Netjack dilema

[Geoff Beasley-2]

Not sure my last email got through so i'll restate my main points again. I use netjack 1 in studio production all the time. My studio is built around it and relies on it.This has been the case since early  2006. It works extremely well. Netjack2 offered some great innovations over 1 : auto detect via multicast, superior stdout to console, control of jack t'sport from any client or the master. The downside is it didn't work very well.  The problem with multicast is you must use a gateway, which is limiting. Much better to be able to bind with ip's directly, but it does work very well. I wanted to run 2   parallel gigabit lans(1 for net 1 for audio) but can't 'cause of the gateway issue with multicast. Transport control from master and slave was too difficult so Stephane disabled it a few weeks ago. The real problem I think goes much deeper as netjack 2 slaves cannot locate at all:they simply return to zero when ever start/stop is sent from the master regardless of masters' act
 ual time.Jack2 is superior to 1 for me here:many fewer xrun events, but netjack1 is far superior as it will locate properly to the master and you can 'hard wire' the ip's easily. It just won't work with Jack2. So there we are:currently the best option would be to add netjack1 to jack2. I know about 'jacknone' but what's the point as it's not currently maintained and Stephane mentioned it was broken as well. Let me know (again i say) if any of the remaining devs need more info or help from me.
best,

g.
 Geoff Beasley
Composer,Arranger,Producer
Laughing Boy Records Australia

_______________________________________________
Jack-Devel mailing list
Jack-Devel@...
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org

[Stéphane Letz]

Le 13 oct. 2009 à 23:01, Geoff Beasley a écrit :

> Not sure my last email got through so i'll restate my main points  
> again. I use netjack 1 in studio production all the time. My studio  
> is built around it and relies on it.This has been the case since  
> early  2006. It works extremely well. Netjack2 offered some great  
> innovations over 1 : auto detect via multicast, superior stdout to  
> console, control of jack t'sport from any client or the master. The  
> downside is it didn't work very well.  The problem with multicast is  
> you must use a gateway, which is limiting. Much better to be able to  
> bind with ip's directly, but it does work very well. I wanted to run  
> 2   parallel gigabit lans(1 for net 1 for audio) but can't 'cause of  
> the gateway issue with multicast. Transport control from master and  
> slave was too difficult so Stephane disabled it a few weeks ago. The  
> real problem I think goes much deeper as netjack 2 slaves cannot  
> locate at all:they simply return to zero when ever start/stop is  
> sent from the master regardless of masters' act
> ual time.Jack2 is superior to 1 for me here:many fewer xrun events,  
> but netjack1 is far superior as it will locate properly to the  
> master and you can 'hard wire' the ip's easily. It just won't work  
> with Jack2. So there we are:currently the best option would be to  
> add netjack1 to jack2. I know about 'jacknone' but what's the point  
> as it's not currently maintained and Stephane mentioned it was  
> broken as well.

jacknone contains several stuff : netjackone (seems ok), improvement  
on "adapters" that do not work for me, some change in JACK server  
engine internals that I don't want in trunk for now.

> Let me know (again i say) if any of the remaining devs need more  
> info or help from me.
> best,
>

The best I can do is to extract "netjackone" stuff from jacknone  
branch and merge in trunk. Since Torben seems to have disappear, I'll  
have to hack a bit to do that. I'll try to do that ASAP.

Stéphane
_______________________________________________
Jack-Devel mailing list
Jack-Devel@...
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org

[Phil Rhodes]

I don't suppose any of this will make it feasible to send audio over a wifi
link, will it?

P

_______________________________________________
Jack-Devel mailing list
Jack-Devel@...
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org

[James Warden]

Hi,

Why isn't jack_(dis)connect() independent of the port argument order ?
One has to pass port arguments in a strict order, namely
(readable, writable). Why not making it independent of this order ?

>From an API user's point of view, the only thing that should matter is that we do pass a pair of readable and writable ports.

Or is there a strong reason for imposing a strict argument order ? Do you prefer leaving the ordering logic to the API user ?

Cheers,
J.


