Hi Rose, Krister, Jenifer, Holly, Sandy, and others,
I'm cc'ing this response to the macvisionaries list because these questions are
likely to come up again, and in the long run, the ability to quickly reference
this post in the Mail Archives will cut down on list traffic. First of all,
I'll preface this by stating that the Prizmo scanning and OCR app for the
iPhone, normally priced at $9.99, is on sale this week for $2.99, as posted in
the App Deals section of the AppleVis web site. The price has never been this
low since the app was introduced late last summer, and I doubt that you will
find it available at such a low price again any time soon. This is the one OCR
app for which a number of viphone list members have reported success, and
probably the only one I would recommend, even at the regular price of $9.99.
International list members might want to know that Prizmo supports OCR in 10
languages. I think Sandy's done a great job of explaining the app (her post
appended below this one), but since there still seem to be questions about the
layout of the Prizmo screen, I'll try to give a slightly different description.
One of the features that contributes to Prizmo's increased effectiveness for
visually impaired users is the ability to take pictures with speech control, by
saying, "Take Picture", which reduces the likelihood of a blurred image due to
moving the phone when you double or split tap a button to take a picture. You
turn this feature on in the Settings menu by double tapping the "Settings"
button in the top right corner of the main Prizmo screen. Flick right to the
"General Settings" button under the "Settings" heading and double tap. On the
"General Settings" screen, flick right past the "Capture" heading and set the
"Alignment Grid" to "Switch button, off", then set the "Speech Control" to
"Switch button, on". I believe both switches are turned on by default, but it
is several months since I configured Prizmo. I'm not sure whether turning off
the "Alignment Grid" is necessary, but you certainly don't need it to be
displayed, since you can't see it while using the camera. Once you have checked
that these settings are in place, double tap the "Settings, back button" in the
top left corner to return to the "Settings" screen, then double tap the "Done"
button in the top right corner to return to the main Prizmo screen.
The main Prizmo screen has a list of types of documents to scan and OCR,
followed by a "Documents" button at the bottom of the screen that are results
you want to store in the app. In general, you will choose to double tap "Text"
(the first listed entry).
The app works in 3 stages, which are indicated by the 3 buttons at the bottom
of the the next ("Photo") screen: (1) Source (obtaining the image using the
Camera -- can also use previous images from your photo album taken outside the
app), (2) Image (options to apply corrections to image before OCR -- generally
requires vision to perform actions, such as cropping to only include the part
of the image that contains text or correct for distorted pictures because you
didn't hold the iPhone flat), and (3) OCR (perform the OCR with an option to
choose the language to improve the OCR), then display the results for editing,
copying, etc. or saving in the app's stored "Documents". You can navigate
through the app without ever using these buttons, just by double tapping the
"Next" button in the top right corner of the screen to move through each stage,
and that's probably the easiest way to use the app when you get started,
especially if you only work with documents in English. However, if you want
to OCR a document in a different language, you can improve your results by
double tapping the "OCR" button (third or three at the bottom right corner of
the screen) and then flicking left to the language button (e.g., "English") in
the row of option buttons for that screen, and double tapping. You'll be able
to double tap another language from the list of 10 available languages. (Note
that you will not hear "selected" announced beside the currently selected
language in the list, but you can only exit the "Language" menu by either
double tapping a new language selection or by double tapping the "Cancel"
button in the top right corner. Once you have changed the language (e.g., to
French) and returned to the "Photo" screen with the OCR button set, you should
hear the new language, "French", announced if you flick left to that button.)
Setting the OCR language before applying the "next" button improves the
recognition of words with accented characters -- otherwise I may get a "6" for
an "e" with acute accent, and other such examples.
Here's the sequence for the iPhone 4, assuming that you have configured the
"Settings" menu as outlined earlier. It is very helpful to use the $0.99
"Light Detector" app to check lighting conditions and also get a sense of the
optimal centering and height of the iPhone above the page before you use
Prizmo. (You can hear when the "Light Detector" signal is loudest if you move
the phone up and down or left/right and forwards/back with respect to the page.
