Sign in with
Sign up | Sign in
Your question

DSP audio equalization for loudspeaker purposes

Tags:
Last response: in Home Audio
Share
Anonymous
April 14, 2004 12:39:45 AM

Archived from groups: rec.audio.tech (More info?)

Hi All,

It seems to me that DSP correction is in the future of loudspeaker systems
and I know of some of the hardware products out there. I'm looking for a
good exchange of theoretical views on the subject. I know there are software
packages that will transform a wav file with a specific filtered response,
but can this be done on the fly with streaming audio yet on a PC? Anybody
know of any good links or books on the subject?

Wessel
Anonymous
April 14, 2004 12:39:46 AM

Archived from groups: rec.audio.tech (More info?)

"Wessel Dirksen" <wdirksen@p-we.com> writes:

> Hi All,
>
> It seems to me that DSP correction is in the future of loudspeaker systems
> and I know of some of the hardware products out there. I'm looking for a
> good exchange of theoretical views on the subject.

I view it as a mixed bag. It's far better to correct the anomalies in other
ways if possible or practical, but if not, DSP is a good alternative.

> I know there are software
> packages that will transform a wav file with a specific filtered response,
> but can this be done on the fly with streaming audio yet on a PC?

It depends on the sample rate and the filter. If you're talking 192 kHz
24-bit precision with an FIR thousands of taps long, heck no. If you're
talking 44.1 kHz 16-bit precision with an FIR a few hundred taps long,
then PCs have been capable of this for decades. Of course DSD is a whole
bag of worms here.

> Anybody know of any good links or books on the subject?

Try searching the rec.audio.pro usenet archives somewhere between 1
and 12 months ago for my name as author (but follow the whole thread)
- I had a "discussion" with another chap there.
--
% Randy Yates % "I met someone who looks alot like you,
%% Fuquay-Varina, NC % she does the things you do,
%%% 919-577-9882 % but she is an IBM."
%%%% <yates@ieee.org> % 'Yours Truly, 2095', *Time*, ELO
http://home.earthlink.net/~yatescr
Anonymous
April 14, 2004 12:39:46 AM

Archived from groups: rec.audio.tech (More info?)

On Tue, 13 Apr 2004 20:39:45 +0200, "Wessel Dirksen"
<wdirksen@p-we.com> wrote:

>>I'm looking for a
>>good exchange of theoretical views on the subject.

comp.dsp

>>I know there are software
>>packages that will transform a wav file with a specific filtered response,
>>but can this be done on the fly with streaming audio yet on a PC?

ASIO and ALSA, to name two with which I have worked. There are
others. The PC offers a phenomenal amount of processing power.

GB
Related resources
Anonymous
April 14, 2004 12:39:47 AM

Archived from groups: rec.audio.tech (More info?)

>>comp.dsp

Also hifi_dsp group on Yahoo, and Music-DSP mailing list.
Anonymous
April 14, 2004 2:15:08 PM

Archived from groups: rec.audio.tech (More info?)

Wessel,

> It seems to me that DSP correction is in the future of loudspeaker systems
<

This has been discussed here and elsewhere in detail over the past year or
so. DSP cannot solve room acoustic problems - not even in theory - unless
you're willing to sit with your head clamped in a vise. The peaks and severe
nulls present in all small rooms are too localized to allow correction that
works beyond one cubic inch. The last time this came up here I did a test
and measured room-induced level variations of 15 dB at 100 Hz across a
physical span of only four inches.

If that's not enough, here are two more reasons DSP cannot correct room and
speaker errors in practice:

* No power amp and loudspeaker can handle the 30 dB or more boost needed to
counter the deep nulls at low frequencies that are present in all small
rooms.

* All the DSP in the world cannot eliminate reverb, echoes, and a room's
modal ringing.

--Ethan
Anonymous
April 14, 2004 2:28:30 PM

Archived from groups: rec.audio.tech (More info?)

"Ethan Winer" <ethanw at ethanwiner dot com> writes:

> Wessel,
>
> > It seems to me that DSP correction is in the future of loudspeaker systems
> <
>
> This has been discussed here and elsewhere in detail over the past year or
> so. DSP cannot solve room acoustic problems - not even in theory - unless
> you're willing to sit with your head clamped in a vise. The peaks and severe
> nulls present in all small rooms are too localized to allow correction that
> works beyond one cubic inch. The last time this came up here I did a test
> and measured room-induced level variations of 15 dB at 100 Hz across a
> physical span of only four inches.
>
> If that's not enough, here are two more reasons DSP cannot correct room and
> speaker errors in practice:
>
> * No power amp and loudspeaker can handle the 30 dB or more boost needed to
> counter the deep nulls at low frequencies that are present in all small
> rooms.

