Drum Scanners: Upgrade from Howtek 4500 to Heidelberg Tango?

[SergeyT]

This is what I have been doing since pretty much day one with NewColor:

All scanner output goes into LAB, no exceptions.
Sharpening is enabled but all values are set to 0.
Both Negative and Positive settings do point to the same Color ICC profile. I have made a variety of profiles scanned off a variety of targets and also with different scanner settings (by that I mean White Point and Exposure ? slider). The rest of controls is disabled.) Keeping the dark point at 5.00 or letting NewColor chose one does not make a difference. So I let NewColor chose it with the added benefit of enabling NewColor to suggest the White point value for me. Which I usually override based on my prefs.
By having multiple profiles I have an extra flexibility to choose one most suitable for the particular emulsion or image content.

BW Negatives are scanned as Color Negatives with inversion done at the time of scanning. In other words, they come out from NewColor the way I expect to see a BW picture - with Whites being White and Blacks Black. Goes into LAB originally. The scanner settings and conversion done for BW Negatives by NewColor are hard to beat, so why bother...

Color negatives are scanned as Color positives. The inversion, conversion to RGB and editing are done in PS.
What's interesting is that for Color Negatives LAB + Invert + "Convert to a profile (RGB)" produces slightly different colors than LAB + "Convert to a profile (RGB)" + Invert

Color Positives are scanned as Color Positives. I keep and edit my Color Positives scans in LAB and only convert them into a particular Printer RGB at the time when I need to print them.

I still feel like the older LinoColor produces better results on darker slides or in darker image areas than NewColor does and sometimes use it for just these images.

SergeyT.

[Grumium]

Sergey,

Thank you very much for your response.

A year ago you stated that the Tango's IQ is also depending on the selected scale and output resolution:

[...] Image size in Tango is a <Scale> x <Output Resolution>. One sets too much of a value for Output resolution and it results in a smaller Scale - read softer image. That alone probably explains some reports of Tango scans being soft.

As I am still looking for maximum performance of my Tango, I run a few test series in order to verify your statement. Unfortunately, my results differ. Please find an example below (135 film, target resolution = 6000 dpi).

Overview:


Comparison of most extreme settings:


Comparison of most extreme settings (sides swapped):


Settings in NewColor (Maßstab = Scale, Auflösung = Output Resolution):


Conclusion:
- No difference in IQ visible to my eye
- Re-scanning using the same settings shows variation in sharpness (sharpness not always spot on)

Sergey, and others, what do you think?

Port du Guilvinec



Contax T2 | Kodak Portra 160

[Grumium]

I performed another run in order to check if varying the parameters <Scale> and <Output Resolution> while maintaining <Scale> x <Output Resolution> = constant effects image quality. Aperture 0 this time in order to allow the Tango automatic adjustment in both directions.



Conclusion:
- No difference in IQ visible

Regarding the aperture setting:

I always select Aperture -5 as I don't trust Tango's automatic (and not visible) determination of the "best aperture" for a given size. I compared Aperture setting -5 to 0 and checked if it is possible to "blur" a -5 scan to a 0 scan. It is by applying a Gaussian blur. I also tried to go the other direction with different methods of sharpening: It didn't work - information is missing.



Conclusion:
- No reason to go for 0 if you long for maximum resolution

Note: This is grainy Tri-X souped in HC110. Guess how this would look like if shot on TMAX 100 ...

[onnect17]

I dont see the point of scanning at 4000 dpi if Blur is applied. The image without it looks pretty good.

[Grumium]

Yes, that's why I always scan at -5.

For those who don't have a Tango and would like to see how a scan at 11000 dpi (aperture -5) looks like:



This section has a width and height of 1.85 mm (or 0.07") on the negative (full scan would have more than 750 megapixel) ;-)

[onnect17]

IMHO grain is the best way to evaluate the scan, to sense what's going on. I love to see all the imperfections in the emulsion (I can clean it later).
I wonder why the software use -5 as the aperture value instead of the physical sample. Have you ever checked the optical output using a resolution target?

[Peter De Smidt]

Are there flexible high resolution targets?

[Grumium]

Lasersoft Imaging aka Silverfast sells them.

[onnect17]

If located in the US, B&H resells Silverfast targets.

[Peter De Smidt]

Interesting. I wish they said more about it. For instance, from their data I can't tell if it's flexible or not. My hi-res target is chrome on glass from Edmund Optics, and it's anything but flexible.