Search inside a PDF, why is this removed?

[Parkaceed 8]

In the old ios version you could open a pdf in evernote and then tap on the magnifying glas to search. the option to search inside a PDF in v10 is now gone? Why???, this was such a useful feature. Please bring this back.

[gazumped 11,663]

Hi.  v10 seems to be a work-in-progress.  AFAIK we have no information on what features will return and when,  but Evernote seem to be indicating "soon".

[r_ziliani 2]

Hi guys, I'm really desperate. Searching the PDFs was the only REASON why I pay the subscription. It is an indispensable and wonderful feature, because it has been removed?

And then I wonder, do we only regret this feature? There should be a riot of all premium and non-premium users.

Why did they take it off? So Evernote becomes a very common notes app that doesn't deserve money unfortunately ...

Do you know if there is any news?

Thanks a lot

[PinkElephant 8,100]

Right question, wrong place.

At least it is not on the list of discontinued features, so there is hope it will be added on a later release.

[r_ziliani 2]

Hopefully, even if I still don't understand why it was 'suspended'. I use Evernote as a second brain, and if you can't do the research it's totally useless ....

thanks a lot

[PinkElephant 8,100]

I think nobody outside of EN understands why they force replaced a working setup by this down-stripped version.

At least search still finds the note with the pdf. But you are lost if the pdf is large, or if there are several pdfs in one note.

If it is a large pdf, a workaround is to download it to files, open it with another app, and search it there. If OCR/text data is embedded this will work. It will not work if the OCR was done by EN, because this OCR data is not exported.

[r_ziliani 2]

Ok thanks, we hope they will be back soon ...

[Greg_in_Evernote 1]

Can I convert my Evernote PDFs to PNG in Evernote?

[PinkElephant 8,100]

No, there is no converter.

Furthermore this often makes no sense. Pdf is a multipaged format, optimized to show and print on every device alike. PNG (or JPEG) are picture formats, single paged only. So instead of one nice attachment you would end up with one picture per page.

Beside this, if you want to search in a pdf, this function has returned to the mobile client since this thread was started a year ago. Just Open the pdf, tap on the magnifying Glas top right, type the search string, and a preview with the hits shows, Tap on one of the previews, and it takes you to the correct page, and even highlights the searched text in the pdf.

For me there is no need to follow up on this thread, since it is already solved.

[Greg_in_Evernote 1]

11 hours ago, PinkElephant said:

No, there is no converter.

Furthermore this often makes no sense. Pdf is a multipaged format, optimized to show and print on every device alike. PNG (or JPEG) are picture formats, single paged only. So instead of one nice attachment you would end up with one picture per page.

Beside this, if you want to search in a pdf, this function has returned to the mobile client since this thread was started a year ago. Just Open the pdf, tap on the magnifying Glas top right, type the search string, and a preview with the hits shows, Tap on one of the previews, and it takes you to the correct page, and even highlights the searched text in the pdf.

For me there is no need to follow up on this thread, since it is already solved.

Thank you @PinkElephant I think I am in the wrong thread.  What I want is to find handwritten documents indexed by any of the words in the document.  I have been saving PDFs generated by the Livescribe integration for a a few years.  Some time last year, it seems, the automatic OCR of handwriting stopped or, at least, the results are no longer visible and the document(s) does not get picked up by the search.  Upgrading to premium seemed to promise "Searchable PDFs" but I think that might be only search within the document itself.  Any ideas?

[PinkElephant 8,100]

I am not sure handwriting in PDFs was ever indexed. I have searchable handwriting PDFs - but they were indexed by the source software, in my case GoodNotes 5, and the text embedded into the pdf on creation.

EN technicaly restricts indexing of handwriting to picture files (JPG, PNG and GIF). The function runs server based. I doubt they would take pdf handwriting indexing away if it ever existed. AFAIK it never was done.

Note that ENs own handwriting app Penultimate neither has an option to save as pdf. It only writes single page picture files into EN.

[gazumped 11,663]

1 hour ago, Greg_in_Evernote said:

Some time last year, it seems, the automatic OCR of handwriting stopped

 

Hi.  I'm not sure that was ever officially possible - How Evernote makes text inside images searchable

• Images containing handwriting should be added to Evernote as JPG images, not PDFs. PDF files are the preferred format for typewritten documents or scanned pages containing typewritten text. Handwriting is not indexed in PDF files.

 

Plus - I don't know whether the process has changed since the app was extensively rewritten for v10,  but handwriting OCR was a little different from character-based OCR.  Because of the variable quality of images (and handwriting!) words are assigned a range of 'possible' meanings - 'horse' might also be 'house' or 'hence' forinstance.  If you need word-for-word transcription of your text,  it would be better to use an app that shows you the results on the fly so you can correct any errors in real time and save a text-based result.

[Greg_in_Evernote 1]

18 minutes ago, gazumped said:

 

Hi.  I'm not sure that was ever officially possible - How Evernote makes text inside images searchable

• Images containing handwriting should be added to Evernote as JPG images, not PDFs. PDF files are the preferred format for typewritten documents or scanned pages containing typewritten text. Handwriting is not indexed in PDF files.

