GB1320970A - Method and apparatus for generating speech - Google Patents

Method and apparatus for generating speech

Info

Publication number
GB1320970A
GB1320970A GB6108870A GB6108870A GB1320970A GB 1320970 A GB1320970 A GB 1320970A GB 6108870 A GB6108870 A GB 6108870A GB 6108870 A GB6108870 A GB 6108870A GB 1320970 A GB1320970 A GB 1320970A
Authority
GB
United Kingdom
Prior art keywords
cycle
sound
bits
store
program
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired
Application number
GB6108870A
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Cognitronics Corp
Original Assignee
Cognitronics Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Cognitronics Corp filed Critical Cognitronics Corp
Publication of GB1320970A publication Critical patent/GB1320970A/en
Expired legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/16Sound input; Sound output

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • General Health & Medical Sciences (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Signal Processing Not Specific To The Method Of Recording And Reproducing (AREA)

Abstract

1320970 Speech synthesizers COGNITRONICS CORP 23 Dec 1970 [23 Dec 1969] 61088/70 Heading H4R In a speech synthesizer a number of characteristic sound cycles of voice sounds are stored, selected sequentially, and each reproduced a number of times to synthesize speech. As described with respect to Figs. 2A and 2B the store 10 comprises a magnetic drum on which information is recorded digitally. Two tracks on the drum are used for timing the operation of a control signal generator and providing basic timing for the synthesizer. Four tracks store program information in respect of each word which the synthesizer is to reproduce. Each word which is to be reproduced by the synthesizer is broken down into a number of elemental sounds each comprising a pitch period cycle which is reproduced a predetermined number of times to synthesize the elemental sound. The pitch period cycles are stored in sectors of 40 sound tracks 14 on the drum 10, each sector comprising 256 bits to define the cycle waveform, and there being 8 sectors on each track. The elemental sounds are reproduced from the sound tracks under the control of the program tracks 12, of which there are four, each track having 128 sectors, each defining an elemental sound, each sector being 16 bits long. Each program sector includes 3 bits defining the sector and 5 bits defining the track of the pitch period cycle to be reproduced for the elemental sound together with 2 bits specifying the number of times the sound is to be reproduced, three bits to define the gap to be left between each pair of repetitions of the cycle, to govern the pitch of the reproduced sound, and a further 2 bits to control the envelope of the reproduced waveform. In operation a computer in response to an optical reader, voiced sound input, or punched cards etc., produces from a store information as to the phonetic equivalents of the words to be synthesized, and program instructions as to their characteristic sound cycle content, in the form of program locations expressed as 7 bits defining the position on a program track defined in a further two bits. The 9 bit program location is fed into a store 68, 70 where the last two bits effect the selection of the appropriate track while the first 7 bits are compared with a program sector counter, controlled by the timing tracks on the drum, to cause read out of the appropriate program sector of 16 bits into the store 42. At the appropriate time, under the control of the control signal generator 40, the information in respect of the particular sound cycle to be reproduced next is fed from the store 42 into the store 78 to be compared with the state of a sound sector counter 86, controlled by the drum timing tracks 20, 22; to effect the transfer of the appropriate sound cycle sector from the sound tracks 40 into a recirculating memory track 16 on the drum, to provide a fast access store. Simultaneously the number of repeats of the pitch cycle is fed from store 42 to the cycle repetition counter 138 the pitch information is fed to pitch store 150 and envelope information to a buffer 172. The read head of memory track 16 is spaced # of a drum revolution from the write head and during synthesis the read head output is fed back to the write head so that the sound cycle information is available every #th drum revolution. For reading out on synthesis the speech cycle in memory 16 is fed into the memory 18 and the shift register 112. The first 32 bits of the cycle are entered into the shift register and the remaining 224 bits into the recirculating memory 18, having read and write heads spaced by the 224 bit interval. Four bit groups, from part 116 of the shift register, representing amplitude samples of the cycle, are fed to a converter 120 and low pass filter 124 to provide analogue output signals. Read out of 32 bits from the register 112 requires the time taken for recirculation of 224 bits in memory 18 so that the complete cycle can be read out continuously through register 112 after the read out of one complete sound cycle, cycle repetition counter 138 is decremented by one and the cycle repeated until counter 138 has been decremented to zero, when the cycle being reproduced will be replaced by the succeeding sound cycle called for by the program. After each successive repeat of a cycle a number of #th drum revolution gaps is left, to enable variation of the pitch of the reproduced sound. The number of #th drum revolutions is determined by the pitch signal stored in location 56 of the store 42 which is fed to the pitch store 150 and compared with the state of a counter 156 counting #th drum revolutions after the end of a sound cycle, and a repeat only being enabled after coincidence is established between the two counters. In addition for natural speech the amplitude of the last few sound cycles before the end of a word may be progressively reduced, this is provided by the AVC circuit 126 controlled by a ramp generator 168 whose operation is initiated through the control signal generator in accordance with the program set into location 58 of the program store 42.
GB6108870A 1969-12-23 1970-12-23 Method and apparatus for generating speech Expired GB1320970A (en)

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
US88749169A 1969-12-23 1969-12-23

Publications (1)

Publication Number Publication Date
GB1320970A true GB1320970A (en) 1973-06-20

Family

ID=25391261

Family Applications (1)

Application Number Title Priority Date Filing Date
GB6108870A Expired GB1320970A (en) 1969-12-23 1970-12-23 Method and apparatus for generating speech

Country Status (1)

Country Link
GB (1) GB1320970A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0307137A2 (en) * 1987-09-11 1989-03-15 Hashimoto Corporation Multiple language telephone answering machine

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0307137A2 (en) * 1987-09-11 1989-03-15 Hashimoto Corporation Multiple language telephone answering machine
EP0307137A3 (en) * 1987-09-11 1989-09-06 Hashimoto Corporation Multiple language telephone answering machine

Similar Documents

Publication Publication Date Title
US3755608A (en) Apparatus and method for selectively alterable voicing in an electrical instrument
US3575555A (en) Speech synthesizer providing smooth transistion between adjacent phonemes
GB1279459A (en) Information storage and retrieval
US3928722A (en) Audio message generating apparatus used for query-reply system
GB1072447A (en) Digital to analogue converter
GB1289999A (en)
CA1087748A (en) Time-compression system
US4356517A (en) PCM Audio reproducing system
US3532821A (en) Speech synthesizer
GB1320970A (en) Method and apparatus for generating speech
JPS5864622A (en) Data reproducing device
US3001180A (en) Data revolving
CA1057849A (en) Data read-write apparatus for a magnetic recording medium
US4134139A (en) Method and arrangement for writing and reading bit sequences
JPS5774809A (en) Voice signal processing device
JPS5886600A (en) Time base converter
GB1408818A (en) Magnetic recording verification
FR2127592A5 (en)
FR2279290A1 (en) Television display of musical notation - by binary signals on magnetic tape and shift register memory
SU417965A3 (en)
SU447760A1 (en) Device for reproducing information
GB857284A (en) Improvements in magnetic recording apparatus
SU767827A1 (en) Device for playing-back data from magnetic tape
SU1094050A1 (en) Device for reproducing magnetic record
SU723672A1 (en) Device for marking and retrieval information on magnetic tape

Legal Events

Date Code Title Description
PS Patent sealed
PCNP Patent ceased through non-payment of renewal fee