A Comprehensive Guide to Voice Command Systems in Portable and Battery-Powered Products, including the top applications and use cases for always-on voice command, the potential challenges preventing always-on voice command in portable and battery-powered products, the core hardware requirements, and the software algorithms that enable always-on voice command.
Our previous paper, “Fundamentals of Voice UI,” explained the algorithms and processes required for a voice UI system. In this paper, we demonstrate how the different microphone types and array configurations affect performance of voice UI systems, and make specific recommendations engineers and product design teams can use to get the best performance from their voice UI products.
Voice UI–or voice user interface–features found in the Amazon Echo and Google Home have captured the attention of consumers. This paper outlines the basic concerns product developers face when creating and optimizing voice UI products; examines ways of measuring and evaluating them; and recommends best practices in voice UI systems engineering.
Smartphones and the Internet of Things have made microphones much more vital in today’s technology. In this paper, the author describes a technology for digital processing of microphone signals that requires no coding skills or DSP expertise.
Audio products are becoming more complex, as are consumers’ expectations of them. Creating these products—from concept to R&D and tuning—can be difficult and time-consuming. In this paper, we will discuss the challenges in the product creation phases and explain how the combination of a graphical audio development tool with a low-power image processing DSP can transform what has long been an inefficient process.
This presentation was given at the Embedded World Conference in Nuremberg in 2015. We present an 8 channel automotive audio system based on the Atmel SAMV7 processor. The SAMV7 is an ARM Cortex-M7 processor running at 300 MHz and is sufficient for most low- to mid-level automotive audio systems.
This is the presentation made at the AES 2014 Conference at the product developer's track. This paper contains benchmarking information for the different processors. There is some surprising results with the new ARM Cortex-M4 and M7 microcontrollers as well as how much can be achieved on the Cortex-A processors. This will help you choose the best processor when developing your audio product.
The recently announced Analog Devices SHARC® ADSP-2146x processor incorporates hardware accelerators for implementing three widely used signal processing operations: FIR (finite impulse response), IIR (infinite impulse response), and FFT (fast fourier transform). The accelerators offload the core processor and have the potential to more than double the computational throughput of the processor. This paper introduces the accelerators using their application in next-generation audio systems as an example.
We describe an asynchronous sampling-rate conversion (SRC) algorithm that is specifically tailored to multichannel audio applications. The algorithm is capable of converting between arbitrary asynchronous sampling-rates around a fixed operating point, and is designed to operate in multi-threaded systems. The algorithm uses a set of fractional delay filters together with cubic interpolation to achieve accurate and efficient sampling-rate conversion.
In this paper, we present a novel algorithm for sampling rate conversion by an arbitrary factor. Theoretically, sampling rate conversion of a discrete-time (DT) sequence can be performed by converting the sequence to a series of continuous-time (CT) impulses. This series of impulses is filtered with a CT lowpass filter, and the output is then sampled at the desired rate. If the CT filter is chosen to have a rational transfer function, then this system can be simulated using a DT algorithm for which both computation and memory requirements are low.
This paper discusses frequency warping, a technique for designing and implementing discrete-time filters using allpass filters. This technique is particularly useful for audio filters because specifications are often given on a logarithmic frequency scale. It is shown that frequency warping allows a class of recursive filters to be designed using standard FIR techniques, and naturally leads to a structure for implementing the filters. The fixed-point behavior of this filter structure is analyzed and is shown to be relatively insensitive to coefficient quantization and round-off noise.