speech to text to notetaking app
Started by jimspoon
on 9/17/2013
jimspoon
9/17/2013 6:22 am
pardon for me for starting another thread from somebody else's post in another thread ... but it might get lost int he original thread ... I was very interested in what Slartibartfarst had to say here:
Slartibartfarst wrote:
Not only am I nuts about getting text from photos of books/documents, I'm also nuts about getting my voice memos into text form. So this part of the post is also very interesting to me. And this function of OneNote and Windows Search is entirely new to me.
I like to whip out my voice recorder and make quick voice memos. I want a transcription of my voice memos in my favorite note manager as soon and as easily as possible. I have good options on my Android phone. In the Evernote app, with the "Dictate" option I can talk and watch the text appear in my note as I talk. It works very well. I can also dictate text into any app using the speech to text button on the keyboard - i think it's Google speech recognition using most keyboards, but Nuance (Dragon) speech recognition when using the Swype keyboard. Not sure which is the best method - will take more experimentation.
But for me there is very much a convenience factor involved. The more difficult it is, the more steps that involved in recording a voice memo, the less likely I am to do it. With the Android Evernote app, or any other Evernote app - there may be several steps involved in waking up the phone, finding the app, tapping on the app, making a new note, stopping recording, etc.
That's how a single-purpose voice recorder can be good. Grab it, press the start button, talk, press stop button. No looking at a display to get to that point. Actually one of the best devices I ever had for making a quick voice note was my old T-Mobile Wing - press button, wait for beep, talk, release button, hear beep. It's a heavy old phone though.
I would love to have a device that I could grab like that - press and hold, beep, talk, release button, auto-upload to cloud via wifi/4g, auto-transcribe.
Alternatively, would like the ability to make a quick recordings through the day and batch upload to cloud service for transcription later in the day. But I haven't found a way to to batch upload audio recordings to Google or Nuance in this manner. So far as I know - the dictation, upload, and speech recognition have to be done in "realtime".
But it sounds like OneNote can do at least part of this - it seems you can put your prior voice recordings into OneNote and they will be recognized and searchable. BUT - does OneNote automatically provide a text transcription of your voice memos that you can read? (instead of just the ability to go to a location in an audio recording that you can listen to). That would be crucial to me.
jim
Slartibartfarst wrote:
Not only images but words/phrases in sound files can be identified and
indexed by OneNote. This can be very handy if you have recorded a
meeting or (say) a radio interview to a sound file. All you need to do
is save the sound file to OneNote, and it will be scanned quite rapidly
in the background. Searches will later show the word or phrase you are
looking for and the time into the recording when that word/phrase
occurs.
Audio phrases in audio recordings saved into OneNote are readily found
via Windows Desktop Search/Index from the Windows Desktop Search bar
(click the Start button), or via the normal Windows Explorer search, or
via OneNote search.
Not only am I nuts about getting text from photos of books/documents, I'm also nuts about getting my voice memos into text form. So this part of the post is also very interesting to me. And this function of OneNote and Windows Search is entirely new to me.
I like to whip out my voice recorder and make quick voice memos. I want a transcription of my voice memos in my favorite note manager as soon and as easily as possible. I have good options on my Android phone. In the Evernote app, with the "Dictate" option I can talk and watch the text appear in my note as I talk. It works very well. I can also dictate text into any app using the speech to text button on the keyboard - i think it's Google speech recognition using most keyboards, but Nuance (Dragon) speech recognition when using the Swype keyboard. Not sure which is the best method - will take more experimentation.
But for me there is very much a convenience factor involved. The more difficult it is, the more steps that involved in recording a voice memo, the less likely I am to do it. With the Android Evernote app, or any other Evernote app - there may be several steps involved in waking up the phone, finding the app, tapping on the app, making a new note, stopping recording, etc.
That's how a single-purpose voice recorder can be good. Grab it, press the start button, talk, press stop button. No looking at a display to get to that point. Actually one of the best devices I ever had for making a quick voice note was my old T-Mobile Wing - press button, wait for beep, talk, release button, hear beep. It's a heavy old phone though.
