Had a pretty cool experience this weekend. I got an order for a print from someone who saw the Capitol display. The image can be seen here.
I was printing an 8x18 from a 35mm slide. I scanned it on my Epson 2400 (don't know the exact settings, but the same as I always do) and got what looked to be a good scan. I printed a good test print and then the 8x18. This has always been a tough print due to the large amount of real estate taken up by solid or graduated colors. The big print was okay, but there were definitely some bands of color in the colored part of the sky, and the windmill wasn't terribly sharp. I decided to have Rich V. do a drum scan and I printed it on my 1280.
There's no way to compare the two prints by posting pics here, but what was interesting is that in a section cropped and shown here at 100% there's much more grain visible on the drum scan file. When I first started printing it large, that grain showed up (printing the .tif from PS). I was using Pixel Genius for sharpening (which rocks, by the way), but I didn't think that was causing the problem. I still don't know what the issue was, but after burning through a lot of paper and ink (and Fat Tire) trying to figure it out, I finally imported the .tif into the print program that came with the 1280 (Film Factory). I printed from there and wa-la. A perfect print!
1) I was impressed at the HUGE difference that a drum scan and an intelligent sharpening program could have when trying to enlarge a 35mm slide to 18". The final print is tack sharp and very nice! I expected a difference, but not that much. 2) I have absolutely NO idea why the grain in the digital file would show up on the print when printing from PS, but not when printing from the Epson program.
Can't rule out operator error, but if anyone has a similar experience or any insight, I'd love to hear it. The final print is fantastic - I'm sure I could go to 24" and have it look very sharp even when viewed up close. I just wish I new why the grain from the drum scan came out through with one priting method and was perfectly smoothed out using another.
P.S. the price of the scan was well worth it, if for no other reason than to see Rich's house; it's a trip. It's a museum of computer, photography and developing technology. I think I saw some glass plates lying around there next to a leaf scanner... And he has more CAT 5 wiring than the Sun Microsystems campus.
Loved having you over yesterday. Once Veronica and I get everything together in the " printing studio" and I fix the drywall issues in the darkroom we'll have an open house for CNPers to come over. Fat Tire and a rotating drum scanner (a great combo).
I am really surprised at the differences you had with he printing -strange -- Nat = you have a good idea?
What really surprised me about this was the fantastic quality of Rick's original 35mm slide. I gave up on 35mm back in 1989 when I switched to 645 on a Bronica. I never really had terrific glass with my 35mm gear. The detail we found in the windmill really surprised me when we zoomed in on it. We used no sharpening tools in the scanning. The lesson I take away from yesterday is to not skimp on 'glass' (which I'll remember as I get ready to move into my DSLR world).
Look forward to helping out again soon! Thanks Rick!
RichieV
PS - I actually do have a few 1880(ish) glass plates lying around near the Leaf scanner area
Posted by Nat Coalson on March 6, 2006 at 11:30 AM
Sounds like the difference has to do with the way each application interpolates the pixel data when sending it to the Epson print driver. The amount and method of interpolation would depend on the resolution of the file and the native resolution of the printer along with whatever automated image enhancement routines the software imparts.
Since Film Factory would be designed for the widest possible user base, including novice users, it stands to reason that Epson would include some "intelligent" algorithms to tweak the output to get the best possible result. Photoshop would not alter output data this way, so strictly speaking, the print from Photoshop would be more "accurate", but of course as this example illustrates, the most accurate is not always the most pleasing.
I would recommend using the Remove Noise filter in Photoshop, or ideally a third-party plug-in like Noise Ninja or Grain Surgery (which I use) to reduce the amount of visible grain to your taste. It's always better to have a file that looks perfect and then prints exactly the same way, without the mystery of differing results.
As for the grain - the Howtek is capturing exactly what's on the film, while the Epson scan software will likely reduce the grain automatically (or not see it due to narrower dnamic range).
Hope this is helpful... great shot by the way, and congrats on the sale!
Wow, those crops are a trip I had no idea the flatbeds would be that bad compared to the drum scanner, but would love to see crops from the same 4x5 on both scanners to see what is being lost by taking the cheaper route and using a flatbed to scan the film. That drum scan looks very close detail wise (down to the film grain) to what I get out of my film scanner. I'm sure the drum has way less digital noise and can dig deeper into the shadows, but I'm glad to see that my 35mm scans are at least getting what detail is there.
Rich/Rick, did you happen to scan any 4x5's on both scanners and it so mind putting up some comparison like Rick did for this one?
To make a more accurate comparison, it might be possible to disable any automatic noise reduction being applied by the Epson scanner software and the Film Factory output options.
Thanks for the info, Nat. Good point about PS not altering data prior to sending it to the printer whereas the Epson software may be tweaking it behind the scenes. Normally, I don't see any difference in the print regardless of what program I print from, but this is the first time I've printed from a digital file with as much detail/data as one gets from a drum scan.
