Tag Archives: abbyy

OCR Smackdown: ABBYY FineReader vs. Adobe Acrobat

A very common request that I get here at DocumentSnap is to compare the Optical Character Recognition (OCR) capabilities of ABBYY FineReader with Adobe Acrobat. Why? Well, for starters, both of them come included with models the Fujitsu ScanSnap as well as other scanners.

I decided to do a quick test comparing the OCR of the two packages using the following criteria:

  • OCR Speed
  • Resulting File Size
  • Accuracy

The Hardware

For a scanner I used my ScanSnap S1300.

I used two computers for the test:

  • Windows: A new cheap Acer laptop with a Core i3 2.40 GHz processor and 4 GB RAM running Windows 7
  • Mac: An old 2.5 GHz Intel Core 2 Duo MacBook Pro with 4 GB RAM running Mac OS X Snow Leopard

The Software

Here are the packages I used:

  • Windows: ABBYY FineReader For ScanSnap 4.1 (called from ScanSnap Manager) vs. Adobe Acrobat 9 Pro
  • Mac: ABBYY FineReader For ScanSnap 4.1 (run standalone) vs. Adobe Acrobat 8 Pro

Yes, I realize that Adobe Acrobat X is out, but since I am not aware of any scanners that come bundled with it yet, I decided to stick with the versions that ship with the ScanSnap. I’ll update Acrobat X in a later post.

The Document

I scanned a magazine article for this test. It probably would have been better to do this with a bunch of different documents to compare, but hey.

In all cases except one, I scanned without OCR so that I could run it standalone later. Here’s some info on the document that I used:

  • Pages: 2
  • Scan Quality: 300dpi, Color
  • Resulting File Size: 1.5 MB
  • Columns: 2, with some images

Maybe I am blind, but I couldn’t figure out a way to run ABBYY FineReader for ScanSnap on Windows standalone. If you know how, please leave a message in the comments. In that test, I re-scanned with “Create Searchable PDF” checked in the ScanSnap Manager settings.

The Settings

I tried not to do too many fancy settings to keep things as “real-life” as possible. There were essentially three configurations:

ABBYY FineReader

ABBYY FineReader OCR Settings

I set Save Mode to “Text under page image” and Quality to High. These were the settings for the Mac ABBYY, and I believe it is what ScanSnap Manager on Windows uses as well.

Adobe Acrobat (Normal)

Adobe Acrobat OCR Settings

I set the output style to “Searchable Image (Exact)” because leaving it just as Searchable Image in my experience has caused some weird things to happen with the resulting PDF. I used these settings on both Windows and Mac.

Adobe Acrobat (With ClearScan)

Adobe Acrobat ClearScan

In Acrobat 9 there is a setting called ClearScan. I used that as an additional test to see what the difference is.

Speed

Windows

  • ABBYY Windows: 20.5 seconds
  • Acrobat 9: 13.9 seconds
  • Acrobat 9 With Clearscan: 17.6 seconds

Mac

  • ABBYY Mac: 44.7 seconds
  • Acrobat 8: 20.2 seconds

Winner: Acrobat!

Since they are different machines, you can’t directly compare the Windows and Mac times, but clearly in both cases Acrobat is faster.

File Size

The non-OCR’ed PDF was 1.5 MB.

Windows

  • ABBYY Windows: 1.7 MB (+.2 MB)
  • Acrobat 9: 1.5 MB (same)
  • Acrobat 9 With ClearScan: 315 KB (-1.16 MB)

Mac

  • ABBYY Mac: 1.4 MB (-.1 MB)
  • Acrobat 8: 1.5 MB (same)

Winner: Acrobat 9 with ClearScan!

With an astonishing 1.16 MB reduction in file size after OCR, Acrobat 9 with ClearScan is the winner. Wow.

Accuracy

Here is a passage from the article:

Article Text Before OCR

Let’s see how each of the packages did:

ABBYY Windows

The spreadsheet has become the virtual “slide rule” for CMAs. It’s used for everything from preliminary strategic plans to financial statements. As with any familiar method, it finds its way into numerous situations where better alternatives are available, mostsignificantly in itswidespread use as a de facto reporting tool.
The appeal of the spreadsheet as the quickest way to get a report out is not hard to appreciate. “Excel is probably the most comfortable environment for a lot of financial professionals,” Alok Ajmera, vice-president, professional services withMississauga, Ont.-basedProphixSoftware, says. “There’s a very little learning curve, you can effectively do whatever you want with the data, and it works fairly well in smaller organizations.”
Periodic and complex reporting in processes like revenue management or cost management, however, is where the spreadsheet model really starts to break down.

Acrobat 9 Windows

T he spreadsheet has become the virtual “slide rule” for CMAs. It’s used for everything from preliminary su·ategic plans to financial statements. As with any farniliar method, it finds its way into numerous situations where better alternatives are available, most significantly in its widespread use as a de facto reporting tool.
The appeal of tlle spreadsheet as the quickest way to get a report out is not hard to appreciate. “Excel is probably tlle most comfortable environment for a lot of financial professionals,” AJok Ajmera, vice-president, professional services with Mississauga, Ont.-based Prophix Software, says. “There’s a very little learning curve, you can effectively do whatever you want witll tlle data, and it works fairly well in smaller organizations.”
Periodic and complex reporting in processes like revenue management or cost management, however, is where the spreadsheet model really starts to break down.

