OCR rotates pages back (after manual rotation)

[guebert]

Hallo,

I'm still testing the new OCR module. Today I found out, that the OCR module is not able to regocnize wrong rotated pages. Therefore I rotated them manually AND SAVED THE PDF.

But after OCR has finished it's job, the content (!) of the pages were rotated back, but the paper orientation has not been changed. So the text was cropped.

You can see this in the attached files.

(BTW, it's quite astonishing, that the OCR file is bigger than the input file, even the input is 300dpi and OCR set t0 150dpi ...)

Michael

[Tracker Supp-Stefan]

Thanks for this Michael,

We will now investigate this and post again here in this topic as soon as we have any news.

Best,
Stefan

[Ivan - Tracker Software]

The issue is fixed. The fix will be available in build 201 which should be released today

[guebert]

I can confirm, the rotation-problem is fixed.

But the increasing file size still persists. I usually scan documents with 400 dpi to get better OCR results (with Acrobat) and let them downsample to 150 dpi after OCR. In my test case, the same doc makes files like this:

Scan 200 dpi - 79 kB
Scan 400 dpi - 216 kB

Scan 200 dpi OCR text layer only - 127 kB
Scan 400 dpi OCR text layer only - 264 kB

Scan 400 OCR image 150 - 370 kB

OCR with Acrobat 6
Scan 400 dpi (Searchable Image exact, downsample 150dpi) - 94 kB

So please enhance the downsample process (and make the function accessible independent from OCR).


BTW, OCR result is nearly the same, so good work!

Michael

[Tracker Supp-Stefan]

Hi Michael,

Glad to hear that you are happy with the actual OCR results, and as for the file size after OCR - I will pass your comments to the developers responsible and you can be sure that we will investigate this.

Best,
Stefan

[John - Tracker Supp]

Hi - any chance we could get some sample files from you to analyze for the size issue ?

Though I should also say - we are doing our best to avoid any major changes at this time so as to concentrate as much as possible on the new Version releases later this Spring - so it could well be you will not see the benefit of these until those releases ...

thanks

[guebert]

John,

here you are sample files. (Settings in Acrobat: Searchable Image exact, downsample 150dpi)

Please keep me informed regarding this compression issue, as it is the very last reason for me not to change from Acrobat completely.

Michael

[Walter-Tracker Supp]

Please watch for updates; we will try to address this sooner rather than later, however as John indicated it may be that we simply put it into the new release.