Scanner comparison: Epson V750 Pro added

[sanking]

It could be both, and/or something else. Lack of detail in an image might result from photographing a scene where atmospheric conditions limit detail, say in mist or fog, or the lack of detail could result from problems with technique such as the ones you mention. Lack of detail in the negative could also result from conscious choice, say in use of film type, f/stop to control DOF, or in diffusion of the image or in choice of camera type, for example a pinhole or zoneplate camera.

Most people would agree that scanning a pinhole negative with a drum scanner would be kind of pointless because no matter how good the scanner may be it can not pull detail from where there is no detail. You could scan the negative with an Epson 2540 or a Tango and it won't make much difference in terms of detail because there is only so much detail in the negative.

So my point was that in comparing prints from scans of a flatbed like the V750 with prints made from drum scanner scans we should remember that the negative may have been the limit to final image quality, even though the final image quality in the print appears to be similar. Or of course many other factors may also limit final image quality, including the resolution of the output device.

Sandy King

Sandy,

When you say, "Many things limit final image quality more than the scanner itself, including how much information is in the negative and the type of output. I would wager that very few 4X5 negatives have the equivalent of more than 1400-1600 spi of real detail..."

Do you mean lack of detail specific to the characteristics of film being scanned, or are you referring to lack of detail specific to photographic technique or hardware limitations, such as lack of scene detail, lack of focusing skill, or lack of film plane/focus screen alignment?

Not challenging at all, just curious.

[Keith S. Walklet]

Thanks for chiming in Kirk. A lot of water has gone under the bridge.

Some more observations. I selected the 35mm film to scan because I figured the smaller piece of film would exacerbate any limitations that the flatbed had, plus I could do side-by-side comparisons with each of the other scanners. My newer tests which will compare the 4870, V750, Nikon w/anti Newton holder, as well as the stock holder, and TANGO will be done with medium format film.

My previous tests with the 4870 showed, (regardless of the claims of the manufacturer of what that actual resolution is), that there were steady improvements in the quality of the image with each bump in scanning resolution 2400spi, 3200spi, 4800spi, until 9600spi, at which point the scanner was running interpolation of the data and the image quality suffered. This observation supported the common wisdom to scan at the maximum optical resolution the scanner could capture.

I made samples to illustrate the progressive improvements in scanner resolution by printing 6" x 24" strips of the image one above the other. The steady improvement was easy to see even for individuals that had not previously been exposed to the concepts. I routinely share these prints with my students to let them draw their own conclusions.

I then ran tests to see the effect of sharpening on each of the files and concluded, in short, (similar to what was posted by another user earlier in this thread) that one needs to approach sharpening differently with a flatbed than with a drum scan. The data has a different look and feel.

Thanks all. I've got to run. I'll try and write more when I have time.

[mandoman7]

Personally, I'm not all the motivated about scanning 35mm or 120 negs. This is, after all, a large format forum. I have tons of old slides and negs, but the calls are not frequent enough to warrant big expenditures. Film flatness and other factors make it something of a different animal for me. Commercially, the digital alternative has been established over those two formats in my workflow.

I'm interested in a reasonable cost way of getting LF negs into the computer. I don't want to drive for an hour, ship back and forth, call 5 times, or alternatively, spend thousands on something that may be beyond what I need.

Here's a question: How many LF shooters who have a drum scanner can say that they've gotten their return on the investment with print sales? If you're a name guy who's writing articles for publications then, of course, the rationale is clearly in place. If you're an MD who's dabbling in LF, then going top notch with your equipment also fits the paradigm. When you read some publications, by the way, there doesn't seem to be a reality beyond those two perspectives. That's their income base, I guess.

Passing acknowledgement has been given for people doing what works for them, but I would give it a bigger priority. For my money, its about getting work done and getting it out there. Its not about having sharper technology than your buddy who's also not selling or showing work.

I would really like to see a display of enlargements by different people working with different scanners to really see how people are using their tools. Back in the darkroom days it wasn't the guys with the fanciest darkrooms that were putting out the best work. It was the dedicated guys who were making their stuff work for them.

JY

[Keith S. Walklet]

Agreed JY.

I used the 35mm chip for my tests so that I could add the Nikon 8000 to the mix of scanners tested since it was considered an excellent dedicated film scanner, and consistently rated higher than the flatbeds. I'm doing the next round with the medium format for the same reason, so that I can include the Anti Newton holder to ascertain what trade-offs exist for that option.

My first round of tests were done using the standard Epson film holder. This round of tests will include wet-mounted scans.

With regard to the spi, the drums have the capacity to achieve higher resolution (reputed to be as much as 10,000spi) but the recommended scans typically seem to be done at 5000spi for 35mm and 67cm, and 2000spi for 4x5, with the reasoning being that anything beyond that was not contributing to improvement of the final output.

So, I am comparing a 5000spi scan with the highest optical sample (regardless of what that resolution actually is) I can obtain from the other three devices (4800spi with 4870, 6400spi w/V750, 4000spi with the Nikon).

