Monday, November 30, 2015

Voice Dictation and Speech-to-Text on Kindle Fire HDX's and current HD tablets (built in and useful for email, notes, and forum comments) -- How it works. Also, an excellent voice-recording app (free, only at Amazon) that's good for making sound-files for keeping recorded notes or for sharing. UPDATED Nov 29-30 with tips on trying voice-dictation-to-text offline and for an Extra 30% Off Print books thru' Nov. 30.


This is an example of using voice-dictation on a Facebook post:

Click on the first screenshot to see where the microphone icon is.  The second one shows the screen during dictation, and the 3rd one shows the completed comment to Facebook.


Kindle Fire's Built-in Voice Dictation for tablets with Fire OS v3.1 (See Updated entry at bottom.)
This capability is available, per Amazon's Language and Keyboard Display help page on the following Amazon tablets:
  "For Kindle Fire HDX (3rd Generation), Fire HD (4th Generation), or Fire HDX 8.9 (4th Generation) devices, tap the microphone icon on your keyboard to use your voice to dictate text instead of typing."

  The Fire HD 4th generation would be the ones released in Year 2014 (HD 6" and HD 7" that year, still current in Year 2015).
  For example, the lowest-cost Fire HD model has it, as does the Year 2014 HD 7 tablet you might have, though that one doesn't show in current tablet offerings.

The thumbnail photos above lead to larger photos for more readability.  These screenshots are from a session using the built-in voice-dictation (or speech-to-text] feature to reply on Facebook to a post.  (I'm using a post from Wildtiger Org's Jack Kinross in connection with their Remote Area Aid for disaster relief in Nepal.  This is a photo he took of a woman carrying wood from recently cut-down trees to her home because of the now two-month-long Nepal Fuel Blockade, which prevents receipt of fuel (for cooking and transportation) and medical supplies as well as health workers for areas of Nepal that need it).  The full photo by Jack Kinross can be viewed here.
  Today, a report by Unicef that a Shortage of essential supplies threatens millions of children this winter in Nepal goes ignored while the mainstream news outlets ignore that story.  ("200,000 families affected by the tremors are still living in temporary shelters, at an altitude above 1,500 metres") facing winter without fuel for food and transport and no medical supplies due to the mentioned blockade there.)

  The first screenshot shows the voice-dictation microphone at bottom left of the keyboard; the second one shows the screen interface during the actual voice dictation with a dotted circle around the microphone and "Done" to be tapped once you've finished your dictated comment.  The third one shows the comment as it appears before you press the Enter key on the keyboard to submit the comment.

According to an Amazon press release, this "Voice dictation converts your speech to text—available in all languages when online; offline support available in US English.."

  (I haven't confirmed that we can use the feature offline when using US English, since my WiFi is always on when I'm working with it.  Let us know what your experience is with that.)

Here's a list of some frequently used voice commands
  and some numerical items the dictation-feature understands:

  (SOME of the may not work in the UK-English version.)
• “Period” – [ . ]
• “Comma” – [ , ]
• “Exclamation mark” – [ ! ]
• "Exclamation point" - [ ! ]
• "New line"
• "New paragraph"
• “Open bracket(s)” – [ [ ]
• "Closed bracket(s)" - [ ] ]
• "Open parentheses" - [ ( ]
• "Closed parentheses" - [ ) ]
• "dash" - [--] {"The weather -- quite cold -- requires a coat."}
• “Nine Euros fifty” – [ 9.50 EUR ]
• “Half past nine” – [ 9:30 ]
• "Eight fifty nine AM" - [ 8:59 AM ]
• “Thirteen forty five” – [13:45]
• “Twelve hyphen 3" - [12-3] {"The Giants won 12-3."}

If the voice-dictation IS based on Nuance's "Dragon" capabilities, some have said that an "utterance" is limited to 30 seconds, but I've been able to talk on and on for well over a minute without a break and it continues and is very accurate.  When I stopped, it waited for me for a short while to press the "Done" icon, having apparently 'learned' that I can be long-winded.  So, this is quite good for a built-in undocumented voice-dictation feature.

