This article refers to the address: http://
Summary  The integrated real-time accent processing chip SD771D is a single-chip voice processor introduced by Taiwan Xiangyin Technology Co., Ltd. The chip can transpose the output speech (such as raising and lowering), and can also convert the voice of male and female voices to each other. The paper introduces the working principle of the chip, and gives the hardware interface circuit of the typical application of SD771D .
Key words  Real-time accent  SD771D  Typical application circuit
1 Â Overview
Speech technology include speech recognition (speech recognition), voice synthesis (speech synthesis) forwarding sound and text TTS (Text-To-Speech) . At present, speech synthesis technology is in a mature development period, both in research and in application, and some have already been put into practical application. Taking TTS technology as an example, this technology and related products have been promoted by many manufacturers, such as the software-only TTS processing software package (ie, SAPI SDK software package) introduced by Microsoft ; the voice processing chip WTS70l and ISD launched by Winbond Electronics Series of voice processing products, etc.
SD771D is a single-chip real-time accented processor introduced by Taiwan Xiangyin Technology Co., Ltd., which uses real human voice as a sample, which can transform the output speech in real time (such as raising and lowering), and can also interpret the voices of male and female voices in real time. Conversion. It can be used to create a variety of virtual characters, create a changing audio environment, and better entertainment.
2 Â SD771 features and working principle
2 . 1 Â SD771 features
â—‡ Built-in microphone amplifier circuit;
◇ Built-in 32 Ω speaker drive circuit;
â—‡ Built-in digital volume control ( 15 levels ) circuit;
â—‡You can select the digital or analog input mode to change the pitch;
â—‡The digital input mode is 15 tone changes;
â—‡The analog input mode is 180 tone changes;
◇ Built-in 16 -bit △ - ∑ A / D and D / A converter;
â—‡ Working voltage range is DC+3.6~+6.0V ;
â—‡ Working current is 25mA ;
You can immediately convert the speaker's voice into a different tone output (eg, the pitch becomes higher and lower, and the male and female voices are converted).
Figure 1 shows the internal structure of the S13771D . The BPF ( Band Pass Filter ) and LPF ( Low Pass Fiilter ) are anti-noise digital filters. In Figure 1 , GAIN_TNC is the volume up button, GAIN_DEC is the volume down button, PITCH_INC is the pitch up button, PITCH_DEC pitch down button, ORIC is the restore tone button, and RESET is the reset button. All of the buttons described above are active low. The duration of each key press must be greater than 10ms . The interval between two different button presses must be greater than 60ms , as shown in Figure 2 . PITCH_VR is the input of the analog mode change tone, and XI and XO are used to connect the input of the system crystal circuit.
2 . 2 Â SD771D working principle and typical application circuit
( 1 ) Change the level of sound tuning
The sound input from the microphone is sent by the DAC after being processed by the SD771D . The SD771D simultaneously sends two signals, one for the digital signal and the other for the analog signal, and used to change the level of the sound tuning. When the level of PITCH - VR is less than 0 . At 1V , the digital signal channel is turned on and the analog signal channel is turned off. The regulation of the sound signal level of the digital signal channel and the analog signal channel is independent of each other.
( 2 ) Digital signal channel
The digital signal path is controlled by the PITCH — INC tone up button or PITCH_DEC (tone down button). When the level of PITCH_VR is less than 0 . At 1V , the level of sound tuning is controlled by PITCH_INC or PITCH_DEC . At this point SD771D repeatedly check PITCH_INC and PITCH - The state DEC. When one of the two buttons has a low-level pulse, the SD771D will immediately change the level of the sound tuning. The level of sound tuning has 7 levels of upshift and downshift , including the normal tone level when no key is pressed, a total of 15 levels, as listed in Table 1 .
( 3 ) Analog signal channel
The SD771D has an 8 -bit analog-to-digital converter ( ADC ) inside . The input analog signal voltage is usually controlled by the PITCH_VR terminal for sound leveling. If the voltage on PITCH_VR is at zero . 3 ~ 2 . Between 2V , then the value of this voltage range will be divided into 180 levels, and each level corresponds to a tone ratio.
Figure 3 shows a typical application circuit for the SD771D .
Knot  language
The SD771D can transpose the output speech in real time (such as raising and lowering), and process the transposition technology according to the level of sound tuning. It can also convert the voices of male and female voices into each other in real time.
Bolted Silo,Grain Dryer Co., Ltd. , http://www.nsgraindryer.com