Wizzard Software Corporation > WizzScribe Speaker Independent for Linux, WizzScribe Transcription for Windows

21st century business has discovered a wide spectrum of reasons for converting voice to text… including call analysis, data-mining, archival, text messaging and a seemingly endless list of new opportunities.

Wizzard’s offering is designed to be simple and easy. Fundamentally, it's an internet based subscription service at a nominal monthly fee, with low usage charges based on a “per audio minute” pricing schedule that varies depending on the clients required “turnaround time”. The technology used is speaker independent with telephony acoustic modeling, thus enabling state of the art recognition accuracy that is good enough (dependent on audio quality) for many emerging uses of text converted voice audio. The flow is just as simple: Voice audio is sent to the Voice2Text interface, processed on secure Wizzard hosted hardware/software, and returned to the client based on the required turnaround objectives. Following the return of the text to the client by Wizzard V2T, all records are deleted at Wizzard.

Input Data Formats that Wizzard supports:

	Audio:
		WAVE PCM (*.wav)

		MPEG (*.mp3)

		WMA (*.wma)

		AAC (*.aac)

Video:

MPEG (*.mp4, *.mov, *.m4a, *.m4b, *.m4v)

AVI (*.avi)

WMV (*.wmv)

FLASH (*.flv)

REAL (*.ra, *.rv)

Output Data formats that Wizzard supports:

		Raw text (words separated by space, sentences separated by<PAUSE> and new line).

		Example: Hello friend. How are you

		Normalized text (words are case-formatted and separated by space, sentences separated by space).

		Example: HELLO FRIEND <PAUSE> HOW ARE YOU

		Time-word table (words have associated time stamps, words separated by new line).

		Example: 000000.72 HELLO 000001.21 FRIEND 000001.72 <PAUSE> 000002.15 HOW 000002.31 ARE 000002.69 YOU

Turn-Around Time

Wizzard will make every effort to meet or beat the Turn-Around Time criteria specified in the Subscription Agreement Turn around time is measured starting when the entire file has been received by Wizzard and ending when the output file is sent to Client.

Accuracy

Wizzard is using state of the art conversational speech technology as the backbone of this service, but results will always be dependent on the quality of the audio. Specialty language models are not supported at this time. We invite prospective clients to talk with Wizzard Sales representatives about accuracy expectations. Our Sales team can setup a “dry run” for clients using their representative voice audio files. In this manner, prospective clients can preview results and evaluate adequacy for their individual needs.

Subscription Agreement

To review the base Wizzard Voice2Text Services Agreement click “agreement”. For inquiries, send click “inquire” or call Wizzard Sales at 954-678-4155.