name conflict for netjack... call for new project name.

classic Classic list List threaded Threaded
8 messages Options
Reply | Threaded
Open this post in threaded view
|

name conflict for netjack... call for new project name.

torbenh

hi...

on a recent discussion on lau it turned out, that included with jackosx
there is netjack.

unfortunately robert and me changed the name of the udpsync project to
netjack also.

before deciding on a new name for our project (and changing sf-project
name etc.)
i would like to see if it was possible to make both projects compatible.
sf cvs service are not working currently and i cant get the jackosx
sources. so i cant analyze netjack-osx.

but if it does not solve the inter jackd synchronisation perfectly like the
linux-netjack i will have to change names.


--
torben Hohn
http://galan.sourceforge.net -- The graphical Audio language


-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php
_______________________________________________
Jackit-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jackit-devel
Reply | Threaded
Open this post in threaded view
|

Re: name conflict for netjack... call for new project name.

Stéphane Letz
Hi,

The osx netjack code was developed by Johnny Petrantonni starting  
from jack-udp code. It does not has any sync feature. From what I  
undertand from the netjack sf project page, it is somewhat similar to  
what you call "jacknet_client", but is used with two machines each  
synchronized on its audio card (thus having the clock drift issue...)

How do you think the 2 projects could be made compatible?


Stephane


>
> hi...
>
> on a recent discussion on lau it turned out, that included with  
> jackosx
> there is netjack.
>
> unfortunately robert and me changed the name of the udpsync project to
> netjack also.
>
> before deciding on a new name for our project (and changing sf-project
> name etc.)
> i would like to see if it was possible to make both projects  
> compatible.
> sf cvs service are not working currently and i cant get the jackosx
> sources. so i cant analyze netjack-osx.
>
> but if it does not solve the inter jackd synchronisation perfectly  
> like the
> linux-netjack i will have to change names.
>



-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php
_______________________________________________
Jackit-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jackit-devel
Reply | Threaded
Open this post in threaded view
|

Re: name conflict for netjack... call for new project name.

torbenh
On Sun, Nov 06, 2005 at 04:36:42PM +0100, St?phane Letz wrote:
> Hi,
>
> The osx netjack code was developed by Johnny Petrantonni starting  
> from jack-udp code. It does not has any sync feature. From what I  
> undertand from the netjack sf project page, it is somewhat similar to  
> what you call "jacknet_client", but is used with two machines each  
> synchronized on its audio card (thus having the clock drift issue...)
>
> How do you think the 2 projects could be made compatible?

well my mail suggested, that i dont think they can be made compatible.
but i have not seen any sourcecode. so this was the question if it was
possible.

when the jackosx cvs is online again, i will think about this again.
perhaps a unification of packetheaders.
my packetheader can be changed because there is no big userbase.

but if you dont take transport into account then i guess...
unification wont be possible...


--
torben Hohn
http://galan.sourceforge.net -- The graphical Audio language


-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php
_______________________________________________
Jackit-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jackit-devel
Reply | Threaded
Open this post in threaded view
|

Re: name conflict for netjack... call for new project name.

Stéphane Letz

Le 6 nov. 05 à 22:20, [hidden email] a écrit :

> On Sun, Nov 06, 2005 at 04:36:42PM +0100, St?phane Letz wrote:
>> Hi,
>>
>> The osx netjack code was developed by Johnny Petrantonni starting
>> from jack-udp code. It does not has any sync feature. From what I
>> undertand from the netjack sf project page, it is somewhat similar to
>> what you call "jacknet_client", but is used with two machines each
>> synchronized on its audio card (thus having the clock drift issue...)
>>
>> How do you think the 2 projects could be made compatible?
>
> well my mail suggested, that i dont think they can be made compatible.
> but i have not seen any sourcecode. so this was the question if it was
> possible.
>
> when the jackosx cvs is online again, i will think about this again.
> perhaps a unification of packetheaders.
> my packetheader can be changed because there is no big userbase.
>
> but if you dont take transport into account then i guess...
> unification wont be possible...


Can you explain how transport is taken in account?

I really think we should try to unify the 2 projects...


Stephane



-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php
_______________________________________________
Jackit-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jackit-devel
Reply | Threaded
Open this post in threaded view
|

Re: name conflict for netjack... call for new project name.

torbenh
On Mon, Nov 07, 2005 at 11:36:04AM +0100, St?phane Letz wrote:
>
> >but if you dont take transport into account then i guess...
> >unification wont be possible...
>
>
> Can you explain how transport is taken in account?

i transmit the current transport position with the packet.
the goal is that the local ardour can record from the remote hydrogen
for example. which is starting and stopping with ardour like it was
connected to the same jackd.

there is one problem i need to solve before though.
i have one period of roundtrip delay.
so i must be able to compute the transport position for the next period.
which is sometimes impossible.

any input on this would be apreciated ;)
the current implementation on the slave side looks whether the master
transport is rolling and starts / stops accordingly.

but this is not sample accurate.
the jacknet_source is obviously a slow-sync client.


>
> I really think we should try to unify the 2 projects...

me too. but only if i dont have to cut down my feature plans.


>
>
> Stephane
>
>
>
> -------------------------------------------------------
> SF.Net email is sponsored by:
> Tame your development challenges with Apache's Geronimo App Server. Download
> it for free - -and be entered to win a 42" plasma tv or your very own
> Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php
> _______________________________________________
> Jackit-devel mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/jackit-devel
>

--
torben Hohn
http://galan.sourceforge.net -- The graphical Audio language


-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php
_______________________________________________
Jackit-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jackit-devel
Reply | Threaded
Open this post in threaded view
|

Re: name conflict for netjack... call for new project name.