This only gives a rough sense of where to position the phone when you use
Prizmo, but if you can't hear any difference in signal with Light Detector when
you move the phone around over the page, chances are that either the background
light level is too low to take a good image, or you're blocking the light
source with part of your body. You should also use a headset connected to the
headphone jack, since as soon as the iPhone camera is accessed in a mode that
is capable of responding to your spoken "Take Picture" command the volume of
VoiceOver through the speaker will drop, and not get restored until you exit
the camera screen either by double tapping the "Use" button in the bottom right
corner of the screen or the "Retake" button in the bottom left corner of the
screen.
1) From the main Prizmo screen, double tap "Text"
2) On the Photo screen, flick right to the "Camera" button that is the first of
the 3 "Get text from picture" source options, near the bottom of the screen,
and double tap
3) On the iPhone 4 Camera screen (no heading), there are three buttons at the
top: one for the "Flash" at the top left and one for "Camera Chooser" at the
top right. In between is an unlabeled "Button" that toggles on and off the
speech control function when the switch for that option is turned on under the
"General Settings" submenu of the "Settings" menu. If the switch for the speech
control option is not turned on in "Settings", this unlabeled button will not
appear on the screen. If, when you enter the Camera screen, your phone does
not vibrate, and VoiceOver's volume through the speaker does not immediately
drop, then the ability to use speech control to take a picture by saying "Take
Picture" on this screen has not been toggled on, and you will have to double
tap that unlabeled button to the left of the "Camera Chooser" button to
activate this function. Once you toggle on the speech control function, your
iPhone should vibrate, and the volume of VoiceOver through the speaker should
immediately drop. I believe that the first time you access the Camera screen,
speech control is toggled off (i.e., you can do a two finger flick down to
"Read All" the controls on the screen and hear everything announced through the
speaker). If you double tap the unlabeled button at the top of the screen to
turn on voice control of the camera shutter, then whenever you enter the Camera
screen your speaker volume will automatically be lowered, requiring use of
headphones. Presumably this is to get a good trigger when you say, "Take
Picture". Your speaker volume will remain lowered until you leave the Camera
screen (either by double tapping the "Use" button in the bottom right corner
after taking a photo, or by double tapping the "Cancel" button in the bottom
left corner to leave the screen without taking a picture), or until you
toggle the speech control feature off by double tapping the unlabeled button at
the top of the screen again. (Note, if you become very confident and
proficient in the layout of the camera screen, so you can double tap the "Use"
button in the bottom left corner without a VoiceOver prompt, you can run this
without headphones. Alternatively, if you're at home in a quiet setting you
may still be able to hear VoiceOver through the lowered speaker volume and also
work without headphones. Also, if you're fast at getting the iPhone to the
correct height and level before you hear the buzz, and then say "Take Picture",
you can work without headphones. But in general, be prepared to use headphones
when you use voice control for the camera shutter.)
4) Revisiting step 3. Assuming you set your speech control on, and exited the
Camera screen (e.g. by double tapping the "Cancel" button in the bottom left
corner), you can start the process again, with your headphones connected.
After double tapping the "Text" button (Step 1) and before double tapping the
"Camera" button (Step 2), place your iPhone on the page you want to OCR using
the edges of the paper to align the sides of the device. Center the iPhone so
that the camera lens (top right corner as you normally hold the device) is
approximately in the center of the page. (If you use a case, make sure that
the back of the case does not block the camera lens). Sandy has mentioned
suggestions like lightly folding the page in half length-wise and width-wise to
guide positioning and aligning the iPhone, in the case of practicing with a
sheet of letter paper. Make sure that you flatten the paper because failing to
keep either the page or the iPhone flat when you take the picture will cause
some of the letters to show up as tildes; this is typical if you scan a book
page and haven't flattened the section near the spine enough.
5) Double tap or split tap the "Camera" button, then, using two hands to make
sure you keep the iPhone level, raise it steadily up from the page to a height
of about 7 to 9 inches. (Nine inches is roughly twice the length of the
iPhone.) You'll feel the iPhone vibrate, and then VoiceOver will announce
"auto-focused". At this point you can say "Take Picture" and hear the shutter
close. (You'll probably start by lifting the iPhone too slowly to reach 7-9
inches before you feel the vibration and hear "auto focused". Say "Take
Picture" anyway -- your image is more likely to be in focus. The actual
distance you lift depends on the size of the page. Small billings only require
about 6 or 7 inches. I practice lifting and then check the distance I've
lifted an iPhone against the handspan of one hand, which for me is 7 inches,
but I have small hands.
