scanning question

[ignatiusjk]

I recently bought a Epson V700 scanner. My question is this. Is the resolution you see on the screen going to be the same as when you print it? I've scanned a tack sharp neg at 400 dpi and 4800 dpi and can't tell the difference between them. I had the same neg scanned and printed at a lab to a 5x7 and the neg is amazingly sharp what is the deal?

[Rob Champagne]

you work it out backwards.

Say you want a 20x16 inch print. You need 360 pixels per inch (on an epson) for approx 7 line pairs per millimeter in the print which is high quality. So 20x360 = 7200 pixels width required across the print. So your scan needs to scan to produce a file 7200 pixels wide. So if you have a 35mm neg, that is 36mm wide, then you need to scan at: (36mm = 1.42 inches) 7200 / 1.42 = 5070 dpi to get a file size big enough to print a 20x16 at 360ppi.

Your screen is much lower resolution and is probably around 90 pixels per inch. So if you view the scanned file at 100%, it will be stretched out and look soft. If you view it at print size, then it will be an iterpretation at approx 90ppi so will still look soft. Only when you print will you be able to see its true resolution.

many people say you only need 300ppi print resolution and others say you can't tell the difference over 200ppi. Take your pick and do the maths above substituting your print size and your neg size and your required print resolution.

[Brian Ellis]

"I recently bought a Epson V700 scanner. My question is this. Is the resolution you see on the screen going to be the same as when you print it?"

No. The resolution of your monitor will normally be much less than the resolution of your print.

[David Luttmann]

Your monitor rez will between 72 and 96ppi on average. That means if you scan at 4800ppi X 5inches on the wide side, you’ll have 24000 pixels wide in your scan. Even at a screen rez of 96ppi, viewing at 100% is like looking at a print nearly 21 feet wide.

Something to keep in mind is that the V700 has a true resolution of approx 2050 to 2150ppi. You won’t find much of any advantage to scanning over 2400ppi on the Epson. Doing that will give you an uninterpolated 360dpi print out to 33” wide. Only when the better optics kick in at 6400 ppi does the rez step up to a real 2400ppi or so. The problem is that you’ll be working with a 4.9gb file to get it. Not worth it in my experience!

[ignatiusjk]

Dear Dave, how many ppi would I need for a GOOD 16x20 b&w print. I am scanning 4x5 negs,and trans.

[ignatiusjk]

What rez is that?

[Ron Marshall]

When I print on an Epson, I usually use 360 ppi, but you can get by with less.

Scan at 2400, then you will have some room for cropping.

[David Luttmann]

Dear Dave, how many ppi would I need for a GOOD 16x20 b&w print. I am scanning 4x5 negs,and trans.

I send everything to my Epson 3800 at 360ppi. From a 2400ppi scan, I get just less than 12,000 pixels. I do all my work on the native size file, and then downsample with bicubic to 360ppi for a 16x20. There is no point in feeding more to the print driver as you won't get any better results.

Hope that helps.

[Lenny Eiger]

I recently bought a Epson V700 scanner. My question is this. Is the resolution you see on the screen going to be the same as when you print it? I've scanned a tack sharp neg at 400 dpi and 4800 dpi and can't tell the difference between them. I had the same neg scanned and printed at a lab to a 5x7 and the neg is amazingly sharp what is the deal?

The reason is optical resolution. This is the problem I have always had with this scanner. One needs to make a distinction between the number of pixels a scanner can generate and how many bars it can actually resolve. This is similar to using 8x10 film with a pinhole camera. The pinhole image isn't going to be any sharper by virtue of it being taken with an 8x10.

The optical resolution of an Epson scanner is likely around 900-1000 (there is some disagreement here). It's true that your image won't improve at 4800, as it won't go past 1000. It's akin to interpolating, which doesn't actually help, in my opinion (or on my equipment).

I sell scanning services, so my bias can be up front here... but you can't get a horse to run in the Indy 500. You probably have a lot invested in cameras, more in good glass. You're doing it a disservice (as you have discovered) with a cheap scanner.

