Just if anyone cares to see another scanner comparison

go here :

 

http://www.shutterflower.com/scanner%20comparison.htm

 

we have a few scanners on there, all scanning the same shot - different actual

frames, but the same shot in 3 media and two formats.

I'm not sure this is a great scanner comparison. The "rules" simply state that the scans were made -

 

"scan at the highest resolution that is reasonable (to create the best possible image for viewing here)"

 

Several questions come to mind immediately.

 

Since viewing on the Internet constitutes a 72 ppi image - is that how you setup the scanners? How is a setup for viewing on the Internet applicable to making a scan for a print? With no documentation or description of how the scanners were setup - there is nothing to evaluate.

 

That makes the "comparison" useless since their is no information as to the ppi for each scanner. Also, without knowing whether the scanner software setup was optimized for each scanner - that too makes the comparison useless as how the scanner is setup greatly affects the final scan.

 

How can this possibly show the "true optical performance" for each scanner without knowing the ppi and software setup?

 

I know for a fact, the the Imacon Flexcolor software setup makes a great deal of difference since it applies unsharp masking and has a setup range from -250 to +500. It even applies some unsharp masking when the setup shows the unsharp masking turned "off." The only way you can get no unsharp masking is by turning it off and setting it to -250. Geee...I wonder how many other parameters with all of the scanners weren't set equally....

 

Also, you have gamma setup on the software - that makes a great deal of difference in shadow detail and highlight detail. Likewise, the Howtek software has a tremendous amount of controls that affect the final image.

 

What you have is not a comparison in any way shape or form since the outputs were apparently neither normalized NOR optimized. Normalization would be difficult as you'd have to know the neutral setting for each scanner. Optimization would be better as you could see the absolutely best output from each scanner.

 

What you have are some images from each of the scanners....that's about it.

This sounds similar to the same exercise going on at rangefinderforum.com. It tells you more about the user's workflow and scanning technique than it ever does about the scanner hardware. Perhaps a fun comparison to make but really tells us little about the scanners themselves.

Review of this sort are basically worthless. You can't compare scanners in an "out of the box" state any more than you can compare films processed and printed "the same way". Scanners (and films) have subtle differences, which do not necessarily make one better than the other. You have to work at getting the best out the process. That's simply not a practical way to conduct a review. Anectdotal tales are even less useful.

 

How do you make a decision then? There are three factors - (1) Brand - certain manufactures seem to stand out, like Epson inkjets, (2) Basic Technology - flatbed scanners will never equal the results of a dedicated film scanner. Only the wildest cheerleader would claim otherwise, and (3) Software - native scanner software ranges from mediocre to really bad. Silverfast stands out from the rest as a third-party source. Vuescan has a following, more for price than performance.

Well, it looks useful to me. That's pretty much exactly what I'd expect and have been seeing, and it's nice to see it all on one page. This is definately a useful new page. It'd be nice to know what the films were and how the B&W film was processed, though.

 

It'd also be nice to have all 6 frames from the Nikon 8000 or 9000, though. And the new Epson V700 or V750. (Sheesh, what a greedy bloke I am.)

 

Thanks for doing the work. I hope it doesn't get deleted, since it's a page I'll be referring people to.

Sheesh. What a bunch of grumpy folks. I'd _much_ rather see raw scans "straight out of the box" than scans processed by various workflows. You really can't make gold out of dung, and if the raw scan is dung, the print is going to reflect that at some level.

 

It sounds like people's favorite scanners did badly. Heck, I was irritated that the Howtech and Imacon did so well, since I had seen only uninspired scans from those scanners and have accused them of being buck-nekked emprors. Oops. I was wrong.

Actually with the multi-scanner support and color management capability Vuescan would make a pretty good tool for things like this as you could really turn off the sharpening, and hopefully get the colors equalized.

The most interesting finding is how superior the E6 scans are to C41. It would be nice to know what films were used.

According to the names : E6 is Velvia, C41 is a 160 film, and B&W is some Neopan 400 (which looks way grainy, what was it souped in ??).

>>And the new Epson V700 or V750. (Sheesh, what a greedy bloke I am.)<<

<p>

Yep. I hope we start seeing some more examples from these as people actually start receiving them. There seems to be quite a bottleneck with the 750 right now in the U.S. Some are actually starting to ship though. It would be great to see a few more scans just to confirm the quality level shown in the few reviews that are out on the net now.

<p> Doug<p>

<a href="http://www.betterscanning.com">New film holder designs for Epson, Agfa and Microtek</a>

"flatbed scanners will never equal the results of a dedicated film scanner"

 

The CreoIQ2smart easily beats a Nikon 9000, although at a cost of around $13000. You won't see a Epson flatbed beating a dedicated scanner for a while only because you can't expect a $700 scanner to produce the same quality as a 9000 costing almost 3 times more.

George, thanks. Well done.

 

The results shouldn't surprise anybody, but the normal online axes always must be ground :-)

 

Presumably all the scanners can do better than shown, operated to their maximum by our fellow geniuses, but I suspect the relationships will remain generally the same.

 

We've seen many apparently credible tests to the effect that the Nikons, as well as some of the defunct Minoltas, can approach or rival many drum scanners and that the various flatbeds can be tweaked to do better than their negative urban legends.

Ok, all you people, the scans are meant to be raw, untouched by sharpening, curving, etc. That was the idea. We weren't looking for the best image you could create, but the best the scanner could do from a hardware standpoint. Forget Vuescan and all that mess.

This IS the rangefinderforum scanner test.

 

I'm Shutterflower.