Achievable resolution with Epson V700/V750/V800/V850

[blockend]

which has a filthy (misty) underside platen -- this needs to be cleaned one day,

Same here. I'm not sure it makes any difference to IQ. At least I'm hoping not, though it could soften things a little.

As chance would have it I'm currently scanning 120 colour negatives at 2400 on a V500. They look ok, excellent in fact for a 60 year old 3-element lens. Not sure I'd print from the scans tho'.

 

[ParkerSmithPhoto]

What scanner were you using? How did you process the scan (sharpen, etc.)? what dod you mean "except for matt/glossy surfaces"? Thanks.

Alan, Sorry about the slow reply. Haven't logged in for a while.

I used a Canoscan 9000 with a large piece of glass to which I fluid mounted the negative.

When comparing the prints, the silver print had a glossy surface and the pigment print had a matt surface. Under glass, however, matt and glossy look essentially the same.

 

[faberryman]

Not from using "the best scanner".
Not from using "the best printer".
But by making a blood good photograph at the start that will scan and print well, without floss, artifice or arguing about resolutions.
There is no mystery to beautiful prints. It all starts with the photographer's knowledge and understanding of the subject, and his/her ability to best bring it across to film.
At that stage it has nothing at all to do with scanners and printers.

You are conflating the aesthetic with the technical. A poor print of a great image is still a poor print.

 

[ParkerSmithPhoto]

Interesting indeed...I know of at least six photographers who still scan with Epson 4990s, but it may be asking too much to also include this venerable veteran in the list...

I previously used an Epson 4990 with a fluorescent tube, took forever to heat up but made beautiful scans. I wet mounted and scanned my 6x7 rodeo negatives, then printed them all with Eboni carbon pigments (OMG! DID I JUST SAY THAT ON APUG???) and everyone who saw them said they were just gorgeous. You could see the threads in the weave of their shirts.
This was all done at 2400 PPI. We all think more is better but when scanning film you soon run out of scannable data. Cranking up the resolution doesn't give you more data after a certain point, just a much larger file. 6x7 at 2400ppi gives around a 16" file at 300 dpi. It's plenty of data.

 

[Down Under]

... the apparent grain gets much smaller scanning at 4800 dpi, and can be sharpened a lot more.

A good point. A film archivist I know and have great respect for (having seen his finished work) once gave me the following advice on scanning which he applied to both 35mm and roll films for a university library archive. I have to say his results certainly spoke for themselves.

1. Use the professional mode but turn off everything on the scanner. especially sharpening. Go for a basic scan. This will (usually) be quite gray.
2. Scan to TIFF.
3. Set 120 at 3600 and 35mm at 4800.
4. Reduce your scan to a sensible size for storage.
5. Never do any work on your original scan.
6. Number your scans, keep a notebook with detailed notes.
7. Be ultra selective. Don't scan every image in your negative files. Life is too short!

I have followed this advice and produced some excellent scans, but have never put any of these points to any in-depth testing.

Anyway, FYI. I am particularly interested in others' comments on points 2, 3 and 4.

 

[Down Under]

I previously used an Epson 4990 with a fluorescent tube, took forever to heat up but made beautiful scans. I wet mounted and scanned my 6x7 rodeo negatives ... and everyone who saw them said they were just gorgeous. You could see the threads in the weave of their shirts. This was all done at 2400 PPI. We all think more is better but when scanning film you soon run out of scannable data. Cranking up the resolution doesn't give you more data after a certain point, just a much larger file. 6x7 at 2400ppi gives around a 16" file at 300 dpi. It's plenty of data.

Good points overall. It makes sense to use whatever equipment you have to the best of its capabilities.

 

[Ted Baker]

said

1. Use the professional mode but turn off everything on the scanner. especially sharpening. Go for a basic scan. This will (usually) be quite gray.

