surround bass management with onepole~

Sep 19, 2008 at 1:08am

surround bass management with onepole~

Hi all,

I’m doing a surround sound installation in quad. Unfortunately, I don’t have access to, or funds to acquire, 4 of the same speakers, so I’ll be using 2 separate home stereo systems. One of these systems is a 2.1 system, so I was thinking I could sum the bass of all 4 channels into the sub (which has a separate input, independent of its satellite speakers).

My question is: how best to apply filters to send the appropriate sub signal to the sub and the rest out the other 4 channels. I’ve been messing around with onepole~ to supply the bass filter, and then subtracting its output from the original signal to supply the other 4. But I was wondering if this would be appropriate given the sample latency that is introduced into the system via the filter.

Does anyone have experience with this or something similar? Is there a better solution for bass management in such a set up? Will the latency introduced by onepole~ be noticeable (i.e. negatively effect the ‘highpass’ signal)?

Of course, time is a little precious, so I’d be grateful for any input.

Thanks

#39911
Sep 19, 2008 at 6:18am

How about adding a few samples delay to the unfiltered signal?

#140737
Sep 19, 2008 at 8:35am

Hi Tristan,

It’s common to use a special kind of arrangement called a linkwitz-riley
filter for speaker crossovers as they ensure consistent delay between
the high and low parts.

Trond Lossius has made a max one:
< http://www.maxobjects.com/?v=objects&id_objet=3054&requested=linkwitz-riley>


Owen

Tristan Coleman wrote:
> Hi all,
>
> I’m doing a surround sound installation in quad. Unfortunately, I
> don’t have access to, or funds to acquire, 4 of the same speakers, so
> I’ll be using 2 separate home stereo systems. One of these systems is
> a 2.1 system, so I was thinking I could sum the bass of all 4
> channels into the sub (which has a separate input, independent of its
> satellite speakers).
>
> My question is: how best to apply filters to send the appropriate sub
> signal to the sub and the rest out the other 4 channels. I’ve been
> messing around with onepole~ to supply the bass filter, and then
> subtracting its output from the original signal to supply the other
> 4. But I was wondering if this would be appropriate given the sample
> latency that is introduced into the system via the filter.
>
> Does anyone have experience with this or something similar? Is there
> a better solution for bass management in such a set up? Will the
> latency introduced by onepole~ be noticeable (i.e. negatively effect
> the ‘highpass’ signal)?
>
> Of course, time is a little precious, so I’d be grateful for any
> input.
>
> Thanks _______________________________________________ maxmsp mailing
>

#140738
Sep 19, 2008 at 9:14am

>Hi Tristan,
>
>It’s common to use a special kind of arrangement called a
>linkwitz-riley filter for speaker crossovers as they ensure
>consistent delay between the high and low parts.
>
>Trond Lossius has made a max one:
>< http://www.maxobjects.com/?v=objects&id_objet=3054&requested=linkwitz-riley>

hi

as far as i am aware this one never was ported to MAX 4.6….. (but I
hope Trond will port it)

kasper

Kasper T. Toeplitz
noise, composition, bass, computer

http://www.sleazeArt.com

http://www.myspace.com/sleazeart

#140739
Sep 19, 2008 at 9:33am

thanks everyone,

yes, i’m just checking out Trond Lossius’ site and it seems his externals are PPC only at the moment. The installation I’m doing is running off an old Mac G4 tower, however, I’m having to build runtime files to run on this computer with my intel mac, so I guess I can’t use Trond’s filter.

Is there another solution that people have tried, or would I be best to just ignore the sub altogether, as using simpler filtering methods will introduce phase cancellation problems?

Thanks again.

