Talk

Bringing Low Power, High Performance Audio and Voice to Market on the i.MX RT600 Crossover MCU

Watch Now

25:24

EOC 2020

Login to Watch

Brendon Slade

Brendon Slade is the Director of the general purpose MCU Ecosystem team. He has more than 25 years of experience in the DSP and microcontroller industry where he has held design, applications and technical marketing roles serving industrial, mobile, automotive, voice communications and audio processing markets. His team focuses on enablement of NXP's ArmÂ® CortexÂ®-M based MCUs, working with partners and NXP's internal software teams to define and deliver complementary development tool and software solutions. A graduate of the University of Plymouth, UK, he lives in Sunnyvale, California and holds patents in debug technology.

View full profile

Mike Vartanian

Mike Vartanian is a Director of Strategic Partnerships at DSP Concepts. He has more than 20 years of experience in DSP and audio holding several positions in embedded software development, business development, and corporate strategy. Mike is a graduate of the University of Michigan and Harvard University and holds several patents in speech enhancement.

View full profile

NXP

NXP Semiconductors N.V. enables secure connections for a smarter world, advancing solutions that make lives easier, better, and safer. As the world leader in secure connectivity solutions for embedded applications, NXP is driving innovation in the automotive, industrial & IoT, mobile, and communication infrastructure markets.

Website LinkedIn

Topics Covered

DSP Low Power Microcontroller Software Design

Talk

Bringing Low Power, High Performance Audio and Voice to Market on the i.MX RT600 Crossover MCU

Brendon Slade + 1 more

25:24 EOC 2020

Brendon Slade

25:24

Join this session to learn how DSP Concept's Audioweaver tools and libraries can unleash the performance of NXP's new i.MX RT600 MCU, allowing audio designers to quickly implement designs from the algorithm level.

1 / 4

Please log in or create an account to test your knowledge and see the answers.

What is the primary architectural advantage of the i.MX RT600 for audio and voice applications as described in the presentation?

A It is a single-core Cortex-M33 MCU optimized for high-frequency audio only.

B It pairs an ARM Cortex-A application processor with a separate external DSP chip for audio.

C It combines a Cortex-M33 core with a Cadence Tensilica HiFi 4 DSP coprocessor to balance performance and energy efficiency.

D It uses only the HiFi 4 DSP and removes general-purpose MCU functionality to save cost.

E It relies on external SRAM and a GPU to offload audio processing for lower MCU load.

How does DSP Concepts’ Audio Weaver change the typical development workflow for getting audio features onto the i.MX RT600 family?

A It forces developers to write low-level DSP assembly manually for optimal performance.

B It provides a graphical, algorithm-level design tool and lets you debug and iterate in real time on the hardware without hand‑coding every change.

C It only provides offline simulation on the PC; all final tuning must be done by firmware engineers on the target.

D It replaces the need for any software tools or IDEs and directly programs the flash from the Designer.

E It requires teams to port a new proprietary DSP engine before any audio algorithms can run.

When you add more processing stages (for example, many IR filters) in an Audio Weaver design running on the RT-685/RT600, what practical method did the presenter demonstrate to understand the impact on runtime resources?

A Estimate CPU and memory impact offline and assume it scales linearly; no runtime measurement is necessary.

B Use the Audio Weaver profiler on the target hardware to view CPU usage and memory (words) and then modify the design while observing changes.

C Only measure on the host PC since target profiling is not available for the HiFi DSP.

D Increase the device clock speed until there is no audible distortion rather than checking CPU usage.

E Remove all debug displays because they consume most of the CPU, then assume remaining load is negligible.

If you find a block in your Audio Weaver design is using more CPU than necessary because it is processing many channels, what practical optimization did the presenter demonstrate?

A Duplicate the block for every channel so each instance has fewer channels to process (this always reduces CPU usage).

B Convert every block to floating-point processing globally since float is always more efficient than fixed-point on HiFi DSPs.

C Move the block to a different point in the signal chain or reduce the number of pins/channels the block processes so it only handles the needed channels.

D Increase the block size to the maximum so the DSP can buffer more samples and thereby reduce real-time CPU usage to zero.

E Process all channels at all times to maintain consistent timing and avoid any channel-based routing optimizations.

Formatting help

italics	surround text with asterisks
bold	surround text with two asterisks
hyperlink	[hyperlink](https://example.com) or just a bare URL
inline code	surround text with single `backticks`
code block (multi-line)	wrap on its own lines with three backticks: ``` your code here ```
~~strikethrough~~	surround text with ~~two tilde characters~~
quote	prefix with >

Upvotes Newest Oldest

Bryce

Score: 0 | 6 years ago | no reply

Hello Mike and Brendon; thank you for the replies to my questions. I will be in touch!

Bryce

Score: 0 | 6 years ago | 1 reply

What are the considerations for the number and arrangement of microphones for a given application?

Mike VartanianSpeaker

Score: 0 | 6 years ago | no reply

There are many considerations when thinking about the number and arrangement of microphones...
1.) Degree of Operation - Does the system need full 360 degree operation or is 180 degrees sufficient?
2.) Microphone Placement - On top of the product, microphones visible to user, properly ported, spaced properly.
3.) Microphone Geometry - Line arrays are last choice an only if necessary. Microphone arrays work best if the microphones are displaced in the horizontal plane.
I would be glad to discuss further specifics.
Thanks,
Mike

Bryce

Score: 0 | 6 years ago | 1 reply

Is MATLAB (and toolboxes) required to run the Audio Weaver Designer tool?

Brendon.SladeSpeaker

Score: 0 | 6 years ago | 1 reply

Bryce - MATLAB is not required to use Audio Weaver, but you can use these tools together.

Mike VartanianSpeaker

Score: 0 | 6 years ago | no reply

As the current version of Audio Weaver Designer was built using MATLAB, the PRO version will enable you to automate and run scripts in Matlab using APIs.

Khaled

Score: 0 | 6 years ago | no reply

Thank you very much for the valuable presentation.

hMonsarrat

Score: 0 | 6 years ago | 1 reply

Hi,
Thanks for this nice presentation.
Is there a specific reason why the audio goes through the M33?
Is it to allow the M33 to do simple audio processing at a low-power, and
use the Hifi4 core only when needed, or any similar scheme?
(Wake word handling for example)

Mike VartanianSpeaker

Score: 1 | 6 years ago | no reply

Hi hMonsarrat,
Thank you for watching the presentation and your question. The solution is flexible and can have audio running only on the HiFi 4, M33, or distributed depending on the use case. You are correct that the more simple audio processing could be done on the M33. We have done designs with the loudspeaker processing on the M33. WWE could be done on either the M33 or HiFi4 also depending on the use case.
Thanks,
Mike

Partners & Sponsors

STMicroelectronics

Arm

Microchip

The Qt Company

Cypress

NXP

IAR

PE Micro

LDRA

TUXERA

Become a Sponsor