| I agree with all that Richard says, especially about viewing your sound
file before playback. While I am not an audio circuit designer, a few
other points should be noted.
A speaker attempting to reproduce a high, yet still safe, level signal
below its frequency response, can distort the audible signal. This
reveals itself as a type of intermodulation. In a power amplifier, as
the subsonic information gobbles up the current available from the power
supply, the desired audible range can be reduced and/or distorted. This
usually does not happen with better designed amplifiers or at lower
levels.
As for DC, many professional and some of the better consumer amplifiers
have switchable DC blocking capacitors. In well-designed circuits, the
audible degradation from these capacitors is minimal, though they do
tend to roll off the very lowest frequencies. Some amps have these
capacitors permanently in the path. This may be why Larry thought the
amp would automatically block VLF.
Another solution for DC, which I have used occasionally, is to insert a
high quality transformer into the signal path. This can be part of
another piece of equipment or as a 1:1 transformer in a stand-alone,
passive unit. However, good transformers are expensive, and cheap ones
sound bad.
Perhaps the most useful quick-fix, especially in the situation Tolve
originally asked about, would be to use a Butterworth filter (butterhp),
or series of them in the instrument design. The manual says that
butterhp is a second order high pass filter. In general terms, this
means it will roll-off asig at 12 dB per octave below kfreq. To get
steeper slopes, add additional copies of the same filter in series for
an additional, approximate, 6 dB/octave of rolloff per copy. Steeper
slopes will allow (require, in fact) you to move kfreq lower, but
probably at the cost of increased side effects in the passband. So
experimentation is in order. Note that kfreq is defined as the -3 dB
point in the curve. That is why steeper slopes will cause more rolloff
at the lower end of the passband, requiring you to move kreq lower to
preserve the rest of your signal.
Of course, all the above remedies will have some audible effect, but,
carefully done, not as much as blown speakers. :-)
Hope this helps.
-dB
> -----Original Message-----
> From: Richard Dobson [SMTP:RWD@cableinet.co.uk]
> Sent: April 19, 1998 10:12
> To: Larry Troxler
> Cc: csound@maths.ex.ac.uk
> Subject: Re: Protecting speakers wes Re: sndinfo
>
> Modern HiFi amplifiers are Direct Coupled, and can easily send a DC
> signal (or
> a VLF one) to a speaker. The speaker coil cannot usually handle the
> high
> current level for very long - especially if the speaker cone is pushed
> anywhere
> close to its limit.
>
> Whenever I generate a sound which I am not fully confident about (eg
> when
> developing a new program), I always view the file first - I do not
> relish
> putting full-amplitude DC or clipping either into my speakers, or
> (perhaps even
> more unpleasant!) into my headphones. Alternatively, I turn the volume
> right
> down, just in case.
>
> Richard Dobson
>
> > On Sat, 18 Apr 1998, tolve wrote:
> > > concerned that subsonic sideband frequencies that result from fm
> might
> > > damage speakers. tried to check out some files with sndinfo, but
> base Frq
> > > doesn't show up in the report on ppc running csound 3.48. how do i
> > > ascertain lowest frequency?
> >
> |