MacSpeech Forums and Discussion » MacSpeech Dictate » Extending Dictate

1. Yosele
   Member

   Hi,
   I can't find any directions for transcribing my iPod Nano voice memos. Would be grateful for any assistance.
   Thanks,
   Yossi

   Posted 1 year ago #

2. Ricky Buchanan
   Member

   To my knowledge, Dictate doesn't currently handle transcribing from any recorders.

   Posted 1 year ago #

3. Martin C
   Member

   http://www.google.ca/search?num=100&hl=en&client=firefox-a&rls=org.mozilla%3Aen-US%3Aofficial&hs=gZp&q=soundflower+macspeech&btnG=Search&meta=

   found in particular:

   Dan Thies on August 12th, 2008 11:43 pm

   Re: the comment on using Dictate for transcription.

   I installed Dictate yesterday, and tested it while recording a podcast with Audacity, using Soundflower and the “Software Playthrough” feature (look in Audio I/O prefs) on Audacity.

   I first had to set up a profile with Soundflower (2ch) as the audio device, but Software Playthrough made that work just fine. I was able to read the text in the Dictate window into Audacity, and Dictate picked it all up seamlessly.

   Simultaneous recording of the podcast along with transcription worked just fine using a Dictate notepad, but playing back older podcasts through Soundflower seemed to give Dictate more than it could handle.

   Adding approx. 5 seconds of silence at the beginning & end of the audio file and adding a couple seconds of silence every 30 seconds or so, appears to improve the results so far.

   Dictate worked well enough that I will be able to generate transcripts with about the same amount of editing time as I spend with my transcription service’s output.

   Next test is the “analog” one - podcast playing through speakers next to the mike.

   So far, Dictate has been far and away the best speech recognition software I’ve used, and I’m now just one good Mac app away from retiring my (virtual) PC for good.

   (From http://macapper.com/2008/03/16/macspeech-dictate-solid-voice-recognition/)

   Posted 1 year ago #

4. Martin C
   Member

   I am having trouble with soundflower in that I can't get it to display its menu bar.

   Their support guy Ben kindly suggested some other routes for people in a hurry:

   Soundflower is one of several utilities like this currently available. Please check out these other internal audio routing projects:

   http://jackaudio.org/
   http://www.rogueamoeba.com/
   http://www.ambrosiasw.com/utilities/wiretap/

   Also, you may want to look into taking advantage of Mac OS X's ability to create aggregate devices out of multiple audio devices, which you can read more about here:

   http://www.apple.com/pro/techniques/aggregateaudio/

   Wouldn't be necessary if MacSpeech accepted audio from any source...

   Posted 1 year ago #

5. Martin C
   Member

   I asked Dan how to insert pauses... he said:

   It would be difficult to automate, because you don't want to insert a pause in the middle of a word. It takes a couple minutes to do it by hand with a 30-minute session.

   Dan then kindly replied again and said:

   It probably isn't very clear how to set up the macspeech profile with soundflower.

   You have to read the sample text to create a profile, and you can't switch inputs on a profile, so it took a little kung fu to create a soundflower profile.

   In the Audacity audio settings, I set soundflower (2ch) as the playback device, my headset as recording device, and checked "software playthrough." In MacSpeech, I set soundflower (2ch) as the input device, and created a profile. Reading the training text into my headset, Audacity piped the audio out to soundflower, which pushed it into Macspeech.

   Once the profile is set up, I was able to play my audio in Audacity with soundflower as the playback device, and that feeds into MacSpeech.

   Note that this won't work with other people's voices, and the recording quality has to be good. Your mileage may vary, I am getting what I need from it but I may just be lucky.

   I haven't tested it, but it might be worth testing, using something like Quicktime to play the audio, and simply using the A/V controls in Quicktime to slow the pacing down a bit so MacSpeech doesn't get overwhelmed.

   Posted 1 year ago #

6. Martin C
   Member

   I confirm that playback of a recording through Audacity -> Soundflower -> new MSDictate profile did indeed start to inject transcribed text. So at least that works. I don't care for perfection, this is simply to help me index audio files. Ideally MSD would insert the transcription track into the audio... I guess that's dreaming off into the future!

   A further issue with this is that Dictate inserts the text it hears into the foreground window. If you were to transcribe a previously recorded document you may well wish to use your Mac at the same time. Two Macs (or running MSD in a Virtual Machine) is not a solution because one can't share the license across two machines.

   Posted 1 year ago #

RSS feed for this topic

Reply

You must log in to post.