JACK2 v1.9.18 release

1. Jack Bridge Software
2. Jack 6 Bridge Software
3. Jack Bridge Software
4. Jack 6.0 Bridge Software
5. Jack 6.0 Bridge Software

Jack is a computer bridge program developed by bridge players for bridge players. One of the developers is former world champion Berry Westra (Bermuda Bowl, Chile 1993). Jack uses state of the art computer algorithms and is the current world champion computer bridge. Jack's possibilities exceed those of all other bridge playing programs. Jack 4.0 Computer Bridge Software (CD-ROM). Over 80 Conventions. Multilingual CD with English, Dutch, French, and German. Jack is by far the strongest and most versatile bridge program available. The most recent version of Jacks bridge engine as well as Jacks Bali 2013 convention card. In Computer World Bridge Championship of 2013 at Bali, Indonesia, Jack played better and faster to keep his 2012 title from Lille, France. Now you can use the same methods with and against Jack. Complete support for Berry Westras teaching methods. Bridge on the computer. You can get Jack through amazon.com, or for as little as $54 at The Bridge World, bridgeworld.com. The Jack 4 demo is available at jackbridge.com. If you want to do more with the program: Jack can read and write PBN files. While it can automatically bid and play a whole deal for you, it cannot do that for a whole file of.

On by falkTX

A new version of JACK2 has just been released.
You can grab the latest release source code athttps://github.com/jackaudio/jack2/releases and macOS/Windows installers athttps://github.com/jackaudio/jack2-releases/releases.

This release focuses on 2 main things: adding zalsa internal client and fixing macOS builds in regards to DB/meta-data.
A few other fixes were made as contributed by developers.

With zalsa now part of JACK2, we are one more step closer to JACK1 feature parity.
Worth noting that this is purely the client, we still need to add the switches to be able to load internal clients directly from jackd startup command-line.
Will be interesting to attempt to make this zalsa client also behave as a “slave adapter” like done with ALSA audioadapter, coremidi and winmme drivers.
Something for a future release.

On macOS, please restart the JACK server after updating to 1.9.18.

The official changelog is:

• Add zalsa_in/out as internal client (based on zita-a2j/j2a and jack1 code)
• Fix jack_midi_dump deadlock on close after the jack server is restarted
• Fix interrupt signal for linux futex waits
• Fix usage of meta-data in official macOS builds (private DB errors)
• Log error message when cleaning previous DB (macOS and Windows)

Now for those unaware of this whole “zalsa” thing…
This is basically zita-a2j/j2a by Fons Adriaensen repackaged as an internal client instead of being an external standalone tool.
JACK1 already added this a few years ago.
This zalsa version is slightly more up to date and with a few new fixes on top.

On supported systems, we can use it by running:

This will activate the “Mio” soundcard as an output in the JACK graph.
You can use jack_unload zalsa_out to deactivate it, or alternatively control the start/stop purely in CLI with:

Which will automatically unload the internal client after Ctrl+C or SIGTERM.

Worth noting that, since internal clients run in the JACK server process,they will log/print information on that process and not on the jack_load one.
So watch out for that if you need to debug anything.

JACK2 v1.9.17 release

On by falkTX

A new version of JACK2 has just been released.
You can grab the latest release source code athttps://github.com/jackaudio/jack2/releases and macOS/Windows installers athttps://github.com/jackaudio/jack2-releases/releases.

This release focuses on improvements to the macOS and Windows binaries, now with macOS universal builds natively supporting Apple M1 hardware.
On M1 hardware, you can now have native arm64 JACK clients talk to Intel/x64 once and vice-versa.

On Windows, JACK-Router (the special ASIO driver that allows any ASIO-compatible application to talk to other JACK clients)has been brought back, now integrated with the installer so it is much easier to manage.
Also, 32-bit and 64-bit clients can now be mixed without any side-effects to the meta-data.

Please restart the JACK server after updating to 1.9.17 on those systems.