Acrobat 9 With ClearScan

The spreadsheet has become the virtual “slide rule” for CMAs. It’s used for everything from preliminary su·ategic plans to financial statements. As with any farniliar method, it finds its way into numerous situations where better alternatives are available, most significantly in its widespread use as a de facto reporting tool.
The appeal of tlle spreadsheet as the quickest way to get a report out is not hard to appreciate. “Excel is probably tlle most comfortable environment for a lot of financial professionals,” AJok Ajmera, vice-president, professional services with Mississauga, Ont.-based Prophix Software, says. “There’s a very little learning curve, you can effectively do whatever you want witll tlle data, and it works fairly well in smaller organizations.”
Periodic and complex reporting in processes like revenue management or cost management, however, is where the spreadsheet model really starts to break down.

ABBYY Mac

The spreadsheet has become the virtual “slide rule” for CiMAs. It’s used for everything from preliminary strategic plans to financial statements. As with any familiar method, it finds its way into numerous situations where better alternatives are available, most significantly in its widespread use as a de facto reporting tool.
The appeal of die spreadsheet as the quickest way to get a report out is not hard to appreciate. “Excel is probably the most comfortable environment for a lot of financial professionals,” Alok Ajmera, vice-president, professional sendees with Mississauga, Ont.-based Prophix Software, says. “There’s a very little learning curve, you can effectively do whatever you want with the data, and it works fairly well in smaller organizations.”
Periodic and complex reporting in processes like revenue management or cost management, however, is where the spreadsheet model really starts to break down.

Acrobat 8 Mac

T he spreadsheet has become the virtual “slide rule” for CMAs. It’s used for everything frorn preliminary strategic plans to financial statements. Aswith any familiar method, it finds its way into numerous situations where better alterna tives are available, most significantly in its widespread use as a de facto reporting tool.
T he appeal of the spreadsheet as the quickest
way to get a report out is not hard to appreciate.
“Excel is probably the most comfortable
environment for a lot of financial professionals,” avaJlaun:.:,JIIU:::’l;)It;IIIULauuy1111l::>WIUC::>PU:C1U uocd::>
a de facto reporting tool. T he appeal of the spreadsheet as the quickest
way to get a report out is not hard to appreciate. “Excel is probably me most comfortable environment for a lot of financial professionals,” AJok Ajmera, vice-president, professional services with Mississauga, Ont.-based Prophix Software, says. “T here’s a very little learning curve, you can effectively do whatever you want with the data, and it works fairly well in smaller organiza tions.”
Periodic and complex reporting in processes like revenue management or cost management, however, is where the spreadsheet model really starts to break down.

Winner: ABBYY FineReader for Mac looks the best to me. Acrobat 8 on the Mac is pretty terrible (in this example anyways).

Conclusion

Is there a “best” choice? It seems that in this example anyways, Adobe Acrobat 9 with ClearScan turned on gives fast results with good OCR while dramatically reducing the file size.

If you don’t really care about speed so much, FineReader produces good OCR results and for ScanSnap users, has the additional benefit of being integrated with ScanSnap Manager.

As with most things, the best software is the one that works the best for you. Have you found similar results? Any other tests of your own to share? Leave a note in the comments.

(Photo by Polina Sergeeva)

Comments ( 21 )

OCR And Orphan Works

As I have written about before, I always find it fascinating to read about different scanning projects, especially when it comes to scanning old stuff.

Over at the GalleyCat blog, Jason Boog writes about using Optical Character Recognition software to dig through orphan works.

What the heck are “orphan works”? I didn’t know either. According to Wikipedia:

An orphan work is a copyrighted work for which the copyright owner cannot be identified and contacted.

Here’s the project that the GalleyCat editor was working on:

While researching an essay about New York City poets and the Great Depression last year, this GalleyCat editor read through hundreds of pages from 1930s novels, periodicals, and self-published materials that couldn’t leave the New York Public Library.

He used his digital camera to take pictures and then ABBYY FineReader Express to OCR the text.

The results were impressive. Check out the GalleyCat post to see more.

(Photo by p0psicle)

Comments ( 0 )

FineReader Mac Update For ScanSnap Works For Older ScanSnaps Too

File this post under the “I will probably regret posting this” category.

I had a tip from DocumentSnap reader Hamad that was too helpful not to share.

Remember back in November 2009, Fujitsu had to release an update for Abbyy FineReader because it was having problems with Mac OSX Snow Leopard?

According to the website, it is an update for the ScanSnap S1500M and S510M.

However, it turns out that this version of FineReader works for even old-school ScanSnaps too. For example, I’ve even had reports of it working with a ScanSnap fi-5110EOXM from 2005!