  Tips on getting voice commands and punctuation to work as expected
  Before asking for 'new line' or 'new paragraph' it works better if a period prefaces it (though it's not usually necessary).

  Also, it's best to pause slightly after saying 'period' when wanting a new line or new paragraph.

  ON the other hand, don't wait too long to say 'period' to end a sentence, as it will add a space before it if you pause for more than a millisecond.


If you have used the voice-dictation feature (available since November 2013 on the mentioned Kindle Fire tablets) and you find new capabilities while trying it out, let us know in the Comments area.


NOTES on a really good audio-recording app that I tried this weekend
(This one is NOT for speech to text but to get dictated material in audio-file format to send to others or for record-keeping.)
  UPDATE2: AFTER I made the demo recordings for 'medium' and "high" quality versions of the original free Easy Voice Recorder app, I found out tonight that the once $4 Easy Voice Recorder PRO is now "Actually Free" under the new Amazon Underground feature, which allows us to use, for free, apps that are normally paid apps at other major app stores.

  Amazon offers these for free when the app developer is paid based on usage (which means that how often you use the app IS tracked for the developer's revenue), and apparently this developer has confidence in the app's usefulness.  I can see why.

The *now FREE* Easy Voice Recorder Pro  (EVR Pro)
.
  For the Easy Voice Recorder Pro app, I added two sound files in WAV format from the original free non-Pro app.  The first audio file is an EVR Medium Quality WAV sample.

  The second audio file uses the "high" quality setting and is my EVR High Quality WAV sample.

  The third file is an EVR Medium quality M4A sample.  The second audio file uses the "high" quality setting and is my EVR High Quality M4A sample.

  Since I recorded these two audio files before finding that the "PRO" version is now free and offers stereo recording, my wording in these files should be ignored.

  While these older free-version samples are in mono, they still give an idea of the clarity you'll hear in the Pro app files (newer version that's free only at Amazon).
   Since Easy Voice Recorder Pro records and STORES, on the tablet, the considerably larger WAV format file, to keep highest quality available (while also offering compressed versions in M4A format when sending the file to others), I've linked the medium and high quality WAV as well as the two compressed M4A files for anyone curious about the differences.

  These should give you an idea of which might work better for you if you're interested in a voice-recorder with sound-file output for sending or sharing.  I noticed that several customer reviews mention that these apps work pretty well for singing as well :-)

  Apologies for sounding so sleepy in them, but that's because I was!

  BONUS feature.  I saw that Easy Voice Recorder Pro allows you to Play a recording WHILE Recording on top of that, which means you are "layering" tracks and can do a duet with the output from the first file to make the 2nd file.  Unfortunately, it is picking up the speaker output of the first file, so the quality of the first sound-file's audio is not good.  But if you're recording music, it's a good way to practice with another part.


  An alternate fully-featured voice-recorder app, with a great interface:
Another very flexible voice recorder app: Hi-Q Audio Recorder also allows you to pause a recording, has several additional, interesting features, and makes files in M4A, WAV, OGG or Flac. ($4~) (m4a sample).  This app doesn't offer an mp3 output file because Hi-Q also offers a very very good, free mp3-recorder app, Hi-Q MP3 Recorder, but this free one does not have a Pause option, which is an important lack (mp3 sample).  The sample from from the $4~ "audio" app is better for sibilants while the sample from the free "mp3" app has better presence. The Hi-Q apps, like the Easy Voice apps, do allow you to Play one recording WHILE recording on top of that.


TIP on making recorded notes when away from WiFi or 3G and then later converting these sound files into text
  I was asked about how to save notes, when using Voice-Dictation (to text) during times that we don't have a WiFi or 3G connection for sending a note or email somewhere.

