Thread View: comp.dsp
2 messages
2 total messages
Started by R Clement
Thu, 15 Oct 1992 10:42
Vocal Erasers - What can we leave in?
Author: R Clement
Date: Thu, 15 Oct 1992 10:42
Date: Thu, 15 Oct 1992 10:42
21 lines
1332 bytes
1332 bytes
I've just implemented a vocal eraser (thanks DAn) but have noticed one main problem. When I cancel out the vocal (usually dead centre), it tends to cancel out the bass as well. Since bass is usually at lower frequencies than vocals, would it not be possible to filter out the bass parts of the stereo signal (hopefully avoiding nearly all of the vocal) and then add this back to the (monoised) vocal cancelled signal. This would (theoretically) add some stereo width back to the signal. Can anyone suggest a suitable frequency cutoff (in Hertz) for a low pass filter to filter out the bass parts of a music signal? What about a high pass filter, what are the highest frequencies present in a lead vocal? Are they low enough so that there's useful stuff in the higher frequency ranges (I'm using 48Khz sampling, and processing output from a CD player). Thanks in anticipation, Ross-c
Re: Vocal Erasers - What can we leave in?
Author: Christopher John Rolfe
Date: Sat, 17 Oct 1992 00:03
Date: Sat, 17 Oct 1992 00:03
46 lines
2536 bytes
2536 bytes
ross-c@scs.leeds.ac.uk (R Clement) writes:
>Since bass is usually at lower frequencies than vocals, would it not
>be possible to filter out the bass parts of the stereo signal (hopefully
>avoiding nearly all of the vocal) and then add this back to the (monoised)
>vocal cancelled signal. This would (theoretically) add some stereo width
>back to the signal.
>Can anyone suggest a suitable frequency cutoff (in Hertz) for a low pass
>filter to filter out the bass parts of a music signal?
>What about a high pass filter, what are the highest frequencies present in
>a lead vocal? Are they low enough so that there's useful stuff in the higher
>frequency ranges (I'm using 48Khz sampling, and processing output from a
>CD player).
The low E-string on an electric bass is usually tuned to about
41 Hz., a bit less than an octave below a bass singer. Male pop
singers seem to be predominantly tenors or baritones which means that their low
note fundamental would be up about an octave from that. So you may
find two octaves more or less between the bass and the vocal.
Problem is that the overtones of the bass are probably quite
necessary to its overall sound and distinctiveness, up to 160 Hz. and beyond.
There's probably also a lot of other program information in the mid-range.
What you might try is putting
the mono version of the original unprocessed tune onto a
track A and B of a two track tape. Then dub the double mono version
thusly:
A ---> voice-eraser -_
_ ---> mono result
B ---> filter ------
Use a band-pass filter on B, set the lower cutoff freq. to 0,
and adjust the upper cutoff by ear for the optimal result ( perhaps
between 80 and 160 Hz.).
I don't expect this to be very transparent, but if you track
the song a bit ( with the upper cutoff ) you might get something
useable.
--
Chris | 2617 East 4th,| If brevity is the soul of
rolfe@sfu.ca | Van., B.C. | wit, what is semibrevity
Simon Fraser University | V5M 1K5 | the soul of?
Burnaby, Canada | Canada |
Thread Navigation
This is a paginated view of messages in the thread with full content displayed inline.
Messages are displayed in chronological order, with the original post highlighted in green.
Use pagination controls to navigate through all messages in large threads.
Back to All Threads