     
_______________________________________________
Jack-Devel mailing list
Jack-Devel@...
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org

[Fons Adriaensen-2]

On Wed, Oct 14, 2009 at 04:33:49AM -0700, James Warden wrote:

> Why isn't jack_(dis)connect() independent of the port argument order ?
> One has to pass port arguments in a strict order, namely
> (readable, writable). Why not making it independent of this order ?
>
> >From an API user's point of view, the only thing that should matter is that we do pass a pair of readable and writable ports.
>
> Or is there a strong reason for imposing a strict argument order ? Do you prefer leaving the ordering logic to the API user ?

And one could ask the same for jack_connect().

It would matter only if  programmer doesn't know
which of his ports are readable or writeable.
And in that case (s)he is serious trouble anyway.

Strict argument order is a sanity check, and it
helps to detect programming errors. I'd leave it
as it is.

--
FA
_______________________________________________
Jack-Devel mailing list
Jack-Devel@...
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org

[Patrick Shirkey]

On 10/14/2009 10:31 PM, Phil Rhodes wrote:
> I don't suppose any of this will make it feasible to send audio over a
> wifi link, will it?
>
>


There was a brief discussion about that earlier this year and the answer
was that it does work quite well on a 802.11n connection.




Patrick Shirkey
Boost Hardware Ltd



_______________________________________________
Jack-Devel mailing list
Jack-Devel@...
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org

[Jörn Nettingsmeier-5]

Patrick Shirkey wrote:

>
> On 10/14/2009 10:31 PM, Phil Rhodes wrote:
>> I don't suppose any of this will make it feasible to send audio over a
>> wifi link, will it?
>>
>>
>
>
> There was a brief discussion about that earlier this year and the answer
> was that it does work quite well on a 802.11n connection.

except when it doesn't.

torben has been using it extensively with good results, but the
fundamental problem remains:

wifi, like old hubbed 10/100baseTX ethernet, is a shared medium. unless
you're the only one on the network, all is well. if somebody else is
using or even just scanning your network, you *will* have collisions,
and each collision is a dropout.

so for all serious studio use, wifi is not an option. you may have
useful results for recreational listening or within the context of a
performance, if the artistic concept is robust wrt dropouts.


--
Jörn Nettingsmeier

Meister für Veranstaltungstechnik

Audio and event engineer
Ambisonic surround recordings

http://stackingdwarves.net
[phone]

_______________________________________________
Jack-Devel mailing list
Jack-Devel@...
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org

[Jörn Nettingsmeier-5]

Jörn Nettingsmeier typed faster than his brain could keep up with:

> wifi, like old hubbed 10/100baseTX ethernet, is a shared medium. unless
> you're the only one on the network, all is well.

s/unless/as long as/


--
Jörn Nettingsmeier

Meister für Veranstaltungstechnik

Audio and event engineer
Ambisonic surround recordings

http://stackingdwarves.net
[phone]

_______________________________________________
Jack-Devel mailing list
Jack-Devel@...
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org

[Phil Rhodes]

I'm thinking of a point to point realtime audio link for open-air
applications. What's important is low latency and letting dropouts go. Many
of the preexisting things for doing internet audio streaming have very poor
latency (because they don't care) and try to correct dropouts, which is the
last thing I want - if it's gonna glitch, just glitch, and keep in sync with
other, hardwired links.

My application would use yagi or dish parabolic antennas, keeping crosstalk
with other RF spectrum users at a minimum.

I am a windows guy and have no experience of linux audio, which is
reportedly a disaster zone - if there's a document anyone could point me to
as to how to get things set up so I can just stream a stereo pair from one
machine to another, I'm all ears.