  The first thing I found was that the voice-dictation-to-text offline is, for all expected purposes, useless.  Apparently the data set that is on the tablet system has to be very pared-down and therefore the app can understand VERY LITTLE of what I say when offline and not connected to the Net's large data servers with the full data sets.  What it typed for me, then, was not understandable by me.  This was not unlike some human-to-human conversations.

  So, don't bother with trying to use Voice to Text when WiFi or 3G are not available.  The Voice Dictation feature is very accurate and capable when you have a fast WiFi connection.  I have Comcast Cable and am unable to test how it would be under slower DSL connections.

BUT, what you can do when in a space without WiFi or 3G is to use a simple voice-recorder app like Easy Voice Recorder or Hi-Q's two apps described above, which all save automqtically to sound files on your device.

  When you're back on WiFi, you can send the sound files you want, to your computer (via EMail, WiFi transfer, or USB) and then play the audio on the computer while using the tablet's built-in Voice Dictation-to-text to convert it to text for emails or text notes.  The conversion-to-text won't be quite as accurate as when you're actually talking to the microphone, but it's workable with more post-editing required.  Using speakers with good treble output to project sibilants will improve the accuracy of the voice-dictation conversion to text.


Reminder for Nov 30: Amazon has extended the "extra 30% off" for PRINT books deal through the end of Nov 30 (11:59 PST) -- you can get any PRINT book at Amazon for an extra 30% off -- just enter "HOLIDAY30" at Checkout time in the promo code field.  HOWEVER, it's restricted to a maximum $10 off.  This was extended to 11:59pm PST Monday, Nov. 30.  See Amazon's "terms and conditions" here.

  Also, Amazon's aggressive-pricing continues during its CyberMonday Week that started today.


(This blog article was first posted Nov 28 and has been revised Nov 29-30.


Below are ways to Share this post if you'd like others to see it.
-- The Send to Kindle button works well only on Firefox currently.

Send to Kindle


(Older posts have older Kindle model info. For latest models, see CURRENT KINDLES page. )
If interested, you can also follow my add'l blog-related news at Facebook and Twitter
Questions & feedback are welcome in the Comment areas (tho' spam is deleted). Thanks!

11 comments:

  1. I just tried the voice dictation to text feature on my Facebook app using the Fire HD 6 and it work even at airplane mode. I am not an English native speaker but it still recognize more than 90% of what I said. The voice command on punctuation and numerical items also work most of the time. This is really a cool feature.
    Unfortunately, I can’t find the same microphone icon when I run the Facebook app in my new Fire 7 (the super low-cost $49.99 Fire tablet that Amazon just least few months ago). It seems that Amazon don’t support this voice dictation to text feature in their new 5th generation Fire tablets. What a pity...

    ReplyDelete
    Replies
    1. Jack, I mentioned which Kindle models -are- supported -- they're all the currently-available HD and HDX ones -- so, they're not getting rid of it. I agree, it's a cool feature :-) . But the $50 Fire is that much cheaper for a reason -- it doesn't do as much and is the budget model for those who prefer to spend less money when buying several for family etc. The voice-dictation also works for the Kindle HD 4th Generation (2014). When I first tried it in Nov 2013, it wasn't as smooth as it is now. I was pleasantly surprised to find that it works so well now ! But they've taken feedback and dealt with some gliiches, apparently (from what I read in forums).

      Thank you for your feedback! Interesting that it works pretty well even in airplane mode on your HD 6 !

      Delete
  2. I'm looking for something that will record voice memos, save them to cloud, and do speech transcription. I'm not sure such a thing exists, closest thing I know of is OneNote, which lacks only speech-to-text. Evernote doesn't record voice memos in the Android app yet, but it is a logical thing for them to do at some point, and adding ability to index voice memos to make them searchable would be a good complement to their ability to do OCR on photos to make those searchable.

