3.2.56 Beta
- Thread starter mikelove
- Start date
Screen OCR uses it for layout analysis but not recognition, and there's no toggle. (we're still debating how to add a toggle to that UI) Better place to test it is with a regular still image - sorry, I assumed from your references to textbooks etc that that was what you were working with.
Usually I use an external reader, as it has many options for more comfortable reading (auto-crop margins, etc.) and is not image/pdf-only. So Screen Reader/OCR is more better option for this case.
But for testing purposes I'll try to use the Still OCR reader.
But will be the new engine used in the future versions for SR/OCR?
But for testing purposes I'll try to use the Still OCR reader.
But will be the new engine used in the future versions for SR/OCR?
Mostly the latter, but I also think the current default approach may be the best one anyway + am not necessarily sure there'll even be a toggle versus a buried setting somewhere.
To be honest, we get very few complaints about mixed-in English text at present, and so while I recognize that for your particular use case it's vital that you be able to hide it + that the OCR accommodate that I'm not yet convinced that's something important enough to clutter up the interface with versus burying it in settings. (the other two screens are already extremely cluttered + adding one more button to those bars was a comparatively minor thing - Screen OCR is one of the few parts of our UI that actually doesn't look like it was designed by 7 committees of people who hate each other and I'm reluctant to mess with that)
To be honest, we get very few complaints about mixed-in English text at present, and so while I recognize that for your particular use case it's vital that you be able to hide it + that the OCR accommodate that I'm not yet convinced that's something important enough to clutter up the interface with versus burying it in settings. (the other two screens are already extremely cluttered + adding one more button to those bars was a comparatively minor thing - Screen OCR is one of the few parts of our UI that actually doesn't look like it was designed by 7 committees of people who hate each other and I'm reluctant to mess with that)
HW60
状元
Pleco and all files I open in File Reader are on my SD card. It happens with new and old files.
It mostly happens after I changed a file I had previously opened in File Reader with an external editor and then try to open it gain with File Reader via List Recent Files.
Sure, as for me - I don't need any toggle 
One good working engine is enough.
Concerning "complaints about mixed-in English text" - if you remember I didn't asked for this All I asked is to make the green box movable (and I'm not alone in this even on the forum) - it will serve most purposes.
You converted it to "mixed text" task, so I just accepted your vision (though I think that "make movable" is more easy task to do).
And use-cases I described above are common for most learners (at least for all beginners) so it's not something rare. May be very beginners just cannot get how to use current OCR for their tasks (I hope coming flashcard auto-creation from OCR-sources will make it more clear).
And the OCR functionality is still usable for described use-cases - it's possible to move the box by resizing, but just isn't convenient. So it's not a new future request, just a question of usability.
BTW, as a new user of Pleco and the very beginner Chinese learner, I don't see any signs that you trying to understand and make much difference between beginners and experienced learners/users I think creating just two bundles (Basic and Pro) are not enough. But it's another thread, if you interested - I could explain my thoughts as a new user.
One good working engine is enough.
Concerning "complaints about mixed-in English text" - if you remember I didn't asked for this
All I asked is to make the green box movable (and I'm not alone in this even on the forum) - it will serve most purposes.You converted it to "mixed text" task, so I just accepted your vision (though I think that "make movable" is more easy task to do).
And use-cases I described above are common for most learners (at least for all beginners) so it's not something rare. May be very beginners just cannot get how to use current OCR for their tasks (I hope coming flashcard auto-creation from OCR-sources will make it more clear).
And the OCR functionality is still usable for described use-cases - it's possible to move the box by resizing, but just isn't convenient. So it's not a new future request, just a question of usability.
BTW, as a new user of Pleco and the very beginner Chinese learner, I don't see any signs that you trying to understand and make much difference between beginners and experienced learners/users
I think creating just two bundles (Basic and Pro) are not enough. But it's another thread, if you interested - I could explain my thoughts as a new user.
@HW60 - ah, interesting; it shouldn't fail entirely in that case but it should put up a warning message.
