Fundamentals of Product Development¶
The development of electronic product solutions primarily involves hardware and software development. Hardware development typically refers to designing circuit boards for the product using chips and other electronic components. Software development involves developing code based on the type of main control chip on the circuit board and the product’s functional requirements. This code is then compiled into firmware that can be downloaded into the main control chip, working with the chip and other hardware to achieve specific product functions. The main steps for developing a new product solution include hardware selection, hardware development, firmware development, and product solution testing. After addressing software and hardware issues identified during testing, the development is completed, followed by subsequent production and manufacturing processes as needed.
1. Basic Concepts¶
This section discusses some terminology and concepts involved in developing product solutions using our company’s chips, as follows:
1.1 Intelligent Speech Chip¶
A chip, also known as an integrated circuit (IC), is a device that miniaturizes circuits and some electronic components, manufactured on the surface of a semiconductor wafer through photolithography machine, and processed to form a device that can be directly soldered and used. Using chips in hardware solutions can significantly reduce the area of the circuit board and the complexity of the design, helping to speed up product development and reduce development difficulty, while also saving the overall cost of the solution. Therefore, a large number of chips have been applied to almost all fields of electronic products, including home appliances, automobiles, consumer electronics, industrial control, etc.
An intelligent voice chip is a type of chip that provides the necessary computing power through neural network technology on the chip side, allowing voice recognition and other voice-related functions to be processed directly at the terminal devices without relying on network servers. Intelligent voice products directly using intelligent voice chips can greatly enhance the intelligent voice experience, improve development efficiency, and reduce the overall cost of the solution. Therefore, more and more intelligent devices are using professional intelligent voice chips to develop products.
1.2 Voice Recognition Process of Intelligent Speech Chip¶
Typically, when an intelligent speech chip performs speech recognition, the process begins with an external microphone capturing the acoustic signal and converting it into an electrical signal. This signal is then sampled by the chip’s built-in Analog-to-Digital Converter (ADC), transforming it into a digital signal. The digitized audio is passed to the chip’s internal voice processing module, where recognition algorithms and control logic are executed. After processing, the voice data is decoded by software, resulting in a text-based output that represents the recognized speech.
Users can click Learn about ASR Principles for more related knowledge.
1.3 Intelligent Voice Module¶
An intelligent voice module integrates the intelligent speech chip and other passive components onto a miniaturized PCB board, which can connect to external devices like microphones and speakers. The module can download the corresponding functional firmware to form a hardware circuit board with basic voice functions.
If users directly use a pre-designed intelligent voice module in product development, they do not need to learn the hardware knowledge required for designing circuits with intelligent voice chips, nor do they need to design the voice circuit part of the product, thus avoiding hardware board issues caused by improper handling of voice-related signal processing. This can also accelerate the hardware development of the product. Therefore, for new users, it is recommended to directly use our company’s existing standard voice modules for product solution design.
1.4 Firmware¶
Firmware generally refers to the binary file generated by the compiler and other tools after the user has developed the functional software, which can be directly downloaded into the master chip. This binary file mainly consists of machine code instructions that the master chip can recognize. After the master chip is powered on, it executes the instructions within the firmware to achieve the software functions desired by the user.
In the development of solutions for our company’s intelligent voice chips, the firmware corresponding to our chips mainly consists of the software control part, acoustic model, language model, and voice prompt (or voice announcement).
The software control part includes the operating system and the instruction sets used to manage peripheral functions of the chip, such as task scheduling, UART communication, and I/O control.
The acoustic model and language model are files required for voice recognition, which users can obtain through our AI development platform.
The voice prompt refers to the sound played after a voice command is executed, providing users with confirmation that the instruction has been processed.
In actual product development, since our company’s intelligent speech chip may serve either as a master chip or a slave chip, the firmware is accordingly divided into: master firmware and slave firmware, as described below:
- Master firmware: When the intelligent speech chip is used as the master of the product, it can control the product’s functions based on the recognized voice commands while performing voice recognition and voice prompt.
- Slave firmware: When the intelligent speech chip is used as a slave chip, it communicates with the master chip (MCU), uploads the voice recognition results to the master chip, and can also accept the master chip’s instructions for voice prompts.
