Review of Text to Speech Systems for Dictation

This will save the next person some time, since most of the programs were unacceptable. I only spent enough time with each one to decide if it would work without being frustrating. The first test was a 100-word sample. They all have free online demos. Please add any other programs you know of.

All research done March 2011 unless otherwise stated. Some of the systems were still under development.

Warning about licenses. At least one online system (didn’t note which) will own the right to do anything they like with any text you send it.

++++

Cepstral SwiftTalker. My favourite. Does everything I want with decent quality vs price vs prep time balance. Easy for speeds over 100 wpm. See http://www.cricket.onebit.ca/Nonfic/DictProg for lower lower speeds using SSML (Speech Synthesis Markup Language). Free works well, but you need to edit out “Please pay” every 90 s or so. Costs $30 for private use to remove the reminder. Several years ago they were willing to quote on an automated shorthand dictation program.

++++

Microsoft Narrator, built into Windows. I couldn’t find how to make sound files. It’s designed for hearing-impaired computer users, not file creation.

++++

++ Paid Programs ++

vozMe – Free version, but sounds terrible.

YakiToMe – Free version, but sounds terrible.

iSpeech – Decent sound, but free is only 1 min. Doesn’t say if it accepts SSML or can change speed.

Ivona – Decent sound. Charges by character.

ZamZar – Not sure why this was listed on the source site. Converts between formats.

imTranslator – Decent sound. No sound files.

++ Open Source, often a research program ++

espeak – horrid sound.

freeTTS – horrid sound.

MARY – demo gave an error page rather than a sound.

Babelfish – not a translator. Lists other sites.

Google – says it uses espeak.

MBROLA – part of a system, not a stand-alone.

Festival – part of a system, not a stand-alone. Demo had horrid sound.

Flite Tools – part of a system, not a stand-alone.

FestVoxTools – part of a system, not a stand-alone.

gnuspeech tools – part of a system, not a stand-alone.

(by
Cricket for everyone)

 

2 comments Add yours

Leave a Reply