 3200. Not sure which yours is doing.

If Epson scan gives you more pixels (whether or not they actually hold more information or not is another argument), then maybe it's a "feature" of Vuescan not to output more than 1600ppi.

This is an interesting thread to me. I get very sharp 8x10 scans with a V750 at 2400 dpi. Are you all saying that if I drop back to 1600 I will get better results?

Edit: So, 1600 is not a default option in Epson scan; you have to type it in. I just did a quick test with MF film. I did one scan at 2400 and another at 1600. I then downsized the 2400 image in PS to the size of the 1600. The 1600 is sharper at 100 percent than the downsized 2400 on screen, but I'm not sure what this means in terms of printing. If I sent the 2400 out for printing at 16x20 would it actually be less sharp than the 1600?

[urs0polar]

To answer optimum focus questions, I have a few comments.

One is that there are 2 different lenses in the 700/750. One is focused at or about the glass, and the other (the "high res" one) is at the height of an "average" film holder.

The one that is on the glass will scan an 8x10 area. The one that is at the filmholder height will max out at 6"x8" or something like that.

So, since there is no such thing as an "average" film holder, pretty much everyone's focus is terrible with the filmholder lens.

With my betterscanning variable height mounting station, my optimum film holder height happens to be at ~1.8mm. Well, at least that seems like the best, afaik.

see here: http://www.flickr.com/photos/urs0polar/3317832213/

Now, the 8x10 on the glass isn't that sharp, but it's a whole lot sharper than 6x7 used to be before I went to the trouble of the mounting station.

Then again, 6x7 at 4800 or whatever vs 8x10 at 1600, who knows. Different cameras, different lenses, different film, different everything

For rdenney,

Yeah, I guess that it just goes to show how good of a scanner the 9000 is that 6x7 looks comparable-ish to 8x10 on the V700. Of course, maybe I need to mount 8x10 to some upside down AN glass and start shimming it to get better focus. Maybe the fact that at 1600 ppi at 100%, I'm still not seeing the grain per se (i think) has something to do with it.

Mark "definitely arm-waving" Hodos

[rdenney]

For rdenney,

Yeah, I guess that it just goes to show how good of a scanner the 9000 is that 6x7 looks comparable-ish to 8x10 on the V700.

I'm more off-topic than that. My comparison was between 6x7 in the Nikon to 4x5 at 2400 in the V750 (using the supplied film holder, which in my case, seems to be miraculously in focus). Those provide about the same number of pixels.

8x10 ought to be better, even at 1600, where it provides considerably more pixels to work with than 6x7 in the Nikon, and with no chance of grain aliasing or other grain effects. Is it all 8x10 can provide? Of course not. The Nikon scanning 6x7 is much closer to the limits of the film. But it's plenty for prints from a 24" printer. And it may be close to the best the Epson can do in any case.

Rick "who has the wet-mount station and needs to try it out" Denney

[John NYC]

Hi John,

I don't know what everyone else is saying, but what I'm saying is that no matter what I set the resolution at above 1600 ppi, the resulting file is 1600 ppi if you divide the resolution by the number of inches it actually scanned (~8" in one case, ~10" in the other). Maybe it does some internal interpolating or something, especially since as I understand it, 2400 is an interpolated resolution anyhow... so, it either interpolates up form 1600, or interpolates down from 3200. Not sure which yours is doing.

If Epson scan gives you more pixels (whether or not they actually hold more information or not is another argument), then maybe it's a "feature" of Vuescan not to output more than 1600ppi.

Nope using Epson Scan on my V750, I get way more pixels at 2400 than I do at 1600. When I scan 8x10 at 2400, I get a HUGE file.

What is bothering me is are the extra pixels worth anything in terms of resolution when going to the printer? When I downsized the 2400 file to be the same size at the 1600 file, the 1600 file was sharper.

It gets even weirder. When I used PS to "upsample" the 1600 file to 2400 (and a respectively larger number of pixels per side) it was STILL sharper than the scan that was originally done at 2400.

EDIT: I just did a bunch of googling and found on flickr that Tim Parkin (who also posts here sometimes) found that 1600 the breakpoint for the V750. I tend to trust that guy as he does an amazing amount of research into his interests (plus we both do software, which makes me immediately like him ;-).

This has actually been a great thread in that I am going to save TONS of hard disk space by switching to 1600 instead of 2400 scans.

[urs0polar]

Nope using Epson Scan on my V750, I get way more pixels at 2400 than I do at 1600. When I scan 8x10 at 2400, I get a HUGE file.

What is bothering me is are the extra pixels worth anything in terms of resolution when going to the printer? When I downsized the 2400 file to be the same size at the 1600 file, the 1600 file was sharper.

It gets even weirder. When I used PS to "upsample" the 1600 file to 2400 (and a respectively larger number of pixels per side) it was STILL sharper than the scan that was originally done at 2400.

EDIT: I just did a bunch of googling and found on flickr that Tim Parkin (who also posts here sometimes) found that 1600 the breakpoint for the V750. I tend to trust that guy as he does an amazing amount of research into his interests (plus we both do software, which makes me immediately like him ;-).

This has actually been a great thread in that I am going to save TONS of hard disk space by switching to 1600 instead of 2400 scans.

Yeah, I think that basically from your experiments, PS upscales from 1600 to 2400 better than Epson scan does, since the 1600 is sharper. Which would make sense with my experiments in that when I tell Vuescan to give me optically 1600 or optically 3200, but don't interpolate, I get 1600.

Interesting. Wow a thread that's actually solved something! ... for two people at the same time!

