Hello,
I just did some research but couldn't find any satisfying answer and I am not
an BT specialist. So before digging deeper in the BT-world I want ask here if
anybody already has an solution and if it could work theoretically ?
Here my state of research:
Since BLuez >= 5.0 dropped ALSA [1] I tried using pulseaudio (can route it
trough jackd). I connected a BK8000L module (sure hifi) via bt-manager and
pulseaudio works fine. I try to connect a second, the first one gets
disconnected so only one is audio device seems to be possible at the same
time.
As a second solution I tried the alsa implementation: bluez-alsa [2], but
I didn't succeed to use it via jackd (didnt try hard) and it has the same
behaviour with aplay on an first try on command line.
Any ideas are appreciated.
mfg
winfried
Purpose: simple distributed audio-system adding BT speaker for sound
installations, the first try two BT-Speaker used each mono, forming a stereo
pair.
[1] https://git.kernel.org/pub/scm/bluetooth/bluez.git/commit/?
id=4ff9b99292eca193dc0c149722328cb0b1ab0818
[2] https://github.com/Arkq/bluez-alsa
--
--
- ao.Univ.Prof. DI Winfried Ritsch
- ritsch(a)iem.at - http://iem.at/ritsch
- Institut fuer Elektronische Musik und Akustik
- University of Music and Dramatic Art Graz
- Tel. ++43-316-389-3510 (3170) Fax ++43-316-389-3171
--
Hello to users of the various Linux sound systems!
I have a rather technical question regarding the implementation of
multichannel recording and playback :
Is the relative alignment of different channels exactly constant? I mean,
can I reliably assess ultra short differences between multiple channels,
e.g. if one channel's signal is a delayed signal of another channel by
about 700 ns and I want to calculate the exact delay of the signals
relative to each other?
I'd also need to pass exactly timed signals to the outputs.
Internally for the inputs and outputs I would probably choose a platform
with I2S, but other embedded protocols would be alright too, if I can
match them to the ADCs chips outputs and DACs chips inputs.
So what I need to know is not about the delay from input to output, but
rather how well and exact the synchronization of multiple channels works.
I would be happy about hints and answers.
Regards
Christian
On Aug 21, 2017 19:13, "Jonathan E. Brickman" <jeb(a)ponderworthy.com> wrote:
>
> On 08/22/2017 12:04 AM, david wrote:
>>
>> On 08/21/2017 06:51 PM, Jonathan E. Brickman wrote:
>>>>
>>>> On 08/17/2017 07:28 PM, David Jones wrote:
>>>>
>>>> And to think all I did to get my cheap USB sound card to work at
>>>> 5-10ms latency was install a low-latency kernel (from KXStudio) and
>>>> change buffer and period settings on QJackCtl's non-Advanced
>>>> settings tab.
>>>>
>>>> That's on two different systems: one Intel i7, the other an AMD
>>>> Phenom 2. No tweaking deadline scheduler or scaling governor (for
>>>> the Intel, Phenom 2 has no such capability).
>>>>
>>>> But I'm using Debian Testing + KXStudio, not ArchLinux or Ubuntu.
>>>
>>> David, what desktop are you using? LXDE? Other?
>>
>>
>> XFCE. One difference that might affect your set up: My card only does
>> 48000Hz max and has 2 stereo channels in and out.
>>
> That's very interesting. This time around, XFCE gave me xruns which I couldn't eliminate, LXDE has been flawless. Am running 48 kHz these days, it looks like I'd have to spend a good bit to get 96.
My card cost 20US$, so I'd have to spend a lot to get 96, too. I'd like more channels, though.
My laptop has 2 USB3 ports (indicated by blue connectors) and 2 USB2 ports. I have a USB3 ext drive attached to one, the soundcard to the other USB3 port.
Every time I've tried LXDE, I've been frustrated by difficulties configuring it. And it would freeze even without RT audio. Glad you've had good luck with it.
I don't remember what background services XFCE installs by default.
David W. Jones
gnome(a)hawaii.rr.com
authenticity, honesty, community
http://dancingtreefrog.com
Hi, been inactive for a while.
I just up/cross-graded to Ubuntu Studio 16.04 -- very much enjoying it so far! I'm getting better USB audio performance than I could ever manage under plain Ubuntu (provided WiFi is off, but I can live with that for shows).
One problem that the upgrade didn't solve is that sometimes jackd gets stuck and it can't be killed.
Initial symptom: No audio, in or out.
Secondary symptom: In qjackctl, the CPU usage number is frozen -- never changes.
At that point:
$ ps x | grep jackd
3140 ? SLsl 0:16 /usr/bin/jackd -dalsa -r44100 -p1024 -n2 -D -Chw:Set,0 -Phw:Set
4458 pts/1 S+ 0:00 grep --color=auto jackd
$ kill -9 3140
$ ps x | grep jackd
3140 ? SLsl 0:16 /usr/bin/jackd -dalsa -r44100 -p1024 -n2 -D -Chw:Set,0 -Phw:Set
4460 pts/1 S+ 0:00 grep --color=auto jackd
"kill -9" has no effect.
"sudo kill -9" has no effect.
If I reboot, it takes several minutes for the system to shut down -- i.e., jackd is even preventing a clean shutdown.
So, the question: How do I regain control when this happens? (Ideally, without having to close everything.)
Thanks,
hjh
V 1.5.3 - Swift
Small, streamlined and fast moving.
We have revised the whole of Microtonal (scales) for better accuracy, and
fixing originally incorrect range limitations. This is now much closer to the
Scala specification, although there seems to be an ambiguity when you have a
keymap defined and have an inverted scale - which key is the pivot point?