Jack O'Quin-2
[hidden email] writes:

> i transmit the current transport position with the packet.
> the goal is that the local ardour can record from the remote hydrogen
> for example. which is starting and stopping with ardour like it was
> connected to the same jackd.

Where can I find your current code?

> there is one problem i need to solve before though.
> i have one period of roundtrip delay.
> so i must be able to compute the transport position for the next period.
> which is sometimes impossible.
>
> any input on this would be apreciated ;)
> the current implementation on the slave side looks whether the master
> transport is rolling and starts / stops accordingly.

I don't understand the whole picture.  But, your client could possibly
declare itself to be timebase master.  That callback is invoked one
cycle before each new position change.  (If that interface does not
work, we could maybe invent a similar one with the right triggers and
information.)

> but this is not sample accurate.
> the jacknet_source is obviously a slow-sync client.

In some ways this situation is similar to a pair of old analog tape
decks chasing each other.  Sample accuracy may be attainable, but even
if not the result could still be useful for some applications.

>> I really think we should try to unify the 2 projects...
>
> me too. but only if i dont have to cut down my feature plans.

votes++;  :-)
--
  joq


-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php
_______________________________________________
Jackit-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jackit-devel
Reply | Threaded
Open this post in threaded view
|

Re: name conflict for netjack... call for new project name.

torbenh
On Mon, Nov 07, 2005 at 03:57:07PM -0600, Jack O'Quin wrote:
> [hidden email] writes:
>
> > i transmit the current transport position with the packet.
> > the goal is that the local ardour can record from the remote hydrogen
> > for example. which is starting and stopping with ardour like it was
> > connected to the same jackd.
>
> Where can I find your current code?


netjack.sf.net in the cvs.

>
> > there is one problem i need to solve before though.
> > i have one period of roundtrip delay.
> > so i must be able to compute the transport position for the next period.
> > which is sometimes impossible.
> >
> > any input on this would be apreciated ;)
> > the current implementation on the slave side looks whether the master
> > transport is rolling and starts / stops accordingly.
>
> I don't understand the whole picture.  But, your client could possibly
> declare itself to be timebase master.  That callback is invoked one
> cycle before each new position change.  (If that interface does not
> work, we could maybe invent a similar one with the right triggers and
> information.)

there can only be one timebase master.
in fact if the jacknet_source got called after the timebase master it
could read the information the timebase master had set. and it would be
able to send it out.
but what if it was before the timebase_master in the process order the
position would not be computed.

the solution i have in my head is calling the timebase master with
current_period+2...

but i am not sure if this would not break any other app.
>
> > but this is not sample accurate.
> > the jacknet_source is obviously a slow-sync client.
>
> In some ways this situation is similar to a pair of old analog tape
> decks chasing each other.  Sample accuracy may be attainable, but even
> if not the result could still be useful for some applications.

well when i implement slow sync support then the framecount on both
machines will be identical. but the reaction to tempochanges and other
information provided by the timbase callback will be interpolated.


>
> >> I really think we should try to unify the 2 projects...
> >
> > me too. but only if i dont have to cut down my feature plans.
>
> votes++;  :-)
> --
>   joq
>
>
> -------------------------------------------------------
> SF.Net email is sponsored by:
> Tame your development challenges with Apache's Geronimo App Server. Download
> it for free - -and be entered to win a 42" plasma tv or your very own
> Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php
> _______________________________________________
> Jackit-devel mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/jackit-devel
>

--
torben Hohn
http://galan.sourceforge.net -- The graphical Audio language


-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php
_______________________________________________
Jackit-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jackit-devel
Reply | Threaded
Open this post in threaded view
|

Re: name conflict for netjack... call for new project name.

Jack O'Quin-2
[hidden email] writes:

> On Mon, Nov 07, 2005 at 03:57:07PM -0600, Jack O'Quin wrote:

>> I don't understand the whole picture.  But, your client could possibly
>> declare itself to be timebase master.  That callback is invoked one
>> cycle before each new position change.  (If that interface does not
>> work, we could maybe invent a similar one with the right triggers and
>> information.)
>
> there can only be one timebase master.
> in fact if the jacknet_source got called after the timebase master it
> could read the information the timebase master had set. and it would be
> able to send it out.
> but what if it was before the timebase_master in the process order the
> position would not be computed.

You are correct that the extended position data managed by the TB
master might not be filled in yet.  But, the frame number is available
for the entire cycle.  I think we could add a new callback to reveal
that information to *any* interested client.

Or, I think you could perhaps solve this problem more cleanly using
slow-sync.  I need to think about that a bit more.  I'll try to take a
look at your code tomorrow.

> the solution i have in my head is calling the timebase master with
> current_period+2...
>
> but i am not sure if this would not break any other app.

I'm not sure quite what you have in mind there.

>> > but this is not sample accurate.
>> > the jacknet_source is obviously a slow-sync client.
>>
>> In some ways this situation is similar to a pair of old analog tape
>> decks chasing each other.  Sample accuracy may be attainable, but even
>> if not the result could still be useful for some applications.
>
> well when i implement slow sync support then the framecount on both
> machines will be identical. but the reaction to tempochanges and other
> information provided by the timbase callback will be interpolated.

That sounds reasonable.
--
  joq


-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php
_______________________________________________
Jackit-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jackit-devel