4 Replies Latest reply on Oct 1, 2011 1:48 AM by Sant-Iago

    Transcribing audio to text


      First I apologize for my bad English ....

      Let's see if I can lend a hand, it is urgent, I want to do is to reproduce audio, and what is being heard in the audio (voice) go see text and goes over.

      It should be from an xml file and if it can be animated, the better, to be easily upgradeable.

      I do not know if I explained well

      Greetings and thank you very much!

        • 1. Re: Transcribing audio to text
          sinious Most Valuable Participant

          There is no built in way to convert a waveform (audio) to text, or detect words.


          You're going to manually need to do this based on custom timing. You can certainly put the text in XML files and read it out of them and display it at the right time. That timing however must be specified by you. Flash has no way to automate this.


          Even analyzing the audio signal with what flash actually can do would be so complicated it's certainly not a forum issue.

          1 person found this helpful
          • 2. Re: Transcribing audio to text
            Sant-Iago Level 1

            Thank you for your answer, the way I do it is as you say, with a xml by hand, but looked for a particular example because I know a lot more little programming.


            What I do is the following:


            I have three buttons, each would load an mp3 audio in a player (which also have to look like once uploaded to...), audio, xml file should be synchronized with the accompanying text with the audio you're listening, and thisis somehow remarcaría and go with animation going on ... ..


            Thanks again to see if anyone can help me, because I programming ... not very well ... as you can see.

            • 3. Re: Transcribing audio to text
              Rothrock Level 5

              I have used Adobe Soundbooth to create a transcript of audio files. The transcript can be output to xml with the word and time code. It only works well if the audio is very clearly spoken and has little/no background noise. Words can't be slurred or run together. It is possible to correct the transcript, but the interface for that is very tedious.


              For another project where we wanted something more like TV closed captioning. In that case we prepared the transcript by hand, broke it into chunks and assigned them to an array.


              var captions:Array=[];


              captions["welcome"]="The welcome message here."

              captions["msg1"]="Some more text goes here. etc."


              Each "key" needs to be unique. We then put labels on the timeline which would match the point at which the message should be shown. Finally we used and enter frame event to check the currentLabels and see where we are and display the appropriate caption.

              • 4. Re: Transcribing audio to text
                Sant-Iago Level 1

                Thanks for your answer Rothrock, the second option is what I need, what happens is that I know very little about programming and see if I can clarify. One more question, the text passes can be viewed indifferent sizes? (which is being heard in one size and that is happening in a smaller size? and different color?


                Greetings and thank you very much!