The GUI representation has been improved so you always see what you entered,
not an approximation with strings of '9's!
Microtonal settings are now fully accessible to the CLI.
Vector control has had a makeover. Amongst other things, the name field is now
editable and is stored when you save. This means it will be retained on patch
sets and states as well, so when these are reloaded you will know what vectors
are embedded.
An entire state file can now be loaded silently in the same way as patch sets
and vectors can.
Further improvements to the CLI are full access to all configuration settings,
and better organisation of command grouping and help lists.
Some data was not ordinarily saved if features were disabled at the time of
saving. We have added a switch to allow all data to be saved regardless. This
makes for larger files of course, but does ensure that you can get an *exact*
recovery if you need it.
You can now directly interact with the formant filter graphic display in a way
that is more intuitive and easier to use.
The Console window now scrolls the right way! One of yoshimi's little helpers
worked out how to scroll the window to keep the most recent line visible at the
bottom.
It is now possible to run Yoshimi stand-alone with both GUI and CLI input
disabled, thus responsive only to MIDI input. In view of this we have added a
new shortform NRPN that will shut it down cleanly. You simply send 68 on both
NRPN CCs (99 and 98).
Under the hood:
As well as additional Gui controls transferred to the new lock-free system,
some of the earlier implimentations of CLI controls have been transferred. One
result is that much of the code is leaner, and easier to follow.
Some needlessly dynamic memory allocations have now been changed to fixed ones.
This gives a noticable reduction to DSP peaks.
There are a few more old and new bugfixes.
To build yoshimi fetch the tarball from either:
http://sourceforge.net/projects/yoshimi
Or:
https://github.com/Yoshimi/yoshimi
Our user list archive is at:
https://www.freelists.org/archive/yoshimi
--
Will J Godfrey
http://www.musically.me.uk
Say you have a poem and I have a tune.
Exchange them and we can both have a poem, a tune, and a song.
On Aug 17, 2017 12:08, Jaromír Mikeš <mira.mikes(a)gmail.com> wrote:
>
>
>
> 2017-08-17 22:10 GMT+02:00 Ralf Mardorf <ralf.mardorf(a)alice-dsl.net>:
>>
>> On Thu, 17 Aug 2017 15:48:18 -0400, Paul Davis wrote:
>> >On Thu, Aug 17, 2017 at 3:21 PM, Fons Adriaensen wrote:
>> >> On Thu, Aug 17, 2017 at 01:59:29PM -0400, Paul Davis wrote:
>> >>
>> >> > parameters which were never really meant to be exposed to users
>> >>
>> >> then the at least the defaults should be OK. They usually are not.
>> >>
>> >I've never adjusted any of the parameters that are now on that tab in
>> >more than a decade.
>> >
>> >There has never been a pull request or a bug report suggesting new
>> >defaults.
>>
>> Why should somebody suggest new defaults? The values could be
>> adjusted, if required and QjackCtl provided most jack parameters, if not
>> all in a single tab, before it was split into two tabs. Much
>> likely LTS release distros still provide the old version of QjackCtl,
>> let alone that jack's defaults provided by package maintainers could
>> vary, I've seen configurations with --clients=foo
>> --ports-per-application=bar.
>
>
> The reason I started this topic is that I bought new USB sound card Focusrite 2i2.
> But I am not happy with latency 15-20ms is not fine for simple task I am using it.
> Just running japa to check RTA.
> Periods/Buffer 3, deadline scheduler, scaling_governor performance
>
> Just trying figure out if I can make somewhat better.
>
> mira
And to think all I did to get my cheap USB sound card to work at 5-10ms latency was install a low-latency kernel (from KXStudio) and change buffer and period settings on QJackCtl's non-Advanced settings tab.
That's on two different systems: one Intel i7, the other an AMD Phenom 2. No tweaking deadline scheduler or scaling governor (for the Intel, Phenom 2 has no such capability).
But I'm using Debian Testing + KXStudio, not ArchLinux or Ubuntu.
David W. Jones
gnome(a)hawaii.rr.com
authenticity, honesty, community
http://dancingtreefrog.com
In the current rebuild of my BNR ( http://lsn.ponderworthy.com ), I
noticed that I could not get qjackctl to cooperate with jackdbus, though
traditional manual startup seems to work very well. The Arch wiki said
that qjackctl is becoming outmoded, and recommended 'cadence', part of
the KXStudio toolset, instead; cadence cooperates with dbus better, but
doesn't autostart, and I'm missing something in the compilation which
will ungray the patchbay autoarrange. Anyone have corroboration and/or
new best practices? I am using Manjaro, but saw the same behavior when I
tested Debian Stable and Testing.
--
/Jonathan E. Brickmanjeb(a)ponderworthy.com
<http://login.jsp/?at=02e47df3-a9af-4cd9-b951-1a06d255b48f&mailto=jeb@ponder…> (785)233-9977/
/Hear us at http://ponderworthy.com -- CDs and MP3s now available!
<http://ponderworthy.com/ad-astra/ad-astra.html>/
/Music of compassion; fire, and life!!!/
> Pretty sure I've seen the same thing a few times, and the solution has been
> to reload the sound card driver. Here's my script to do that:
>
> [kjetil at localhost ~]$ more bin/startdelta.sh
> ...
Interesting. One thing I see is:
> /home/kjetil/bin/alsasound start # ccrma dist script
... but I'm not using Planet CCRMA, so I don't have this script.
Here?
https://ccrma.stanford.edu/~nando/linux/sounddriver
Just want to cover all the bases before doing something that I can recover from only by rebooting.
Thanks,
hjh