[Coco] reducing size of scanned magazines

[Aaron Wolfe]

On Wed, Aug 12, 2009 at 1:14 AM, J.P. Samson
<coco+list at jeanpaulsamson.com> wrote:
>
> On Aug-11-09, at 10:08 PM, Aaron Wolfe wrote:
>>
>> On Tue, Aug 11, 2009 at 11:42 PM, J.P. Samson
>>>
>>> On Aug-11-09, at 9:38 PM, Aaron Wolfe wrote:
>>>>
>>>> I've been experimenting with the magazine scans on maltedmedia.  Playing
>>>> around with settings in Acrobat, I've managed to get the file sizes down
>>>> drastically without losing readability.
>>>>
>>
>> I've been using two different options in Acrobat.  First the "Optimize
>> Scanned PDF" tool which has several settings and a general slider for
>> quality vs size, this will get you down a factor of 8-15 and can take out
>> the yellowed paper effect, and then the "Reduce file size" option to
>> compatibility with version 9, which takes about 10-20% more off.   I've
>> played around a bit with the settings on the first tool,  but I'm sure
>> further tweaking could yield even better results.  It seems a 10x reduction
>> is size without compromising readability of any text is quite realistic.
>
> Assuming you have Acrobat Professional 9, I'd be interested in seeing how the ClearScan OCR does for further compressing things.  It seems to work surprisingly well for reducing the size of documents, with the added bonus of making the text searchable.  For those not in the know, ClearScan tries to retain the original look of the text by synthesizing new fonts to replace the source image text.  Actual graphics and background are extracted and compressed to your requested resolution (e.g. 150 dpi), probably using JPEG.

I experimented quite a bit with the OCR possibilities.  Clearscan did
not reduce the size of the file, but it does increase the readability
of the text to nearly that of the original, maybe even better in some
cases.  After refining the process a bit, I now have a 40MB file that
seems to be just as readable as the 550MB original.  The
graphics/pictures/artwork are not as good as the original, but
certainly fine enough for me.

I put the 40MB I've made here:
http://coco.spogbiper.com/rainbow.june.1985_opt1.pdf

The original 550MB is on malted media
ftp://maltedmedia.com/coco/MAGS0/RAINBOW/1985/
There is also a 100MB version there, but its readability is too low
for comfortable reading, at least with my bad eyes.

If anyone has comments or ideas to make it even smaller/better, I'd
love to hear them before I start shrinking more issues.
-Aaron