1 / 10

Echo cancellation on Cell Broadband Engine

Echo cancellation on Cell Broadband Engine. Jaromír Škorpil (skorpj1@fel.cvut.cz) Tomáš Uko (ukot1@fel.cvut.cz). Goal of project. To i mplement Echo cancellation on Cell Broadband Engine with using of parallerism which this architecture offers. Motivation.

cianna
Download Presentation

Echo cancellation on Cell Broadband Engine

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. IBM - CVUT Student Research Projects Echo cancellation on Cell Broadband Engine Jaromír Škorpil (skorpj1@fel.cvut.cz) Tomáš Uko (ukot1@fel.cvut.cz)

  2. Goal of project • To implement Echo cancellation on Cell Broadband Engine with using of parallerism which this architecture offers

  3. Motivation • Echo cancellation has many applications • Telecommunications - removing(reducing) of sounds in background during conversation • Hands-free car phone systems • Standard phone in speakerphone or hands-free mode • … • Sound systems - Microphone – Reproductor back-coupling • CELL implementation offers possibility to compute more acoustic channels in parallel IBM - CVUT Student Research Projects 3

  4. Acustic echo IBM - CVUT Student Research Projects 4

  5. Structure of AEC IBM - CVUT Student Research Projects 5

  6. Main problems we had to solve • Double talk detection • SIMD implementation of adaptive filter on SPU • Implementation of DMA transfers for maximal data throughtput • Obtain test data IBM - CVUT Student Research Projects 6

  7. Implementation IBM - CVUT Student Research Projects 7

  8. Benefits of SIMD implementation of adaptive filter • 4 samples processed by single instruction in parallel IBM - CVUT Student Research Projects 8

  9. Project results • Working and tested SIMDized code of AEC for CELL SPU including DMA based data transfers • 60 seconds sample -> takes 2,6 seconds to compute (measured on Playstation 3) • NonSIMDized code working on CELL P(S)PU and other x86 platforms • 60 seconds sample -> takes 8,9 seconds to compute (measured on Playstation 3) IBM - CVUT Student Research Projects 9

  10. Future improvements • Direct connection with Soundcard • Real-time AEC • processing of multiple independent audio channels in parallel on multiple SPU (modification of PPU code is only necessary, SPU code remains the same) IBM - CVUT Student Research Projects 10

More Related