I'm in software too -- this sort of thing doesn't happen that often

I've seen some of your scans in other threads, and they are pretty sharp. Are you wet mounting or raising the negative off the glass at all? I'm not into wet-mounting (tiny nyc apartment as it is), but i could be convinced to start shimming a little if I'm feeling masochistic.

[jim kitchen]

Gentlemen,

I conducted an experiment last year to see what the difference might be with a scanned 8X10 negative that had sharpening applied to the scan, using Epson's various unsharp masks, while the negative was scanned, compared to not...

The resulting "low level" scan impressed me, regarding the scan's sharpness, so much so, that I do consider this low level setting to be another scanning option. The resulting "low level unsharp mask" file was quite pleasant to work with, but I would not use this setting as a continuous setting going forward, since I really do not need to do this with my 8X10 negatives.

I can only state that the resulting scanned file was sharper, but the noise level did increase with minimum effect, which could be removed later, if required...

If you have a moment, I would suggest you try this low level unsharp mask setting to see whether it suits your work flow. This setting, initiated during a scan, surely goes against most, if not all written literature to affect a scanned image, but I know that since I use an 8X10 negative, any resulting errors were minimized. Smaller scanned negatives may see the effect more dramatically, compared to an 8X10 negative, therefore generating a less than positive result. This setting does not compensate for blocked shadows, so a drum scan is still the king the hill when that issue arises.

Just my two pennies...

jim k

[John NYC]

I've seen some of your scans in other threads, and they are pretty sharp. Are you wet mounting or raising the negative off the glass at all? I'm not into wet-mounting (tiny nyc apartment as it is), but i could be convinced to start shimming a little if I'm feeling masochistic.

I'm not wet mounting, but the negs are always lifted off the glass in one of two ways, For 35mm, medium format and 4x5, I have always used the film holders that came with the V750. For 8x10, I use a variation of a technique I learned here. I'll copy in a description I wrote about it over on the 8x10 group on flickr:

"1. Get a piece of ANR glass that is larger by at least a half inch than 8x10.

2. Tape the negative to the glass. Put the emulsion side against the ANR glass. Make sure you tape the neg so there is no slack in the middle when you turn the glass upside down. Otherwise, you will get a Newton ring in the middle of your picture (see next step).

3. Place the glass with the negative (backing side) down on the glass, but don't rest it on the glass. Do one of two things: 1) Rest the edge of one long side along the rim of the scanner bed and place a quarter on the other side of the glass to allow air space between the scanner glass and the back side of the neg or (2) accomplish the same thing by placing pennies or other coins on all four corners of the ANR glass.

Scan away! As long as your neg is not sagging and touching the scanner glass in the middle you will have zero newton rings. The Epson 750 I use seems to understand how to focus when the neg is just a little bit off the glass."

[John NYC]

Gentlemen,

I conducted an experiment last year to see what the difference might be with a scanned 8X10 negative that had sharpening applied to the scan, using Epson's various unsharp masks, while the negative was scanned, compared to not...

snip

Just my two pennies...

jim k

The Epson sharpening can definitely look good. But what I am in the habit of now is to sharpen in PS afterward (using smart sharpen or more often the regular USM filter). Sometimes after downsizing the neg in PS for posting online, I don't even sharpen it as just the downsizing created a sharp enough image.

Now that I've been doing this a little longer, I really should do a comparison of the two techniques again though.

[jim kitchen]

Dear John,

I do sharpen my 8X10 scanned negatives in PS after I set the finished image size too, but I use a very minimal USM setting, such as 80,3,0. If I marry two 8X10 scanned negatives together, sharpening is not employed resulting from the compressed data present in the completed printer file.

Using USM during a scan produces a sharper original digital file, but as I mentioned earlier, the noise level increases too, which is not a problem for an 8X10 negative, and any incremental noise can be removed successfully. My first impression happens to be that using Epson's low level USM with the 750 during an 8X10 scan, produces a scanned file that looks as though it was a drum scanned negative, where the negative's edge sharpness surprisingly emulates the drum scanned information. Is all the finite detail there compared to a drum scan? Certainly not, but the perceived sharpness is surely there at this low setting, compared to the softer normal Epson scan without low level USM.

Finding the optimum focal point above the scanner's glass bed for any Epson scanner, was my first priority, and once I discovered that location, my images snapped into focus with excellent results. Determining whether I could improve upon that result was left to my experimentation with USM during a scan. I still believe that any USM applied against a smaller negative will deteriorate the results, accordingly, since I do not believe that USM is applied with any programmable linearity, and matter of fact I would bet money that the USM function is not linear, or scaled to the negative size.

The resulting finished image with low level USM during the scan, presents a different look, and feel to the finished image, complete with any applied post processing, compared to not using low level USM during a scan. I have not done any further testing to determine whether I would pursue the low level USM setting during a scan going forward, but it looked tempting. For the moment, I achieve the desired sharpness level within my finished images, while using the tools present with PS, and without using USM during a scan...

jim k

[John NYC]

For the moment, I achieve the desired sharpness level within my finished images, while using the tools present with PS, and without using USM during a scan...

jim k

Yes, exactly what I do currently. To see what I get there, have a look at my 8x10s on Flickr at the largest "original" size. You have to be a flickr member for the "All sizes" button to show up to see these sizes.

http://www.flickr.com/photos/33946021@N04/tags/8x10/

What I am kind of excited and disappointed about at the same time is that the Epson seems to max out on resolution at 1600, I see now. So, smaller files = good for my hard drive. Less resolution than I thought I'd ultimately get out of the Epson = bad for prints(?).