On Sun 2012-Jun-10 15:59, Don Y writes:

First, thanks for the explanation of your networked
speakers. That's what I thought I recalled from the thread, but wanted to be sure I was on the right page of music

For our example, we're doing a radio commercial. We want
the music bed up full until the announcercomes in to tell
us about the great deal you can get if you call before midnight

That's the problem with the analogy, though. You're already
saturating the channel before you add the "announcement".
So, you have to attenuate that signal if you want anything
to come across "above" it.

Okay, can relate. Ducking, or full blown muting is probably
the simplest to implement, especially muting. See below.

Think, instead, of how you would add that announcement over
a soft, slow classical piece. Chances are, you wouldn't
further attenuate it as it would fall into the noise.
Rather, you would mix in your announcement with that signal
unchanged. Probably even attenuating your announcement so
that it isn't overpowering in the context of the other signal!

True, but then I need to be able to make subjective
judgments, I think others have kicked that one around here,
average level, etc.

Instead of manually manipulating the faders we set a
compressor to reduce the level of the music bed to where we
wish it when the voiceover is present. Then, we feed the
signal from the voiceover channel to a "side chain" on the

When the voice channel has signal the compressor acts on the
signal which is the music bed. otherwise, the music bed is
"up full" and we can then just play our multitrack and bang
out our commercial. Saves the mixer time, etc.

Yes. You also typically have a choice of what signals you want to
use to control the process. I'm stuck with just the two.

True, wehre if it chooses, in your case, what level to do
the door annunciator just as the piece you're listening to
does a nice crescendo, say byebye doorbell annunciator
signal. This is why I'm thinking ducking as one viable

The musicians are (or can be) intelligent agents in the
process. The mixing console, by itself, can't.

Uh huh, my point exactly!

By contrast, a doorbell has no awareness of the music playing
under it. Nor does the music have any awareness of the
doorbell trying to sound *over* it.
HEre's the problem, and then there's all the different
speakers in all sorts of environments in your house. What
works for the bathroom won't work for the living room, etc.
This is why I keep saying this is a tall order. You're
asking your delivery system to do the jobof a mastering
engineer as much as a mixing engineer, to massage all that
sound to not be uncomfortable or hard to deal with for every
room in your house. NOt saying you can't come up with
usable, but something's going to give, and if it's too
unpleasant to deal with you'll end up disabling it all

But you do this currently. Just in an ad hoc fachion at
the mercy of the particular devices that you happen to

Indeed I do, but *I* get to make the choices. YEs it means
i have to get out of my chair maybe so I can stand up and
turn that one down, or close the door so that what Kathy's
doing in the other room doesn't intrude. You'll note in a lot of places that have public address along with that
dreaded muzak or whatever underneath when the paging
function is activated it mutes the background music
completely. But, this is a single space, and a one size
fits all solution can work. But, let's say you want your
media server to deliver that classical piece to your nice
amp/speakers combination so that you can kick back in your
chair and really absorb the music. The way it processes
what aannunciators, i.e. doorbell, telephone, alert signal
is going to be quite a bit different than it might in the
workshop, or in the bedroom while you're folding and hanging up clothes.

I.e., you turn down the volume of the T V if you are
trying to listen for the clothes dryer's signal.
You locate telephones where you can hear their ringers
(or, god forbid, *carry* them on your person!). You
locate your HiFi in whichever room you most want to
listen to it. You adjust your behavior to accommodate
the constraints imposed by those devices. I don't like
having to adjust my lifestyle to the misconceptions of some
faraway product designer.

Ah which means you want to build the processing in to each
individual networked speaker. HOw you'd implement that one
i have no clue! <big grin>. I'm an audio guy, but that's
the only way I can see to get there from here <grin>.

Also for Don, a little mental exercise for you related to
your quest that I hope will help.

I'm mixing sound for a live performance, and for the butts
in the seats in that room to consume. I'm listening to the
sounds in that room and creating a blend which delivers the
blend I wish to most of the seats there, at least as much as it
can be. i can hear the speakers that are delivering the
sound to those butts in seats.

Now consider the alternative ...

I'm in my remote truck, mixing the sound to be consumed by
folks listening via a radio or television broadcast. Some
are going to be listening on higher end systems, 2 and 3 way
speaker systems, stereo, etc. etc. Others may be listening
in mono on a tabletop radio or television with a small

My approaches are quite a bit different. In the remote
truck I'm going to be switching my listening chain every now and
then to judge all these variables, akin to doing the job of the
mastering engineer at the same time I'm mixing live.
I'm also going to switch from stereo to monaural to check
for mono compatability, etc.

<are we having fun yet?>

But that's a different sort of goal. Do you spend an
equivalent amount of time ensuring that your doorbell
"sounds right" in each room in your house under the
same number of different environmental conditions?
Or, do you just say, "Well, at least I can *hear* it
in this room..."?

IN some ways yes, but, were I living with such a system I
still don't want it to degrade the primary program I'm
listening to, whether that be the maritime mobile service
net on 14300 khz or MIles Davis, or the NPR afternoon news.
i also don't want the annunciator to blow my head off.

At the live show I can walk the room, see how decisions I
make at the mixing console translate to the butts in seats
in different parts of teh room. Mixing in the remote truck, say for that sporting event, when we go from the play by
play announce team to the cute blonde who's interviewed the
star player who's about to break the record I want to make
sure that the audio from that interview enters the sound
stage seamlessly while the players take their time out, then when we go back to the announce team with the cheering crowd there's teh obvious transition, a little higher energy ...

The human at the controls still gets to make the choices
though, instead of letting the machine do it.

Consider for example those tape recorders and camcorders
with automatic level control on the recording. When you
lsiten to recordings made with those devices you can hear
the electronics 'hunting" at times, and that pumping sound
as the noise floor comes up, then somebody talks close to
the pickup, the noise floor drops, etc. etc.

So let's consider you're sitting at your desk. the doorbell rings. yOur wife answers the door, gives the Jehovah
witness the bum's rush. About the time the automation has
returned you to your regularly schedule program at the level you wish to listen the annunciator tells you that your
baking project has three minutes to go, the automation
probably has to still do a little bit of adjusting, <hmmm>

But, long and short of it is, that mixing capability needs
to be incorporated in every one of yoru network speakers.
That requires more software writing guru than I'll ever hope to be <grin>.

Guys like Hank and i aren't so much telling you that you
can't get there from here, just that there's lots of bumps
on the way to it even if you'd code it up yourself.

NOte here your discussion with Ron, it appears simple to the end user, such as your car. But, to get that simple
interface that delivers what you need, and what you want is
a major coding undertaking. You're basically wanting to
build a full blown digital audio workstation into each
networked speaker's firmware.
| Remove .my.foot for email
| via Waldo's Place USA Fidonet<->Internet Gateway Site
| Standard disclaimer: The views of this user are strictly his own.