It seems that, as long as the PDF is created by a ScanSnap (any Mac ScanSnap), it will work.

So, if you are running Snow Leopard and have an older ScanSnap, give it a try. It will probably work for you too. If you’re using something other than an S1500M or S510M and it works, drop us a comment and let us know.

Comments ( 5 )

Doing OCR Batch Processing Using The ScanSnap And ABBYY FineReader

Sometimes, when you have to scan a large number of documents at once, the step of doing OCR (making the PDF searchable) after each document can really slow things down. It may be preferable to scan them all in and then OCR them all in one big shot.

In the past I have posted about how to do batch OCR using Adobe Acrobat and have posted an Acrobat Applescript.

Over at the Optimality! blog, Tobi has posted a walkthrough of using ABBYY Finereader, which comes with the ScanSnap S1500M (and S1500 for that matter) to do batch OCR.

The problem is that in the default setup, each scan is OCRed right after the scan and depending on the age your machine (my G5 is getting a little long in the tooth) in can take quite a while. When you’re in the process of scanning many hundred’s of pages of paper documents, you don’t want to have to wait for the computer to do it’s OCR recognition, you’d rather feed it all the documents and let it do OCR while you’re doing something else.

Fortunately, this is possible. Reading all the way through the handbook as well as through the ABBYY online help I found out that you can scan to PDF only, and then afterwards convert the PDFs with ABBYY FineReader.

Check out the post here. Do you have any other tricks for doing batch OCR?

Comments ( 16 )

ABBYY FineReader For ScanSnap Update For Snow Leopard OSX 10.6 Now Available

When it rains it pours.  When Fujitsu released their ScanSnap Update For Snow Leopard, the missing piece was the OCR provided by FineReader.  They said it would be released by ABBYY soon, and as of today, it’s out.

The update is for the ScanSnap S1500M and S510M.

Click Here To Download The FineReader Snow Leopard Update. It’s down at the bottom.

You know the deal.. let us know in the comments how the update worked out for you!

Comments ( 5 )

ABBYY Finereader And Snow Leopard – File Not Created With ScanSnap

One issue with the Fujitsu ScanSnap and OSX 10.6 Snow Leopard that I forgot to mention the other day is the ABBYY FineReader that comes bundled with it.

When scanning with the version of Finereader that ships with the ScanSnap S510M and S1500M, you may get an error message like “File not created with ScanSnap”.

This is a known issue and according to this bulletin from Fujitsu Support, it will be fixed “within 2009″.

Fujitsu has assured me that they’re working on it, so hopefully we’re not talking December 31 here!

I personally do not use FineReader.. anyone have any workarounds for the Snow Leopard issue that they use? Leave a note in the comments.

Update: Thanks to reader Spike in the comments for the tip, ABBYY has released a version of FineReader Express Edition that supports Snow Leopard. More info here.

Update #2 Nov 19/09: The ABBYY FineReader for ScanSnap Snow Leopard Update is now available.

Comments ( 50 )

Abbyy Finereader and Adobe Acrobat – Why Does Fujitsu Include Both?

finereadervsacrobat.gif

I have received a number of questions recently about the software that is included with the Fujitsu ScanSnap. For example, why does the ScanSnap come with both Abbyy FineReader and Adobe Acrobat? Aren’t they both for doing OCR?

I suspect part of the reason that this question comes up is because of my posts about my ScanSnap workflow and my Adobe Acrobat OCR Applescript. Is all that necessary?

Let me start by saying that I personally have the ScanSnap S300M. The S300M comes neither with Abbyy FineReader not with Adobe Acrobat. If you have the S1500 or S1500M, your scanner will come with both and doing OCR is much more integrated than with the S300M, so my post-scan processing fun may not be necessary.

So What’s The Difference?

The ScanSnap comes with a special version of Abbyy FineReader called FineReader for ScanSnap. They’ve integrated that with ScanSnap Organizer, so if you are using the built-in automatic OCR’ing, that is what it is using.

If all you care about is having your PDFs searchable and don’t mind performing the OCR right after scanning, then the supplied FineReader is probably all you need.

To my mind, there are basically two main reasons why you will want to use Adobe Acrobat:

  • You want to do PDF editing after the fact
  • You want to batch your OCR after the fact

PDF Editing

So you have your scanned PDF. Now what? If you want to remove/rearrange pages and do a whole ton of other editing functions, Acrobat is a great tool. It is most definitely not just for making a PDF searchable.

You can see a bunch more information for Adobe Acrobat 9 (included with the ScanSnap 1500) and Acrobat 8 (included with the ScanSnap 1500M). You can see from the price that it’s a pretty good deal that this software is included with the ScanSnap.

Batch OCR

If you have a whole bunch of documents to scan in, it may be annoying to scan, sit there and wait for it to OCR, scan, OCR, scan, OCR, and so on. Some people prefer to scan all their documents to PDF in one shot, and then OCR them all at once. You can use Acrobat to do that instead of the included FineReader.

So there you have it, some of the differences between the two. What are some of the reasons you use one over the other?

Comments ( 12 )