    Amazon's Notes app will do voice memos and sync to Amazon cloud, but I'd prefer something cross platform.

    ReplyDelete
    Replies
    1. Tom, I used the Kindle fire keyboard Voice Dictation to make notes in Evernote, and I'm pretty sure they're searchable. I like that we can photograph a receipt and use it in Evernote, where it makes the content searchable.

      Hi-Q also has an option to save to Dropbox. Both Hi-Q and Easy Voice Recorder have a lot of sharing options. Will try to detail them sometime, since I don't see that on their product pages...

      Delete
    2. Tom, I used the Kindle fire keyboard Voice Dictation to make notes in Evernote, and I'm pretty sure they're searchable. I like that we can photograph a receipt and use it in Evernote, where it makes the content searchable.

      Hi-Q also has an option to save to Dropbox. Both Hi-Q and Easy Voice Recorder have a lot of sharing options. Will try to detail them sometime, since I don't see that on their product pages...

      Delete
    3. I'm not talking about voice dictation, but voice memo recording, so I can listen to what I actually said later on. Not particularly interested in generating text for its own sake. Evernote for Android will not capture audio. I can do this part somewhat now with OneNote or Fire Notes or I guess one of the third party apps you mention (record and save to cloud). But I don't, because it is so much fiddling around to do it. The state of things.

      The complete use case is that I'm driving or walking or hiking and have this idea. I push a button, say 'record voice memo', dictate, say 'stop recording' and it saves to the cloud, while I keep driving or walking or hiking. Then something up there transcribes it so I can search for it later (I'm going to have hundreds of these ideas of course, if only the friction to capture one is reduced to near zero).

      It seems like such an obviously useful thing, but as far as I know there is nothing that does this right now (but I'm just starting to research it).

      Delete
    4. Agreed on that first paragraph.

      On your 2nd paragraph, Iwhere you want to 'search' your ideas (that would be more easily done by saving to text and not to audio for a simple, non-expensive app). I did something like that about 2 hours ago. One Note now has an option to use One Note Badge app. It's memory resident on the tablet, and has an icon in the top-right corner (which I'm not sure at all that I want that around but I'm just trying things out).

      With a real, physical memo recorder, no app button or icon taps would be necessary to start recording.

      I did have to tap on it of course, and a little form pops up, and I didn't bother giving it a title (it then uses your first sentence as a 'title.' I did have to tap on the 'Add a quick note' and it brought up my keyboard so I tapped the microphone and then just started talking and rambled away. You won't like having to do all those taps though -- And I think a dedicated voice recorder is better for you as long as it comes with 3G :--). if you want it saved to Cloud (and device for safety) A dedicated voice recorder that could hold a better data set and would be more accurate with fast talkers.

      What I dictated, I checked later by opening One Note and it was there. Once I turn Wireless On I can sync it to the Cloud. BUT, I don't know OneNote. I am on Ofc 2010 and hardly use it and my OneDrive space is only what I get for my laptop. Am quite clueless on how it works. I use Evernote. The voice-memo apps I mentioned are great for making sound files you can just listen to though And you can play them on one device while the voice-dictationon another turns it into text.

      Kindle's built-in voice dictation is not very smart when Wireless is Off. It's almost flawless when it's on WiFi but when it's off, it misunderstands quite a bit unless I talk very slowly;and clearly (not very likely in a car). "Wireless off" becomes "wireless sauce" when I'm talking fast.

      As far as saving sound files, the voice memo apps have added PAID features to have your voice memos forwarded to them transcribed, so of course I wasn't interested in that. But they make good sound files for notes to one's self.

      I'm sure that was a frustrating read since it's not what you want in capbilities, but it's just what I was trying.

      Delete
    5. Agreed on that first paragraph.