The official changelog is:

• Add jack-router to Windows installer, opt-in
• Fix jack_control stopping after first command iteration
• Fix library compat/current version under macOS
• Fix registry keys for Windows, add 32bit compat ones on 64bit
• Fix return codes of jackd on success
• Ignore DB_VERSION_MISMATCH error on windows, it is expected
• Remove old workaround for asio4all, as it breaks with jack-router
• Support for arm64 macOS builds
• Show welcome and license pages on windows installer
• Update QjackCtl used in macOS/Windows installers to v0.9.0, with some commits cherry-picked from develop branch

JACK2 v1.9.16 quick hot-fix (for v1.9.15)

On by falkTX

An issue was found on the last JACK2 release that caused a few applications to crash on start.
This is a quick hot-fix for that release. If you are running v1.9.15, please update.

Because of a new release being needed, we took the chance to remove the jack-router Windows code, as it now belongs in its own repository.
These two items are the full release.

You can grab the latest release source code athttps://github.com/jackaudio/jack2/releases.

You can also grab the latest pre-built installers for macOS and Windows athttps://github.com/jackaudio/jack2-releases/releases.
On other systems, ask your jack2 package maintainer to update to the latest release.

JACK2 v1.9.15 release and current status

On by falkTX

A new version of JACK2 has just been released.
You can grab the latest release source code athttps://github.com/jackaudio/jack2/releases.

This version of JACK2 brings back binary builds for macOS and Windows.
You can grab the latest pre-built installers athttps://github.com/jackaudio/jack2-releases/releases.

A few notes about the new installers:

• This is the author’s first time on doing such installers, there might be some small oddities
• If you have old JACK2 installations on macOS or Windows, please uninstall those first (they are not 100% compatible)
• JackRouter is not included, as it does not work on macOS 10.15 and can be problematic on Windows.
  New maintainers welcome! Existing code moved to separate jack-router repository.
• macOS 11 (the new arm64 “apple-sillicon”) is not supported at the moment
• Report any issues to jack2 repository (the jack2-releases repository is for automated builds only)

The official changelog is:

• Automated builds for macOS and Windows (see jackaudio/jack2-releases repository)
• Adapt wscript Windows build configuration to match old v1.9.11 installer
• Bump maximum default number of clients and ports (now 256 clients and 2048 ports)
• Delete various macOS and Windows-related files from the source code (no longer relevant)
• Mark JACK-Session as deprecated, please use NSM instead
• Remove unnecessary GPL include from LGPL code
• Split example-clients and tools, as done in JACK1 many years ago (WIP)
• Write Windows registry key during installation, so 3rd parties can find jackd.exe (as HKLMSoftwareJACKLocation)
• jack_control: Fix handling of dbus bytes
• jack_control: Return a proper exit status on DBus exception
• jack_property: Fix possible crash with “-l” argument usage
• jack_wait: Add client name option -n/–name
• Fix compilation of documentation
• Fix compilation of mixed mode with meta-data enabled
• Fix compilation with mingw
• Fix client-side crash if initial meta-data DB setup fails
• Fix macOS semaphore usage, so it works again
• Several fixes for Windows (with contributions from Kjetil S. Matheussen)
• Several minor fixes and grammar corrections (with contributions from Adam Miartus and Timo Wischer)

This release is a nice good step towards what was planned back in 2017 with the v1.9.12 release.
JACK1 remains in bug-fix mode, while JACK2 is slowly incorporating changes from it.
The examples and tools directory was split in the same way that JACK1 is,so that in the future we can use those as submodules instead of having duplicated code between JACK versions.
Once we have matching headers, examples and tools we can begin the switch.
If you are reading this and want to help, pull requests for syncing those repositories code to what JACK2 has will be greatly appreciated!

For those that missed it, JACK-Session has been marked as deprecated. Please use “New Session Manager” instead.
You can find more about the project here.This includes links to the source code, bug/issue tracker and documentation.

For the future, short-term plan is to continue the feature parity between JACK1 and JACK2.
We are pretty much there now besides JACK2 missing the a2jmidi and zita internal clients (and the special options to use them).
Support for macOS arm64 is in progress, and there is a whole range of pull requests and issues to investigate and fix.
Focus will be on those before any features are even thought of, unless they serve to improve something.

JACK-Session has been marked as deprecated

On by falkTX

A small heads-up for developers, JACK-Session API has now officially been marked as deprecated.