#140740
Sep 19, 2008 at 9:52am

Well, Trond provides the source for them. I’ll try and build UB versions
tonight…

Tristan Coleman wrote:
> thanks everyone,
>
> yes, i’m just checking out Trond Lossius’ site and it seems his
> externals are PPC only at the moment. The installation I’m doing is
> running off an old Mac G4 tower, however, I’m having to build runtime
> files to run on this computer with my intel mac, so I guess I can’t
> use Trond’s filter.
>
> Is there another solution that people have tried, or would I be best
> to just ignore the sub altogether, as using simpler filtering methods
> will introduce phase cancellation problems?
>
> Thanks again. _______________________________________________ maxmsp
>

#140741
Sep 19, 2008 at 10:01am

wow,

thanks owen.
cycling74 must have the best community on the web!

#140742
Sep 19, 2008 at 10:06am

>Well, Trond provides the source for them. I’ll try and build UB
>versions tonight…
>

!!!!!!!!!

if you could post them somewhere, that would be terrific!!!

best

kasper

Kasper T. Toeplitz
noise, composition, bass, computer

http://www.sleazeArt.com

http://www.myspace.com/sleazeart

#140743
Sep 19, 2008 at 12:06pm

hello folks!

i started a thread on 5.1/multichannel/immersion a couple days ago, and
whilst reading this thread and pondering on what i can up with in my thread,
i was wondering if a 5.1 patch would work for you. i don’t know which
patch, and i still don’t know a hell of a lot about 5.1, but some patches
may calculate the subwoofer output for you, at least that is my impression.
someone suggested one of the examples that comes with max/msp…”5 point 1
fun”? let me know. there was also mentioned a nice $20 – 5.1 interface in
that thread, which also got me to thinking about using 2 – 2.1 cambridge
soundworks computer amp set with and extra speaker.

best,
john

On Fri, Sep 19, 2008 at 6:06 AM, Kasper T Toeplitz
wrote:

> Well, Trond provides the source for them. I’ll try and build UB versions
>> tonight…
>>
>>
> !!!!!!!!!
>
> if you could post them somewhere, that would be terrific!!!
>
> best
>
> kasper
> –
> Kasper T. Toeplitz
> noise, composition, bass, computer
> http://www.sleazeArt.com
>
> http://www.myspace.com/sleazeart
>
>

http://www.johnhudak.net

#140744
Sep 19, 2008 at 8:50pm

Hi Kasper and Tristan,

Here are just the audio ones for now (incl. the crossover). They seem
fine on my PPC with both 4.6 and 5, but I’ve not got access to an
intel machine to try them out.

If you could confirm that they load and seem to function ok on intel
(you’ll need the help files from the distro…), I’ll go ahead and do
the rest, and see if Trond wants to stick them in the distro.

< http://www.owengreen.net/tl.objects-UB.zip>


Owen

>> Well, Trond provides the source for them. I’ll try and build UB
>> versions tonight…
>>
>
> !!!!!!!!!
>
> if you could post them somewhere, that would be terrific!!!
>
> best
>
> kasper

#140745
Sep 20, 2008 at 4:16am

looking forward to trying the UB versions…but I think they’re not attached to your last post.

if you could re-post with them attached, that would be great. or email me. i’ll forward my address to you.

tristan

#140746
Sep 20, 2008 at 9:07am

Don’t know why the link’s not showing up, it was definitely in the original:

Trying again:

http://www.owengreen.net/tl.objects-UB.zip


O

Quoting owen@owengreen.net:

> Hi Kasper and Tristan,
>
> Here are just the audio ones for now (incl. the crossover). They seem
> fine on my PPC with both 4.6 and 5, but I’ve not got access to an
> intel machine to try them out.
>
> If you could confirm that they load and seem to function ok on intel
> (you’ll need the help files from the distro…), I’ll go ahead and do
> the rest, and see if Trond wants to stick them in the distro.
>
> < http://www.owengreen.net/tl.objects-UB.zip>
>
> —
> Owen
>
>
>>> Well, Trond provides the source for them. I’ll try and build UB
>>> versions tonight…
>>>
>>
>> !!!!!!!!!
>>
>> if you could post them somewhere, that would be terrific!!!
>>
>> best
>>
>> kasper
>
>
>

#140747
Sep 20, 2008 at 4:42pm

>Hi Kasper and Tristan,
>
>Here are just the audio ones for now (incl. the crossover). They
>seem fine on my PPC with both 4.6 and 5, but I’ve not got access to
>an intel machine to try them out.
>
>If you could confirm that they load and seem to function ok on intel
>(you’ll need the help files from the distro…), I’ll go ahead and
>do the rest, and see if Trond wants to stick them in the distro.
>

I only had a few minutes – just enough to download them and try 2 or
3 help files

_I blew (totally) one of the filters which shut the Max-audio in a
heavy way (restarting the dac was not enough, i had to restart max
to have audio again) – but well, due to a sharp trackpad movement i
gave it a negative value……….