 

Plus - I don't know whether the process has changed since the app was extensively rewritten for v10,  but handwriting OCR was a little different from character-based OCR.  Because of the variable quality of images (and handwriting!) words are assigned a range of 'possible' meanings - 'horse' might also be 'house' or 'hence' forinstance.  If you need word-for-word transcription of your text,  it would be better to use an app that shows you the results on the fly so you can correct any errors in real time and save a text-based result.

Thanks for the suggestion.  It certainly worked well enough for a number of years.  I've now switched to PNG.  When I get Livescribe to actually send the latest scans to Evernote I'll find out whether it works.

That leaves me with about a year of un-indexed PDFs, hence my original question about getting them converted.

[PinkElephant 8,100]

Difficult - OCR of handwriting is not a staple.

You can ask Dr.Google for online services that may do it (if you want to trust your notes to an external website). The installed options I found cost money (and not a small amount), like Omnipage Ultimate or the Abby FineReader.

GoodNotes 5 OCRs handwriting, but only the notes created with the software. It analyses the writing itself to create the OCR. Other handwriting apps I know work on the same principle.

A pragmatic approach (if it is not too many pages) could be to take screenshots of every page, put them into one EN note and let EN do the rest. Just make sure the screenshots are saved in one of the 3 formats that EN will OCR.

For cases like this I find it practical to insert a table into the note, and put the pictures into the table cells. It allows to control the visual picture size by setting the column width, without changing the picture file size or properties. And it keeps the sequence of the pages, plus allows to use a second column to annotate.

[gazumped 11,663]

Hmmn.  Internet searches can be interesting.  Check out: https://www.wired.com/2010/06/google-docs-can-now-convert-images-and-pdfs-to-text/

[PinkElephant 8,100]

From the wired article (from 2010, BTW) it is not clear if this means OCR in general, or OCR for handwriting. I doubt it does handwriting OCR, but have no proper knowledge about it since I avoid Google services wherever possible.

If somebody is not affected from G-phobia like me (add to that FB-phobia), giving it a try may be an option.

[Greg_in_Evernote 1]

At the moment Evernote indexing is not working on handwriting in PNG either.  I waiting a few days to see if it would happen in batch but there is nothing.  Is anyone else experiencing the same problem?

All that is needed is for some of the words to OCR well enough to create some indexable text.  For instance, I always write the title in capitals.  This has worked well for years and I don't see anywhere where its withdrawal has been annouced.

[gazumped 11,663]

43 minutes ago, Greg_in_Evernote said:

Is anyone else experiencing the same problem?

Whilst you're waiting for any responses,  I'd suggest raising this with Support - you'll at least get a definite answer whether the feature is still officially available...

These Forums are mainly user 2 user - subscribers can raise support queries here - https://help.evernote.com/hc/requests/new and Free users here - https://twitter.com/evernotehelps (or use the feedback option in the mobile client). 

 

[Greg_in_Evernote 1]

5 hours ago, gazumped said:

Whilst you're waiting for any responses,  I'd suggest raising this with Support - you'll at least get a definite answer whether the feature is still officially available...

These Forums are mainly user 2 user - subscribers can raise support queries here - https://help.evernote.com/hc/requests/new and Free users here - https://twitter.com/evernotehelps (or use the feedback option in the mobile client). 

 

Thank you very much @gazumped

[PinkElephant 8,100]

Tested OCR today again, on handwriting created with different writing tools using the app Memopad 2 with Apple Pencil on my iPad (to do the graphical job independent from EN, avoiding Penultimate or Sketch). I then send the PNG files into EN, attaching them to new notes.

OCR was done within appr. 2hrs, the searched text was found and highlighted.

It worked better on the digital ballpoint pen than on the pencil imitation. The pencil created a trace of single points, sort of shaded grey. Little wonder the software had trouble to recognize it as handwriting. With the solid lines from the ballpoint tool OCR was pretty good. So make sure the writing is solid, good contrast and as legible as possible (not my strongest side, sorry).

To show what I mean, here is an enlarged cutout, top Pencil tool, bottom Ballpoint tool:

 

[Greg_in_Evernote 1]

12 hours ago, PinkElephant said:

Tested OCR today again, on handwriting created with different writing tools using the app Memopad 2 with Apple Pencil on my iPad (to do the graphical job independent from EN, avoiding Penultimate or Sketch). I then send the PNG files into EN, attaching them to new notes.

OCR was done within appr. 2hrs, the searched text was found and highlighted.

It worked better on the digital ballpoint pen than on the pencil imitation. The pencil created a trace of single points, sort of shaded grey. Little wonder the software had trouble to recognize it as handwriting. With the solid lines from the ballpoint tool OCR was pretty good. So make sure the writing is solid, good contrast and as legible as possible (not my strongest side, sorry).

To show what I mean, here is an enlarged cutout, top Pencil tool, bottom Ballpoint tool:

 

Thanks for taking the trouble to test that @PinkElephant.  Your experience is similar to mine in the past.  So, it seems maybe I need to look at the Livescribe - Evernote integration.

[PinkElephant 8,100]

My best results come from handwriting created in GoodNotes 5. It builds the OCR right while writing - it observes sort of the writing process to generate additional information about what might be written. The handwriting is searchable in GN directly after writing it.

When I close a note, I can export it as a pdf to EN. If selected it exports the search index with it - making it searchable in EN right away. EN would not index it anyhow, because they only index handwriting in picture files, not in pdf.