I have been working together with other members of linuxaudio community to get a better session-management API out there.We previously had difficulties with this due to issues regarding the upstream project.It was obvious to us that the NSM API is superior to all others that came before it (thanks to the initial hard work of its author, Jonathan),but working with upstream has proven immensely difficult or even impossible..

As of a few weeks ago, a few members of the linuxaudio group decided to fork the original NSM project and make a new one.The acronym remains, so we still call it “NSM”, but the new project is now “New Session Manager” while the original is “Non Session Manager”.We are working together with the community to make this session-management API and tools the best as they can be.(The original was a bit lacking on the user-experience side of things..)

You can find more about the “New Session Manager” project here.This includes links to the source code, bug/issue tracker and documentation.

Please give it a look, or better, a chance.For now it is specific to POSIX systems like Linux and BSD, but it can be extended in the future.If we manage to make (Audio) Session Management really work in Linux, that is already a big step up from other tools/APIs that came before it.

We recommend all JACK application developers to implement NSM in their software.Thank you.

[ News Archive ]

By Thorgal - May 3rd 2010

Contents

• 2 The ALSA Loopback 'Sound card'
• 3 Building an asoundrc file
• 4 The Jack Bridge
• 5 Alternative Setup: hardware and software based solution
• 6 Measuring the latency introduced by the Loopback device

Introduction

Some people may use PCs where the Jack Audio ConnectionKit is running all the time. As one of theseusers myself, my DAW PC uses a very lightWM (fluxbox) without any of the audio layers provided by the morefeature rich WMs (like KDE or Gnome). I have jack started at login andhopefully, it never goes down until the next PC shutdown. Since I do notwant any other audio layer such as Pulseaudio, etc, in between Jack andALSA in my case (but it could well be FFADO for firewire devices), thishas the slight disadvantage of making non jackified applicationsunusable.

So how can one provide a permanent bridge between non jackifiedapplications and Jack? Well, there are different ways. One can bepurely hardware: enable another soundcard (e.g the onboard sound chip)and physically link it to your DAW sound-card if you are like me with adedicated audio h/w for DAW operations. Let this extra soundcard be thedefault (ALSA index 0) so that apps like flashplayer, skype, etc, use itby default. However, this may limit the number of h/w IOs of your DAWsoundcard for your pro work. I do not like this physical link because myRME Multiface II has only 8 analog mono INs, while the onboard soundchip (Intel HDA) has no digital output that I can link to the Multifacedigital input. I would be forced to patch two INs of the Multiface tothe stereo output of the onboard chip. It is too expensive to considerin terms of physical IOs.

The alternative is a software solution or a mix h/w - s/w (as used in myfinal setup). So, I was looking for a solution in the form of permanentJack clients, playback and capture ideally, or at least playback since Ican use the capture device of a second soundcard like the onboard chipor anything else (I will clarify this further down but it is notnecessarily an average setup). In terms of software solutions, the ALSAjack PCM plugin is in my opinion not ideal because the Jack client willdisappear as soon as the application stops outputting audio.Furthermore, this PCM plugin has not been updated (except lately byTorben Hohn but the patch is not widespread) and I found the pluginquite buggy / unstable in many situations.

It is not until recently, as I was fiddling with the ALSA Loopbackdevice, that I saw a way to achieve what I needed.

The ALSA Loopback 'Sound card'

The ALSA Loopback sound card is a virtual soundcard that is created oncethe ALSA kernel module snd-aloop is loaded. This virtual soundcarddevice, as its name indicates, sends back the output signal ofapplications using it back to itself, so one has a chance to e.g. recordthis signal from the same device. Simply imagine that you have aphysical link between one OUT and one IN of the same device.

By default, the Loopback sound card consists of 2 devices, each composedof 8 subdevices. Once the kernel module snd-aloop is loaded, you canverify that the sound card has been created:

Note that you can control the number of subdevices with the moduleoption pcm_substreams (8 by default). You can always set it to 2 onlyif you wish at loading time. As an example, here is my ALSA moduleconfig file (/etc/modprobe.d/sound.conf on my debian-based DAW)

As you can see, I do fix indexes even though ALSA and Jack can work withnames only. It is motivated by the special position that is index 0, theALSA default device that flashplayer will try to use.

