Orpheus is the multilingual text to speech synthesiser from Meridian One
for Laptop, Notebook and Desktop computers running Microsoft Windows Windows 7,
Vista or Microsoft Windows XP. Orpheus is available as Orpheus TTS Plus or
Orpheus TTS plus and Orpheus TTS speak 25 languages with synthetic
voices capable of high intelligibility at the fastest talking rates.
Orpheus TTS Plus adds natural sounding voices for UK English,
US English and Swedish.
Orpheus TTS Plus and Orpheus TTS are compatible with Microsoft
SAPI 4, SAPI 5 and Dolphin SAM.
Download Orpheus for a free 30-day full evaluation
Benefits of Orpheus
Orpheus is renowned for the intelligibility of its synthetic voices when
speaking at the fastest talking rates, a benefit favoured by many regular
users of TTS, for instance blind and partially sighted users of screen readers.
When we speak quickly we shorten the speech sounds that we make and try to speed up the
transitions between them. How much we shorten each sound depends on the nature of the
sound and its context, or position, in the speech. However, the maximum speed of transition
between the speech sounds is restricted by the speed that we can move the articulators, our jaw,
tongue, lips and velum. Eventually, as the speech gets even faster, we are unable to speed up
the transitions any further. If we keep trying to speak faster, because we can't move any faster,
then eventually we fail to reach the target positions for the articulators needed for each sound.
The speech then becomes less intelligible. It's a bit like someone who has had a little too much
to drink and so is unable to move their articulators at a normal speed and their speech becomes
So, how does Orpheus maintain intelligibility at the fastest speaking rates?
Orpheus has design features at all levels to do this. For instance:
- Synthesizers like Orpheus do not have a physical restraint on the speed of transition
between the speech sounds. They have no articulators to move. So, as well as speeding up the sounds
by making them shorter, Orpheus speeds the transitions at unnaturally fast talking rates.
- The Orpheus formant synthesizer has features, special to it, that prevent the production
of unnatural spurious 'transient' sounds that are not part of the speech at all, but that can
occur during the synthesis of very fast transitions.
- Other features like the skim reading function allow fast talking voices to read out
just the important words in a phrase.
You can read a summary of the
Fast Speech paper presented at CSun 1999. Fast clear speech is a real practical benefit
for users of TTS in the workplace and in education.
Orpheus TTS Plus is unique in that it can speak voices made in two
different methods and intended for two different scenarios.
It speaks using formant synthesis voices,
which model the human voice production, and
concatenative voices, which join up and
manipulate fragments, or units, of speech taken from real human speech and
stored in a database of sounds.
Orpheus synthetic voices are great for fast-talking workaday applications
where speed and intelligibility matter. Orpheus TTS Plus concatenative voices
offer a much more natural sounding complement for applications where
naturalness is the major requirement.
Fast Voice Switching
Orpheus is designed to switch its voices really fast. This is a major requirement
for working well with a screen reader. Using the Dolphin SAM driver you can instantaneously
switch to any voice or language, even from a synthetic to natural sounding voice or vice versa.
Orpheus responds to the fast voice changes needed for screen reading to support users who
need one voice, or language, for a document's text and another for their screen reader's commands.
Both Orpheus TTS Plus and Orpheus TTS offer unique and special benefits. You can create a
skim reading voice to skip through text speaking only the most important words. You can even
set the level of importance of words that will be spoken.
Orpheus provides the correct pronunciation of many heteronyms - words spelt
the same but pronounced differently - depending on their context.
Orpheus TTS uses a small memory footprint; it does not include the databases
for the natural sounding voices.
The languages supported are:
- Mandarin Chinese (beta)
- Cantonese Chinese (beta)
- US English
- UK English
- Brazilian Portuguese
- European Portuguese
- Latin American Spanish
- European Spanish
Voices for Hindi are currently being added to the portfolio and is
expected by late 2009.
British English, American English and Swedish are also available with
natural sounding concatenative voices in Orpheus TTS Plus.
Orpheus TTS Plus and Orpheus TTS are compatible with Microsoft SAPI 4,
Microsoft SAPI 5 and Dolphin SAM applications.
Dolphin SAM, available from Dolphin Computer Access
is a great way of driving a TTS engine for screen reading and allows superb control
of TTS voices.
Microsoft SAPI 4 and SAPI 5 are the established interfaces for TTS for
Microsoft Windows and are compatible with most talking applications.
Meridian One is grateful for all help and advice received in the development of Orpheus now
and in the past from members of the speech community and from the many blind and visually impaired
users of Orpheus.
Meridian One wishes to acknowledge Mark Huckvale and the department of phonetics at UCL for their
permission to use SFS tools in voice creation for