P

_______________________________________________
Jack-Devel mailing list
Jack-Devel@...
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org

[Geoff Beasley-2]

 > The best I can do is to extract "netjackone" stuff from jacknone
> branch and merge in trunk. Since Torben seems to have disappear, I'll
> have to hack a bit to do that. I'll try to do that ASAP.


fantastic Stephane. Ley me know when it's done and I'll get it up and tested.
;)

best,

g.
_______________________________________________
Jack-Devel mailing list
Jack-Devel@...
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org

[torbenh]

On Thu, Oct 15, 2009 at 04:15:15PM +1100, Geoff Beasley wrote:
>  > The best I can do is to extract "netjackone" stuff from jacknone
> > branch and merge in trunk. Since Torben seems to have disappear, I'll
> > have to hack a bit to do that. I'll try to do that ASAP.
>
>
> fantastic Stephane. Ley me know when it's done and I'll get it up and tested.
> ;)

have merged jack-1.9.3 into my codebase, and upped jacknone-0.3                                                              
seems to work... dunno.                                                                                                        
                                                                                                                               
i cant reproduce audioadapter fails.                                                                                          
and i dont think geoff is using audioadapter anyways.                                                                          
                                                                                                                               
gonna do the trunk merge now. is there anything useful in trunk ?                                                              
or are people fine with a 1.9.3 merge for now ?      

>
> best,
>
> g.
> _______________________________________________
> Jack-Devel mailing list
> Jack-Devel@...
> http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org

--
torben Hohn
http://galan.sourceforge.net -- The graphical Audio language
_______________________________________________
Jack-Devel mailing list
Jack-Devel@...
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org

[Stéphane Letz]

Le 24 oct. 2009 à 19:15, torbenh@... a écrit :

> On Thu, Oct 15, 2009 at 04:15:15PM +1100, Geoff Beasley wrote:
>>> The best I can do is to extract "netjackone" stuff from jacknone
>>> branch and merge in trunk. Since Torben seems to have disappear,  
>>> I'll
>>> have to hack a bit to do that. I'll try to do that ASAP.
>>
>>
>> fantastic Stephane. Ley me know when it's done and I'll get it up  
>> and tested.
>> ;)
>
> have merged jack-1.9.3 into my codebase, and upped jacknone-0.3
> seems to work... dunno.
>
> i cant reproduce audioadapter fails.
> and i dont think geoff is using audioadapter anyways.
>
> gonna do the trunk merge now. is there anything useful in trunk ?
> or are people fine with a 1.9.3 merge for now ?

Please don't do any merge on SVN.

Where is this jacknone-0.3 branch? I don't see anything in  http://repo.or.cz/w/jack2.git

Stephane

_______________________________________________
Jack-Devel mailing list
Jack-Devel@...
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org

[Stéphane Letz]

Le 24 oct. 2009 à 19:15, torbenh@... a écrit :

> On Thu, Oct 15, 2009 at 04:15:15PM +1100, Geoff Beasley wrote:
>>> The best I can do is to extract "netjackone" stuff from jacknone
>>> branch and merge in trunk. Since Torben seems to have disappear,  
>>> I'll
>>> have to hack a bit to do that. I'll try to do that ASAP.
>>
>>
>> fantastic Stephane. Ley me know when it's done and I'll get it up  
>> and tested.
>> ;)
>
> have merged jack-1.9.3 into my codebase, and upped jacknone-0.3
> seems to work... dunno.
>
> i cant reproduce audioadapter fails.
> and i dont think geoff is using audioadapter anyways.
>
> gonna do the trunk merge now. is there anything useful in trunk ?
> or are people fine with a 1.9.3 merge for now ?
>

OK I found jacknone-0.3 on http://netjack.sourceforge.net/.

But as I said in a previous mail about this netjack1/netjack2 story,  
this jacknone-0.3 branch is not going to be merged in SVN thunk in the  
current state for at least 2 reasons:

- it has some improvements is "audioadapter" components that makes  
things even worse compared to the initial code

- JACK engine (JackEngine.cpp) contains some experimental change  
regarding connections  (written ny Nedko).

So my position stays the same:

- either Torben prepares a patch (or a branch or whatever) against  
JACK2 SVN trunk that contains netjack1 code *only* and I'll test and  
merge.

- or I will do that myself when I have time, that is after some work I  
did and still have to finish rapidly for JACK2 on OSX, particularly  
after Harrisson MixBus recent release (that creates "some" interest  
for Ardour/Mixbus, JACK kind of tools in the OSX community...)