6) Double tap the "Use" button in the bottom right corner of the screen.
This is the last element on the screen, so you can also just use a four finger
tap on the bottom of the screen and then double tap. (I preferred the four
finger swipe down, but this is the current gesture for moving tot he end of the
screen.)
7) On the next screen flick right the "Next" button in the top right corner and
double tap. You'll move to a screen labeled "Processing"
8) The results of the OCR will show up in an "Edit" screen which will let you
review and edit your results.
9) Double tap the "Next" button in the top right corner when done. You can
Save your results, or copy them, mail or upload them. Options such as copy or
email are listed as buttons along the bottom of the screen. Do a three finger
flick left to hear the second page of options, which include MobileMe, DropBox,
etc.
HTH. Also, if Krister is not using the "Take Picture" command, he'd probably
do better to use the "ClearCam" camera app that takes several pictures in a
row, and chooses the clearest of them for the Picture Roll. His split tap is
probably still introducing motion and blurring. Prizmo lets you choose
pictures from the Photo Album, so he could take the picture with ClearCam, then
choose the photo album option for the image source in Prizmo, and select the
last image in the Album. ClearCam was made by the same people who made the Red
Laser barcode scanning app.
Cheers,
Esther
On Jun 2, 2011, at 01:43, Krister Ekstrom wrote:
> Hi,
> I wouldn't trust Prizmo on a restaurant, the opera etc, too much fiddling and
> too iffy. and i would probably get the info too late anyway since i only get
> a third of the document in a readable format no matter what i do and the
> readable format still needs a lot of decryption and desiphering so trust it?
> no definitely not. I'm glad you get as good results as you do, and i don't
> understand how, but that doesn't necessarily have to do with the program, but
> with my own stupidity.
> As i use split tapping to take pictures, i don't have a problem with the
> phone shaking. I also have the flash on at all times, which introduces a
> smallish delay between button tap and the actual picture being taken. Having
> said that though, i can agree that Prizmo is by far the best iphone scanner
> out there, and i don't think the knfb reader does a much better job although
> i could of course be wrong, so i don't diss the program entirely i just want
> some sence of proportion here so people don't get unnecessarily dissapointed.
> /Krister
>
> 2 jun 2011 kl. 11.41 skrev Sandy Tomkins:
>
>> Hi Krister and Rose,
>>
>> Re the screen curtain: I don't find it makes any difference, I get the
>> same results either way, so one less thing to worry about. I would, however,
>> suggest that enabling speech, by doubletapping the unlabelled button is
>> worth it as it both gives you the vibration and the ability to take the shot
>> without having to tap the screen, thus taking up one hand which could be
>> better used keeping the phone steady, plus, the act of tapping shakes the
>> phone. Sighted users don't need it because they don't have to tap, merely
>> touch the screen.
>>
>> Krister, if I remember aright, a little while ago you were having no
>> success at all with Prizmo, whereas, now, you seem to be getting some useful
>> info, even if it is limited, from the app. Of course, the flat bed scanner
>> will always do a better job and is the thing to use for perfect results.
>> However, for "on the fly" info, enough to know, at least, what you are
>> dealing with, Prizmo will do the trick. Also, it would be a little
>> inconvenient carrying your P.C. and scanner around with you when you go to a
>> restaurant, college, the opera etc! And for those of us working solely with
>> the iPhone, this app can represent a saving of about a grand or more, if you
>> factor in the cost of a screen reader and Kurtzweil! I suppose I am
>> defending the shortcomings of Prizmo and the obvious difficulties that
>> people are having getting it to work for them because I think we sometimes
>> forget just what it is! that is: something for on the fly reading, something
>> to tell you what and for whom the post is, for accessing short documents
>> immediately, even if you need to ask another person to take the shot, you
>> still get the info when you need it and not when it is too late.
