muski;462955 Wrote: 
> Caleb --
> 
> Thanks for the interesting post.  I put Floyd Toole's book on my amazon
> wish list.
> 
> From a implementation standpoint, I took away two things:
> 
> 1) At least two subs
> 
Yes, if you care about more than 1 listening position.
muski;462955 Wrote: 
> 
> 2) That, at a minimum, the subs be equalized (< ~100Hz, though <500Hz
> would be better).
> 
Ah, careful there -- 500 is definitely not necessarily better than <300
or < 100.  You need to take measurements all around your listening area,
then make a decision where it makes sense to EQ.  

The procedure would go something like this:

1 Set up your subs in the locations you can stand having them.  side of
room, middle of the wall is best in square room calculation, but ymmv,
depending on room shape.  

2 decide on a few 'target listening positions'.  Whatever is right for
your room.  These locations will probably be pretty limited due to your
particular living arrangement.

3 Without EQ, take measurements at each location.

4 try adjusting the relative gains of each.

5 Move the speakers and try again.  Goto 3 a few times :-)

6 For setps 3-5, you're looking for the locations and levels that gives
you the most *consistent* measurements, not the locations/levels that
give you the *flattest* response.  In an ideal world, at every listening
position, you'll get the SAME, non-flat response.  This is the hard
part.

6a You will notice that it becomes completely impossible and hopeless
to get any consistancy above some frequency.  Maybe 200Hz.  Maybe 300Hz.
It depends on your room.  Don't touch any frequencies above this range
-- you will only ruin a perfectly good speaker response.

7 Now that you have a consistent frequency response across all your
listening positions, simply EQ out the bumpies in the frequency response
with whatever EQ method you choose.  Frankly, whether you do it with FIR
or IIR is relatively unimportant (yes there *may* be some audible
artifacts of one method over another, but BY FAR the most important
thing is frequency response.)  Use whatever method is convenient for
you.  And to prevent trouble with crossovers, you might find it easiest
to do the EQ on the full spectrum signal.  Don't push the EQ too far --
If you find yourself adding or subtracting 30dB, something's probably
broke in the setup or EQ system, or you have horrible resonance/dip that
needs to be resolve physically.  +- 10-15 dB is probably reasonable.

muski;462955 Wrote: 
> 
> Now, if I understand correctly, you are suggesting that IIR filters are
> better than FIR filters.  I know of three options for doing the sub EQ,
> the built-in PEQ in a Velodyne DD-10, or a Behringer DEQ2496, or Inguz.
> 
Well, not necessarily.  My suggestion is that since speakers are
fundamentally a 2nd or 4th order system, in order to compensate
properly, you'll need an IIR-like response.  However, FIR filters are
extremely flexible, much more so than IIR.  Really, for the home user,
it really comes down to what you get get your hands on and use.

muski;462955 Wrote: 
> 
> I assume that the PEQ of the Velodyne EQ & the DEQ2496 are IIR-based? 
> (I guess I plug the DEQ2496 into the 'digital loop' of my Transporter?)
> 
Uh.  Dunno.  

muski;462955 Wrote: 
> 
> With Inguz, aren't there two ways of doing this?  One is to build a
> filter from a sweep using DRC (wish I could figure out which DRC
> parameter would only apply this to <500Hz).  But this would be a FIR
> filter.  The other way is using the EQ function in Inguz.  I'm not quite
> sure if this method is FIR or IIR.  Inguz describes it as:
> 
'EQ' and 'PEQ' functionality is usually implemented as IIR, but IIRC,
Inguz is based on BruteFir, which is fundamentally an FIR system.  So,
then I'd guess that it's a truncated IIR response implemented as an FIR.
But I'd have to actually look at their web site to figure that out :-)

muski;462955 Wrote: 
> 
> "The InguzDSP EQ filters aren’t like that.  They’re “semi-parametric
> linear-phase shelving filters”.  By “semi-parametric” I mean that you
> only have two adjustable parameters: frequency and amplitude.  And
> “shelving” means that the frequency response in between two adjacent
> bands is a gentle (cosine-based) slope." 
> 
Ah!  Linear Phase, that means it's *not* based on an analog IIR
prototype filter.  That's fine too.  Don't get hung up on that.  All you
care about is that your frequency response is flat :-)

muski;462955 Wrote: 
> 
> The other question I have is about the pros/cons of active vs passive
> crossovers for subs?
> 

There really are 3 classes of filters you might want to think about: 
Passive (low signal), Passive (high signal), Active Analog, Digital.

This is my 5-minutes-of-thought answer.  Somebody correct me if I'm
wrong:

Passive, High signal
*  These are built with large, air-core inductors and capacitors. 
These would go after your power amp, to a passive sub, and passive
speakers.   I'd stay away from anything like that like the plague.  This
is the same technology that's in your speakers already (if it's a
passive-multi-way system).  In a 'black-box' system like a speaker, this
is just barely manageable, and not very well at that.  The impedance
variance of drivers cause the crossover to misbehave.  Passive
crossovers are very difficult to get right in the best of times.

Passive, Low Signal
* This would do an active crossover, somewhere before the power
amplification stage.  Any passive crossover is bad, but this would be
much less bad than trying to do the crossover on the power PA signal. 
The reason passive is bad is because the frequency characteristics of
the filter will be highly dependent on *exactly* what the output
impedance of the driving device, and the input impedance of the driven
device are.  It's a recipe for disaster.  It's true that active filters
have the potential to add noise, but on balance, I'd take a little noise
over completely mucked up frequency response.  I'd stay away from these
like the swine flu.

Active, Low signal.
Well, this is better than passive, low signal for sure.  A bunch of
stuff falls into this category:  graphic equalizers, PEQ boxes, etc.  It
will be much more effective, and you can build slightly complex EQ
filters.  (Passive filters are very limited to very simple filters.  a
single 4th order filter is probably about as good as you can get).  I'd
stay away from these like the common cold.

Digital:
This includes any DSP based system, including Inguz, and other PC based
systems.  This gives you *everything* you need to do the job right, in
principal.  One thing to take care of is that the math needs to be high
enough precision so as to not create quantization noise.  The
professional systems give you a nice easy interface to use, and you
don't need to think too much.  Since I'm writing the stuff for Baby Boom
in C, I need to worry about everything.  Hopefully, you don't have to do
that :-)

I hope this gives you some where to get started :-)  Let me know what
your plans are, and I can probably give you more pointers.  I'm quite
interested to hear how the system comes out if you go for the full-blown
room treatment.

(Speaking of room treatment -- that's a whole other story.  Less is
more... I'll write on that some day too.  Or you can just read
'Loudspeakers and Rooms')

-Caleb


-- 
ccrome2

Caleb Crome
Sr. Hardware Engineer
Logitech SMBU  (i.e. the Squeezebox people)

<B>The future is here.  It's just not widely distributed yet.</B> 
<I>-William Gibson</I>
------------------------------------------------------------------------
ccrome2's Profile: http://forums.slimdevices.com/member.php?userid=18023
View this thread: http://forums.slimdevices.com/showthread.php?t=67022

_______________________________________________
audiophiles mailing list
[email protected]
http://lists.slimdevices.com/mailman/listinfo/audiophiles

Reply via email to