Forum Posts Feed
how to get voice recordings and transcribed and into a PIM as text notes
< Next Topic | Back to topic list | Previous Topic >
Posted by jimspoon
Apr 18, 2011 at 04:08 AM
I want to way to convert digital recorder voice files into PIM text notes with as little work as possible. More specifically, I want to be able to use voice recognition as a time tracker tool. Whereever I am, I want to be able to dictate what I am doing or have just done into a voice recorder, and later end up with a series of time-stamped text notes - ideally it would go straight into a personal information management program, or could easily be imported into one.
I’ve been looking at reviews of Dragon Naturally Speaking Premium, which will attempt to recognize speech in audio files imported from a voice recorder, and transcribe them into text. DNS apparently is not as accurate in recognizing recorded speech as it is in recognizing real-time speech dictated into a microphone at the computer. One user says the correction process isn’t easy, either. I don’t know anything about how one would get text transcriptions out of DNS and into a PIM. There are lots of complains about tech support provided by Nuance, which sells DNS.
I am looking at alternatives for accomplishing the same end result. One thing you can do is set up a Google Voice account, and call your Google Voice number, and leave a message. Google Voice will transcribe your message. When you go to your Google Voice account web page, you will see the text along with a playback button. As you play back the audio message, the corresponding part of the transcribed text is highlighted. Very neat. But Google Voice does not allow you to edit the transcribed text during this process. In my brief testing, the accuracy of the transcription was decent but not great. Also - it’s not very convenient to have to make a phone call to make a note - you have to wait for the phone to ring and rollover into voicemail, etc. Also it might not be very easy to export transcribed text from emails into a PIM.
Another option is to use Evernote with Voice2Note / Dial2Do. Voice2Note gives you a phone number you can dial. When you dial that number from a designated phone, you are quickly allowed to dictate your recording. After a short time - it shows up in Evernote as a text note. I tried it once and the accuracy was quite good. I think it’s possible to use Evernote mobile apps to make voice recordings and Voice2Note will transcribe them into a text note. Voice2Note could be a viable option.
Another approach I am looking at is Sonocent Audio Notetaker. It doesn’t do voice recognition - says it’s just not accurate enough - but does a lot to manage and navigate your voice recordings and add notes and slides. But what I would want to do is to get the text notes out of Audio Notetaker and into my all-purpose PIM.
Anybody have any thoughts or ideas on this subject?
JIM
Posted by MadaboutDana
Apr 18, 2011 at 10:59 AM
Take a closer look at OneNote. Two good articles here:
http://blogs.msdn.com/b/onenotetips/archive/2008/05/29/audio-and-video-recording-in-onenote.aspx
and here:
The latter article in particular shows why OneNote doesn’t support voice recognition straight out of the box (although it does support speech searches), but can be adjusted to support voice recognition (question of adjusting compression to voice recognition-compatible levels). And of course OneNote time-stamps as a matter of course. You’ll find quite a lot of people claiming OneNote doesn’t support voice recog - in fact, it can.
Posted by Dr Andus
Apr 18, 2011 at 02:08 PM
jimspoon wrote:
>I want to way to convert digital recorder voice files into PIM text notes with as little
>work as possible. More specifically, I want to be able to use voice recognition as a
>time tracker tool. Whereever I am, I want to be able to dictate what I am doing or have
>just done into a voice recorder, and later end up with a series of time-stamped text
>notes - ideally it would go straight into a personal information management program,
>or could easily be imported into one.
If these are just short notes (2 or 3 sentences), you could use the Dragon app on iPhone/iPod Touch and then paste them into the Notes app, which timestamps them for you and also syncs with Gmail, where you can add labels (tags) to organise them further. This is a very quick and easy process. You can also correct the text in Notes (and to some limited extent in the Dragon app). Longer texts wouldn’t work as well because I think there is a limited amount of text the app can handle and also you can’t see the text while dictating (as opposed to the desktop version), so there could be more mistakes to correct afterwards.
>I’ve been looking at reviews of Dragon
>Naturally Speaking Premium, which will attempt to recognize speech in audio files
>imported from a voice recorder, and transcribe them into text. DNS apparently is not
>as accurate in recognizing recorded speech as it is in recognizing real-time speech
>dictated into a microphone at the computer. One user says the correction process
>isn’t easy, either. I don’t know anything about how one would get text transcriptions
>out of DNS and into a PIM. There are lots of complains about tech support provided by
>Nuance, which sells DNS.