      On your 2nd paragraph, Iwhere you want to 'search' your ideas (that would be more easily done by saving to text and not to audio for a simple, non-expensive app). I did something like that about 2 hours ago. One Note now has an option to use One Note Badge app. It's memory resident on the tablet, and has an icon in the top-right corner (which I'm not sure at all that I want that around but I'm just trying things out).

      With a real, physical memo recorder, no app button or icon taps would be necessary to start recording.

      I did have to tap on it of course, and a little form pops up, and I didn't bother giving it a title (it then uses your first sentence as a 'title.' I did have to tap on the 'Add a quick note' and it brought up my keyboard so I tapped the microphone and then just started talking and rambled away. You won't like having to do all those taps though -- And I think a dedicated voice recorder is better for you as long as it comes with 3G :--). if you want it saved to Cloud (and device for safety) A dedicated voice recorder that could hold a better data set and would be more accurate with fast talkers.

      What I dictated, I checked later by opening One Note and it was there. Once I turn Wireless On I can sync it to the Cloud. BUT, I don't know OneNote. I am on Ofc 2010 and hardly use it and my OneDrive space is only what I get for my laptop. Am quite clueless on how it works. I use Evernote. The voice-memo apps I mentioned are great for making sound files you can just listen to though And you can play them on one device while the voice-dictationon another turns it into text.

      Kindle's built-in voice dictation is not very smart when Wireless is Off. It's almost flawless when it's on WiFi but when it's off, it misunderstands quite a bit unless I talk very slowly;and clearly (not very likely in a car). "Wireless off" becomes "wireless sauce" when I'm talking fast.

      As far as saving sound files, the voice memo apps have added PAID features to have your voice memos forwarded to them transcribed, so of course I wasn't interested in that. But they make good sound files for notes to one's self.

      I'm sure that was a frustrating read since it's not what you want in capbilities, but it's just what I was trying.

      Delete
    6. I've been saving all the echo update emails, but I haven't done much with the new stuff. The thing I really want is for Alexa to take dictation. For me that would be way useful.

      Tom mentioned OneNote as being able to utilize audio information. I don't use OneNote much (even though I'm very MS centric -- OneNote as a spoken todo list is on my list of ToDo's). I wonder if audio on OneNote is searchable in the same way that "ink" is searchable on OneNote.

      Also I wonder if there's some way to hook Cortana into voice recognition in a OneNote environment (Cortana is another of those things on my list of ToDo's :grin).

      Based on your posts here, dictation using my Kindle Fire HDX 8.9 (last year's model) is another thing just added to my (approaching endless :grin) ToDo list.

      Delete
    7. I'm sorta shocked, Ed, by how good the built-in voice dictation to text is.

      I very much doubt that ONeNote listens to a voice recordin and transcribes that to searchable text (but Google can do it for phone calls received on my Google phone) and I doubt OneNote even transcribes direct from sound).

      Have not tried Cortana yet at all.

      Delete
  3. As one who does not have Kindle Fire, are there any other options? Have tried youtube's translation, Google's in Chrome, and several free speech to text Windows software. All are rubish....Have audio recordings of several hours that I want to convert to text.

    ReplyDelete

NOTE: TO AVOID SPAM being posted instantly, this blog uses the blogger.com "DELAY" feature.

Am often away much of the day, and postings won't show up right away. Posts done to use referrer-links may never show up.

Usually, am online enough to release comments within a day though, so the hard-to-read match-text tests for commenting won't be needed this way.

Feedback and questions are welcome. Thanks for participating.

Technical Problems?
If you're having problems leaving a Comment, Google's blogger-help asks that you clear the 'blogger.com' cookies on your browser's Tools or Options menu bar and that will fix the Comment-box problems (until they have a permanent fix).

IF that doesn't work either, then UNcheck the "keep me signed in" box -- Google-help says that should allow your comment to post (it's a workaround to a current bug).
Apologies for the problems.

TIP: There's a size limit. If longer than 3500 characters or so, in a text editor, make two posts out of it.

[Valid RSS]