@Bvo - true, but my understanding was that you wanted the box to be movable due to mixed-in English text bothering you, correct? It's not actually that hard to add a movable box, I'd just rather not force people to resort to that - if the problem is that there's too much English on the screen then a movable box is a middling solution while an OCR engine that lets you remove that text entirely is a better one. But perhaps it would be better to proceed on the assumption that we won't be able to fix this on the algorithm side to everyone's satisfaction and just make the box movable.
Regarding bundles: the problem there is confusing users with too many choices; Basic and Pro and an option to go from the former to the latter for just the difference in price makes for a pretty easy decision for most users. Asking people to segment themselves into beginning / intermediate / expert users adds unwanted friction into the process, not to mention raising all sorts of edge case concerns - what if I'm an expert but my spouse is a beginner and we share a tablet, what if I'm a beginner now but I intend to get better - is my purchase going to be useless once I do, etc.
@Bvo - true, but my understanding was that you wanted the box to be movable due to mixed-in English text bothering you, correct? It's not actually that hard to add a movable box, I'd just rather not force people to resort to that - if the problem is that there's too much English on the screen then a movable box is a middling solution while an OCR engine that lets you remove that text entirely is a better one. But perhaps it would be better to proceed on the assumption that we won't be able to fix this on the algorithm side to everyone's satisfaction and just make the box movable.
Regarding bundles: the problem there is confusing users with too many choices; Basic and Pro and an option to go from the former to the latter for just the difference in price makes for a pretty easy decision for most users. Asking people to segment themselves into beginning / intermediate / expert users adds unwanted friction into the process, not to mention raising all sorts of edge case concerns - what if I'm an expert but my spouse is a beginner and we share a tablet, what if I'm a beginner now but I intend to get better - is my purchase going to be useless once I do, etc.
Mike,
thank you very much!
I think many people (thousands, may be millions?) should be happy now
And yes, the new engine is much improved in skipping non-Hanzi strings - for standard (common printed) font it's petfect.
But movable box is just universal and easy-to-use solution for any cases.
And a text will look more clear if chars to skip won't be highlighted with green (i.e. if only Hanzi chars will be highlighted).
As it's beta, there are some bugs
1) Sometimes the app is crashing while SR/OCR is in use.
You can easily reproduce this bug if you will try to change the green box size several times.
2) I tried to check (1) with Still OCR (w scanned PDF) - the same behavior + I cannot resize the box's width to the screen width (it's shorter on the right and don't want to move farther).
Should I send some logs?
thank you very much!
I think many people (thousands, may be millions?) should be happy now
And yes, the new engine is much improved in skipping non-Hanzi strings - for standard (common printed) font it's petfect.
But movable box is just universal and easy-to-use solution for any cases.
And a text will look more clear if chars to skip won't be highlighted with green (i.e. if only Hanzi chars will be highlighted).
As it's beta, there are some bugs
1) Sometimes the app is crashing while SR/OCR is in use.
You can easily reproduce this bug if you will try to change the green box size several times.
2) I tried to check (1) with Still OCR (w scanned PDF) - the same behavior + I cannot resize the box's width to the screen width (it's shorter on the right and don't want to move farther).
Should I send some logs?
BTW, the new engine is a bit slower than the old one, so it "thinks" for a few seconds and nothing happens on the screen during this time - it may confuse new users (especially if they do several resizes in consequence to get it work), so I think adding some progress indication isn't a bad idea
HW60
状元
3.2.56 Beta is very slow on my SD card. If I load a user dictionary from exported flashcards, it takes 9 minutes for 2000 flascards. Pleco installed in internal memory takes 2 minutes 10 seconds.
Is this a matter of my SD card (I had no problems before 3.2.56 with Pleco on the SD card) or can it be a matter of 3.2.56?
Is this a matter of my SD card (I had no problems before 3.2.56 with Pleco on the SD card) or can it be a matter of 3.2.56?
HW60
状元
Unchanged in 3.2.56 Dec 27. Reopen Last File and List Recent Files still does not work after I changed a file - this did work before 3.2.56.