9 Feb
2014
9 Feb
'14
10:08 p.m.
Hi,
I started the thread "How is the bass mixed? Per-channel frequency
analysis? Histogram?" a few weeks ago. One suggestion that came up a
few times was to simply downmix the bass to mono before sending it
out.
How would I go about this? Currently, this system doesn't have any
audio. What I intend to do is send audio out over USB to a DAC.
Ideally, I'd like to do on-the-fly remixing of only "bass frequencies"
to mono. Ideally the frequency should be configurable, but I think
about 150Hz or below is a good start.
If possible, I'd like to keep things simple, and have only ALSA and
MPD running. As far as I can tell, I don't think I can do this at the
MPD level.
I saw an example where I can easily downmix all output to mono. I'm
not opposed to that, but it would be nice to selectively downmix only
the lower (subwoofer) frequencies.
Thanks again!
Matt
9 Feb
9 Feb
10:27 p.m.
On Sun, Feb 09, 2014 at 03:08:41PM -0600, Matt Garman wrote:
I started the thread "How is the bass mixed?
Per-channel frequency
analysis? Histogram?" a few weeks ago. One suggestion that came up a
few times was to simply downmix the bass to mono before sending it
out.
How would I go about this? Currently, this system doesn't have any
audio. What I intend to do is send audio out over USB to a DAC.
Ideally, I'd like to do on-the-fly remixing of only "bass frequencies"
to mono. Ideally the frequency should be configurable, but I think
about 150Hz or below is a good start.
If possible, I'd like to keep things simple, and have only ALSA and
MPD running. As far as I can tell, I don't think I can do this at the
MPD level.
You really need to provide a bit more information here.
If your sub is an active one (with a built-in amplifier)
it will have the required filter. And probably it will
have two inputs which it will add to mono. So you just
connect the two outputs of your DAC.
Many active subs also provide outputs for the amplifier
of the main speakers, which will have the bass (which is
used by the sub) filtered out. In that case just connect
your main amplifier or active speakers there.
If your equipment is different, then at least provide a
clear and complete descriptn, nobody will be able to
help without that.
Note that it is normally more important to remove the
bass signal for the main speakers than to remove the
mid/high part for the sub.
Ciao,
--
FA
A world of exhaustive, reliable metadata would be an utopia.
It's also a pipe-dream, founded on self-delusion, nerd hubris
and hysterically inflated market opportunities. (Cory Doctorow)
10 Feb
10 Feb
12:52 a.m.
On Sun, Feb 9, 2014 at 3:27 PM, Fons Adriaensen <fons(a)linuxaudio.org> wrote:
On Sun, Feb 09, 2014 at 03:08:41PM -0600, Matt Garman
wrote:
Sure, no problem. What I have is:
(1) Linux server with a media collection in a basement network closet
(2) A powered subwoofer on the main floor, with only a single coax
run between the sub and network closet in the basement
(3) In-ceiling speakers on the main floor, all wires terminating
in same network closet
So the fundamental problem is that there's no physical locality
between the server and the speakers. What I was thinking was to use a
USB DAC that has dual outputs. One output would go to a power amp,
which in turn would go to a speaker selector (so I can power multiple
sets of speakers).
The DAC's other output needs to go to the sub. But there's the
problem: I only have a single coax run, so I can only send one
channel. This means, either pick one and hope most music doesn't pan
the bass (the point of the other thread I started), or come up with
some way to mixdown the signal to mono so I can pass it over the
single conductor. I thought about doing this in hardware, see this
thread:
http://www.audiocircle.com/index.php?topic=123314.0
But it occurred to me I could possibly spare some expense by mixing
down the signal at the server level.
In reality, based on the previous thread, I think I could probably
send just one channel and be OK most of the time. But if I can easily
get it 100% of the time, why not? :)
Note that in this solution, I'm sending the full-range signal to both
the in-ceiling speakers and the sub. The sub has a builtin crossover,
and the speakers can handle the full signal.
As a side note, if you read the AudioCircle thread above, you'll see I
already have an AVR that can send a bass signal over coax... my goal
is to get rid of the AVR though, as I have more speakers than it has
outputs (would rather not use the AVR and speaker selector), plus I
want to move control off the AVR and onto the Linux server (mpd).
Let me know if anything is unclear or if you have any other ideas!
Thanks,
Matt
How would I go about this? Currently, this
system doesn't have any
audio. What I intend to do is send audio out over USB to a DAC.