-Sharpening can't be reversed, if you want the option to decide to use a different amount or a different algorithm, then do it later. Same applies if the image is clipped.​

2. Scan to TIFF.

For the epson 16bit if you intend to do any post processing, it's essential if you intend to do all adjustments later.​

3. Set 120 at 3600 and 35mm at 4800.

3 and 4 are related. The scanner always scans at the same resolution, it also uses a technique called pixel shifting. The software/firmware in the scanner itself resizes the image. Unfortunately these scanners use USB 2.0 and old technology, that could be easily improved, but has not for 10 years or more. Using a different resizing algorithm will change the final result. It is however very likely the epson algorithm is optimal for these scanners.​

4. Reduce your scan to a sensible size for storage.

see point 3​

5. Never do any work on your original scan.

Good advice​

6. Number your scans, keep a notebook with detailed notes.

Let the computer do that for you​

7. Be ultra selective. Don't scan every image in your negative files. Life is too short

This is old old advice, not applicable for negative scanning, when you don't have access to contact sheets.​

 

[Alan Edward Klein]

A good point. A film archivist I know and have great respect for (having seen his finished work) once gave me the following advice on scanning which he applied to both 35mm and roll films for a university library archive. I have to say his results certainly spoke for themselves.

1. Use the professional mode but turn off everything on the scanner. especially sharpening. Go for a basic scan. This will (usually) be quite gray.
2. Scan to TIFF.
3. Set 120 at 3600 and 35mm at 4800.
4. Reduce your scan to a sensible size for storage.
5. Never do any work on your original scan.
6. Number your scans, keep a notebook with detailed notes.
7. Be ultra selective. Don't scan every image in your negative files. Life is too short!

I have followed this advice and produced some excellent scans, but have never put any of these points to any in-depth testing.

Anyway, FYI. I am particularly interested in others' comments on points 2, 3 and 4.

2. I scan to TIFF also to avoid jpeg artifacts.
3. If you think you're getting resolution at 4800 for 35mm, why not use 4800 on 120 film instead of 3600? I use 2400 because someone once said I'm not going to get more than 2400 on my Epson V600.
4. Storage is cheap. Scan for max resolution that the scanner will actually give you not their sales max figure.
7. You can do a quick "contact" print scan by selecting 300 resolution. The scanner works quickly at lower settings. Scan flat with all auto adjustment turned off. Then you can select the best and re-scan that one at the higher resolution.

 

[Doug Fisher]

>> I'm not sure it makes any difference to IQ. At least I'm hoping not, though it could soften things a little.<<

If it has been quite a while or forever, cleaning off the outgassing haze is definitely worth the effort!

Doug

 

[faberryman]

3. If you think you're getting resolution at 4800 for 35mm, why not use 4800 on 120 film instead of 3600? I use 2400 because someone once said I'm not going to get more than 2400 on my Epson V600.

At least with my film scanner, I have to scan at 6400 to get the maximum actual resolution, which is 4300. If I scanned at 4800 or 3600, I would actually get less.

 

[Ted Baker]

At least with my film scanner, I have to scan at 6400 to get the maximum actual resolution, which is 4300. If I scanned at 4800 or 3600, I would actually get less.

I think that applies universally with almost all of these film scanners that use a CCD.

If anyone has any scans of resolution targets taken with a v700/v800 series of scanners at different "scanner resolutions" they would be happy to send me (drop box or google drive) I would like to try some of the different re-sampling approaches. Just PM me.

 

[faberryman]

I was responding to point 3 of Alan Edward Klein's post. If he is scanning at 2400, he likely isn't getting the maximum resolution of his V600. According to filmscanner.info, the maximum resolution of the V600 is 1560, and is achieved by scanning at 3200 or higher. However, there is no benefit of scanning higher than 3200; you won't get more resolution than 1560.

 

[Ted Baker]

I was responding to point 3 of Alan Edward Klein's post. If he is scanning at 2400, he likely isn't getting the maximum resolution of his V600. According to filmscanner.info, the maximum resolution of the V600 is 1560, and is achieved by scanning at 3200 or higher. However, there is no benefit of scanning higher than 3200; you won't get more resolution than 1560.

It's confusing, however the point I am trying to make is that for example even though Alan Edward Klein may set 2400dpi in his scanning program, the scanner still works the same, or at least almost the same (as I am not sure if pixel shifting is used below 3200spi).

i.e. it scans at its maximum resolution, then the image is re sampled to the new resolution.

 

[MattKing]

It's confusing, however the point I am trying to make is that for example even though Alan Edward Klein may set 2400dpi in his scanning program, the scanner still works the same, or at least almost the same (as I am not sure if pixel shifting is used below 3200spi).

i.e. it scans at its maximum resolution, then the image is re sampled to the new resolution.

I've seen this information before, but it doesn't seem to make sense to me.
The reason? With my scanner(s) if I scan at lower resolution they actually scan more quickly - the scanning head travels more quickly to the end of its travel.

