Epson 4990 scan depth (and is linearity important?)

[pellicle]

Hi

I just recently got an Epson 4990 used on Ebay. Having used a 3200 and a 4870 I feel that the 4990 is quite comparable with the 4870. I feel that it gets into the dense areas of film. I've recently adopted scanning a Stouffer wedge for testing this and then plotting the results on a spread sheet to see what I get. If I select a section of each "step" Photoshop gives me a mean / median / standard deviation of the tones for that (nifty).

I was surprised that after I took the results I got indicate that while the scanner penetrates more into the depth it does so at the expence of linear responce (and greater error [noise?] in the sample)

So, I guess that Epson have dialed up the gain in this area which has duffed the linear responce. Since this doesn't really pose a problem for film scanning (as films aren't linear in responce in the dense areas anyway I suppose its not a problem for that purpose, and I guess that makes the 3200 a better scanner for densitometry in its range of linearity than the 4990 does.

 

[scovell001]

Hi Pellicle,

Well done you for plotting a linear tone curve. This is always done when setting up a drum scanner before making a profiile.

I could right a book on the Epson 4990. I have one in the loft which I took about as far as I could go with mods. I owned a 4870 previously.

To get the Epson 4990 to deliver the optimum results you need to do the following:

1) remove the glass & replace with 2mm perspex or similar with a hole cut in the size of the film you are scanning (in my case 120)

2) paint the inside black

3) use betterscanning.com height adjustable holder & anti newton glass or even better a piece of frosted glass cut to size. Set the scan height to about 2.5mm above the glass

4) use Silverfast with 4x oversample & a gamma set to 3.0

5) scan wide open with no adjustments to highlight or shadow

For scanning transparencies, this scanner (or any of the other Epson flatbeds including the V700/V750) has no sharp point of focus. Instead it has a zone, around 1-1.5mm where the image is less soft & just about infocus.

The heat-sink is on the back of the scan head which is just nuts! You will also find the 1st scan after you've turned it on to be the best, with any subsequent scans after it has warmed up in any way being noisier.

There is some more info on scanning here: Dead Link Removed

 

[pellicle]

Hi

Well done you for plotting a linear tone curve. This is always done when setting up a drum scanner before making a profiile.

thanks ... perhaps its because I started in chemistry doing stuff with spectrophotometers and other optical sensitometry stuff, I'm honestly a little confused by the "gosh who cares as long as it looks good" views I normally get

I could right a book on the Epson 4990. I have one in the loft which I took about as far as I could go with mods. I owned a 4870 previously.

I started with a 4870, moved to Finland and needed a scanner, bought a 3200 and now I've bought a 4990. To be honest the 3200 is well enough for 4x5 black and white work. Problem is that I've just started using 120 colour in a 6x12 holder and the 3200 is well out of its depth there ...

To get the Epson 4990 to deliver the optimum results you need to do the following:

all good points and pretty much exactly what I was intending to do (although I hadn't thought seriously about silverfast).

The heat-sink is on the back of the scan head which is just nuts! You will also find the 1st scan after you've turned it on to be the best, with any subsequent scans after it has warmed up in any way being noisier.

now that is an interesting observation, I will be looking into that and seeing if its moddable (at the very least a larger heatsink may be fittable).

thank you

There is some more info on scanning here: Dead Link Removed

and thanks for the link too

 

[sanking]

There is someone on the LF forum who describes a procedure whereby he fluid mounts a piece of plexi directly on the bed glass, and then fluid mounts the negative on top of that, with mylar over. This would appear to eliminate the problem of refraction, which I assume is the reason you replaced the glass with plastic and cut a hole in it.

Any comments on how the above procedure might compare to the method you used.

Sandy King

To get the Epson 4990 to deliver the optimum results you need to do the following:

1) remove the glass & replace with 2mm perspex or similar with a hole cut in the size of the film you are scanning (in my case 120)

 

[pellicle]

well my feeling is the less (non lens focusing type) glass in the path between the film and the sensor the better.

 

[sanking]

OK, but do you have any results to support your feelings. Someone whose opinion I respect found in actual tests that results with the Microtek M1 were better with the scan on the glass than in the no-glass holders. Since refraction is the major issue it is not clear to me what method would give best results because I don't know the refractive difference between the various materials being used.

Focus is not an issue with the M1 since it is a focusing scanner, or at least that is my understanding as I write.


Sandy King

well my feeling is the less (non lens focusing type) glass in the path between the film and the sensor the better.

 

[pellicle]

Hi

OK, but do you have any results to support your feelings.

none. The reason for procrastination is that I am presently in Finland living in an apartment with no workshop. Compared to Australia or Japan (other places I've lived) there is less than nothing here when it comes to resources of this nature (I say less than none because what is here is scattered all over the countryside and poorly advertised and or costs a fortune).

When I get back in Australia and can fabricate what I need to get rid of my glass and make up a support mechanism for the film holders that is reliable and accurate I will do exactly that. Then you can expect to see that on my blog ;-)