Ideally, I'd like to do on-the-fly remixing of only "bass frequencies"
to mono. Ideally the frequency should be configurable, but I think
about 150Hz or below is a good start.
You really need to provide a bit more information here. 
1:52 a.m.
I know you said you wanted to keep it simple, but I'd like to point out
that I have a similar setup running at home. I listen to all my music with
a guitar amplifier for the treble and a bass amplifier for the bass.
Obviously, I have to downmix the signal to mono (I don't miss the stereo
sound, I just mind hearing every single instrument as clear as possible).
The signal coming from my music player is routed into 2 separated EQs which
I use to filter the high frequencies from the bass amp and the low from the
guitar amp. You could use highpass and lowpass filters for this, but using
EQs also serves the purpose of removing troublesome high harmonics in the
guitar amp, and also, equalizing the output.
All this I do with the calf plugins (never bothered to try any other,
because they've suited me good enough so far) but you can also do that with
any gui-less plugin.
If I had to downmix the bass I'd get the player's output through a lowpass
and then downmix it by connecting left and right outputs of the filter to
the channel that's output to the subwoofer. At the same time, you can route
the untouched signal to the treble speakers adding (possibly adding some
plugin in the middle to compensate the latency difference).
I know this is a bit more involved than you originally planned but I doubt
there is a simpler way with the ALSA tools available, and this way you will
end up with a completely customized setup.
Some of the stuff I described is only based on my trial and error
experience. The enlightened people on the list will surely know how to
polish this method though I use what I have described everyday and it just
works.
Sorry for the long text, have a good one!
On Sun, Feb 9, 2014 at 8:51 PM, Matt Garman <matthew.garman(a)gmail.com>wrote;wrote:
On Sun, Feb 9, 2014 at 3:27 PM, Fons Adriaensen
<fons(a)linuxaudio.org>
wrote:
On Sun, Feb 09, 2014 at 03:08:41PM -0600, Matt
Garman wrote:
Sure, no problem. What I have is:
(1) Linux server with a media collection in a basement network closet
(2) A powered subwoofer on the main floor, with only a single coax
run between the sub and network closet in the basement
(3) In-ceiling speakers on the main floor, all wires terminating
in same network closet
So the fundamental problem is that there's no physical locality
between the server and the speakers. What I was thinking was to use a
USB DAC that has dual outputs. One output would go to a power amp,
which in turn would go to a speaker selector (so I can power multiple
sets of speakers).
The DAC's other output needs to go to the sub. But there's the
problem: I only have a single coax run, so I can only send one
channel. This means, either pick one and hope most music doesn't pan
the bass (the point of the other thread I started), or come up with
some way to mixdown the signal to mono so I can pass it over the
single conductor. I thought about doing this in hardware, see this
thread:
http://www.audiocircle.com/index.php?topic=123314.0
But it occurred to me I could possibly spare some expense by mixing
down the signal at the server level.
In reality, based on the previous thread, I think I could probably
send just one channel and be OK most of the time. But if I can easily
get it 100% of the time, why not? :)
Note that in this solution, I'm sending the full-range signal to both
the in-ceiling speakers and the sub. The sub has a builtin crossover,
and the speakers can handle the full signal.
As a side note, if you read the AudioCircle thread above, you'll see I
already have an AVR that can send a bass signal over coax... my goal
is to get rid of the AVR though, as I have more speakers than it has
outputs (would rather not use the AVR and speaker selector), plus I
want to move control off the AVR and onto the Linux server (mpd).
Let me know if anything is unclear or if you have any other ideas!
Thanks,
Matt
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user(a)lists.linuxaudio.org
http://lists.linuxaudio.org/listinfo/linux-audio-user
How would I go about this? Currently, this
system doesn't have any
audio. What I intend to do is send audio out over USB to a DAC.
Ideally, I'd like to do on-the-fly remixing of only "bass frequencies"
to mono. Ideally the frequency should be configurable, but I think
about 150Hz or below is a good start.
You really need to provide a bit more information here.
2:39 a.m.
On Sun, Feb 09, 2014 at 05:51:54PM -0600, Matt Garman wrote:
So the fundamental problem is that there's no
physical locality
between the server and the speakers. What I was thinking was to use a
USB DAC that has dual outputs. One output would go to a power amp,
which in turn would go to a speaker selector (so I can power multiple
sets of speakers).