This is a good point but the assertion is a bit excessive.

It only takes common sense to see that the amplifier is going to have
to work harder to correct room responses, and it is true that if you're
already pushing your system, you may have problems attempting to
equalize. For example, a pair of relatively inefficient electrostatic
speakers in combination with a modest amp of 100 WPC is going to have
a hard time reproducing equalized sound at >100 dB SPL.

However, there are systems that well-exceed their normal operating
requirements, especially those with efficient
(e.g., > 90 dB SPL @ 1W @ 1M) speakers. My system is capable of
reproducing well over 120 dB SPL at the listening position using
150 WPC.

Also, the "30 dB" term is misleading. A particular null may be 30
dB down as measured in a very small bandwidth, but likewise since
that bandwidth is small the corresponding total power in that band
is small. Remember that you have to integrate power spectral density
over bandwidth to get total power in that bandwidth.

> * All the DSP in the world cannot eliminate reverb, echoes, and a room's
> modal ringing.

If you mean "at all places in the room" then you're correct. However, it
is not true that DSP cannot remove these things at a fixed position. Thus
your statement is, at best, misleading.
--
Randy Yates
Sony Ericsson Mobile Communications
Research Triangle Park, NC, USA
randy.yates@sonyericsson.com, 919-472-1124
Anonymous
April 14, 2004 4:22:42 PM

Archived from groups: rec.audio.tech (More info?)

Randy,

> A particular null may be 30 dB down as measured in a very small bandwidth,
but likewise since that bandwidth is small the corresponding total power in
that band is small. <

Maybe with 1/3 octave pink noise played in a lab, but certainly not with
real music. At the listening position in my partner's control room we
measured a huge dip at 82 Hz, which aligns exactly with an E note on an
electric bass. So any time that note is played, the amp and speakers need to
play it at any level up to full volume. Here's a link to my recent article
in EQ magazine that includes the response graph I'm talking about:

www.ethanwiner.com/eq_vibe.html

Note that this is not some unique anomaly in one room. This horribly skewed
response is absolutely typical for typical small listening rooms.

> If you mean "at all places in the room" then you're correct. However, it
is not true that DSP cannot remove these things at a fixed position. Thus
your statement is, at best, misleading. <

I was careful to qualify that by mentioning the results of my test from last
year, where I reported a 15 dB difference across a physical span of only
four inches. Hence the "head in a vise" analogy.

We've been through all this before. Yes, you can fix any room problem with
DSP, as long as that room remains inside an ivory tower. In the real world,
with real listeners playing real music through real loudspeakers, I maintain
that trying to solve room acoustics problems with DSP is futile.

--Ethan
Anonymous
April 14, 2004 5:34:16 PM

Archived from groups: rec.audio.tech (More info?)

"Ethan Winer" <ethanw at ethanwiner dot com> writes:

> Randy,
>
> > A particular null may be 30 dB down as measured in a very small bandwidth,
> but likewise since that bandwidth is small the corresponding total power in
> that band is small. <
>
> Maybe with 1/3 octave pink noise played in a lab, but certainly not with
> real music.

Perhaps it is the case that you get some pretty nasty "spikes" in the
spectrum (i.e., the power spectral density skyrockets in a very narrow
bandwidth), in which case you would have a significant power in a small
bandwidth. I concede this point, Ethan.

> At the listening position in my partner's control room we
> measured a huge dip at 82 Hz, which aligns exactly with an E note on an
> electric bass. So any time that note is played, the amp and speakers need to
> play it at any level up to full volume. Here's a link to my recent article
> in EQ magazine that includes the response graph I'm talking about:
>
> www.ethanwiner.com/eq_vibe.html
>
> Note that this is not some unique anomaly in one room. This horribly skewed
> response is absolutely typical for typical small listening rooms.

Didn't Audio magazine used to include a plot of the response at the
listening position for many of their speakers reviews? I don't remember
seeing anything like a 35 dB dip.

> > If you mean "at all places in the room" then you're correct. However, it
> is not true that DSP cannot remove these things at a fixed position. Thus
> your statement is, at best, misleading. <
>
> I was careful to qualify that by mentioning the results of my test from last
> year, where I reported a 15 dB difference across a physical span of only
> four inches. Hence the "head in a vise" analogy.