I would love to have a device that I could grab like that - press and hold, beep, talk, release button, auto-upload to cloud via wifi/4g, auto-transcribe.
Alternatively, would like the ability to make a quick recordings through the day and batch upload to cloud service for transcription later in the day. But I haven't found a way to to batch upload audio recordings to Google or Nuance in this manner. So far as I know - the dictation, upload, and speech recognition have to be done in "realtime".
But it sounds like OneNote can do at least part of this - it seems you can put your prior voice recordings into OneNote and they will be recognized and searchable. BUT - does OneNote automatically provide a text transcription of your voice memos that you can read? (instead of just the ability to go to a location in an audio recording that you can listen to). That would be crucial to me.
jim
Slartibartfarst
9/17/2013 10:00 am
OneNote is unable to provide a text transcription of your voice memos that you can read.
I am not aware of any desktop/client application that can do that. I would certainly want to use it if there were one!
I am not aware of any desktop/client application that can do that. I would certainly want to use it if there were one!
WSP
9/17/2013 12:48 pm
My recollection (admittedly a bit hazy) is that in OneNote you can type notes and simultaneously record; then afterwards you could go to a particular place in the notes and hear the relevant part of the recording. Or something like that. I never tried out that feature, but it struck me as potentially very useful for students and journalists.
Bill
Bill
jimspoon
9/17/2013 3:03 pm
WSP wrote:
My recollection (admittedly a bit hazy) is that in OneNote you can type
notes and simultaneously record; then afterwards you could go to a
particular place in the notes and hear the relevant part of the
recording. Or something like that. I never tried out that feature, but
it struck me as potentially very useful for students and journalists.
Bill
thanks for that reminder Bill ... http://gigaom.com/2008/09/02/coolest-technol/
This reminds me of the Livescribe Sky Wifi smartpen. It records audio while you write, uploads the handwritten notes and audio to Evernote via wifi, and the audio is synchronized with your notes. http://www.livescribe.com/en-us/smartpen/wifi-smartpen/
Audio Notetaker also streamlines navigation through audio - the audio is visualized as colored bars, and you can annotate it, search through the annotations. http://www.sonocent.com/en/the_software/audio_notetaker#software-what-does-it-do
But with these methods, in the end you're still going back to audio.
Usually I just want to get my short voice memos into text form, so that I never have to go back to the original audio any more.
I do have a rather complicated method of doing this. I dictate time stamps in my voice memos, copy the voice memos to my PC, merge them into one big mp3 file using mergemp3, then drop the merged file into Dragon Naturally Speaking DragonPad. That way I end up with a stream of text with time stamps throughout. Then I correct the text as necessary in DragonPad, and cut and paste the text into my notetaking program. As there are a number of steps - i haven't been keeping up with this well lately. BUT - at least with this crazy method - making the initial voice recording is simple and automatic - and I do end up with a stream of text with timestamps. Of course I am always looking for something better!
jim
Gary Carson
9/19/2013 1:54 am
The Dragon Naturally Speaking Professional Edition has an auto-transcribe folder that does what you're looking for, but the pro version costs a small fortune (Knowbrainer.com sells the DNS Professional full version for around $600).
If you can't afford that, there's only two ways I know of to transcribe your voice files and get them into a PIM of some kind. You have to either transcribe them manually or use Dragon (or some other voice-recognition program) to transcribe them, then you have to go through them, correct recognition errors and manually copy the transcripts into your PIM.
Some outliners and PIMs allow you to import files that include special markup codes which tell the program to create new entries, use certain lines as entry titles and so on. SuperNoteCard does this and it works pretty well. You can import a text file into NoteMap and it will create a new outline entry every time it encounters a new paragraph mark. The problem with these methods is that you have to dictate the codes as you're recording your memos or notes or whatever. I've found that it's more trouble than it's worth.
At this moment, Audio Notetaker is the best application out there for managing spoken-word audio files. It won't transcribe your files, but it makes it very easy to annotate them or transcribe them manually and you can break your files up into multiple sections by subject and so on.