 

[faberryman]

I've seen this information before, but it doesn't seem to make sense to me.
The reason? With my scanner(s) if I scan at lower resolution they actually scan more quickly - the scanning head travels more quickly to the end of its travel.

That is my experience is well. The lower the selected resolution, the faster the mechanism physically scans.

 

[Ted Baker]

The reason? With my scanner(s) if I scan at lower resolution they actually scan more quickly - the scanning head travels more quickly to the end of its travel.

Yes but the sample size does not actually change, the sensor does not change size and there is no zoom lens. Epson use a technique called pixel shift, I am not sure exactly how it's implemented, but it will defnitely be in use at 6400dpi.
At the higher resolutions it requires the stepper motor to move in smaller increments, plus more processing power, and sent over USB 2.0 to your pc.

 

[faberryman]

Yes but the sample size does not actually change, the sensor does not change size and there is no zoom lens. Epson use a technique called pixel shift, I am not sure exactly how it's implemented, but it will defnitely be in use at 6400dpi.
At the higher resolutions it requires the stepper motor to move in smaller increments, plus more processing power, and sent over USB 2.0 to your pc.

Are you suggesting at a 300 setting it scans at maximum resolution and downsamples so you end up with less information in the resulting image?

 

[Ted Baker]

Are you suggesting at a 300 setting it scans at maximum resolution and downsamples so you end up with less information in the resulting image?

Yes, as the sample size does not change. The algorithm maybe well be simpler.

 

[faberryman]

Yes, as the sample size does not change. The algorithm maybe well be simpler.

That is certainly counter-intuitive.

 

[alanrockwood]

Regarding image quality when scanning at different ppi, are there other benefits to scanning at higher ppi? For example, even if resolution does not increase, do you get other benefits, such as less grain aliasing or better signal/noise? Does the picture look less pixelated? Does sharpening work better if you scan at a higher resolution than the ppi limit?

 

[Ted Baker]

Regarding image quality when scanning at different ppi, are there other benefits to scanning at higher ppi? For example, even if resolution does not increase, do you get other benefits, such as less grain aliasing or better signal/noise? Does the picture look less pixelated? Does sharpening work better if you scan at a higher resolution than the ppi limit?

In summary yes, however the image is already scanned at the highest resolution in the first place. What I am interested in if you can get a better result if you do the re-sampling outside of epson scanner/software.

 

[Alan Edward Klein]

In summary yes, however the image is already scanned at the highest resolution in the first place. What I am interested in if you can get a better result if you do the re-sampling outside of epson scanner/software.

Ted, What do you mean by re-sampling outside of Epson scanner/software?

 

[Ted Baker]

Ted, What do you mean by re-sampling outside of Epson scanner/software?

Pretty much what I said

Take the scan data from the ccd in the rawest form available. In the case of v700 that would at 6400dpi after it has applied it's pixel shift algorithm, and try different resizing algorithms in different software, to see if you can get a small improvement. The main problem is the USB 2.0 interface is very slow so it adds a few extra minutes to the scan.

This is old diagram that shows the internals of this type of scanner. http://content.epson.it/maco/technology/scanners/ccd.htm

 

[alanrockwood]

This is old diagram that shows the internals of this type of scanner. http://content.epson.it/maco/technology/scanners/ccd.htm

Thanks for the link. I have been looking for something like that.

 

[jeffreyg]

For what it is worth... I've read many threads and various articles on scanning resolution and decided to do a somewhat test with my equipment. Here goes. I photographed a tree in my back yard with a Hasselblad on a tripod with the 50mm lens, Ilford Delta 400 at box speed developed in Ilford DDX as per mfg time/temperature. I scanned the negative with an Epson 4870 and Silverfast Ai Studio software as Tiffs. Without changing the negative and its position I did four scans 1) 4800ppi which took 5min 10 sec including scan + processing + saving, 2) 2400ppi 1min 45 sec for the same, 3) 1200ppi 45 sec for the same and 4) 300ppi 20 sec for the same. In PhotoShop CS5 (with no editing). I enlarged each to 6 inches, then down sized all but the 300ppi to 600ppi and printed them with an Epson 3880. The printer was set to ultra luster, mode ACCU HD2,no color management and superfine 1440 dpi. At arms length viewing other than the 300ppi there is no appreciable difference in sharpness but with a magnifying glass the 4800ppi scan is definitely sharper. IMOP scann at high resolution, edit then down size in increments before printing especially if making larger prints.

http://www.jeffreyglasser.com/