[paperboy317]

I'm wanting written transcripts of podcasts and audio files I have. Is there a reliable (even semi-reliable) software out there that can 'read' an audio file and genereate a text file of the conversation?

[Stream Recorder]

There are three services I know that convert speech to text to allow you to search within a podcast: Blinkx, Podzinger, Podscope.

All these services run podcasts through a speech recognition engine to convert speech to text, that is why it becomes easy to search within podcasts.

Blinkx Inc. is using speech to text software from Autonomy Corp. It shows a transcript but starts from the beginning of the podcast.

Blinkx scans not only podcasts but also millions of hours of TV broadcasts.


Podzinger is based on the technology from BBN Technologies Inc.
that created services for the U.S. intelligence services, namely developed speech recognition that has been used for real-time analysis of foreign broadcasts in Arabic and other languages, now it allows you to mine podcasts.

Podzinger provides the time at when the search phrase was found, a text transcript of that time period that you can scan quickly to figure out if they are talking about what you are actually looking for.


TVEyes makes Radio & TV searchable by keyword, phrase or topic - just as you would use a search engine for text. TVEyes has been indexing television and radio broadcasts since 1999. And now TVEyes offers Podscope - a service to search within podcasts.

TVEyes crawls the web with Podscope looking for podcasts and creates an index against every word, thereby making the contents searchable. The user can search on a term, generate a list of results ranked by a variety of ods to find the most relevant podcast and click to play or click to download. Although Podscope doesn't produce a transcript you can read.


As you can see there are podcast searching services that successfully use speech recognition engines. However ATM I can't suggest you a standalone program that can convert your downloaded podcasts to text.

[Mohsin_Khan]

Also will the below mentioned software work in a transcription industry with different people using the the software?

[lcarlson]

[moderator's notice]spam is not allowed
You can use a professional to transcribe for you. Some of them use the following realtime technology, which means that while they listen to your tape or audio file, they input steno into a stenomachine; the steno gets converted into English using court reporting software and appears in realtime on a computer screen. It is quick turnaround.

[Jodi Harmon]

Though voice recognition software (described in prior posts) is getting better and more accurate all the time, it’s always a single trained voice that gets the best results. A multi-voice environment such as a podcast is one of those instances where a human needs to facilitate or perform the service, and that’s where realtime transcriptionists come in. Realtime transcripts can be streamed live at the time of the podcast or emailed later, they are searchable and can even be synched with the audio and/or video content.

Jodi Harmon
Realtime Transcription, Inc.
www.realtimetranscription.com