I finally wanted to test the OCR features, but DTPO V1.3beta2 is unable to detect my Scanner, CanoScan 3200F.
Usually I scan with Image Capture, I still can do so and then import to DTPO, but I cannot set the options for this import, for instance to save the scanned pdf to a certain folder. I always get an error message, that there is no scanner detected, I should try to open Image Capture, I do so, choose the Scanner, but nothing changes for DTPO.
I tried to delete and reinstall the driver, disconnected the scanner, rebooted the computer, tested some other Voodoo (repairing permissions…), nothing helps.
yes, I finally realized that I’m using ScanGear via Image Capture. I spent this day with scanning and OCRing and searching the best way to do so.
The results are not so good as I expected, the best way for me is scanning to a folder outside DEVONthink and then importing.
I tried different settings, because scanning with colour and 600dpi did not get the best results. What was really confusing, are the different wordcounts for the same document, depending on scanning it with 150 or 300 or 600 dpi, from 316 to 734 words! And at the end, when importing this text in a rtf-document, I got a nice salad of letters and signs.
Are there any differences between the ‘normal’ version of ReadIris and the licensed version fo DTPO? I’m asking, because I don’t need EMail-archiving in DTP, but really need OCR.
For the moment Image Capture won’t recognise at all the scanner…
Maybe tomorrow.
The best overall result you get with 300dpi and colour (according to the IRIS people). We are using the same OCR engine as ReadIRIS. But with ReadIRIS you can customise the behaviour and the output to your heart’s content. So if you have papers that require a lot of finetuning to get the best results or very specific requirements for the output, that would be the better choice.
Thank you for the information. I will give a try to ReadIris. Unfortunately I missed the special offer in the AppleStore.
Do you have an explanation for the different wordcount for the same document, scanned with different dpi?
What are your experiences with 300 or 600 dpi? I scanned articles from a newspaper, they looked better when using greyscale, and the best OCR result I got with 150 dpi.
Your word count will vary based on the accuracy of the OCR. In general with a higher dpi this should go up, but you will also get more noise so YMMV.
The problem with a lot of newspaper scans is that the paper is very thin so the letters printed on the other side may come over on the scan and confuse the OCR. In such cases you need to fiddle with the scanner contrast settings and thus it is better to use the manufacturer provided scan software. That allows you to tweak everything that your scanner can do.