>>
>> Rose, re your situation, da sola, as you are while your hubby is in
>> hospital, have you tried VizWiz at all? This just might be perfect timing
>> for you, coming out, as ti has, just now. Also, if you are having difficulty
>> getting camera apps to work for you, perhasp there is something really
>> simple that you are doing wrong re holding the phone, I am sure there are
>> people on list who can give you some tips or, maybe, if you have a sighted
>> friend who could watch what you are doing, they could tell you the best way
>> to hold and direct the phone?
>>
>> Lastly, I really would suggest that people who are finding this app
>> absolutely useless that they find a way of making a controlled environment
>> at home, somewhere the light is good, with some sort of stand, even getting
>> assistance re setting this up, so that all you need do is place the phone on
>> your stand, putting your target text in it's place and then taking the shot.
>> After all, this is what is required for a flatbed scanner to work.I am going
>> to experiment with this to see if I can get Prizmo to work, flawlessly set
>> up this way. I actually think that Prizmo can give the big boys a run for
>> their money, with the right set up, apart from columns, that is. Maybe that
>> would be something to suggest to the developers.
>>
>> Best to you all,
>>
>> Sandy.
>>
>> Sent from my iPhone
>>
>> On 2 Jun 2011, at 09:38, Krister Ekstrom <kris...@kristersplace.com> wrote:
>>
>>> I don't think we should bother too much about the vibration. It's just to
>>> let you know that you can start taking photos, not that it's focused
>>> correctly or something, it just indicates that the camera is now active. I
>>> for one have deactivated speech alltogether, because i didn't get any
>>> better results with or without speech. I'm now at the stage where a third
>>> of the document is readable text, sorta and the other two thirds are
>>> gibberish. I use Prizmo for very quickly determining where a letter comes
>>> from and to get a tiny hint of what it is and if i need to act on it, that
>>> is if i need to type referrence numbers or ocr nummbers for bills etc, i
>>> still use Abbyy finereader express on the Mac, because it works 100% all
>>> the time.
>>> /Krister
>>> 2 jun 2011 kl. 05.17 skrev Holly Anderson:
>>>
>>>> I dont' get the vibration either. maybe it's because I don't have the
>>>> speech settings on?
>>>> On Jun 1, 2011, at 8:33 PM, Rose Combs wrote:
>>>>
>>>>> I have not gotten one image with this app and I can't even get the phone
>>>>> to
>>>>> vibrate, have been all through the settings in the app and otherwise and
>>>>> something just does not register for me. I have used oMoby but, other
>>>>> than
>>>>> a can of diet Coke and my dishwasher, have not had a lot of luck with it,
>>>>> but LookTell worked great.
>>>>>
>>>>> This is frustrating me because my husband is in the hospital so there are
>>>>> things I need to be able to read that normally I could take to him and say
>>>>> "what is this?" or, read directions for me.
>>>>>
>>>>> Hard to look up the directions on something if you have no clue what it is
>>>>> to begin with.
>>>>>
>>>>>
>>>>> -----Original Message-----
>>>>> From: viph...@googlegroups.com [mailto:viph...@googlegroups.com] On Behalf
>>>>> Of Sandy Tomkins
>>>>> Sent: Wednesday, June 01, 2011 3:45 AM
>>>>> To: viph...@googlegroups.com
>>>>> Subject: Re: Prizmo rides again!
>>>>>
>>>>> Hi Jen,
>>>>>
>>>>> No problem! I'll try to give a blow by blow account of what I do,
>>>>> however, the trick with Prizmo seems to be to experiment and find one's
>>>>> own
>>>>> best way.
>>>>>
>>>>> I'll start by saying that I do have good light recognition, so I can
>>>>> tell, at a glance, if the conditions are right for using the app with no
>>>>> extra light. In general, I would say that using it during the main part of
>>>>> the day will get you the best results. Electric light is doable, but not
>>>>> as
>>>>> good. Prizmo suggests that for best results one should enable the flash,
>>>>> but
>>>>> I find that the flash has to be enabled each and every time, so I don't
>>>>> bother. Plus, if the daylight is good, then Prizmo works well without the
>>>>> flash.
