You should be able to handle that in Linux. Seeing, that it's traffic noise,
a typical fft-based denoising tool is out. Your tools of choice should be EQ
You can of course remove the very low end regardless. You could try to find
the bass frequency of the voice and set the low end even higher than 50Hz, not
too close though, since no filter is perfect.
Then put an EQ on your recording, Activate only one band, set it as narrow
as possible and the gain about 12Db or even higher. Then sweep the bands
frequency slowly. You will find the first point where it hurts. Just lower the
gain drastically below 0Db.
Another idea: If it's a stereo recording with the voice bang centre, you
might perhaps go a stranger route. There was a sort of karaoke LADSPA plugin,
which tried to lower something in the centre and only leave the rest. If that
works, leaving mostly the traffic, you can use some graphical wave-editor or
display to line the two files up, in case, that there is some delay introduced
by the plugin. that done, you can use csound, PD or some other allround synth
to subtract the mainly traffic only signal from the original. This is an
outside chance, but it may end up giving a good result.
Such Is Life: Very Intensely Adorable;
Free And Jubilating Amazement Revels, Dancing On - FLOWERS!
====== Find my music at ======
"If you live to be 100, I hope I live to be 100 minus 1 day,
so I never have to live without you." (Winnie the Pooh)
Linux-audio-user mailing list