Epson V700 v Nikon 9000

[bez]

Here's a strange thing – compared to the Epson the Nikon doesn’t seem able to render this small dense part of a 6x7 b&w neg. We're looking at the clouds, so ignore sharpening, dpi, and my extreme attempts with exposure & tones to get the same highlight detail on the Nikon.

Nikon 1

Nikon 2

Epson V700


Any ideas - light source?

[Kevgermany]

Sorry, missed this before.

Not sure about the Nikon, but do you have control over the exposure - in terms of how fast the scanning head moves? Looks as if it's going too quick, I've had the same problem with the 4870 epson. Needs more exposure, but can't control it becuase the head speed/lamp brightness are fixed.

[bez]

Thanks Kev. I certainly have control over exposure in the Nikon (and contrast, much the same as Epson Scan, although the histogram doesn’t work quite the same way - you can't drag the sliders outside the graph) but I’m not sure how this affects the scanning motor. I could try minimum and maximum exposure, while timing it.

Both the Nikon scans are x2 multisample, although I believe this affects shadows more than highlights, and of course both machines were set to 16bit grayscale.

I think the Nikon uses LEDs while my oldish Epson V700 is a cold cathode lamp, which is the only explanation I have come up with so far. Unlike the Epson you can't see what's going on inside the closed Nikon box …

Any suggestions where I might get help, or is the Epson actually superior in this limited area!?

[Kevgermany]

It's a neg, so the longer/slower scans help the highlights, which is where you see the most grain and washout.

I didn't really compare properly, but subjectively the microtek seems better than the 4870 at the darker areas of negs and slides. Especially Velvia/Kodachrome where things get really black, but there's a still lot of detail. Performance seems to be limited by noise, which is really down to insufficient light - just like a digital camera. Here I think traditional darkroom is better. You can increase the exposure without incurring the noise penalty. Drum scanning is another possibilty... But the Epson results look good, so probably not warranted.

I think the real answer is a design change on the scanners - dual pass, one with a higher lamp output setting to resolve the dark areas of film. And a pass with a lower lamp setting, then blend. Software wise we're there, but the hardware needs a variable level light source that doesn't change colour as it changes intensity.

Jo-1 scanned one of my negs on his minolta 5400. That was able to resolve all the dark areas of the neg without blowing the rest. Really impressive.

One other thing that may be affecting you is if you're using 16bits per channel or not. I'd guess you're doing that as a matter of course, only switching down to 8 bits for web use. If you check the specs, I think you'll find that internally the Nikon uses fewer bits/channel than the Epson (24/48 maybe), and this would also affect the darker areas of the film adversely. (There was a really good article on bit depth and how it affects dark areas of an image on Luminous landscape a few years ago, not sure if it's still there.). Net summary, the smaller the bit depth, the higher the level of posterisation, which is most noticeable in the darker areas of the image (here read film).

Edit - doesn't look as if the article is still there. Antoher thought is the colour space you're using. as soon as you move an image into a big colour space (like prophoto), your small colour space image will be compressed and lose detail. This will tend to increase the posterisation.

[bez]

Kev – your mention of exposure put me on the right track. The Nikon has an Analogue Gain setting of up to 2 stops either way, which I’ve never needed to use before (perhaps because my exposures are generally quite good )
In this case of overexposure it slows down the stepping motor, almost doubling the scan time, and has solved the problem thanks
I guess this is different to the software simply altering the digital output from the sensor, when adjusting the histogram etc.

[Kevgermany]

Good to know. Did it increase noise much?

Must see if I can find a setting like this for the microtek.

[bez]

No obvious increase in noise - I'll check more carefully, and time the V700 on variable exposures.
Interesting the Epson got it right first time ..