Neither the Epson scanner software nor the Print Factory provide much in the way of sophisticated control, and no options releative to automatic noise reduction. I agree that there probably is something taking place, but nothing that can be turned on/off by the user. The main difference in the scans here is likely the different resolutions and the limited dmax of the Epson. When I get some time, I'll rescan the tranny - it's a 35mm - on the Epson using the same settings as Rich used on the Howtek (or as close as possible). I scanned for a 20" print at 200 or 250 dpi; Rich scanned for a larger print (24 or 30"?) at 360 dpi. Hence the different crops for the same number of pixels above.
Discussion
5 repliesI was printing an 8x18 from a 35mm slide. I scanned it on my Epson 2400 (don't know the exact settings, but the same as I always do) and got what looked to be a good scan. I printed a good test print and then the 8x18. This has always been a tough print due to the large amount of real estate taken up by solid or graduated colors. The big print was okay, but there were definitely some bands of color in the colored part of the sky, and the windmill wasn't terribly sharp. I decided to have Rich V. do a drum scan and I printed it on my 1280.
There's no way to compare the two prints by posting pics here, but what was interesting is that in a section cropped and shown here at 100% there's much more grain visible on the drum scan file. When I first started printing it large, that grain showed up (printing the .tif from PS). I was using Pixel Genius for sharpening (which rocks, by the way), but I didn't think that was causing the problem. I still don't know what the issue was, but after burning through a lot of paper and ink (and Fat Tire) trying to figure it out, I finally imported the .tif into the print program that came with the 1280 (Film Factory). I printed from there and wa-la. A perfect print!
1) I was impressed at the HUGE difference that a drum scan and an intelligent sharpening program could have when trying to enlarge a 35mm slide to 18". The final print is tack sharp and very nice! I expected a difference, but not that much. 2) I have absolutely NO idea why the grain in the digital file would show up on the print when printing from PS, but not when printing from the Epson program.
Can't rule out operator error, but if anyone has a similar experience or any insight, I'd love to hear it. The final print is fantastic - I'm sure I could go to 24" and have it look very sharp even when viewed up close. I just wish I new why the grain from the drum scan came out through with one priting method and was perfectly smoothed out using another.
P.S. the price of the scan was well worth it, if for no other reason than to see Rich's house; it's a trip. It's a museum of computer, photography and developing technology. I think I saw some glass plates lying around there next to a leaf scanner... And he has more CAT 5 wiring than the Sun Microsystems campus.
Thanks, Rich - great work!
Loved having you over yesterday. Once Veronica and I get everything together in the " printing studio" and I fix the drywall issues in the darkroom we'll have an open house for CNPers to come over. Fat Tire and a rotating drum scanner (a great combo).
I am really surprised at the differences you had with he printing -strange -- Nat = you have a good idea?
What really surprised me about this was the fantastic quality of Rick's original 35mm slide. I gave up on 35mm back in 1989 when I switched to 645 on a Bronica. I never really had terrific glass with my 35mm gear. The detail we found in the windmill really surprised me when we zoomed in on it. We used no sharpening tools in the scanning. The lesson I take away from yesterday is to not skimp on 'glass' (which I'll remember as I get ready to move into my DSLR world).
Look forward to helping out again soon! Thanks Rick!
RichieV
PS - I actually do have a few 1880(ish) glass plates lying around near the Leaf scanner area
Since Film Factory would be designed for the widest possible user base, including novice users, it stands to reason that Epson would include some "intelligent" algorithms to tweak the output to get the best possible result. Photoshop would not alter output data this way, so strictly speaking, the print from Photoshop would be more "accurate", but of course as this example illustrates, the most accurate is not always the most pleasing.
I would recommend using the Remove Noise filter in Photoshop, or ideally a third-party plug-in like Noise Ninja or Grain Surgery (which I use) to reduce the amount of visible grain to your taste. It's always better to have a file that looks perfect and then prints exactly the same way, without the mystery of differing results.
As for the grain - the Howtek is capturing exactly what's on the film, while the Epson scan software will likely reduce the grain automatically (or not see it due to narrower dnamic range).
Hope this is helpful... great shot by the way, and congrats on the sale!
Rich/Rick, did you happen to scan any 4x5's on both scanners and it so mind putting up some comparison like Rick did for this one?
Thanks Craig
Neither the Epson scanner software nor the Print Factory provide much in the way of sophisticated control, and no options releative to automatic noise reduction. I agree that there probably is something taking place, but nothing that can be turned on/off by the user. The main difference in the scans here is likely the different resolutions and the limited dmax of the Epson. When I get some time, I'll rescan the tranny - it's a 35mm - on the Epson using the same settings as Rich used on the Howtek (or as close as possible). I scanned for a 20" print at 200 or 250 dpi; Rich scanned for a larger print (24 or 30"?) at 360 dpi. Hence the different crops for the same number of pixels above.
Thanks again everyone!