Dragon on the desktop is a very complex and sensitive animal, so it requires some initial investment of time and effort to learn about it and to train it and mould it to your needs. The quality of the transcription will depend on a lot of things, including how clearly you speak, the type of vocabulary and the amount of background noise. Once the recording is fed into the software, it outputs the file as RTF, so any outliner or notetaking software that can import it would work. But clearly there would be a discrepancy in terms of timestamping between the time of recording and transcription (unless you record that as well).
If you use Windows 7, you may not need additional software to organise the notes. You could just save the transcribed RTF files into a Windows folder, and then in Windows Explorer select the file preview pane and presto!
However, beware: the latest Dragon version requires an up-to-date computer, lots of RAM and computing power and it generally doesn’t like many other programmes running at the same time (especially not Skype).
As for Nuance tech support, I had a positive experience. However, I think only the first 2 weeks are free. So sort out all your problems ASAP with them :) But they were very responsive and helpful and that’s how I figured out that I had to quit Skype.
Posted by Dr Andus
Apr 18, 2011 at 02:13 PM
Dr Andus wrote:
>If these
>are just short notes (2 or 3 sentences), you could use the Dragon app on iPhone/iPod
>Touch and then paste them into the Notes app, which timestamps them for you and also
>syncs with Gmail, where you can add labels (tags) to organise them further. This is a
>very quick and easy process. You can also correct the text in Notes (and to some limited
>extent in the Dragon app). Longer texts wouldn’t work as well because I think there is a
>limited amount of text the app can handle and also you can’t see the text while
>dictating (as opposed to the desktop version), so there could be more mistakes to
>correct afterwards.
Forgot to add that the Dragon app on the iPhone/iPod only works if you have an internet connection. The desktop version doesn’t need an internet connection to work.
Posted by Gary Carson
Apr 18, 2011 at 05:26 PM
You don’t need software to enter a timestamp. Just dictate the date and time at the beginning of each recording.
I just tried transcribing a nine-minute file from my Olympus DS5000 voice recorder (using Dragon v.10 Preferred) directly into OneNote 2007 and it worked fine despite the fact that Dragon sees OneNote as a non-standard window.
You could also transcribe your recordings into a text file and then copy the text into your PIM. Requires an additional step, though. Maybe multiple steps, depending on what you’re trying to do.
The amount of work you’ll have to do depends on how you want to organize these notes. If you’re going to dictate notes throughout the day and then create ONE note in your PIM for each day’s notes, then the work isn’t so bad. Just use one of the methods above.
On the other hand, if you want a separate PIM note for each dictated note, you’re going to have to create a new PIM note for each entry and either transcribe each dictated note separately or do a bunch of copying and pasting.
The other method is to use a PIM that lets you import text files with special formatting characters that break the text into separate PIM notes with titles, etc. All of these special formatting characters can be dictated. You would then transcribe the dictation and import it into your PIM. I don’t know if you can do this with OneNote or Whizfolders, but I’ve tried it with SuperNoteCard and it works perfectly. I dictated the content for half a dozen cards, including all the formatting characters, punctuation, etc., transcribed the file, imported it into SuperNoteCard and ended up with half a dozen properly titled cards.
I can’t remember the formatting characters SNC uses, but let’s say the pound sign (#) separates the title from the body of the card and the at-sign (@) creates a new card. The dictation would go like this:
at-sign January 12 2011 10:15 a.m. new paragraph pound-sign
I walked down to the grocery store comma robbed the place comma then rode off on my bicycle period at-sign
January 12 2011 12:30 p.m. new paragraph pound-sign
Had lunch with the President period
etc. etc.
Another method is to forget about the PIM altogether and use your voice recorder as your system of record. Just keep everything on the recorder. The major problem with this is that you can’t search the files. I have yet to find a good application for archiving and searching sound files. AudioNoteTaker probably comes the closest, but it can’t handle DSS files (which my own recorder produces) and I don’t want to go through the extra step of converting my DSS files to WAV. Also, you don’t end up with transcripts with AudioNoteTaker, just a collection of sound files with annotations.
Conclusion: there’s no really good way to deal with large collections of sound files or transcripts of sound files, but the most efficient method is probably to use a PIM that allows the importing of text files. Dictating all those formatting characters and so on sounds like a major pain, but it really isn’t all that bad once you get used to it. I guess it all boils down to whether your PIM allows importing and how complicated the formatting characters are. With Dragon, you can create spoken versions of various text strings that simplify things.
Does anybody know if you can import files into OneNote or Whizfolders?