SkryBot S2T (speech to text) is a software for recognition and decoding of speech to text installed on customers’ servers.

Speech recognition is the basis of new attractive services offered by multimedia content providers – radio, TV, internet:

• search engines for searching in audio and video archives,

• publication of the decoded content on websites.

SkryBot S2T speech recognition software package comprises:

• 1. A speech decoder integrated with the acoustic and language model, continuous speech recognition with the use of LVCSR (Large Vocabulary Continuous Speech Recognition) dictionary.
  2. A package of tools that automate the speech recognition process and the decoding of audio files to text.
  3. SkrySearch 1.0 (no sound editing)

Speech recognition and decoding software:

SkryBot S2T one licence per server, 1-2 processors, price: 20 000 PLN* (one-year update included) updating in the following year: 50% of the licence price

Recommended hardware:

Dual-processor servers 2Ghz, 2×4 cores, 8GB RAM (INTEL or AMD)

Tested performance:

1. Dual-processor server:
fast mode: 3:1 (3 hours of audio in 1 hour of server work) accurate mode: 1:1 (1 hour of audio in 1 hour of server work)
2. Two dual-processor servers:
fast mode 8:1 (8 hours of audio in 1 hour of server work) accurate mode: 2:1 (2 hours of audio in 1 hour of server work)

It is possible to increase performance by increasing the amount of computational resources. Performance may vary due to constant software updates. * Net prices.