gekopaca
French photographer
It depends on what you mean by "information"—the errors introduced by JPEG compression are generally far smaller than the errors (noise) introduced by the scanner. So another way of looking at it is that the TIFF preserves your scanner's noise better than the JPEG.
So, if I understand well, there's two good reasons to scan in TIFF (I'm a newbie in scanning but an "old" user of HDR apps) :
- results of scanning are better
- HDR apps works better
Isn't it?
bensyverson
Well-known
Nah, I would say JPEG would work just as well, as long as you're scanning at 4000 DPI or higher.So, if I understand well, there's two good reasons to scan in TIFF (I'm a newbie in scanning but an "old" user of HDR apps) :
- results of scanning are better
- HDR apps works better
Isn't it?
The biggest advantage of TIFF for me is that the files open and save slightly faster than JPEG, since the computer doesn't have to do any compression.
C_R
Established
This were two 6x9 Velvia 50 shots, stitched for a panoramic. Hasselblad-Imacon X1 scanner, fff RAW negatives each converted twice with Flexcolor 4.8.8 software, kind of double scan for dynamic range.
http://www.flickr.com/photos/38068178@N08/sets/72157624361125930/with/5286007161/
http://www.carstenranke.com

http://www.flickr.com/photos/38068178@N08/sets/72157624361125930/with/5286007161/
http://www.carstenranke.com
willstar
Newbie
Heyy guys, all great info, but my original question is dealing with silverfast and the hdr scan option for kodak tri-x. When I'm scanning in, it does not turn the negs to positive automatically and when I looked up how to switch to positive in LR, they give a workaround which reverses all the controls.
My question is, is anything lost if I take the images into photoshop, invert and resave the tif as positive? I'm guessing no... but it wasn't even brought up as an option in any of the things I found in my googling.
2nd part, when scanning, do I want to do level adjustments in silverfast for each frame, or scan as is, and apply all curve and exposure adjustments and sharpening in LR?
Thanks,
Will,
My question is, is anything lost if I take the images into photoshop, invert and resave the tif as positive? I'm guessing no... but it wasn't even brought up as an option in any of the things I found in my googling.
2nd part, when scanning, do I want to do level adjustments in silverfast for each frame, or scan as is, and apply all curve and exposure adjustments and sharpening in LR?
Thanks,
Will,
willstar
Newbie
Since i found Silverfast doesn't let you adjust anything about the image if you're scanning 16bit HDR Grayscale, I've decided to scan everything like that and create a Photoshop droplet that does a simple Invert, then batch process all the negatives. I then make all my adjustments in LR. Seems like the best way to go, but if anyone has a better idea, please let me know.
Thanks,
Will
Thanks,
Will
venchka
Veteran
Nah, I would say JPEG would work just as well, as long as you're scanning at 4000 DPI or higher.
Unless you own a drum scanner, that is a waste of time, storage and computer horsepower.
bensyverson
Well-known
I have a fast computer and a Coolscan that operates at a native resolution of 4000 DPI... What is your suggestion?Unless you own a drum scanner, that is a waste of time, storage and computer horsepower.
venchka
Veteran
Since i found Silverfast doesn't let you adjust anything about the image if you're scanning 16bit HDR Grayscale, I've decided to scan everything like that and create a Photoshop droplet that does a simple Invert, then batch process all the negatives. I then make all my adjustments in LR. Seems like the best way to go, but if anyone has a better idea, please let me know.
Thanks,
Will![]()
I tried a few things last night. Got exactly NOWHERE!
1. Scan 6x7 negative as 48 bit positive. I set the black & white points as usual. After several abortive attemps at adjusting the gamma, I gave up. I can't see in "negative".
2. Imported to Lightroom. Edit In Photoshop CS4. Channel Mixer to drop red & blue channels and saved green channel. Invert. Save. Back to Lightoom.