Stephane
_______________________________________________
Jack-Devel mailing list
Jack-Devel@...
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org

[torbenh]

On Sat, Oct 24, 2009 at 07:23:57PM +0200, Stéphane Letz wrote:

>
> Le 24 oct. 2009 à 19:15, torbenh@... a écrit :
>
>> On Thu, Oct 15, 2009 at 04:15:15PM +1100, Geoff Beasley wrote:
>>>> The best I can do is to extract "netjackone" stuff from jacknone
>>>> branch and merge in trunk. Since Torben seems to have disappear, I'll
>>>> have to hack a bit to do that. I'll try to do that ASAP.
>>>
>>>
>>> fantastic Stephane. Ley me know when it's done and I'll get it up and
>>> tested.
>>> ;)
>>
>> have merged jack-1.9.3 into my codebase, and upped jacknone-0.3
>> seems to work... dunno.
>>
>> i cant reproduce audioadapter fails.
>> and i dont think geoff is using audioadapter anyways.
>>
>> gonna do the trunk merge now. is there anything useful in trunk ?
>> or are people fine with a 1.9.3 merge for now ?
>
> Please don't do any merge on SVN.

hmm ? i dont plan to touch your svn. never will.

> Where is this jacknone-0.3 branch? I don't see anything in  
> http://repo.or.cz/w/jack2.git

i have reverted the merge of the no-self-connect branch now, and pushed
it as "no-self-connect-revert"

jacknone was never intended to be merged back, by the way.

>
> Stephane

--
torben Hohn
http://galan.sourceforge.net -- The graphical Audio language
_______________________________________________
Jack-Devel mailing list
Jack-Devel@...
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org

[torbenh]

On Sat, Oct 24, 2009 at 10:34:36PM +0200, Stéphane Letz wrote:

>
> Le 24 oct. 2009 à 19:15, torbenh@... a écrit :
>
>> On Thu, Oct 15, 2009 at 04:15:15PM +1100, Geoff Beasley wrote:
>>>> The best I can do is to extract "netjackone" stuff from jacknone
>>>> branch and merge in trunk. Since Torben seems to have disappear, I'll
>>>> have to hack a bit to do that. I'll try to do that ASAP.
>>>
>>>
>>> fantastic Stephane. Ley me know when it's done and I'll get it up and
>>> tested.
>>> ;)
>>
>> have merged jack-1.9.3 into my codebase, and upped jacknone-0.3
>> seems to work... dunno.
>>
>> i cant reproduce audioadapter fails.
>> and i dont think geoff is using audioadapter anyways.
>>
>> gonna do the trunk merge now. is there anything useful in trunk ?
>> or are people fine with a 1.9.3 merge for now ?
>>
>
> OK I found jacknone-0.3 on http://netjack.sourceforge.net/.
>
> But as I said in a previous mail about this netjack1/netjack2 story, this
> jacknone-0.3 branch is not going to be merged in SVN thunk in the current
> state for at least 2 reasons:
>
> - it has some improvements is "audioadapter" components that makes things
> even worse compared to the initial code

in all jacknone releases "audioadapter" works fine. almost as good as
alsa_in/alsa_out ...

http://repo.or.cz/w/jack2.git?a=commit;h=1b701347538c473a232f624e11d23445c4a8e28a

^^ this works fine.

http://repo.or.cz/w/jack2.git?a=commit;h=5f19e38bf6af75eeae76eae103c152b573e799b2

^^^ AND THIS COMMIT BREAKS IT.

reverting this commit later on doesnt seem to fix it. and i dont want to
digg inside this commit.
oh... well... i guess i am back to "i dont care"


>
> - JACK engine (JackEngine.cpp) contains some experimental change regarding
> connections  (written ny Nedko).

fixed. (well... regarding that some of the above mentioned commit is
still in there, its not fixed... sigh)

i cant believe i was so stupid to actually put some work in this junk.

