At 19:55 04/12/01, Rob wrote: >"bubble" shaped. If you measured the focus at the edge of the frame you >would be in the wrong place. If the curve is bubble shaped, ISTM the best >place would be halfway between the center and one corner.
I am really not doing well here. One last go, altho I realise I am speaking largely to empty space... I too now use the film strip holder all the time except for very unimportant stuff. Even for bubble shaped slides, if you look at the way the film curves, the mean focus position occurs at a location approximately 1/8th of film width from the top or bottom (in landscape) edge. Not at the edge, but 3/4 of the way towards the edge travelling from the centre. This position occurs because the curve is relatively flat at the centre and accentuates curl towards the edge. Now you can also measure this focus mean position at a location away from the centre towards the corner, as Ed is doing. My point, and I was only trying to be helpful to Ed, was that by going towards the corner you will generate focus errors if the slide/neg has a front-to-back position bias as it does under several situations: - when using the SA20 auto strip feeder which some people do some of the time - when your scanner is misaligned as mine is ("no fault found" from Maxwell service) - when at the end frame of a longitudinally curled strip, even when using the manual film strip holder. What I was trying to say, and this is my last go, was that you can find the mean focus distance somewhere along what I call the "y" axis, that is the line through the centre of the film, perpendicular to the long axis. You don't need to go off centre in the longitudinal dimension (i.e towards the corners), and indeed by doing so you can introduce unwanted errors. On the web page, I do show the measurements for the manual strip holder as well as the SA20, and you can see that my manual holder still is not holding the film parallel "front to back", although it is better than the SA20 by a mile. BTW you can speed up the scanning process significantly by purchasing a second film strip holder - so you can be inserting the next film strip while the previous one is scanning. I do this and it is quite useful, much less time sitting frustratedly looking at the progress bar! Cheers, Julian At 19:55 04/12/01, you wrote: >"Julian Robinson" <[EMAIL PROTECTED]> wrote: > > I must not have explained myself well. I understand that the problem is > > bowed film - I have a web page devoted to the issue. > >OK, but having reviewed your web page, you're only talking about colour >negative strips in the motorised SA20 adapter. The way the adapter >operates - pulling the strip the whole way in to measure the number of >frames, then feeding it back out when you scan frame 1 - tends to make the >end of the strip curl longitudinally. What this means for me in practice is >I get a scan which is 3/4 in focus and 1/4 out of focus at one end. This is >the *only* kind of noticeable problem I've had with DOF and my LS30, and >only if I have film that is curled before I put it in the scanner, or I >leave it too long so that that the heat of the scanner makes the plastic >"remember" its curl. > > > I am only saying that while it seems intuitively that a diagonal offset > > from the centre should be best, I think that in practice an offset along > > the "y " axis, not far from the top or bottom edge is a better choice. > >In the case of longitudinal curl of an unmounted strip, I agree. But AFAICR >this discussion about DOF problems began with people who had curved *slides* >in old mounts, especially cardboard ones, which would be a hassle to >remount. In a mounted slide, the curve (in my experience - YMMV) is >"bubble" shaped. If you measured the focus at the edge of the frame you >would be in the wrong place. If the curve is bubble shaped, ISTM the best >place would be halfway between the center and one corner. > >I avoided this whole issue with film strips by using the film strip holder, >but it is painfully slow to use. The IA20 APS adapter doesn't seem to have >the problems with curling that the SA20 does. > >Your focussing measurements may have been affected by the amount of time it >would have taken to do. As I mentioned above, I find the film curl tends to >increase the longer the film strip is in the scanner. > >I can understand Ed not wanting to provide a point and click method of >determining the focal position when few scanners support it - Vuescan >supports an awful lot of scanners! Perhaps one idea would be to have a drop >down list with at least three options - center, diagonal, edge. Obviously >the options should only appear when the scanner supports setting the focal >position. The problem with offset is which direction? When I've seen the >curl problem, it has been the end frame of a strip, and the curl is closest >to the end. This could happen at *both* ends of a strip, but I think it >mainly happens at the "front" end of frame 1 because that is the part of the >strip which is curled the most inside the SA20 in order to fit the strip >into the adapter. > >I wish I had time to do more tests. :( > >Rob