3. Couldn't see any real improvement for all the effort.
4. Enfuse refused to blend any of my scans last night. Even virtual copies. I got so frustrated, I forgot to check to see if Enfuse would blend digital files.
5. My intent was to improve my scans for a panorama. After getting nowhere after about 3 hours, I used 3 old scans to make a new version of the panorama. 70+ megapixels later, I was pleased with the results.
Bottom line: Either I'm lucky or good. My vote is for lucky. I don't think that the negatives I was working with last night needed any help from HDR. If you expose and develop properly, film has all the dynamic range you need. Besides, after viewing a number of St. Ansel's prints, there is absolutley nothing wrong with either paper white or inky black. In the right places. That is the trick.
venchka
Veteran
I have a fast computer and a Coolscan that operates at a native resolution of 4000 DPI... What is your suggestion?
Assuming that 4000 number is accurate for the true optical resolution, continue on. Anything more is software added educated guesses.
Do you often print large enough to require the 4000 DPI? Or do you do it "becasue you can"? I get nice 16x20 prints from 6x7 negatives scanned at 2400 DPI.
bensyverson
Well-known
Attached is a 100% crop from a 4000 DPI scan, no sharpening added... You can see the markings on the clock on that building. So you can see that it's not just interpolated data.Assuming that 4000 number is accurate for the true optical resolution, continue on. Anything more is software added educated guesses.
Do you often print large enough to require the 4000 DPI? Or do you do it "becasue you can"? I get nice 16x20 prints from 6x7 negatives scanned at 2400 DPI.
I print 35mm up to 16x20, and having that detail really does make a difference. For MF and LF I scan at 2400 DPI on a flatbed, since the enlargement isn't being stressed as much as 35mm.
Here's the full image, for reference:

Attachments
venchka
Veteran
Fair enough. I wish I had a Nikon scanner for 35mm.
ColSebastianMoran
( IRL Richard Karash )
Coolscan at 4000 per inch
Coolscan at 4000 per inch
The tests I've seen, and my own experience, the recent Nikon Coolscans do resolve 4000 PPI. I scan 35mm at 4000 PPI. The 100% crop from your posted scan looks good.
I think you are on the right track.
For others reading, while the Nikons do resolve their quoted PPI, flatbeds do not. My Epson v500 actually resolves 1350 PPI in one direction and 2000 PPI in the other. With these scans, I'm happy with 12x18" prints from 6x9 color negatives.
-=- Rick
Coolscan at 4000 per inch
I have a fast computer and a Coolscan that operates at a native resolution of 4000 DPI... What is your suggestion?
The tests I've seen, and my own experience, the recent Nikon Coolscans do resolve 4000 PPI. I scan 35mm at 4000 PPI. The 100% crop from your posted scan looks good.
I think you are on the right track.
For others reading, while the Nikons do resolve their quoted PPI, flatbeds do not. My Epson v500 actually resolves 1350 PPI in one direction and 2000 PPI in the other. With these scans, I'm happy with 12x18" prints from 6x9 color negatives.
-=- Rick
craygc
Well-known
I have a fast computer and a Coolscan that operates at a native resolution of 4000 DPI... What is your suggestion?
As has been mentioned, the Nikon 4000dpi scanners are superb, whether it be C41 or B&W negs or chromes. All I use for 35mm is a Nikon 4000ED; its just easier to use than my Nikon 9000ED. I use Vuescan, for which I never understand why people think its difficult to use; just don't try and adjust everything in an attempt to get a good looking image out of it. Its really about just getting good data. For negatives, your maximum dynamic range is going to be about 2.0 which is well inside the Nikon's capabilities (this is why HDR has no effect here) so without setting the black and white points the scan will look flat; just fix it in post-processing.
Personally, I set the black point for negatives in the scanner with Vuescan by locking the exposure on clear leader first for every different emulsion I scan. The result is that when you fix the white point in post, you stretch the light side of the histogram much more than the dark side and from experience that seems to produce less noise in the shadows.
Share:
-
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.