>
> So my position stays the same:
>
> - either Torben prepares a patch (or a branch or whatever) against JACK2
> SVN trunk that contains netjack1 code *only* and I'll test and merge.
>
> - or I will do that myself when I have time, that is after some work I did
> and still have to finish rapidly for JACK2 on OSX, particularly after
> Harrisson MixBus recent release (that creates "some" interest  for
> Ardour/Mixbus, JACK kind of tools in the OSX community...)
>
> Stephane_______________________________________________
> Jack-Devel mailing list
> Jack-Devel@...
> http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org

--
torben Hohn
_______________________________________________
Jack-Devel mailing list
Jack-Devel@...
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org

[torbenh]

it seems that ./waf only got a bit confused.
the audio adapter code contained in no-self-connect-revert is working
fine.

a test over loopback is also working fine.

still... nobody cares for it, nobody tests it. i dont even use jack
anymore. so whats the point ?

i only get pissed when i touch this stuff.


On Sun, Oct 25, 2009 at 01:32:43AM +0200, torbenh@... wrote:

> On Sat, Oct 24, 2009 at 10:34:36PM +0200, Stéphane Letz wrote:
> >
> > Le 24 oct. 2009 à 19:15, torbenh@... a écrit :
> >
> >> On Thu, Oct 15, 2009 at 04:15:15PM +1100, Geoff Beasley wrote:
> >>>> The best I can do is to extract "netjackone" stuff from jacknone
> >>>> branch and merge in trunk. Since Torben seems to have disappear, I'll
> >>>> have to hack a bit to do that. I'll try to do that ASAP.
> >>>
> >>>
> >>> fantastic Stephane. Ley me know when it's done and I'll get it up and
> >>> tested.
> >>> ;)
> >>
> >> have merged jack-1.9.3 into my codebase, and upped jacknone-0.3
> >> seems to work... dunno.
> >>
> >> i cant reproduce audioadapter fails.
> >> and i dont think geoff is using audioadapter anyways.
> >>
> >> gonna do the trunk merge now. is there anything useful in trunk ?
> >> or are people fine with a 1.9.3 merge for now ?
> >>
> >
> > OK I found jacknone-0.3 on http://netjack.sourceforge.net/.
> >
> > But as I said in a previous mail about this netjack1/netjack2 story, this
> > jacknone-0.3 branch is not going to be merged in SVN thunk in the current
> > state for at least 2 reasons:
> >
> > - it has some improvements is "audioadapter" components that makes things
> > even worse compared to the initial code
>
> in all jacknone releases "audioadapter" works fine. almost as good as
> alsa_in/alsa_out ...
>
> http://repo.or.cz/w/jack2.git?a=commit;h=1b701347538c473a232f624e11d23445c4a8e28a
>
> ^^ this works fine.
>
> http://repo.or.cz/w/jack2.git?a=commit;h=5f19e38bf6af75eeae76eae103c152b573e799b2
>
> ^^^ AND THIS COMMIT BREAKS IT.
>
> reverting this commit later on doesnt seem to fix it. and i dont want to
> digg inside this commit.
> oh... well... i guess i am back to "i dont care"
>
>
> >
> > - JACK engine (JackEngine.cpp) contains some experimental change regarding
> > connections  (written ny Nedko).
>
> fixed. (well... regarding that some of the above mentioned commit is
> still in there, its not fixed... sigh)
>
> i cant believe i was so stupid to actually put some work in this junk.
>
> >
> > So my position stays the same:
> >
> > - either Torben prepares a patch (or a branch or whatever) against JACK2
> > SVN trunk that contains netjack1 code *only* and I'll test and merge.
> >
> > - or I will do that myself when I have time, that is after some work I did
> > and still have to finish rapidly for JACK2 on OSX, particularly after
> > Harrisson MixBus recent release (that creates "some" interest  for
> > Ardour/Mixbus, JACK kind of tools in the OSX community...)
> >
> > Stephane_______________________________________________
> > Jack-Devel mailing list
> > Jack-Devel@...
> > http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
>
> --
> torben Hohn
> _______________________________________________
> Jack-Devel mailing list
> Jack-Devel@...
> http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org

--
torben Hohn
http://galan.sourceforge.net -- The graphical Audio language
_______________________________________________
Jack-Devel mailing list
Jack-Devel@...
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org