Compiling snd-aloop if needed

Update: it may not be needed any longer as of kernel 2.6.38 ...

It may well be that the ALSA Loopback kernel module was not included inyour distribution's kernel package (it is the case in e.g. debian, asfar as I know). This is no bother as we can easily compile it. Note thatthere is no way around since the loopback ALSA module is not part of thekernel baseline in general. So unless your kernel packager had done thefollowing work, you will have to do it yourself ...

Warning: I tried alsa-driver 1.0.21 against 2.6.33.5-rt22 and while itcompiled fine, it would not load at all, even when forced. So don'twaste your time with this version combo.

Make sure you really don't have it installed. Better check that not :)

If modinfo reports nada, time to check that you have installed thekernel headers corresponding to your presently running kernel. I'llleave this to you as this is very distro dependent. In debian baseddistros, the package is called something like linux-headers-xxx andmust match the installed kernel (package linux-image-xxx).

Time to make a backup of the installed kernel modules. Example;

Prerequisite: you of course need a compiler and other tools. In debianbased distros, you can check that you have a package calledbuild-essential installed:

If not, just get it:

Now grab the alsa-driver source code (same version as your installedALSA, in my case 1.0.23 which I will use in my description) from theThe ALSA website, uncompress, untar itand cd to the alsa-driver top dir. Here is a command summary ...

Now you have to configure the source package for compilation. To helpyou, look at what ALSA modules are currently loaded:

And check what card they correspond to by typing

You will see a big list of possible cards. Pick the ones you areinterested in. As an example, this is how I configured the alsa-driversource on my DAW system:

and on my laptop:

So, feel free to configure it the way you want it. Once you haveconfigured the ALSA driver source, you just go through the usualsequence:

It will normally install all the compiled modules into the correctlocation of your kernel installation. Now check that the kernel knowsabout the loopback module:

Allrighty, time to load it. But before that, shut down all audio apps(including firefox). Once done, do this:

Now, see if it works:

If all was cool and dandy, just add snd-aloop in /etc/modules. (Ifyou wish, you can give the loopback soundcard another name than'Loopback' in a modprobe option but I kept the default throughout theentire HOWTO and there is no need to change it.)

In case anything went wrong and you wish to go back to your previousALSA installation, no problem:

Understanding the ALSA Loopback sound card structure

Well, this is not too difficult to grasp. This virtual sound cardconsists of 2 devices:

• hw:Loopback,0
• hw:Loopback,1

If an application outputs its audio to one of the subdevices e.g. sayhw:Loopback,0,0 the audio will be available as input in thecorresponding subdevice hw:Loopback,1,0 because the whole point forthis card is to send the signal back to itself.

So the generic principle is that an output signal to subdevicehw:Loopback,i,n becomes an input signal from hw:Loopback,j,n with

Building an asoundrc file

The goal is to create a default ALSA plug device out of the Loopbackcard. For a complete software solution, we need one PCM playback, soALSA apps can send audio to it, one PCM capture, so ALSA apps can getaudio from it, and combine these 2 PCMs into a nice full duplex 'plug'device.

Note that the underlying goal is this: I want the audio of my jacksystem capture ports (from my RME card) to be available at the ALSAcapture device and vice-versa: hear from my jack system playback portswhat ALSA apps are playing back to the ALSA playback device. Tricky...

asoundrc definition

The asoundrc below should work in most situations.

In summary:

This asoundrc is very generic and one can of course tailor it in termsof sample rate, audio format, buffer size, etc. One can find therelevant parameters in the ALSA-libdocumentation.

Some html5 browsers (i.e. Firefox 30; Chrome 38) will fail to open the pcm.!default device for audio playback. This can be fixed by using pcm.card0 instead:

Here is an example applicable to my DAW. I left some notes so youunderstand the extra stuff I also removed unnecessary dsnoop's anddmix's because when you analyse things a bit more, you realize that someof the ALSA PCMs will only be used by one single client (alsa_in/out)so there is no need to use dmix / dsnoop. Dmix only makes sense for theALSA playback PCM because you can have more than one client outputtingto ALSA at the same time. Anyway, note the hardware parameters I haveadded so that it matches my RME Multiface II requirements. For the dmixbuffering parameters, read on below.

