Home General Asylum

General audio topics that don't fit into specific categories.

[ Follow Ups ] Thread: [ Display  All  Email ] [ General Asylum ]
[ Alert Moderator ]

Re: Some thoughts on filters

> The sin x/x filter I referred to in my article depends on there being
> both pre- and post echo for its operation. However, pre-echo is about an
> unnatural an effect as can exist and while a continual waveform will be
> reconstructed correctly, the pre-echo might well be heard as a
> degradation with a discontinuous waveform, ie, transients. This is what
> I believe you dislike about the sound of filters like this.

Especially considering that music largely consists of transients and
not continuous waveforms, especially at high frequencies where these
effects occur in digital systems. (We are starting now to see the first
arguments against brickwall filters for music.)

> One solution is to use a low-pass filter that is better-behaved in the
> time domain, the so-called "spline" function types made popular by
> Wadia then Pioneer, and currently offered as a option by many
> manufacturers. These filters are more "leaky" when it comes to content
> outside the CD's passband and suffer from in-band rolloff, which is
> perhaps why they are not as popular as the first sort.

> Another solution is to use an analog-domain filter, which is what early
> CD players did. Problems are that good filters are very expensive, the
> unit-unit variability is unacceptable, and the difficulty of
> controlling the inevitable ringing.

The above is not quite correct. What causes ringing is the sharpness
of the filter's corner. It doesn't matter if it's an analog filter or
a digital filter. A steep filter with a sharp corner will ring when
presented with a transient, period.

The only difference between the analog filter and the digital filter
is that the ringing in the analog filter always happens *after* the
transient, while the ringing for a digital filter (assuming the use of
an FIR filter) will be equally divided between pre-ringing and
post-ringing.

This latter is because the digital has no phase shift. Now this is
normally touted as a "good thing". But in truth it is probably a very
*bad* thing, because in nature there are always effects (ie, echoes)
after a transient, but *never* before a transient. That is, a digital
filter *by definition* will sound unnatural.

A gentle filter (as exemplified by Wadia) is not *necessarily* leaky.
It just becomes a trade-off between high frequency rolloff and
out-of-band energy. Most manufacturers choose a compromise with some
high frequency rolloff and some out-of-band energy. But one could
easily choose to have no out-of-band energy, especially when using
a high sample rate system.

> Whichever filter the designer chooses to use is a matter of trading
> tradeoffs. There is no right or wrong. However, if you choose not to
> use a low-pass filter _at all_, then you will no longer be
> reconstructing the original waveform. You are instead passing along to
> the downstream components a series of pulses, or more correctly pulses
> that have been integrated into steps by whatever bandwidth limitation
> exists in the player's output stage. This waveform includes the
> original but also includes all the ultrasonic images of that original.

Again this is not quite correct. For some reason nearly all of the
digital audio texts discuss digital waveforms as a series of pulses.
However, this is a mathematical abstraction that really only occurs
inside the digital filters themselves. Virtually every DAC has a step
output and not a pulse output. This fact is clearly recognized in
digital *video* texts, and changes the picture considerably.

For starters, most music has very litte energy close to the Nyquist
frequency (as was pointed out by Keith Howard's recent articles in
Stereophile). Secondly when a step output DAC is used, the out-of-band
energy falls off with a sin(x)/x function. These two factors combine
to make a practical system such that any theoretical problems with
out-of-band energy are probably no worse than the out-of-band energy
found on LPs when played back with MC cartridges. Therefore, the
system compatibility of a non-OS D/A shouldn't really be much of
a problem in practice.

In conclusion I would like to point out that the signal after being
brickwall filtered during A/D conversion has been significantly
altered from the original. Instead of trying to perfectly recreate
this significantly altered signal by using another brickwall filter at
playback, perhaps a different filter (or no filter) would allow a
better recreation of the *original* music signal before the A/D
conversion. Food for thought...

Charles Hansen

This post is made possible by the generous support of people like you and our sponsors:
  Parts Connexion  

Follow Ups Full Thread
Follow Ups
You can not post to an archived thread.

[ Top Top of Page ] [ Contact Us ] [ Support/Wish List ] [ Copyright © 2024, Audio Asylum, all rights reserved ]
[ General ] [ Speakers ] [ Tubes ] [ Vinyl ] [ Digital ] [ Hi-Rez ] [ Video Asylum ] [ Cables ] [ Tweaks/DIY ] [ Music ] [ Films ]