I think the problem here is that most people using recorders are still thinking in terms of converting their recordings into text. Audio is a completely different medium with its own advantages and disadvantages. The solution is to start thinking in terms of audio only. Instead of reading your information, you're LISTENING to it.
I've been using Audio Notetaker as a PIM. You can create multiple "notebooks" for different topics and each notebook can contain hundreds of audio files which you can then divide into manageable sections that don't take very long to listen to. You can annotate each section, transcribe parts of each section if necessary, highlight sections, search through them very quickly using keyboard shortcuts. It's really an incredible program.
If you can't afford that, there's only two ways I know of to transcribe your voice files and get them into a PIM of some kind. You have to either transcribe them manually or use Dragon (or some other voice-recognition program) to transcribe them, then you have to go through them, correct recognition errors and manually copy the transcripts into your PIM.
Some outliners and PIMs allow you to import files that include special markup codes which tell the program to create new entries, use certain lines as entry titles and so on. SuperNoteCard does this and it works pretty well. You can import a text file into NoteMap and it will create a new outline entry every time it encounters a new paragraph mark. The problem with these methods is that you have to dictate the codes as you're recording your memos or notes or whatever. I've found that it's more trouble than it's worth.
At this moment, Audio Notetaker is the best application out there for managing spoken-word audio files. It won't transcribe your files, but it makes it very easy to annotate them or transcribe them manually and you can break your files up into multiple sections by subject and so on.
I think the problem here is that most people using recorders are still thinking in terms of converting their recordings into text. Audio is a completely different medium with its own advantages and disadvantages. The solution is to start thinking in terms of audio only. Instead of reading your information, you're LISTENING to it.
I've been using Audio Notetaker as a PIM. You can create multiple "notebooks" for different topics and each notebook can contain hundreds of audio files which you can then divide into manageable sections that don't take very long to listen to. You can annotate each section, transcribe parts of each section if necessary, highlight sections, search through them very quickly using keyboard shortcuts. It's really an incredible program.
Gary Carson
9/19/2013 2:06 am
Note:
For a long time now, I've been looking for a program that will let me search audio files for a specific spoken word or phrase and take me to that section of the file. As far as I know, nothing like this exists.
OneNote lets you embed and index audio files. I've experimented with this using OneNote 2007, but it didn't do what I wanted it to. Not exactly. OneNote will search the audio files for a word or phrase (that you type in), but if it finds the relevant string, it will only tell you which file it was in. You still have to listen to the whole file to find the information.
Audio Notetaker doesn't search audio files, but it does search its written notes, so you could use its search function to find certain annotations if they exist. How well this works depends on how thoroughly you annotate the files.
Ultimately if you're dealing with audio files, you end up listening to them. You should be as concise as possible when taking notes, but you can divide the files into multiple sections in Audio Notetaker. My Audio Notetaker PIM is full of short audio clips. A lot of them are only a few seconds or a few minutes long.
For a long time now, I've been looking for a program that will let me search audio files for a specific spoken word or phrase and take me to that section of the file. As far as I know, nothing like this exists.
OneNote lets you embed and index audio files. I've experimented with this using OneNote 2007, but it didn't do what I wanted it to. Not exactly. OneNote will search the audio files for a word or phrase (that you type in), but if it finds the relevant string, it will only tell you which file it was in. You still have to listen to the whole file to find the information.
Audio Notetaker doesn't search audio files, but it does search its written notes, so you could use its search function to find certain annotations if they exist. How well this works depends on how thoroughly you annotate the files.
Ultimately if you're dealing with audio files, you end up listening to them. You should be as concise as possible when taking notes, but you can divide the files into multiple sections in Audio Notetaker. My Audio Notetaker PIM is full of short audio clips. A lot of them are only a few seconds or a few minutes long.
jimspoon
9/19/2013 4:09 am
Gary, that is an interesting approach. I have watched videos about audio notetaker and installed it once - didn't give it much of a workout. I was really put off by the price - $150 USD !! I might buy if it were a lot less - but I guess for some people, what it does is so helpful that it is worth it.
