Some very basic questions about using an Epson V700 Photo please.

[sparq]

There's no 8x10 holder available for V700 because the "above the glass scanning area" is not big enough. 8x10 films need to be scanned off the bed. W/o an ANR glass, I always ended up with horrible Newton rings when scanning off the bed; irreparable in post processing. The ANR glass on top evens out the film pressure and usually helps but it does not eliminate the rings completely, either; especially when there are not enough "teeth" in the emulsion (larger darker areas).

The only guaranteed way is to get the bed messy with fluid mounting but it is a somehow risky business.

[Brian Ellis]

I used to scan 8x10 Ilford HP5+ on a 4990. No holder, I just put it on the glass. The first time I tried it I had a Newton ring problem but when I reversed the side of the film that was on the glass it went away. Can't recall whether emulsion side down or up was the solution but one of them was. I've heard others say it didn't matter for them which side was on the glass, they had the problem either way. So I don't know, maybe there was some synergy between HP5+ and my 4990 or for some other reason I just got lucky.

[Jiri Vasina]

I sometimes have Newton rings regardless of the side I place the emulsion on (up or down) and they keep reappearing no matter what I try. With some other negs, the rings never occur. Sometimes I have to warm the negative up - scan the negative twice and then it scans without the rings... Most often I scan emulsion side up (at least as a starting point).

[seepaert]

Off course you can scan with the supplied epson sioftware, as you can ride th tour de france on a childrens bike. But it has nothing in common with a real scanning program like for instance linocolor.

Silverfast can run on it self, at least it does in my systen 10.4 It is a lot better than the epson software, but still I was very disappointed after the good stories I heard about it. Main problem is that the prescan shows you a result that doesn't resemble the real scan. After scanning the result is not shows, which causes you to file a lot of rubbish, when you should have rescanned. I would not advice it for the difficult corrections on a color negative, but it works on BW and slides.

[newmoon2night]

OK, back to a Lightroom question.
I followed Richards instructions and used it with a 4"x5" transparency, set at 3200, and ended up with a TIFF file of 473MB. Great I thought ... until Adobe Lightroom wouldn't import the file because it was 'too big'.
Some searching on Google revealed that Lightroom limites file size to a maximum of 10,000 pixels per side, and this irritates quite a few people who scan images or stitch panoramas. Opinion seems to be that at some time this limit will be removed, but until then the only option if using Lightroom is to reduce the file size.
So am I right in thinking that in this case dpi does equate to pixels, and therefore with a 5 inch side and 10,000 pixels maximum, I need to limit the scan resolution to below 2000 (which means the maximum has to 1200dpi)?
Or is that a far too simplistic analysis of the situation?
Looking at the file properties in Explorer I see that it defaults to opening with HP Photosmart Premier, and does open with the B9180 software. I suppose I could forget about Lightroom all together for this, and just use the HP software, though I haven't looked to see what it will do, and whether I can clean the file up or anything.
Thanks for any further advice.

[Jiri Vasina]

Yes, you are right. Set he scanning to 1200dpi, you'll get a more manageable filesize. And I don't go above 2400dpi with sheet film.

(image size = [size in inches] x [resolution in dpi] )

[Kirk Gittings]

3200 is well above the true optical resolution on a 700. It is closer to 2200. Scanning above that only gives you a larger file size and no discernible advantages. Like Jiri I scan at 2400 max.

[newmoon2night]

Yes, you are right. Set he scanning to 1200dpi, you'll get a more manageable filesize. And I don't go above 2400dpi with sheet film.

(image size = [size in inches] x [resolution in dpi] )

Thanks Jiri and Kirk. I feel like I'm starting to understand this a little now.

I will ask one other question though, as my understanding of this area is limited to what I have read on Wikipedia - what's the difference between 24 bit color and 48 bit color?

I know a bit is a binary digit (and read that a byte is generally 8 bits). From Wikipedia's article on color depth I also see that 1 bit is monochrome, and 2 bits early greyscale, so 24 and 48 bits are increases in color spectrum I imagine.

24 bit is 256x256x256 color range, and 48 bit must be quite a bit more (not sure how many).

Do computer screens, printers and the human eye perceive at this level or at 48 bit level?

Does scanning at 48bit instead of 24bit considerably increases the file size?

[Jiri Vasina]

I'll explain it on the difference first on black and white images only, and first between 4bit and 2bit image:

if you have a 2bit image, you have 4 possible values of darkness/lightness - you only have white, lighter grey, darker grey, black. Any transition in tonality can be represented only by those 4 values. if on the other hand you take 4bit image, you have 16 possible values - the transition is much smoother.

With 8bits per channel, you have 256 possible values. So the transitions in this case is pretty smooth (in most cases). With 16bits per channel, you have something like 65 000 steps in the transition - way more than your eye can discern.

But now imagine you need to do some modifications to the image, darken some parts, lighten some parts, change contrast or whatever. By any such modification you effectively lose the number of steps in the transition: when boosting contrast, in the source image you may have 2 adjoining steps in the transition, lets say at value 128 and 129. After you boost the contrast, from the 128 you get for example 120, and from the 129 you get for example the value 135. There is nothing in between. So from the smooth transition between 128 and 129, you suddenly have a rough step between 120 and 135. Therefore it's better if you initially have more steps to start with (you have 16bits per channel), because the changes are less discrete, more precise.

That is the reason why almost everyone working seriously with images will try to work with 16bits per channel (and in scientific applications even more), because you don't loose that much information in the process (at least visibly).

And for color image it's the same: you either have 256 steps for each of the channels (Red Green Blue) in 24bit image - 8bits per channel. Or you have the 65536 steps for each channel in 48bit image - 16bits per channel.

And yes, scanning at 48bit will increase the file size over 24bit - the computer has to store twice the amount of information, so the TIFF file would generally be roughly twice as large.

Hope this helps...

[Greg Lockrey]

You would have to have one heck of an eye to discern a 48 bit print verses a 24 bit print practically speaking. Using 48 bit in your processing helps in getting your colors closer to a desired result but if the 24 bit happens to be at the same value, you wouldn't be able to tell the difference. A lot of processing programs won't handle 48 bit nor a lot of printers.