Computer Based Compression/Speech Processing

VK2MEV » Sun Oct 21, 2012 12:37 pm
Today I worked out how to use a Linux based machine and a couple of ALSA plugins to perform speech compression and equalisation in real-time. The basic idea is to use compression/gate on a recording stream and equalisation on a playback stream and record and play back simultaneously, with the smallest buffer possible.

The specific implementation is a PC headset (Beyerdynamic MMX2, not that it matters) with the mic plugged into a cheap USB sound card. The sound card is connected to a Raspberry Pi which processes the audio in real time (ok, there’s a 10ms buffer/delay) and plays the compressed and equalised audio back out the USB sound card. This is then directly connected to the „data“ input of my FT-857d (yes, it should go through an isolation transformer).

Brief details on the ALSA configuration can be found in a blog post on my site: http://www.vk2mev.net/?p=121

An on-air recording comparing the stock mic to the processed audio is here: http://www.vk2mev.net/wp-content/uploads/2012/10/vk2mev-sw-comp-test2-8k.mp3
http://www.youtube.com/watch?feature=player_embedded&v=NqClalkRfTI
http://www.vklogger.com/forum/viewtopic.php?f=41&t=10939&sid=7f2d4d585fc4251562a43a8ed7fcf0f4#p28254
________________________________________________

Informace o Raspberry Pi najdete například zde http://www.dps-az.cz/cs/component/search/?searchword=Raspberry+Pi&ordering=&searchphrase=all