Testing our new default ALSA device

Save this asoundrc config into $HOME/.asoundrc but make sure beforethat you are not overwriting an existing asoundrc file (back up whateveryou have if it already exists).

OK, now we can test it from the command line. If jack is running on yourother hardware (RME card in my case), you of course will not hearanything since we have not bridged yet our default ALSA device to thejack graph.

You can use another app (aplay for example). The idea is that an ALSAapp using the default device we have just created will not spit errormessages and will play along nicely. Try for example lmms using theALSA default :)

The Jack Bridge

OK, this is where it will get a little bit confusing because of theloopback nature of the virtual device ;)

Creating permanent Jack clients using alsa_in and alsa_out

Since we used subdevice 0,0 for playback and subdevice 0,1 for capture,remember that signals from these subdevices will be available byloopback to the corresponding subdevices, respectively 1,0 and 1,1 inthis case. So the trick for jack is to use alsain and alsaout on thelatter subdevices :) Brilliant ins't it ? :D

Let's do it from the terminal

I hope you start to see the underlying idea. Once these clients show upin the graph, when an ALSA app plays back to subdevice 0,0 (default ALSAdevice defined in our asounrdc), the signal will be available insubdevice 1,0, which alsa_in listens to. The 'cloop' client we createdcan now be connected to the jack system output ports and o miracle, youwill hear your ALSA app :)

In order to avoid the warning messages from alsa_in/out, you can addthe relevant parameters, e.g. (my case):

On the other hand, if you connect a jack system input port to the'ploop' client created by alsa_out, the signal is sent to loopbacksubdevice 1,1 which will be looped back to subdevice 0,1. This subdeviceis nothing but our ALSA capture device, defined in asoundrc :). So nowyou can record say your bass or guitar or voice (from your jackhardware) to an ALSA app that does not support jack. I tried skype, andit works just fine. You can also try the command line app calledecasound, which does support jack but also ALSA. It is a VERY convenienttool to have around (see further down).

The beauty of it is two-fold:

• permanence of the 'cloop' and 'ploop' clients (if you shut down yourALSA app, cloop and ploop will remain, always listening
• if jack crashes, it will bring down cloop and ploop but will notdisrupt the ALSA apps since they only talk to the loopback soundcardwhich is completely independent of the jack environment :D

Create scripts to automate bridge initialization via QjackCtl

The creation of the 'p/cloop' clients can be automated, together withtheir connection to jack system ports. Here is my script:

Note that I used -q 1 as an option to alsa_in/out. This has to dowith the resampling quality. At 2.3ms latency, 96kHz s.r. on a 2 x 2.4GHz dual core CPU system and using Jack2, I get a low CPU usage (1-2%)and the quality is reasonable. If you push it to 2, 3 or 4, the CPU willincrease quite a lot at small buffering / latency

In qjackctl (which I use, YMMV), go to Options -> Execute afterserver startup and add

OK, now that you have added all this, save the qjackctl config, quit andrestart it. Start jack, you should see the ploop and cloop clients inthe graph with the connections between the ports we chose in theloop2jack script.

Test it: open say lmms, load a demo project, play it :)

And voila! try skype, which you can record in ardour if you want (don'tforget to connect your jack system mic directly to the ardour trackwhich you had connected to the 'cloop' client or you will miss recordingwhat you are saying to the other person ;)

So this was a pure software solution and this has the benefit that allyour jack input ports are available to the ploop jack client so thatALSA apps can record the audio coming from these jack ports via thelooped-back device. Of course, the loop stuff has latency (the defaultdmix and dsnoop buffering is quite big), but who cares ? ... Wellactually, I did care a little so I revisited some things and estimatedthe latency added by the Loopback device. I also tweaked a hybridsolution where the ALSA capture PCM is using a real hardware (onboardchip or extra soundcard). Just read below.

Alternative Setup: hardware and software based solution

As mentioned in my introduction, I happen to have an onboard chip (IntelHDA) but also a USB webcam with a built-in mic. It would be a shame notto use their recording capability in some way, especially since I tendto use skype from my DAW PC quite often.