On a final note, since this thread began with Leigh's announcement that the V750 had been added to what is a really interesting comparative tool, my own tests show the V750 does a much better job with the darker tones than the 4870. That observation is borne out in the shadow portion of Leigh's comparison, which illustrates the much more detail and smoother data in the lower tonal range.

[aphexafx]

Keith, I have to congratulate you on your dedication to all of this. I, for one, am extremely interested in your wet mount test results, especially if you will prepare comparison images to view.

Of course a drum scanner will be able to pull more (substantially more) available information out of a piece of film. But I am mostly interested in what you will be able to pull with these lesser scanners - more so since you are going to wet mount the source. It will be interesting to finally compare sharpness and tonality from these scanners using a wet-mounted source, with the focus optimized.

Please proceed!

[kurthbousman]

Well , now that I've read thru all of the pontificating and since my pg 1 comment drew so much of said type of commentaries , there are a few points . Make an analogy with printing . I always print at 720 ppi . Why ? Because if I'm not looking at the print with a loupe I can't see a difference , and because I use half as much ink or a quarter as much ink depending . So then it becomes an "economic" decision. I sell prints . This isn't a hobby. I sell my prints for $250 each which means I'm not getting rich and it also means I have to watch my capital investments esp. in equipment. The v700 allowed me to access my archived images in all formats for a minimum investment and the results have been acceptable enough where , w/o looking at the prints with a loupe , I don't see a difference at a normal viewing distance. I'd love to own a Creo but at 3-5 grand used and outdated software ,it made the decision to go for the v700 simple. Also I don't like working with chemicals so that even made the choice between the v750 and v700 simple much less a drum scanner where the time , expense , difficulty and mess are simply not worth the difference in quality. And that's what I see on the scanner comparison chart too. By the way , I think that the v700 files can be sharpened better with a stepped sharpening routine , at least that's my experience. I guess it comes down to if you want to be an artist or a technician . I've been both , spending the first half of my life as a research engineer for univ. of texas . Now I'm an artist and I've learned it's all about what works and the ease of arriving at that point.
-you can see my art and photos at galeria san miguel on the jardin and ..
www.kurthbousman.com
www.hcggallery.com

[Lenny Eiger]

Now I'm an artist and I've learned it's all about what works and the ease of arriving at that point.

I went and looked at your sites. I don't think it matters what scanner you use. I think you ought to be working with a digital camera, as it would be even less expensive and less hassle to get what you want... I would say that your work isn't traditional photographic work. It doesn't look like Paul Caponigro's work, for example. You don't depend on subtle tonalities, it isn't crucial that you have everything sharp, lots of depth of field, or that you can hold everything together to go larger.

There is work that is appropriate for higher end equipment and work that isn't. It isn't a value judgement. It's just different. There is a purpose for every tool, and you clearly don't need a drum scanner. Other folks work in a different style, especially a lot of people who take the trouble to lug around a view camera, and their work does. You can't dismiss either piece of equipment outright (except for yourself).

It depends on what you want to accomplish.

Lenny

[mandoman7]

On a final note, since this thread began with Leigh's announcement that the V750 had been added to what is a really interesting comparative tool, my own tests show the V750 does a much better job with the darker tones than the 4870. That observation is borne out in the shadow portion of Leigh's comparison, which illustrates the much more detail and smoother data in the lower tonal range.

This may be late in the discussion, but do we know what efforts were made to adjust the film to the proper height in those tests? If he was using the standard holders without testing for right height then his results are not that reliable. My first sense in looking at the images was that that was possibly an issue.

JY

[sanking]

I don't recall that information being given for the V750 or any of the flatbeds that use non-focusing lenses. Certainly if it were done for one of the scanners and not the others the comparison would be unfair.

My own take is that the results of the V750 compared to the those of the other scanners tested in the comparison are pretty consistent with what I would have expected from my own testing.

Sandy King

This may be late in the discussion, but do we know what efforts were made to adjust the film to the proper height in those tests? If he was using the standard holders without testing for right height then his results are not that reliable. My first sense in looking at the images was that that was possibly an issue.

JY

[harrykauf]

I would like to add that with a flatbed like the V750 you have the option of doing
multipass scans to increase detail in dense areas. Maybe it would be interesting
to add a multipass scan to those comparisons.

I had some good results and I also use a technique in Lab to reduce the luma noise
that you get when you scan dense areas. (I am not talking about blurring ab to
remove chroma noise)
The V750 generates a per pixel noise in dark areas (or bright in negatives)
that is independent of the actual film grain. Getting rid of it helps a lot.

All this with careful post can increase the quality a lot in terms of shadow detail.

Here is an example:
top row multipass from left to right. raw scan, process steps, final
bottom row the same processing on a single pass scan.

http://www.sooshee.com/tmp/eyes02.jpg

here the same image but with worst and best next to each other for better comparison.
the left is what you get by default the right is what you can actually achive.

http://www.sooshee.com/tmp/eyes03.jpg

another:
A raw
B chroma blur
C classic destructive method for luma noise
D my non destructive method
http://www.sooshee.com/tmp/abney/noiseABCD2.jpg

I am in the UK so its not really practical for me to scan the slide from the comparison
chart but if somebody could send me a single and multipass raw I could
do the processing.