_the other help files do work

_________it was on an intel 2×2.5, max 4.6, os X.5.4 (and only 5
minutes or less)

will try later on, but at least they load without problem

bast – and many thanks

kasper

Kasper T. Toeplitz
noise, composition, bass, computer

http://www.sleazeArt.com

http://www.myspace.com/sleazeart

#140748
Sep 20, 2008 at 8:10pm

Hi,

I’ve gradually been moving more and more of my externals to be part of
Jamoma, with most of the MSP externals now being part of TTBlue. So
all of the filters are available, in a somewhat different dressing, by
installing the Jamoma 0.5 beta version.

Look for the tt.filter~ object, that’s growing into a filter swiss
army knife of ome kind, with quite a few filters implemented, e.g. 1st
- 4th order Butterworth filters and 2nd and 4th order Linkwitz-Riley
filters.

If you check out the externals that come with Jamoma, you’ll see that
quite a few of the other tl.externals have been reincarnated there as
well.

http://www.jamoma.org

http://www.ttblue.org

Best,
Trond

On Sep 19, 2008, at 12:14 PM, Kasper T Toeplitz wrote:

>> Hi Tristan,
>>
>> It’s common to use a special kind of arrangement called a linkwitz-
>> riley filter for speaker crossovers as they ensure consistent delay
>> between the high and low parts.
>>
>> Trond Lossius has made a max one:
>> < http://www.maxobjects.com/?v=objects&id_objet=3054&requested=linkwitz-riley
>> >
>
>
> hi
>
> as far as i am aware this one never was ported to MAX 4.6….. (but
> I hope Trond will port it)

#140749
Sep 22, 2008 at 12:26am

Be careful recompiling Trond’s source for MacIntel. There are some denormal issues inherent with the Intel architecture when implementing feedback in filters.

I have the lowpass source with a fix. I needed it for an old piece that I was reviving a few months ago. I am attaching it if you want to take a look at my fix. Lowers the CPU usage dramatically when silence is being processed.

And I am sure that Trond’s work in Jamoma exceeds this and compensates for the denormal issue. Would still be nice to have the full set in cases like mine with old patches. But I guess that’s what the open source is for.

–Nathan

#140750
Sep 23, 2008 at 4:45pm

Thanks for the fix to the code, Nathan.

> Be careful recompiling Trond’s source for MacIntel. There are some
> denormal issues inherent with the Intel architecture when
> implementing feedback in filters.
>
> I have the lowpass source with a fix. I needed it for an old piece
> that I was reviving a few months ago. I am attaching it if you want
> to take a look at my fix. Lowers the CPU usage dramatically when
> silence is being processed.
>
> And I am sure that Trond’s work in Jamoma exceeds this and
> compensates for the denormal issue. Would still be nice to have the
> full set in cases like mine with old patches. But I guess that’s
> what the open source is for.

I’d rather say that it’s Tims effort, but yes, one of the reasons for
moving the code to TTblue was to make it easier to add denorm, as well
as cater for future transitions to new platforms and processors we’ve
not yet even heard of. It might also be able to compile AudioUnits and
objects for other applications such as SuperCollider and Pd this way.
The work Tim’s been doing lately makes TTBlue a very exciting and
promising library.

I see the point of backward compatibility, and over the next months
I’ll see if I can make Max patch wrappers around tt.filter~ into
providing the same behavior as the tl.objects used to have.

Best,
Trond

#140751

You must be logged in to reply to this topic.