>>>>>
>>>>> One thing to mention for people with no light perception is this: you
>>>>> should position yourself away from the nearest window. What I mean is, if
>>>>> the window is behind you and your target text is in front, you will
>>>>> create a
>>>>> shadow, this doesn't help Prizmo, so, stand, or sit, with the text between
>>>>> you and the window, or any source of light, even electric.
>>>>>
>>>>> Before I began, I went into settings in Prizmo, and disabled the grid
>>>>> lines. These are an aid for those who can see them to align the camera
>>>>> with
>>>>> the text, but for us, I think they make things more difficult. I think,
>>>>> while you are in settings, you might check that the "speech" is enabled,
>>>>> but
>>>>> I think it is by default.
>>>>>
>>>>> Now, back on the main Prizmo page, doubletap "text" and on your new page
>>>>> you will find various options for tapping, go along till you hear "flash"
>>>>> flick once more and you will hear "button" this is the only unlabelled
>>>>> button that I have found in Prizmo, and it is the one you need to
>>>>> doubletap
>>>>> to enable speech command.
>>>>>
>>>>> Now, Choose what you are wanting to scan. while you are learning, choose
>>>>> something that isn't a challenge for Prizmo. So, no newspaper pages,
>>>>> magazines etc, go for something like a nice page of printed text, perhaps
>>>>> a
>>>>> typed letter with the paper nice and uncrumpled. Prizmo, by the way, is
>>>>> not
>>>>> that fussy, but to begin with, it is best to clear the decks of other
>>>>> variables. It would help too, to know which side of the paper is the text
>>>>> and which way up it is. Again, once you get proficient, you won't need to
>>>>> know these things, but baby steps is the order of the day.
>>>>>
>>>>> I am assuming we are now talking of using Prizmo freehand, this is, by
>>>>> the way, not the only way to do it and I will mention a few alternatives
>>>>> later. A sure fire way to position the phone, I am assuming you are using
>>>>> an
>>>>> iPhone 4, both the 3Gs and the Touches have less powerful cameras and so I
>>>>> cannot atest to their capacity to use Prizmo with much success, is to fold
>>>>> your paper one length wise, open it out again, then fold it again
>>>>> widthwise.
>>>>> Always keeping the text on the out side, in this way you get the folds
>>>>> standing proud on the top surface of the paper. Flatten it out well, you
>>>>> will still be able to feel the creases and they shouldn't interfere too
>>>>> much
>>>>> with the scanning. You can now position your phone in the bottom left
>>>>> quarter of the paper. Make sure that the top righthand corner of the
>>>>> phone,
>>>>> where the camera is located, is right up against the cross of your folds.
>>>>> Make sure, also, that the top of the phone is running along against the
>>>>> fold
>>>>> at the top of the phone and that the righthand side is doing the same. In
>>>>> this way, you can be sure that your phone is absolutely square with the
>>>>> paper and that the camera is withing a quarter of an inch of the centre,
>>>>> this is a perfectly acceptable margin of error.
>>>>>
>>>>> Remember, to use the speech command, you must use your earphones, so be
>>>>> plugged in! On the Prizmo screen, select "text" and double tap. On the new
>>>>> page that opens, select "camera" and doubletap.
>>>>>
>>>>> Now, here comes the knack! I lift the phone with both hands, one at each
>>>>> end, making sure not to cover the front face, thereby obscuring the camera
>>>>> aperture. I lift to a distance of about 8 inches, keeping it as flat as I
>>>>> can. I tend to find that if I do this as one fluid action, so raise it
>>>>> straight up from the flat plane it is on on the paper,, don't try and
>>>>> correct or think, is this flat? Just go for it! and once I feel the
>>>>> vibration say "take picture!" Now, Ester has written about the autofocus
>>>>> in
>>>>> the iPhone 4 and, for sure, it makes all the difference, but the question
>>>>> is, when is the autofocus working and when is it not? I have tried both
>>>>> shhooting the pic straight after the vibration and waiting for the
>>>>> "autofocussed" statement. I find that I do better taking the pic straight
>>>>> after the vibration, this, I think, is because the autofocus has kicked in
>>>>> before VO says it, and the extra time involved in waiting for it allows me
>>>>> to jiggle around and lose my fluid action.
