RSC-364QFP中文资料

RSC-300/364Recognition••Synthesis••ControlFromtheInteractiveSpeech™LineofProductsSpeechRecognitionMicrocontrollerGENERALDESCRIPTIONTheRSC-300/364,fromtheInteractiveSpeech™familyofproducts,isan8-bitmicrocontrollerdesignedspecificallyforspeechapplicationsinconsumerelectronicproducts.TheRSC-300/364isasinglechipsolutionthatcombinestheflexibilityofamicrocontrollerwithadvancedspeechtechnology,includinghigh-qualityspeechrecognition,speechandmusicsynthesis,speakerverification,andvoicerecordandplayback.ProductscanuseoneoralloftheRSC-300/364featuresinasingleapplication.TheRSC-300/364supportsSensorySpeech™5.0,thelatestspeechrecognitiontechnologyfromSensory,whichincludesanumberofnewtechniquesthatsignificantlyimproverecognitionperformanceoverpreviousversions.Usingasophisticatedneuralnetworktechnology,on-chipspeechrecognitionalgorithmsreachanaccuracyofgreaterthan97%forspeaker-independentrecognitionandgreaterthan99%forspeaker-dependentrecognition.Inadditiontotheimprovedrecognitionperformance,theRSC-300/364providesfurtheron-chipintegrationoffeatures,includingapreamplifier,multiplier,watchdogtimer,and2.5KbytesofRAM.Acompletesystemmaybebuiltwithfewadditionalpartsotherthanabattery,speaker,microphone,andafewresistorsandcapacitors.TheRSC-300isdesignedforROM-lessforapplicationsthatneedmoreROMspaceandconsequentlyuseoff-chipmemory.FEATURESFullRangeofSensorySpeech™5.0Capabilities•Speaker-independentspeechrecognition•Speaker-dependentspeechrecognition•Highqualityspeechsynthesisandsoundeffects•Speakerverification•Four-voicemusicsynthesis•Voicerecord&playbackIntegratedSingle-ChipSolution•4MIPS8-bitmicrocontroller•On-chipA/DandD/Aconverters,andpre-amplifier•32kHzclockfortimekeeping•Internal64KbytesROM;2.5KbytesRAM•Internal32kHzwatchdogtimer•Externalmemorybus:16-bitAddress,8-bitData•24x24MultiplierforrapidrecognitionprocessingLowPowerRequirements•2.4–5.25Voperationfor2or3batteryapplications•~10mAoperatingcurrentat3V•Powerdownmode;5μAstandbycurrentRSC-300/364BlockDiagramADCMultiplexerDigitalLogicAGCMicrocontrollerPreampandGainControlRAMROM(RSC-364only)RSC-364MicrophoneSpeakerGeneralPurposeI/OExternalMemoryOscillatorAMPDACMultiplierWatchdogTimerRSC-300/364DATASHEET2FromtheInteractiveSpeech™LineofProductsRSC-300/364OVERVIEWTheRSC-300/364isamemberoftheInteractiveSpeech™lineofproductsfromSensory.Itfeaturesahigh-performance8-bitmicrocontrollerwithon-chipA/D,D/A,preamplifier,RAMandROM(RSC-364only).TheRSC-300/364isdesignedtobringahighdegreeofintegrationandversatilityintolow-cost,power-sensitivetoyapplications.VariousfunctionalunitshavebeenintegratedontotheCPUcoreinordertoreducetotalsystemcostandincreasesystemreliabilitywithoutdegradingsystemperformance.TheRSC-300/364delivers4MIPSofintegerperformanceat14.32MHzprovidingmaximumperformanceatminimumcost.TheCPUcoreembeddedintheRSC-300/364isan8-bit,variable-length-instruction,microcontroller.TheinstructionsetissomewhatsimilartotheZilogTM78,andhasavarietyofaddressingmodemovinstructions.TheRSC-300/364processoravoidsthelimitationsofdedicatedA,B,andDPTRregistersbyhavingcompletelysymmetricalsourceanddestinationsforallinstructions.Ofthe2.5KbytesofinternalRAM,2KbytesareorganizedasaDataSpace,with0.5KusedforRegisterSpace.SPEECHRECOGNITIONTheRSC-300/364usesaneuralnetworktoperformspeaker-independentorspeaker-dependentspeechrecognition.Speaker-dependentrecognitionrequiresexternalmemorytostorespeechrecognitioninformation(e.g.,SRAM,optionalSerialEEPROM,FlashMemory).Speaker-independentrecognitionrequireson-chiporoff-chipROMtostorethewordstoberecognized.TheRSC-300/364hasseveraladditionalspeechrecognitionfeaturesasdescribedbelow.Continuouslisteningallowsthechiptocontinuouslylistenforaspecificword.Withthisfeatureaproductcanbeusedinanormalenvironmentandonly“activates”whenaspecificword,precededbyquiet,isspoken.SPEECHANDMUSICSYNTHESISTheRSC-300/364provideshigh-qualityspeechsynthesisbyusingahybridofatime-domaincompressionschemethatimprovesonconventionalADPCMandacustomizedreuseofsounds.Speechsynthesisrequireson-chiporoff-chipROMtostoreaudiosoundsforsynthesis.TheRSC-300/364provideshigh-quality,low-costfour-voicemusicsynthesiswhichallowsmultiple,simultaneousinstrumentsforharmonizing.TheRSC-300/364usesaMIDI-likesystemtogeneratemusic.RECORDANDPLAYBACKTheRSC-300/364canperformaudiorecordandplaybackatvariouscompressionlevelsdependingonthequantityandqualityofplaybackdesired.Dataratesofunder14,000bitspersecondareachievablewhilemaintainingveryhighqualityreproduction.TheRSC-300/364alsoperformssilenceremovaltoimprovesoundqualityandreducememoryrequirements.SPEAKERVERIFICATIONTheRSC-300/364canalsoperformtext-dependentspeakerverification.Afteraspeakertrainsthechiponaspecificword,thechipisabletoidentifywhetherthatwordisspokenbytheoriginalspeaker,thusprovidingbiometricsecurity.POWERThetypicaloperatingcurrentis10mAoperatingat14.32MHzand3V.Loweringclockfrequencyreducespowerconsumption,althoughspeechrecognitionrequiresa14.32MHzclock.Standbycurrentis5μAinpowerdownmode.RSC-300/364ArchitectureDiagramPRE-AMPEXTERNALMEMORYINTERFACECPUTIMINGANDCON