Developing your own DSP framework
Would You, Could You, Should You?
For any company creating digital audio devices, there is always a time when a fatal decision must be made: should you create your own audio framework, use an existing one, or outsource it to a more cost-effective team in a different time zone?
In this talk, I will share the key considerations that led my current employer to develop a bespoke framework, starting with general design goals and possibilities that it opens.
After establishing the general context, I will focus on the technical side and provide concrete examples that illustrate the possibilities of following C++ and good software practices to make your code safer, speeding up your development and reduce your time to market, enabling code re-use between projects and allowing different teams to work in parallel.
To wrap up, I will summarize some common challenges that can be faced when developing your own framework from scratch. And more importantly, how to be clever enough to survive the fall thanks to a bit of that over-engineering that you frowned upon initially.
Miguel Jimenez-Carvajal
Head of Software
Kef Audio Ltd
Software Mad Hatter with almost 20 years of experience in R&D, working across different industries (transport, telecoms, air traffic control, medical) and stack levels, from desktop to embedded c, and anything in between. This has given me regular exposure to different languages, mainly: C, C++, Java, Matlab and python.
For the last 7 years I have been working for Kef Audio, developing internal desktop tools, embedded systems, and doing a bit of devops. I am currently juggling management and technical work depending on the hour as Head of the Software Department in R&D.