I. Overview
1. Introduction chip
LD3320 is based on a non-independent speech recognition: Speech recognition (SI-ASR Speaker-Independent Automatic Speech Recognition) technology / voice chip. It provides a true single-chip speech recognition solutions.
Integrated high-precision A / D and D / A interface LD3320 chip, no need for external auxiliary Flash and RAM, i.e., the voice recognition can be achieved / voice / interactive function. In addition, knowledge of the key words list is dynamically editable.
Based LD3320, at any electronic products, including even the most simple 51 as a master chip system, easy voice recognition / voice / interactive function. Increase VUI (Voice User Interface) voice user interface for all electronic products.
2. Voice Recognition Guide
ASR speech recognition technology is based on a list of key words knowledge of technology. Only you need to set a good list of key words to be recognized, and these key words to the internal LD3320 in the form of characters, you can identify key words spoken by the user. Users do not need to make any recordings training.
ASR technology is the most important practical significance lies in providing a user interface VUI one kind of out of key, keyboard, mouse, voice-based: Voice User Interface
Every time knowledge of the process is the user speech and audio, converted by the frequency spectrum for voice features, and the key word entry in the list one by one match, the best match as a recognition result. For example, in mobile phone applications, the contents of this key word list is the song name in the phone book names / phone menu commands / T card. Whatever this list entry is, the user need only set the relevant registers, can be identified, the corresponding contents to the entry in the character recognition engine form.
LD3320 can know the list of keywords, the user's voice can be said that this list any key words, and does not require any user training prior knowledge. Knowledge engine does not care about the content of key words in the list of key terms, may be a command, names, song names, operating instructions, etc. any string of characters. Each can support the maximum number of words key words, from the perspective of the algorithm is limited to 30 characters. But from a practical point of view, users in one breath said more than eight characters or more entries, there will almost certainly say typo / slip of words / multi-word to say / burp / pause, etc., these will cause serious impact on knowledge and knowledge error. Thus in general, if you want to get a good recognition results, we recommend each word key phrases not too long, to avoid affecting the results.
3. Technical Data
1. Internal mono mono 16-bit A / D analog-digital conversion
2. Built-in two-channel stereo 16-bit D / A digital to analog converter
3. Built-in two-channel headphone amplifier output 20mW
4. Built 550mW mono speaker amplifier outputs
The supporting parallel interface or SPI interface
6. Built phase locked loop PLL, the input master clock frequency is 2MHz - 34MHz
7. Operating voltage: (VDD: for internal core) 3.3V
8. 48pin QFN 7 * 7 of standard package
9. The power saving mode: 1uA
4. scenarios
Cooker / microwave / smart appliances operating
Navigator
MP3 / MP4
Digital Photo Frame
STB / TV remote control
Smart toys / toy dialogue
PMP / gaming machines
vending machine
Subway ticket vending machine
Guide machine
Building television advertising demand
Public lighting systems / health system / intelligent home voice
Two, LD3320 information
1. Pin