Sorry, it didn't look all that careful to me. Since you made it a point
distinct from the other assertions, it sounded to me like you were saying
that *in no way* could DSP correct these types of room acoustics, and that
just isn't true. The fact that such a DSP correction may only operate in
a certain physical range is exactly the same phenomenom that limits its
response-correction ability to a certain range, namely, the change in the
room's impulse response.

> We've been through all this before.

Yes, we have, and I will continue to repeat my clarifications as long as I
feel statements are being made that could mislead.

> Yes, you can fix any room problem with
> DSP, as long as that room remains inside an ivory tower.

No, as long as the listener's position remains fixed. But I concede this
is a severe constraint.

> In the real world, with real listeners playing real music through
> real loudspeakers, I maintain that trying to solve room acoustics
> problems with DSP is futile.

I don't know that I'm disagreeing with you here. I would be careful
perhaps to ensure that folks know you're talking about room acoustic
problems in contrast to speaker problems such as smoothing out the
power response of the speaker.
--
Randy Yates
Sony Ericsson Mobile Communications
Research Triangle Park, NC, USA
randy.yates@sonyericsson.com, 919-472-1124
Anonymous
April 15, 2004 1:09:24 AM

Archived from groups: rec.audio.tech (More info?)

Thanks guys, I'll check out the links. I'm mainly interested in DSP'ing the
loudspeaker to be as ideal as possible prior to amplification using bi (tri)
amping. Thanks for the input. I had already given up the idea of even trying
to nail room correction,

Wessel


"Wessel Dirksen" <wdirksen@p-we.com> wrote in message
news:407c33f2$0$430$a0ced6e1@news.skynet.be...
> Hi All,
>
> It seems to me that DSP correction is in the future of loudspeaker systems
> and I know of some of the hardware products out there. I'm looking for a
> good exchange of theoretical views on the subject. I know there are
software
> packages that will transform a wav file with a specific filtered response,
> but can this be done on the fly with streaming audio yet on a PC? Anybody
> know of any good links or books on the subject?
>
> Wessel
>
>
Anonymous
April 15, 2004 1:43:10 AM

Archived from groups: rec.audio.tech (More info?)

"Greg Berchin" <Bank-to-Turn@mchsi.com> wrote in message
news:gpio705frm6qnqvbr8sbfdgtcs0cqa5spp@4ax.com...
> >>comp.dsp
>
> Also hifi_dsp group on Yahoo, and Music-DSP mailing list.

How does one apply to be a member of hifi_dsp?
Anonymous
April 15, 2004 1:43:11 AM

Archived from groups: rec.audio.tech (More info?)

On Wed, 14 Apr 2004 21:43:10 +0200, "Wessel Dirksen"
<wdirksen@p-we.com> wrote:

>>How does one apply to be a member of hifi_dsp?

hifi_dsp-subscribe@yahoogroups.com
Anonymous
April 15, 2004 3:14:17 PM

Archived from groups: rec.audio.tech (More info?)

Randy,

> Didn't Audio magazine used to include a plot of the response at the
listening position for many of their speakers reviews? I don't remember
seeing anything like a 35 dB dip. <

The big problem with most room analysis, especially those published years
ago, is they report the 1/3 octave response. This completely misses the
peaks and severe nulls present in all small rooms. Look again at the graph I
linked earlier, and notice the A note peak at 110 Hz and the severe null
less than a musical whole step higher. If you plot that data in 1/3 octaves
the peak and adjacent null are averaged together, so the response will
appear basically flat!

Note that this graph was plotted at 1 Hz increments. Since then I bought ETF
from www.acoustisoft.com and I can now resolve LF response to even less than
1 Hz. This level of detail is the ONLY way to measure the true LF response
in small rooms.

> it sounded to me like you were saying that *in no way* could DSP correct
these types of room acoustics, and that just isn't true ... as long as the
listener's position remains fixed. But I concede this is a severe
constraint. <

To me "not practical" and "no way" are identical. I realize many people do
not want to look at acoustic treatment in their living room. I'm sure they'd
reject having to sit motionless just as vigorously. So a better direction to
take is coming up with acoustic treatment that looks nicer! :->)

> I don't know that I'm disagreeing with you here. <

Right - you're just being a stickler for accuracy, and that's cool. So let
it be known I acknowledge that in theory DSP can correct for rooms and
speakers. But since it's not practical I will continue to explain to newbies
and others the futility of trying to fix their rooms with EQ. This is the
real issue - people who don't know any better believe an EQ or other piece
of gear can fix their room because the hi-fi salesman told them so.

--Ethan
!