Built a "new standard"

[JJJM]

In your design, it seems to me you have not given importance to the horizontal movement of the center axis (spine?) of the book when you turn pages. Does it obbey to any reason?

Hi JJJM, I'm thinking about your question. Do you mean the motion of the spine of the book (ie the central axis) to the left and right? I've built this using Dan's (and others') 'block clamp' + sliding base.
The base is on drawer sliders and can be moved right or left as necessary to make sure the spine of the book is in the exact center, at the point where the center of the platen descends. The block clamp allows the book support to be widened, to allow for thicker books; when scanning a wider book I'll unclamp the block, move the right-hand support to the right a bit, move the entire base to the left a bit to compensate for that initial adjustment, and then the center of the book will again be aligned with the center of the platen. Currently I don't have a plan for securing the drawer slide, but the resistance seems pretty good--it doesn't just move around willy-nilly. If it does, a quick solution is just to use a small c-clamp on the 2x4 base to prevent movement.

As for manual focus, I was basing my ideas about that on comments in the forum, ie here and in this thread. After playing around with it a bit, I don't know that there is a huge advantage to manual focus over auto focus. But here's the thinking: the distance between the camera lens and the platen when it is its 'down' position is almost constant through the entire book--the position of the book itself is being adjusted a bit as the pages turn, but the center of the platen descends to the same point each time, and thus the page image pressed up against the glass is also in pretty much the same place over the course of the scan. By using manual focus initially, I avoid the camera re-focusing for each shot. If you were using auto-focus mode, and the camera was focusing on the center of the page (or even on multiple focal points), there is a possible problem with blank pages, or pages that have indistinct features at those focus points, since the camera will automatically focus in and out looking for a distinct line upon which to focus. With manual focus, that should theoretically not be a problem, since the focal length will be the same for all shots. However, this is all in theory, I've never used AF, and indeed have yet to use this scanner.

Thanks very much for the points on lighting and the angle--if I'm understanding correctly, since the page is at a 45 degree angle, the lights should shine directly down from above, with the beam hitting the platen at a 45 degree angle. This places the beam of light at a 45 degree angle to the camera (as you say, "Position them at 45º respect to camera"), ie., like this:

Lighting directly above.jpg

That is to say, the lights should not be at a 45 degree angle to the X-Y axis of the whole scanner, which would mean that the lights shine directly at the platen. This latter option, depicted here, seems like it would cause more of a reflection, and would also perhaps create a shadow of the camera on the platen, yes?

Lighting directly across from platen.jpg

I'm sure this will all come out in the wash. Camera #2 arrives tomorrow. Library book is due back on Saturday. A note on that--it will be a fair use copy of a portion of the book for research!

Jon

Thanks for your reply. I just looked at the platen and I did not see the sliding system. Yes, you are right.

About 45º the first image is right, to avoid reflections. Anyway, I would paint everything in black to avoid any reflection. And also, I would cover the front part of the camera with sth in black.

So far I have a simple scanner for one single camera and no sliding movement. In my opinion, taking double number of pictures was not a problem because from my experience postprocessing is much more time consuming than scanning, so I have focused on improving ocr rather than on diy for scanner. But your design is giving me doubts about this because it seems achievable for my diy skills...

[JJJM]

Most of the scanner designs on this site will not require manual focus to be adjusted every few pages, either because the book rests on a cradle which moves to keep the book centered, or because the cameras are attached to the platen itself. In either instance, the distance from a camera to the page it's photographing should be the same throughout the book.

In my opinion, manual focus is better because it is more reliable. Autofocus will typically fail for pages at the end of a chapter (which have a bit of text at the top of the page, while most of the page is blank), on title pages (which have only a line or two of text somewhere off-center because it's more aesthetic that way), and sometimes at the beginning of chapters and on pages which have a lot of white space for some other reason. Since I'm shooting two pages at once, even if I catch the failure while I'm shooting, I have to take the time to re-shoot and do something with the "extra" image to keep the left and right views in sync. If I don't catch it while I'm shooting (and since I just use the cameras' LED displays rather than a monitor, that would often be the case), I have to set up the book again (hoping I match the zoom I used the first time) re-shoot the pages that failed, and merge them somehow with the good images from the earlier shoot.

Of course, I'm using a "custom" setting in my cameras, so I don't have to focus the images at the beginning of each new book. I found the proper manual focus for my setup one time several months ago, and it's been working for me ever since. Even if I didn't have the custom setting though (and I understand that it's not available on some cameras), the hassle of re-shooting the inevitable out-of-focus AF failures would outweigh the hassle of manually focusing each camera at the start of a shoot.

I skip these problems placing a written piece of paper at the center of the page and focus well.

Anyway, could you post your configuration? Do you zoom?

Thanks for your interest.

[jck57]

Most of the scanner designs on this site will not require manual focus to be adjusted every few pages, either because the book rests on a cradle which moves to keep the book centered, or because the cameras are attached to the platen itself. In either instance, the distance from a camera to the page it's photographing should be the same throughout the book.

In my opinion, manual focus is better because it is more reliable. Autofocus will typically fail for pages at the end of a chapter (which have a bit of text at the top of the page, while most of the page is blank), on title pages (which have only a line or two of text somewhere off-center because it's more aesthetic that way), and sometimes at the beginning of chapters and on pages which have a lot of white space for some other reason.

The Canon A480 camera has Auto Focus Lock (ATL) and Auto Exposure Lock (AEL). To lock the focus, you partially depress the shutter button until the camera focuses. Then you push left on the four way switch. This locks the focus until you unlock it or turn the camera off. So, blank pages are not a problem and you can take pictures faster in succession. Locking the exposure works the same way except you press the up button on the four way switch. Many other P&S cameras have this feature.

[JonEP]

Hi,

Just thought I'd update this discussion regarding Manual Focus versus Auto-Focus Lock. The short answer, for me, after a bit of playing around with my a530s, is that I find Auto-Focus Lock a bit easier to use than Manual Focus. Because of this, I'd recommend the a480 over the a530 (which I have), because the a480 allows Auto-Focus Lock, and also is a 10 megapixel camera, which will provide a better image for post-processing than the 5 megapixel a530.

I've posted on image quality and focusing in a different thread, see this link.

[daniel_reetz]

Thanks for that, Jon! Really great to see your methodical approach.