Automatic Speech to Text Conversion

Speech to Text Conversion in Microcone Recorder

To easily index your meeting recordings for subsequent search and retrieval, Microcone® Recorder includes a Speech to Text Conversion feature. This feature is delivered using a Nuance Dragon pay-per-use web service. The service consumes one Conversion Credit for each segment of audio that is converted to speech. This segment may be either a single Highlight, or a Speaking Turn for the full indexing option. Conversion Credits can be purchased in Pre-Paid Bundles or Subscription Plans - see details below.
Stacks Image 21
Note that the default speech-to-text web service described here is suitable for general usage. Larger corporate or government clients may benefit from a custom tailored audio indexing solution with customised domain models and different hosting options. Dev-Audio can offer customised corporate systems working with a range of audio indexing or transcription partners, including Quorate Technology and Koemei. Please contact us for further details on custom corporate audio indexing solutions.

How it Works

Stacks Image 22
Before doing any Speech to Text Conversions, you should first go to the application Preferences. Here you can change the language and accent region, as well as specify the minimum segment (speaker turn) in the transcript that should be converted (Minimum allowable setting is 2 seconds).
Stacks Image 23
Select the meeting from the list of recordings, then click the Speech to Text button in the top toolbar. This will open the Speech to Text panel.
Stacks Image 24
Sign In to your Microcone Recorder account (requires internet connection).
Click on Create Account if you do not already have one - 100 free Conversion Credits will be provided for your first account.
Stacks Image 25
View your balance of Conversion Credits. Click on Buy Conversions to purchase Pre Paid Bundles through iTunes In-App Purchase. Alternately, login to our Accounts portal to purchase Pre Paid Bundles or Subscription Plans directly.
Stacks Image 26
Select whether you want to convert the All Highlights that you have tagged, or else just a single selected Highlight. The number of Conversion Credits required for each option will be shown - this equates to the number of short audio segments (either a Highlight, or a speaker turn) that will be converted.

You may also opt to Index the Full Recording. This will convert all speech segments longer than the Minimum Segment to Convert setting under Preferences. You can convert only the longer segments if desired, giving a lower cost in terms of Conversion Credits. Note that this is not intended to give an accurate meeting transcript, but may be used to help index your meeting content for subsequent search and retrieval.
Stacks Image 27
Press the Confirm button to commence the Speech to Text Conversions. This will start sending audio segments to the Nuance Dragon web service.

Text results should start returning one by one after approximately 30 seconds, depending on your internet connection. The text will appear in the Highlights panel, or on the Transcript tab, depending on the option you selected.

Accuracy

Speech to Text Conversion in Microcone Recorder has been designed as a powerful tool for automatically tagging your meetings, facilitating search and retrieval from a large collection of meeting recordings, and helping you quickly find the key sound bite you want to review. While this feature uses the state-of-the-art in speech recognition technology, as a fully automatic service it will not match the accuracy of an experienced human transcriptionist. Whether it is being done by a person or a machine, recognising speech in group meetings is a significantly more complex task than single voice dictation due to a range of factors. For these reasons, a reasonable expectation is that a fully automated transcription will be approximately 70% correct at a word level, on average. This will be lower for non-native speakers whose accent varies from the selected language option.

While meetings are conversational and unconstrained in nature, you should understand that the Speech to Text Conversion will give best results if you pronounce as clearly as possible and try to speak in sentences with some grammatical structure. Accuracy will also improve over time as the service learns the voices of people in your recordings.

Supported Languages

  • Danish
  • Dutch
  • English (Accent Options: US, UK, AU)
  • French
  • German
  • Italian
  • Spanish (Accent Options: ES, MX or US)
  • Swedish

Pre-Paid Bundles

Available through iTunes In-App Purchase within the Microcone Recorder app or our Account Login.

S

Small
50 conversion credits for $4.99.

M

Medium
125 conversion credits for $9.99.

L

Large
500 conversion credits for $29.99.

Subscription Plans

Available through our Account Login. Corresponding annual subscriptions offer 1 month free.

S

Small
600 conversion credits per month for $25.00.

M

Medium
1500 conversion credits per month for $45.00.

L

Large
5000 conversion credits per month for $79.00.