>>>>>
>>>>> As soon as I hear the camera click I can move, doubletap the "use"
>>>>> button at the very bottom left of the screen, then flick twice to the
>>>>> right
>>>>> to hear "next", doubletap it, and I hear "is processing". Now I wait for
>>>>> the
>>>>> verdict!
>>>>>
>>>>> Jen, I can say that most often I get a usable amount of info. Often I
>>>>> get a perfect result, but there are days when it hardly works for me, but
>>>>> there are not many such days and my expectation has grown, so I expect to
>>>>> get good results, so am disappointed if I get too many tildas!
>>>>>
>>>>> In short: [when is she ever that, I hear you howl!] Lighting,
>>>>> positioning of yourself with regard to that light source, Positioning the
>>>>> phone on the paper and, finally, lifting in one fluid motion with two
>>>>> hands
>>>>> to keep it as flat as you can is the trick. I don't find that getting the
>>>>> distance, I.E. 6 or 9 inches is critical, distance is more about the size
>>>>> of
>>>>> the paper rather than the phone's ability to focus on the text. I would
>>>>> suggest that you try this while standing at a table, also, it might help
>>>>> to
>>>>> have your paper running lengthwise across the table, in other words, 90
>>>>> degrees turned from the way one would read it. the phone would then go in
>>>>> the top lefthand quarter, still butted against the central cross of your
>>>>> fold and fitted nicely along the two folds running up and across to your
>>>>> left. In this way, using two hands to lift the phone will feel more
>>>>> natural
>>>>> and you will have a better chance at keeping the phone nice and flat on
>>>>> the
>>>>> horizontal plane.
>>>>>
>>>>> The great thing about Prizmo is that it does not use the cloud for the
>>>>> OCR, everything is don onboard, so you can have flight mode on if you
>>>>> like.
>>>>> Prizmo, for this reason, works really quickly, so even if you get a
>>>>> "sorry"
>>>>> message, just try again. On the odd occasion I get that message, take the
>>>>> pic again and get a perfect result, so always try again! Afterall, it
>>>>> doesn't take an age and the more you practise, the better you will get.
>>>>>
>>>>> In the meanwhile, not everyone is interested in jumping through all
>>>>> these hoops, though I an say that once you get the knack, you just place
>>>>> the
>>>>> phone and say "take Picture" and it will, generally, work for you! But for
>>>>> those who want some results they can rely upon: try finding something like
>>>>> the open grid metal trivet that both Scott and I have mentioned, and try
>>>>> it!
>>>>> You can use the bars of the trivet to align the phone and the legs to make
>>>>> sure the paper is square, remembering always, place yourself on the side
>>>>> away from the source of light whether window or electric.
>>>>>
>>>>> Well, Jen, this is the best I can do for now, if I think of any other
>>>>> tricks that I use without noticing, I will get back to you and the list.
>>>>>
>>>>> Keep trying! It is like riding a bike, took me about 2 years, when I was
>>>>> a little girl, to get the courage to take my foot off the ground, but once
>>>>> done, I was off and flying!
>>>>>
>>>>> Salute!
>>>>>
>>>>> Sandy.
>>>>>
>>>>> Sent from my iPhone
>>>>>
>>>>> On 31 May 2011, at 22:37, "Jenifer Gilley" <claudas...@gmail.com> wrote:
>>>>>
>>>>>> Hi Sandy.
>>>>>> I know you've given, several times, hints/tips for prismo... but I always
>>>>> seem to get hit or miss. I always seem to get the message, could not
>>>>> extract text from image, or a lot of garbage. Any tips you could give me
>>>>> and anyone else who seems to have this problem as to how you seem to
>>>>> always
>>>>> get it working so often?
>>>>>> I'm sorry for the question, again.
>>>>>> Jenifer Gilley
>>>>>> email
>>>>>> claudas...@gmail.com
>>>>>> MSN:
>>>>>> claudas...@cox.net
>>>>>>
--
You received this message because you are subscribed to the Google Groups
"MacVisionaries" group.
To post to this group, send email to macvisionaries@googlegroups.com.
To unsubscribe from this group, send email to
macvisionaries+unsubscr...@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/macvisionaries?hl=en.
