Digitizing texts -

A while back we had some discussion going on about what can be done in the realm of quickly digitizing texts on a limited budget.

There are basically two ways to go about it, you can scan or you can take photos. Scanning takes a whole lot more time, and also the risk of damaging the text is greater when bending it open and pressing it against the scanner surface. Eliminating shadows is also harder.

I needed to grab a couple of books with me for the upcoming visit to Vraja, and some of it also needs to be handed around with the Padavali-project to have people transliterating the text.

The setup below is basically built of whatever scrap materials I could find in the house. The "thing" on the top is a piece of foam material where the camera sits still and comfortably. The camera isn't in its place for obvious reasons.

The result would be better if there were two persons doing this, as it'd free both hands to hold the book and make sure the pages are leveled out. However since my dear assistant happens to be at Radha-kunda since yesterday, I had to take care of it myself. (No, don't use your feet for this!)

Below, a view of the setup with two lights (I wish I had stronger), two carton boxes, two chunks of styrox (is that what you call it in English?), two wooden sticks and a sturdy piece of foam material. Budget $0. Speed of digitizing: 200 shots (400 pages) in 25 minutes (8/16 ppm).

[attachmentid=1944]

Attached, a "raw" image and a processed image to show the final outcome. Sample text: Manohara-bhajana-dipika.

[attachmentid=1945]
NB: The original print is pretty blurred at places, a concept I'm sure isn't alien to those familiar with Bengali books.

The setup could use tuning, the camera could probably be tuned better. It's configured for bulb lamp light and aperture has been manually maximized to eliminate shadows, focus is tuned manually to ensure consistent results.

A while back we had some discussion going on about what can be done in the realm of quickly digitizing texts on a limited budget.

There are basically two ways to go about it, you can scan or you can take photos. Scanning takes a whole lot more time, and also the risk of damaging the text is greater when bending it open and pressing it against the scanner surface. Eliminating shadows is also harder.

I needed to grab a couple of books with me for the upcoming visit to Vraja, and some of it also needs to be handed around with the Padavali-project to have people transliterating the text.

The setup below is basically built of whatever scrap materials I could find in the house. The "thing" on the top is a piece of foam material where the camera sits still and comfortably. The camera isn't in its place for obvious reasons.

The result would be better if there were two persons doing this, as it'd free both hands to hold the book and make sure the pages are leveled out. However since my dear assistant happens to be at Radha-kunda since yesterday, I had to take care of it myself. (No, don't use your feet for this!)

Below, a view of the setup with two lights (I wish I had stronger), two carton boxes, two chunks of styrox (is that what you call it in English?), two wooden sticks and a sturdy piece of foam material. Budget $0. Speed of digitizing: 200 shots (400 pages) in 25 minutes (8/16 ppm).

[attachmentid=1944]

Attached, a "raw" image and a processed image to show the final outcome. Sample text: Manohara-bhajana-dipika.

[attachmentid=1945]
NB: The original print is pretty blurred at places, a concept I'm sure isn't alien to those familiar with Bengali books.

The setup could use tuning, the camera could probably be tuned better. It's configured for bulb lamp light and aperture has been manually maximized to eliminate shadows, focus is tuned manually to ensure consistent results.

Six downloads. Any thoughts on the quality?

The quality is not bad if you zoom in on it. With photoshop or a similiar program I'm sure you could get these to look even better. Also, jpeg isn't the best file extension to save it to either. TIFF is better, but it is a lot larger.

I like your round mat. Do you chant japa on it?r

Well, unfortunately my digital camera doesn't produce raw images, I only get JPEG. You'll see that the processed version has been saved as a 8-bit PNG, which is lossless in every other respect aside colors, and that really isn't an issue as we are basically supposed to have black on white. Just so that it's clear, the pictures in the zip-file are before and after Photoshop. I'm sure I could make the final a bit better with more time to tweak around.

The round chair isn't especially for japa, it's for just about anything. The highest chair we have in the house.

I think that would be adequate resolution Madhava. There were a few characters that I wasn't sure about but that could probably be put down to the quality of the book that you are photographing and not the resolution of the scan