The DAC's other output needs to go to the sub. But there's the
problem: I only have a single coax run, so I can only send one
channel.
I can't make sense of this.
One output of the DAC goes to the power amp (for the main speakers)
The other output goes to the sub.
So you won't have stereo. In that case, just send the mono signal
(L+R) to both channels of the DAC. If you don't do that, you'll
just be listening to one channel. And then I really wonder why
you bother about the bass - you'll be missing the entire second
channel anyway.
And why do you use *coax* cable between the DAC and the sub ?
A normal screened cable will do...
If you're a bit handy with a soldering iron, all it takes to
mix the DAC outputs to mono would be two resistors and some
short pieces of wire and connectors.
Ciao,
--
FA
A world of exhaustive, reliable metadata would be an utopia.
It's also a pipe-dream, founded on self-delusion, nerd hubris
and hysterically inflated market opportunities. (Cory Doctorow)
8:51 a.m.
On Mon, 2014-02-10 at 01:39 +0000, Fons Adriaensen wrote:
If you're a bit handy with a soldering iron, all
it takes to
mix the DAC outputs to mono would be two resistors and some
short pieces of wire and connectors.
If the OP doesn't use tube amplifiers it should look like this:
http://1.bp.blogspot.com/_BGB-FuSV33o/S7qQlfl_1aI/AAAAAAAAAHY/QO2t7rZP5N0/s…
When using tube amplifiers you might want to add transformers.
9:08 a.m.
On Mon, 2014-02-10 at 08:51 +0100, Ralf Mardorf wrote:
On Mon, 2014-02-10 at 01:39 +0000, Fons Adriaensen
wrote:
PS:
"Thanks for the useful post. I have one suggestion, though: I think it
might be better to use 1K resistors rather than 10K. The input impedance
of a typical amp should be in the region of 20K. At audio frequencies
(where impedance-matching doesn't matter) it is best if the input
impedance is at least x10 the output impedance. Assuming an output
impedance in the range 100-600 ohms (plausible for an MP3 player
headphone output), that would keep the total output impedance below 2K
which should result in more signal voltage across the amp's input." -
http://marcusmorris.blogspot.de/2010/04/stereo-to-mono-converter-how-to-do-…
If you're a bit handy with a soldering iron,
all it takes to
mix the DAC outputs to mono would be two resistors and some
short pieces of wire and connectors.
If the OP doesn't use tube amplifiers it should look like this:
http://1.bp.blogspot.com/_BGB-FuSV33o/S7qQlfl_1aI/AAAAAAAAAHY/QO2t7rZP5N0/s…
When using tube amplifiers you might want to add transformers.
4:49 p.m.
On Sun, Feb 9, 2014 at 7:39 PM, Fons Adriaensen <fons(a)linuxaudio.org> wrote:
I can't make sense of this.
One output of the DAC goes to the power amp (for the main speakers)
The other output goes to the sub.
So you won't have stereo. In that case, just send the mono signal
(L+R) to both channels of the DAC. If you don't do that, you'll
just be listening to one channel. And then I really wonder why
you bother about the bass - you'll be missing the entire second
channel anyway.
I should have said, "One *stereo* output of the DAC..."
I overloaded and loosely used the term "output" in my original
description. What I meant was, the DAC has two *stereo* outputs. In
other words, four total conductors (two "L" outputs and two "R"
outputs).
I intend for one stereo output pair to go to the power amp.
That leaves another stereo pair left for the sub. But that stereo
output pair for the sub requires two conductors, and I only have one
(a single coax run).
And why do you use *coax* cable between the DAC and
the sub ?
A normal screened cable will do...
That's the way the house was built. Pulling more wire is of course
another option, but not one I'm willing to entertain. :)
If you're a bit handy with a soldering iron, all
it takes to
mix the DAC outputs to mono would be two resistors and some
short pieces of wire and connectors.
That's not out of the question, but I just wanted to see if I could
maybe hack it at the software level first.
5:52 p.m.
On Mon, 10 Feb 2014 09:49:05 -0600
Matt Garman <matthew.garman(a)gmail.com> wrote:
I overloaded and loosely used the term
"output" in my original
description. What I meant was, the DAC has two *stereo* outputs. In
other words, four total conductors (two "L" outputs and two "R"
outputs).