Lenny
EigerStudios.com

[hassiman]

Lenny,

What is your opinion of the CoolScan 9000?:confused:

[David Luttmann]

The reason is optical resolution. This is the problem I have always had with this scanner. One needs to make a distinction between the number of pixels a scanner can generate and how many bars it can actually resolve. This is similar to using 8x10 film with a pinhole camera. The pinhole image isn't going to be any sharper by virtue of it being taken with an 8x10.

The optical resolution of an Epson scanner is likely around 900-1000 (there is some disagreement here). It's true that your image won't improve at 4800, as it won't go past 1000. It's akin to interpolating, which doesn't actually help, in my opinion (or on my equipment).

I sell scanning services, so my bias can be up front here... but you can't get a horse to run in the Indy 500. You probably have a lot invested in cameras, more in good glass. You're doing it a disservice (as you have discovered) with a cheap scanner.

Lenny
EigerStudios.com

Lenny,

The optical resolution of the v700 and v750 are around 2050 to 2100ppi....not 900. If you are obtaining figures that low, check your focus. These two scanners have be tested a fair bit and with perfect focus, 2100ppi is the agreed typical resolution.

[Rob Champagne]

So how do you determine the optical resolution of a scanner since that is what seems to be what is being claimed as the limiting factor of a flatbed scanner?

I know lots of people have tested for this but how exactly have they tested for it?

[Kirk Gittings]

David, one of the issues I have encountered with this issue on scanners like the 750 is that interpreting how well a scanner resolves the bars on a target is very prone to bias. It seems that people who want to see higher resolution overestimate and the drum scanners underestimate the same samples from a flatbed target. I can say this. Good drum scanners tend to hold themselves and their equipment to a much higher standard than most. in terms of dynamic range, d-max and real resolution. even at 2050m DPI? Give a drum scan any day.

[hassiman]

Graham at Nash editions told me that Mac had tested their Scitech flatbed against the best drum scanners and the Scitech had prevailed... but admittedly their Scitech is an $80,000 :eek: flatbed scanner.

[Kirk Gittings]

The Scitech scanners are very very good, I own one myself, but I have yet to see a scan from a pro flatbed by anyone that was quite as good as the very best drum scan.

[sanking]

So how do you determine the optical resolution of a scanner since that is what seems to be what is being claimed as the limiting factor of a flatbed scanner?

I know lots of people have tested for this but how exactly have they tested for it?

You can test a scanner's resolution very easily. Just scan a high resolution target at the maximum optical resolution of the scanner, enlarge the scanner image on screen and note the smaller group of horizontal and vertical bars that are clearly separated. There is some subjectivity in this process but it should not amount to much over about 10-15%.

Bear in mind that resolution is greater with high contrast subjects than with low contrast subjects. Resolution values for film are often given at low contrast and high contrast.

My resolution tests of the Epson V700 indicate that it gives something on the order of 2000 - 2200 spi in real resolution. That is the equivalent of about 40 lines/mm.


Sandy King

[scott russell]

I just want to throw my 2 cents on the printing resolution end. I normally print stuff at 160-200 dpi on an epson and I can hardly tell its a digital print. It seems like overkill to NEED to print something at like 300 or 400 dpi. If you had the resolution, it doesn't hurt to print it that high, but i personally wouldn't worry if i was printing something big and it was only 160dpi or so.

[Lenny Eiger]

Lenny,

What is your opinion of the CoolScan 9000?:confused:

Sorry, I missed this question. I think the 9000 is ok. It often does justice to small film, 35mm, etc. I would still rather have a drum scanner. It's going to be sharper, certainly edge to edge. I recently tested the Premier against a Minolta film scanner (in the same category, maybe or maybe not as good). The drum was definitely sharper, and there was more distinction of subtle colors. The person (scanee?) said there were all sorts of colors he had never seen on his transparency before. This is, of course, anecdotal, and not necessarily scientific fact.

I think the real question is whether the 9000 is working for you - or whether there are occasions when you want every last bit of detail. Then you can send it off to someone like myself, or make the investment for yourself...

A lot of folks here have the opinion that consumer flatbeds do just fine for them with large format. It depends on your aesthetic. I wouldn't do it, but I have a very specific look that I am after.

Lenny