AI


never a realistic outlook. Every company has financial and logistical limits. It’s here that versatile technologies like DSP show their worth. For an engineer facing real uncertainty as to how a project is going to come together, DSP is a proven reference point for incorporating mission-critical functions, regardless of the market.


Approaches to DSP


Given the versatility of DSP, it’s exciting to see new approaches to its deployment taking the inherent flexibility of the technology even further.


engineers to keep up, they need to be able to rely on some ubiquitous technologies that can stay consistent as plans and markets evolve, and that can facilitate connection and understanding between different kinds of data and product.


Introducing Digital Signal Processing One of the most effective examples of this is Digital Signal Processing (DSP), which can take an analogue signal and, as it sounds, digitise it. It takes a real-world input, converts it into data, and then manipulates the numbers behind that data to have a specific effect. DSP has proven itself successful as it directly addresses some straightforward limitations of alternative analogue filters. A physical filter, for example, will degrade over time – and as components begin to wear or fail, the output becomes less accurate. A digital equivalent offers a far more consistent level of performance, for longer, which makes it much better suited to receive and interrogate important data.


There’s also the inherent flexibility that comes with DSP to consider. As you can change the ways in which a DSP filter might affect any input signal, you have much greater agency over the result compared to an analogue filter. It can clean the signal in different ways, and introduce effects like modulation, equalisation, and compression to produce a higher quality output.


www.cieonline.co.uk


These qualities all speak to the versatility of DSP – and, in turn, the products that use it. It enables them to reliably turn data into action, cleaning and transforming input into something precise and meaningful.


A market of niches


Interestingly, that versatility speaks to the diversification that the electronics market has undergone in the past few years. Once upon a time, it was commonplace for semiconductors manufacturers to commit a huge amount of resource and development to one specific market, safe in the knowledge that there would always be adequate demand. Digital cameras, personal computers and the like were well-established categories, with regular and manageable consumer interest. But as our devices have become more intelligent, new markets have emerged and the speed of technological development has increased. The huge markets that once dominated chip supply are smaller, comprising part of a ‘market of niches’, with more bespoke opportunities to target.


Whereas companies could invest heavily in those larger markets before, they can’t do so in smaller markets with the same confidence. Ideally, they would have the resource, time, and expertise to carefully tailor the perfect product to each individual niche. But given how many there are, and the huge variety within each, this is almost


Conventionally, DSP required a specialist component; a ‘pure’ DSP solution, purpose- built to deliver the most efficient and capable signal processing possible. While this will lead to better quality output that the user has more control over compared to a general-purpose solution, it also comes at a higher cost, and with more significant integration challenges – not to mention simply taking up more space. These things are often linked to a higher Bill of Materials, and a slower time to market. By contrast, general-purpose hardware can be coded to offer DSP as part of a wider set of functions. These allow for greater flexibility when it comes to integration, allowing engineers to combine DSP with elements like compute, or input/output functions. In some leading solutions, that balance can be customised in software, allowing the balance between these functions to be reconfigured without new components – which, in turn, reduces both development time and overall cost. This versatility is taken to greater heights with the advent of AI, which can effectively supercharge the standard capabilities of DSP. Not only can AI enhance signal processing, improving the signal quality even in noisy environments, but it can also enable a certain level of personalisation. It can, to take video conferencing as an example, associate a certain voice with a speaker in the room. AI is also the driving force behind deriving insights from the data processed by DSP, passing that context on to other components. Usually this is done via the cloud, but given the inherent security and performance issues here, some disruptive designers are creating solutions that allow on-device data processing, bypassing the cloud entirely.


Obviously, it’s important to remember that not every gadget using DSP is necessarily AI-ready. Developers need to account for AI in the design phase, bringing together a composition that incorporates all the foundational elements required for AI, such as I/O interfaces.


If that balance can be found, DSP and AI working in tandem can represent a real leap in responsiveness and intelligence.


Examples in action


So, where will that leap be felt most? Consumer electronics are perhaps the most obvious example, with interest in voice control having risen significantly in recent years and benefitting from the non-contact rules of pandemic restrictions. Smart TVs, video bars, personal computers, smart watches, headphones, and more will all see a marked rise in the quality of voice control they can offer, not only in terms of accuracy, but also in terms of understanding and actioning commands.


Similarly, the combination of AI and DSP is driving the emergence of new product categories. One such niche is equipment for content creators or influencers.


With the speed at which modern audiences expect content, many influencers and creators are no longer satisfied with recording live audio, taking it home to their PC, and finally uploading on their platform like YouTube. Devices that can record, produce, process and upload content along those lines in real time, to a social platform of one’s choice, are coming to the fore – and they’re doing so on the foundations of AI and DSP.


Moving beyond personal devices, similar improvements are set to be felt in modern vehicles, where voice control can prevent the driver from taking their hands off the wheel, or their eyes off the road. Not only will DSP act as both a noise suppressant for ambient noise like wind, or tyres on a rough road, but it can again enable the gadgets entering modern vehicles – touchscreens, sound systems, etc – to be operated hands-free.


Manufacturing environments will benefit, too. When combined with an appropriate AI, DSP can lend itself to industrial analytics, potentially flagging equipment or component failures by identifying the tell-tale noises that can indicate maintenance is required. It can also integrate with emergency stop controls, automatically shutting a machine down if certain dangers are detected, and increasing the range at which emergency stop commands can be issued verbally.


All these examples speak to a simple but profound improvement across markets with seeming no relation to one another. DSP might only be a set of numbers and algorithms, but its real-world impact can genuinely enrich our experience with electronics of every kind. When combined with AI, we’ve only just started to discover the potential applications.


www.xmos.com Components in Electronics May 2024 27


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42  |  Page 43  |  Page 44  |  Page 45  |  Page 46  |  Page 47  |  Page 48  |  Page 49  |  Page 50  |  Page 51  |  Page 52  |  Page 53  |  Page 54  |  Page 55  |  Page 56  |  Page 57  |  Page 58  |  Page 59  |  Page 60  |  Page 61  |  Page 62  |  Page 63  |  Page 64