CereProc Tagset

CereProc CereVoice provides native markup tags.

usel variant

The usel tag variant attribute allows the user to request a different version of the synthesis for a highlighted region. This is a very useful tag that can be used to make sections of speech sound more appropriate. The variant number can be increased to produce more and more different versions of the speech. The original version is equivalent to variant 0.

The two price cuts will together save UK households an average £132 on their gas and <usel variant='1'>electricity</usel> bills.

usel speed

The usle tag speed attribute biases the selection of units to lower or higher speeds. For example, <usel speed='2.0'> doubles the duration target of a word or phrase and thus causes longer units to be selected where available.

There has been a <usel speed='1.6'>further</usel> fall in the number of crimes in Scotland, according to the latest figures.

usel pitch

The usel tag pitch attribute biases the selection of units to lower or higher pitch. For example, <usel pitch='0.5'> lowers the target pitch of a word or phrase and thus causes lower-pitched units to be selected where available.

The deal <usel pitch='0.5'>will</usel> result in a stock market listing for the bank.

spurt audio

The spurt tag audio attribute selects a vocal gesture to included in the synthesis output for non-speech sounds such as laughter and coughing.

<spurt audio="g0001_004">cough</spurt>, excuse me, <spurt audio="g0001_018">err</spurt>, hello.

lexicon entry

The lexicon tag entry attribute option allows 'function words' such as 'and', 'the', and 'to' to be switched to their 'full form' or 'reduced form'.

In a deal announced on the second day of <lexicon entry='full'>the</lexicon> Paris Air Show.

The company cut its gas prices, also by 10%, on 19 February this year, benefitting more <lexicon entry='reduced'>than</lexicon> seven million households.

lex phonemes

The lex tag phonemes attribute allows the user to define and modify the phonemes used to render a selected word. Each language and accent has its individual phoneset, so the user should use the correct phoneset for the current voice.

<lex phonemes='ii0 n ii0 e1 s t @0'>Iniesta</lex>beat <lex phonemes='p e1 t @0 r'>Petr</lex><lex phonemes='ch
e0 x'>Cech</lex> from the edge of the<usel variant='1'>area</usel>to send Barcelona into the final on away goals.</lex>

sig f0

The signal tag f0 attribute may be used to raise and lower the pitch of the speech output. The pitch value is converted based on the average realised pitch of the speech within the region of highlighted text. The pitch can be changed in three ways: relative, absolute or multiplier. To change the pitch by an exact amount, for example 50Hz : '+50.0' will increase it by 50, '-50.0' will increase it by 50; if you want it to be an absolute value, for example 200.0Hz, enter that fixed value preceded by ?a?: 'a200.0'. To multiply the pitch by a fixed amount, simply enter a multiplier, for example to double it enter '2.0', to halve it enter '0.5'.

But like British Gas they all raised their gas and electricity prices <sig f0='+10'>twice</sig> last year.

But like British Gas they all raised their gas and electricity prices <sig f0='a220'>twice</sig> last year.

But like British Gas they all raised their gas and electricity prices <sig f0='1.1'>twice</sig> last year.

sig rate

The signal tag rate attribute may be used to increase and decrease the speed of the speech output. The default rate is <sig rate='1.0'> and it may be varied on a range between 0.5 and 2.0 by changing the value between the quotes. The speech rate will then be multiplied by that factor.

British Gas are the last of six major energy suppliers to cut their electricity prices <sig rate='0.8'>this year</sig>

British Gas are the last of six major energy suppliers to cut their electricity prices <sig rate='1.2'>this year</sig>

sig amplitude

The signal tag amplitude attribute may be used to increase and decrease the amplitude of the speech output. The default amplitude is <sig amplitude='1.0'> and may be varied on a range between 0 and 50. The amplitude will then be multiplied by that factor.

The cut was made possible by a fall in the cost of <sig amplitude='2.5'>wholesale electricity.</sig>

The cut was made possible by a fall in the cost of <sig amplitude='0.5'>wholesale electricity.</sig>

break type='4'

The sentence break tag may be used to enforce a sentence break or to insert a long (by default, 0.6s) pause into the text. The duration of the pause can be modified by changing the value time='0.6', where 0.6 is the length of the break in seconds.

Some are warning the cuts could be temporary <break type='4' time='0.6' />the cost of wholesale electricity is expected to rise again next year.

break type='3'

The phrase break tag may be used to enforce a phrase break, by default, 0.3s. The duration of the pause can be modified by changing the value time='0.3', where 0.3 is the length of the break in seconds.

Some are warning the cuts could be temporary <break type='3' time='0.3' />the cost of wholesale electricity is expected to rise again next year.

break type='2'

The short break tag may be used to insert a short pause into the text, which can often be useful for improving clarity around important items in a sentence.

The Bank of England has kept interest rates on hold at 0.5% and announced that it will pump an extra <break type='2' time='0.1'/> 50bn<break type='2' time='0.1' /> into the UK economy.

voice

The Emotion tags: happy, sad, calm, and cross, can be applied to word(s) to to vary the emotion of the speech as follows: Note that some applications may remove the <voice> tag (i.e. Voice XML browsers, SAPI) - they should use the <speechstyle> tag instead.

It's <voice emotion='happy'>beautiful</voice> today.

The outbreak<voice emotion='sad'>cast a shadow</voice> over the former Victorian holiday resort.

The beautiful gardens which have been restored in all their <voice emotion='calm'>eccentric Victorian splendour.</voice>

When people leave a tip for staff, in a restaurant or anywhere else, they have a right to know that it will<voice emotion='cross'> not be used</voice> to make up the minimum wage.

Source: CereProc Tagset