2. Hardware Selection¶
In hardware selection, the choice involves either chips or modules. If users intend to design their own voice-related circuits, the focus is on selecting the appropriate chip. Conversely, if users only design peripheral circuits while directly using our company’s modules for voice-related circuits, the emphasis is on selecting the right module. The following explanations use our company’s products as examples:
2.1 Chip Selection¶
Users can organize the hardware specifications corresponding to the product based on the requirements in the application scenario (such as product type, number of voice commands, number of voice prompts, noise environment selection, Direction-of-Arrival (DOA) on/off, Offline Natural speech interaction on/off, serve as master chip or slave chip, etc). This includes determining the number of microphones needed, the required memory size, and the number of control IOs. Then refer to the specifications of our company’s chips in the Hardware Selection Guide to choose the appropriate chip for design.
If you do not know how to choose, we recommend first checking the relevant documents in Product Solution Development to obtain more information to help choose the appropriate chip. If it is still not clear, please contact our company for assistance through the following contact information.
Contact phone: 86-28-61375925
Contact email: support@Chipintelli.com
2.2 Voice Module Selection¶
After sorting out all requirements in the application environment and hardware specifications, please refer to the specifications of our company’s modules in the Hardware Selection Guide to choose the appropriate module for design.
If you do not know how to choose, please first check the relevant documents in Product Solution Development to obtain more information to help choose, or contact our company for assistance through the following contact information.
Contact phone: 86-28-61375925
Contact email: support@Chipintelli.com
2.3 Obtaining Samples¶
Our company provides purchase channels for chip and module samples. After choosing the appropriate chip or module, then click Sample Purchase to obtain samples of chips or modules.
If you are new to this field, it is recommended to directly use our company’s Gen 3 voice chips such as CI1302 or CI-D02GS02S modules for product development. Our company also provides a evaluation kit CI-D06GT01D for fast development. Users can purchase through the above link.
3. Hardware Development¶
After completing hardware selection, users can obtain hardware reference documents, including schematics and PCBs, from our company’s Chipintelli AI Speech Development Platform, and refer to the design methods in these documents for hardware board-level design.
Users can also refer to the Hardware Design Reference to obtain essential design knowledge. This resource helps minimize errors in self-design, enhances development effectiveness, and speeds up development processes.
4. Firmware Development¶
Our company provides a software SDK development package and corresponding tools based on intelligent voice chips. For detailed content, please refer to the SDK section in the document center. New users is suggested first register an account on our AI Speech Development Platform (AI Platform) Click to enter the Chipintelli AI Speech Development Platform, and then obtain the SDK and tools corresponding to the selected chip or module from the platform. We recommend new users first use our company’s development board or standard modules, directly compile the SDK provided by our company, and use the tools to produce firmware, completing the entire development process. Click to view an example of using the platform to create firmware and download for the CI13XX series chips. The document center contains the specifications of the selected module, which includes hardware connection and handling methods for module upgrades, and users can operate according to this document.
For new users, our company recommends directly using the firmware creation function of the AI platform to directly generate downloadable firmware. This can avoid using IDE tools for code development, helping to quickly get started. For this step, please click the Product Development Process Guidance document and follow the corresponding operation methods in that document.
For users familiar with our development process who need to individually synthesize firmware components and package them into firmware, while also engaging in some code development, the appropriate SDK can be downloaded based on the selected chip. The SDK development instructions can be referenced from:
5. Product Testing and Feedback¶
After successfully flashed the firmware onto the hardware board, the hardware can be powered on. Users can then speak wake-up words or command words to confirm and test the recognition effect. The testing methods can be referred to in the relevant documents in Product Testing.
If users encounter poorly recognized command words during testing, they can compress (zip/rar) the audio file and submit them along with the test results through the AI platform’s support ticket system, or email to support@Chipintelli.com, etc. We will quickly respond after analysis.
We wish you a pleasant journey in designing excellent product solutions using our company’s products, platform and tools. We will provide comprehensive technical support to help you develop competitive products in the market.