Jack Bridge Software

Adding extra h/w inputs in asoundrc

So instead of using the Loopback device for the ALSA capture (all thestuff related to 'ploop' in the previous asoundrc), I simply declaredthe extra h/w in the asoundrc. So I removed all the ploop stuffincluding the now useless Loopback subdevices used for ALSA capture andalsa_out, and added hw PCM devices on the Intel device and USB webcam.

Testing the new ALSA capture

This one was easy to test. I made sure that my asoundrc default cardused the 'usb' pcm capture (see above). I then fired up skype, set it touse the 'Default' device for everything. I fired QasMixer (a nice QT4based mixer if you don't know it) and controlled the capture level ofthe USB webcam from there (I do not let skype control my levels). Then Itried the skype test call and made sure it recorded my voice. The webcamis by the way a Logitech Webcam C310 which I am satisfied with. Worksout of the box in linux due to its class compliance with USB vid.

Note that the same thing can be done with the Intel HDA capture ('intel'pcm capture defined in the previous asoundrc) provided that you plug amic to its input jack of course :)

Measuring the latency introduced by the Loopback device

For measuring the latency, I had to be able to provide both Jack andALSA with a common audio source.

Playback only

First, I fired up jackd and alsa_in on 'cloop' (just as before). Then,I used ecasound as the middle-man for allowing the measuring of theeventual delay in ardour (which I am comfortable with, you can of courseuse another jack enabled recording software if you want).

Here is the ecasound command line, very simple:

Then in qjackctl, I connected a sound source like my microphoneavailable at system:capture_3 to ecasound:playback_1/2.

In ardour, I created two tracks: one mono track accepting audio from thesame system capture port, and a stereo track connected to the 'cloop'client. Indeed, since ecasound outputs to the default ALSA device, thecloop client should have the audio by loopback. I then recordedsomething in ardour so both tracks contained data coming from the sameaudio source. I compared the resulting waveform and observed a delay of120 ms when the dmix parameters are set to default.

Another way is to use the click sound ardour provides, instead of amicrophone as mentioned above. Just connect the ardour click outputports to ecasound's input ports, and connect the click ports to one ofthe ardour tracks, the other one should still receive the cloop clientdata. If you fiddle with the dmix parameters in the .asoundrc, you willobtain various delays. It is therefore up to you to decide how theperiodsize and buffersize params must be set.

A 120 ms delay is not bad at all considering the huge buffering dmixconfigures by default. But remember that dmix really sucks at smallbuffering so you have to choose a reasonably large one. I ended up usingthe following:

This setting gives me a final Loopback latency of ~ 35ms, while dmixdoes a good job without choking.

Capture and Playback

If you are using the complete software solution (ALSA playback andcapture via cloop and ploop), then you can still use ecasound as anintermediate tool. Just fire it up in this way:

In qjackctl, connect the ardour click ports to the 'ploop' ports. Thiswill allow ecasound to record the ardour click via the looped-back ploopaudio. Then ecasound will output it to the default ALSA pcm playbackwhich alsa_in collects via the cloop client.

In ardour, just like the setup above, have two tracks, one receivingthe internal ardour click directly, the other connected to the cloopports. Arm the tracks for recording, enable the click, activate thetransport. You will see audio data in both tracks, one is delayed ofcourse (the one connected to cloop). With my tuned asoundrc, I get anoverall Loopback latency well below 100ms (approximately 75ms). It isnot bad at all for the whole purpose of the Loopback bridge.

Jack 6 Bridge Software

If low latency is a concern, don't use ALSA only apps, use jackifiedapps :D

Jack Bridge Software

Final word

I hope all this was clear enough. The idea behind this was to use a h/wcapture device instead of the Loopback device. This reduces the role ofthe Loopback device to ALSA playback only, and removes the need ofalsa_out, sparing some CPU and jack process cycles. At the moment, I amusing my USB webcam for capture because I only need ALSA capture forskype. The Intel HDA is available as well but I don't really need it. Itis connected to my patch panel though, so I can always use it if theneed comes (unlikely).

Jack 6.0 Bridge Software

Troubleshooting

Jack 6.0 Bridge Software

If you have pulseaudio on you machine, better kill it, overwise it doesn't work.