It is strange that OneNote recognizes the words that are spoken, but doesn't transcribe it (probably because of the inevitable misrecognitions and need for correction) or at least take you to the specific part of the audio where the searched terms are found. I wonder if that has changed in later versions of OneNote?
At your suggestion I did start dictating the time into my voice memos so that the time will show up in the Dragon transcription - that helps. Because of the nature of my notes, I am constantly saying the words "new line" so that little bits of speech are put on separate lines in DragonPad. Beyond that I don't try to put in much of the Dragon dictation commands.
jim
It is strange that OneNote recognizes the words that are spoken, but doesn't transcribe it (probably because of the inevitable misrecognitions and need for correction) or at least take you to the specific part of the audio where the searched terms are found. I wonder if that has changed in later versions of OneNote?
At your suggestion I did start dictating the time into my voice memos so that the time will show up in the Dragon transcription - that helps. Because of the nature of my notes, I am constantly saying the words "new line" so that little bits of speech are put on separate lines in DragonPad. Beyond that I don't try to put in much of the Dragon dictation commands.
jim
Alexander Deliyannis
9/19/2013 7:01 am
jimspoon wrote:
If you are interested, I suggest you keep an eye on it at Bits du Jour; it should eventually reappear there.
Gary, that is an interesting approach. I have watched videos about
audio notetaker and installed it once - didn't give it much of a
workout. I was really put off by the price - $150 USD !! I might buy
if it were a lot less - but I guess for some people, what it does is so
helpful that it is worth it.
If you are interested, I suggest you keep an eye on it at Bits du Jour; it should eventually reappear there.
Slartibartfarst
9/19/2013 8:42 am
Gary Carson wrote:
You may have been omitted something necessary there.
For example, refer: Searching for information in audio notes in OneNote - http://www.donationcoder.com/forum/index.php?topic=31755.msg302855#msg302855
"It takes you straight to the audio clip and the position (time) in the audio clip that the searched item occurs:"
For a long time now, I've been looking for a program that will let me=================================
search audio files for a specific spoken word or phrase and take me to
that section of the file. As far as I know, nothing like this exists.
OneNote lets you embed and index audio files. I've experimented with
this using OneNote 2007, but it didn't do what I wanted it to. Not
exactly. OneNote will search the audio files for a word or phrase (that
you type in), but if it finds the relevant string, it will only tell you
which file it was in. You still have to listen to the whole file to find
the information.
You may have been omitted something necessary there.
For example, refer: Searching for information in audio notes in OneNote - http://www.donationcoder.com/forum/index.php?topic=31755.msg302855#msg302855
"It takes you straight to the audio clip and the position (time) in the audio clip that the searched item occurs:"
Gary Carson
9/19/2013 8:24 pm
I was put off by Audio Notetaker's cost as well, but I finally picked it up at half-price the last time it was on Bits du Jour.
Gary Carson
9/19/2013 8:34 pm
“It takes you straight to the audio clip and the position (time) in the audio clip that the searched item occurs:”
I'll have to test this again. Maybe I missed something. If I remember right, OneNote didn't actually open the audio file. Come to think of it, I'm not even sure if OneNote has an embedded audio player. I'll have to check that out. I think you can make recordings in OneNote. As for transcribing the files, I wouldn't expect OneNote to include this capability. Voice recognition is a complicated business. You have to train your voice in Dragon, for instance.
I always dictate the date, time and subject at the beginning of each new recording. It simplifies things when I'm searching through the files on my recorder. Still, managing audio files requires a lot of administrative overhead. Dictation's so fast and efficient you can easily create hundreds of files before you know it.
I'll have to test this again. Maybe I missed something. If I remember right, OneNote didn't actually open the audio file. Come to think of it, I'm not even sure if OneNote has an embedded audio player. I'll have to check that out. I think you can make recordings in OneNote. As for transcribing the files, I wouldn't expect OneNote to include this capability. Voice recognition is a complicated business. You have to train your voice in Dragon, for instance.
I always dictate the date, time and subject at the beginning of each new recording. It simplifies things when I'm searching through the files on my recorder. Still, managing audio files requires a lot of administrative overhead. Dictation's so fast and efficient you can easily create hundreds of files before you know it.