Voice user interfaces - you can scarcely avoid the current hype in the media as giants like Amazon, and Google jostle to exploit the explosion of possibilities that advancements in natural language technologies are providing. Today’s neural networks use algorithms to process language through ever-deeper layers of complexity. Machines can now understand the meaning and intent of spoken words with unprecedented levels of accuracy. This has sparked a revolution for the power of voice.
We love the Amazon Echo at XMOS. It’s a new-to-the-world category of product, brimming with possibilities as a digital assistant, a hub for home automation as well as a point of presence to allow us to access all of Amazon’s goods and services.
At its heart is a piece of technology known as a smart microphone. This enables the Echo to capture voice samples with a high degree of accuracy before transmitting them to the Amazon Voice Services in the cloud where the query is processed before the answer is returned to the device in the form of Alexa’s soothing tones contained in an MP3 file.
In a previous blog I discussed the new DSP capabilities of the xCORE-200 architecture at a very high level. I'd now like to look at some of those DSP capabilities in more detail.
XMOS xCORE-200 devices are built around a 32-bit fixed point architecture, which includes a 64-bit accumulator for enhanced dynamic range. This architecture was chosen because it gives flexibility to implement many different forms of fixed-point and integer arithmetic efficiently.
Accurate, accessible and economical Voice User Interface (VUI) design has been slow to emerge. Our ability to interpret dialogue of varying pronunciation, accent, speed and grammatical correctness is something that we humans take for granted from an early age, but that has proved elusive for machines. Nonetheless, VUI has become a mainstay of our predicted future - who can forget HAL 9000 or the "Talkie toaster" (immortalized by Grant Naylor in the BBC TV series "Red Dwarf"? http://www.youtube.com/watch?v=LRq_SAuQDec)
Designing electronic products today is a highly complicated challenge that goes well beyond the old question of "How many MIPS does the microprocessor have?" Designers must solve the fundamental requirements of their products, while meeting strict international standards, and integrating innovative features to support IoT (Internet of Things) technologies. And consumer products also need to be fashionable as well.
Welcome to our latest guest blog, where XMOS customers discuss their experiences of using xCORE multicore microcontrollers.
Joining us this time is Dominic Baker, Technical Director at Cambridge Audio.
For more than 40 years, UK-based Cambridge Audio has been at the forefront of hi-fi innovation, designing and building a range of award winning products for music lovers. Each product is designed to deliver the best performance and sound quality from a range of sources.
Back in March of this year we announced a first for XMOS … a dedicated audio processor, called xCORE-AUDIO Hi-Res 2. Well we’ve had the samples back in our lab for a while now and the device is now ready to be let out into the wild. Here at the factory we are very excited about the new development.
So what is xCORE-AUDIO Hi-Res 2?
Simply put it’s a dedicated audio processor which offers high resolution stereo (hence the 2 … clever eh?) audio interfacing for consumer audio applications such as DACs and headphone amplifiers.
Welcome to the second in our series of featured guest blogs, where XMOS customers discuss their experiences of using xCORE multicore microcontrollers.
This time out we talk with Thorsten Loesch, Director of Technology for iFi Audio.
Welcome to the second instalment of our Hi-Res Audio Formats overview. In the fist instalment we looked at the Linear PCM (Pulse Code Modulation) encoding scheme used on CDs and which continues to dominate the digital audio world today. If you missed it, you can read the blog here.
Today we are going to look at the other primary format for high resolution audio files, DSD or Direct Stream Digital.
PCM remained as the dominant encoding technique for digital audio until the introduction of the Super Audio CD (SACD) in 1999 which used DSD to encapsulate a higher resolution layer alongside an optional CD layer with the familiar 16-bit, 44.1kHz PCM encoding.