DSP Concepts says its GUI enables any engineer, whether familiar with audio processing or not, to quickly produce studio quality results.
California-based audio digital signal processing company DSP Concepts has wrapped a graphical-user interface (GUI) around its 400+ audio algorithms, compiled them for every major audio processor, and is releasing them in a standard product called Audio Weaver.
For 12 years DSP Concepts and its founder Paul Beckmann provided consulting services for audio digital signal processing (DSP). During that time his team has amassed expertise helping clients solve difficult audio problems, which surprisingly are more important to fix than flawed video.
"People will tolerate grainy video, but poor audio quality is unforgivable by most consumers," Paul Beckmann, chief executive officer at DSP Concepts told EE Times.
The company over the years accumulated all the standard audio functions in modules, which it hand writes for all available processors from Atmel to Analog Devices to ARM to TI's DSPs, ending up with over 400 modules today. Its library also includes advanced modules just coming into play today, such as microphone beam forming and different types of noise detection. For example, one module turns up the audio when it gets noisy on the highway but turns it back down for relatively low-noise situations, such as sitting at a stop light.
Now, with coaxing from professional sound labs like Sennheiser and BareFoot Sound, DSP Concepts released Audio Weaver, which the company claims can solve almost every audio problem—its core algorithms plus advanced concepts previously unheard of—from front-end tasks like managing multiple microphones to audio-zoom into the one principle voice, to middle management like the automatic volume control adjuster mentioned above to back-end management of the frequency curves of speaker to emulate both high- and low-end models on the same studio monitors.
__Figure 1:__ *DSP Concepts says its graphical user interface enables any engineer, whether familiar with audio processing or not, to quickly produce studio quality results. (Source: DSP Concepts)*
DSP Concepts is still adding advanced capabilities to its 400+ library, as well as improving its GUI and interface it more tightly to standard tools like Matlab. It's also still consulting and helping clients build complicated audio products, often promising to cut time-to-market by 90 percent. For instance, recently it consulted with Sirius/XM to do volume levelling across all of their stations to improve the user experience while surfing around the dial. Another recent consulting jobs at Sirius/XM includes codec preconditioning to reduce annoying artifacts created by codecs used in the transmission chain.
The instigator to releasing DSP Concepts secret sauce as a standard product—Audio Weaver—was Sennheiser, who was using it so often that it wanted to use it in-house to help its U.S. design team transfer their designs more easily to their German-based manufacturing branch.
*The video “Audio Weaver Setup for the STM32 F407 Discovery Board” demonstrates how to get started, but Audio Weaver runs on almost every available microcontroller, even home-grown ARM boards. (Source: DSP Concepts)*
"We realised there was a big gap between the algorithms we came up with, and what the product engineers in Germany were able to implement in a shippable way. Often the algorithms we developed couldn’t be implemented, or were difficult to implement, on the platforms that the hardware engineers were using. So we realised we needed some software tools that would facilitate cross-platform development. We looked at what was available, and considered developing something ourselves, and decided Audio Weaver would be the best solution," according to Dan Harris, Chief Technology Integrator at Sennheiser. "Also Audio Weaver…allows us to develop DSP-based products without adding DSP programmers to the staff. Audio Weaver makes it possible for audio products to be developed by acoustics guys or audio engineers who are comfortable with Matlab but don’t know how to implement, say, a digital filter in DSP."
Audio Weaver also lets users develop custom modules in Matlab, C or C++ that are not currently in the DSP Concept repertoire, and if approved, they can be added to DSP's repertoire for others to use with a royalty revenue stream going back to the inventor.
New products using Audio Weaver "inside" include Amazon's Echo, GoPro's Hero-4, Blue Microphone's Nessie.