This becomes a matter of a simple Y-adapter to combine the two outputs
into one. Don't worry at all about filtering out the signals at the
server end, the subwoofer will have a low-pass filter that will cut out
anything above a certain frequency anyway. On some, the frequency of
the low-pass filter is adjustable.
I believe you're adding a lot of needless complexity to this project.
How bass frequencies are panned is irrelevent, as a simple Y adapter solves
that. Filtering at the server is irrelevant, since the subwoofer contains
a low-pass filter. It's all about getting the signal to the sub using the
available coax run, and if it's a standard 75 ohm cable, you should be fine
as long as you get the right adapters and make sure they're securely fitted
at each end. You could also put new adapters on the cables, as there are
compression RCA connectors but that may require additional know-how.
--
======================================================================
Joe Hartley - UNIX/network Consultant - jh(a)brainiac.com
Without deviation from the norm, "progress" is not possible. - FZappa
6:28 p.m.
On Mon, Feb 10, 2014 at 10:52 AM, Joe Hartley <jh(a)brainiac.com> wrote:
This becomes a matter of a simple Y-adapter to combine
the two outputs
into one. Don't worry at all about filtering out the signals at the
server end, the subwoofer will have a low-pass filter that will cut out
anything above a certain frequency anyway. On some, the frequency of
the low-pass filter is adjustable.
I'm not sure this is right. Based on what I was able to glean from
searching the web, there are a couple issues with this:
(1) Depending on how the DAC outputs are wired, using a Y-adapter
on one set of stereo outputs might also cause the other set of stereo
outputs to be mono as well.
(2) At least a basic resistor network (one has already been
described in this thread) is needed to "correctly" sum the two
channels. Now I can't find the link that gave an explanation of why
this is, but my takeaway was that a simple Y-adapter isn't the way to
go.
11 Feb
11 Feb
8:33 a.m.
Matt Garman <matthew.garman(a)gmail.com> wrote:
I started the thread "How is the bass mixed?
Per-channel frequency
analysis? Histogram?" a few weeks ago. One suggestion that came up a
few times was to simply downmix the bass to mono before sending it
out.
How would I go about this? Currently, this system doesn't have any
audio. What I intend to do is send audio out over USB to a DAC.
Ideally, I'd like to do on-the-fly remixing of only "bass frequencies"
to mono. Ideally the frequency should be configurable, but I think
about 150Hz or below is a good start.
Instead of mpd, you could try my dtas-player component of dtas
(duct-tape audio suite)[1]. dtas-player is nearly as much a *nix shell
as a music player. It is written in Ruby, but runs arbitrary shell
commands of your choosing (sox by default).
I suggest starting with a shell script using sox like below:
(My sox knowledge is rusty, so the following may not be 100% correct:)
------------------------ /path/to/sum-mono.sh ------------------------
#!/bin/sh
set -e
: note: TRIMFX, RGFX, INFILE, and SOXFMT vars come from dtas-player
COMMON="sox \"$INFILE\" -p $TRIMFX"
MONO="remix 1,2 1,2"
: stacking two filters should give a Linkwitz-Reily crossover at 150Hz
XO=${XO-150}
lo="$COMMON lowpass $XO lowpass $XO $MONO"
hi="$COMMON highpass $XO highpass $XO"
exec sox --combine=mix -v1 "|$lo" -v1 "|$hi" $SOXFMT - $RGFX
----------------------------------------------------------------------
Then, after dtas-player is started in a different terminal:
$ dtas-ctl source ed sox command=$HOME/bin/sum-mono.sh
$ dtas-ctl play
At the core, dtas-player is really like running the following in a shell:
source-command | sink-command
Except it's possible to replace source-command without restarting
sink-command (and vice-versa) while preserving the same pipe.
To restore default behavior:
$ dtas-ctl source ed sox command=
[1] - http://dtas.80x24.org/ - a bunch of scripts by me, a GUI-phobe :)
Hopefully I'll have more time for this project soon.
Disclaimer: I was also an mpd developer and even project leader for a
short while, but decided to go off on my craziness. Anyways I find it
much more enjoyable to just glue sox/ecasound/whatever commands together
with pipes than deal with maintaining a music player in C (or C++).
3939
days inactive
3941
days old
linux-audio-user@lists.linuxaudio.org
10 comments
6 participants
participants (6)
-
Eric Wong -
F Tux -
Fons Adriaensen -
Joe Hartley -
Matt Garman -
Ralf Mardorf