The noise reduction method is applicable worldwide. It can be used in FM receiver chips, smartphones, car radios, rebroadcast receivers, Hi-Fi stereo receivers and as a standalone device, either hardware-based or as a software solution. The method is patent-protected in Germany and the U.S.A.. In Canada the patent is pending.
patent No. US 10,003,422
title „METHOD FOR PROCESSING AN FM STEREO SIGNAL“.
granted on 19 June 2018 by the U.S. Patent and Trademark Office (USPTO)
patent No. DE 10 2016 107 799
title „Verfahren zur Verarbeitung eines FM-Stereosignals“.
granted on 1 February 2017 by the German Patent and Trademark Office (DPMA)
patent application number is CA 3021918
national phase entry on 28 October 2018
Companies that are interested in using the patented method, and that want to acquire patent utility rights, or get a license, are welcome.
C Source Code
The patented method is currently implemented as C source code.
Three applications are available:
- On Windows: The application reads a WAV file and creates another WAV file with the noise-reduced content.
- On Windows: A streaming application that uses a sound card as input/output.
- On Raspberry Pi: A streaming application that uses a sound card as input/output.
On a Raspberry Pi 3 Model B+, the task manager shows a RAM usage of 3.5 MB and a CPU usage of 20%. There are still several ways to optimize the code in terms of computing time or memory usage. Systematic latency is dependent on the block size processed by the algorithm and the sample frequency. With a block size of 4096 samples and fs = 44.1 kHz the systematic latency is 93 ms.
For further information, please feel free to contact me any time.