| zequence | OvenWerks: Call me slow, but it seems mixdowns don't work on ardour if the PA bridge is active | 15:36 |
|---|---|---|
| zequence | Haven't yet updated to latest, so maybe it was fixed - but that is pretty serious. We need to make that problem go away somehow | 15:37 |
| OvenWerks | zequence: possible | 15:37 |
| OvenWerks | There are a few possible reasons for this: | 15:38 |
| OvenWerks | 1) PA is also connected to another audio interface. | 15:38 |
| OvenWerks | 2) PA is connected to the audio device. | 15:38 |
| OvenWerks | 3) there is a newer version of jackdbus that allows freewheel to disconnect all outputs (I don't think so) | 15:39 |
| zequence | PA is only connected to the sink and source of the jack module | 15:39 |
| zequence | But, it's still PA causing the problem through the bridge | 15:40 |
| OvenWerks | desktop audio should not be able to be heard while doing mixdown | 15:40 |
| zequence | If I disable PA, ardour mixdowns work fine | 15:40 |
| OvenWerks | if it is, then see 2) | 15:40 |
| zequence | PA is connected to the jack module sink and source. Nothing else | 15:41 |
| OvenWerks | even internal to jack? | 15:41 |
| zequence | Yes, that is what I'm saying | 15:42 |
| zequence | I'm not saying the module is connected to various jack apps | 15:42 |
| zequence | Though, it is that too | 15:42 |
| OvenWerks | Then we need a button to disable the bridge in _controls. | 15:43 |
| zequence | HOw about we fix the problem in the first place instead? | 15:43 |
| OvenWerks | zequence: that would require changing pa_jacksource/sink | 15:43 |
| zequence | Either it is a design flaw in jack, or each client needs to be versitaile somehow | 15:43 |
| OvenWerks | I think the PA jack module is not honering the freewheel signal properly. | 15:44 |
| zequence | Yeah, this probably only occurs with PA being connected as a jack client | 15:44 |
| zequence | Well, let's make sure it is known and try to find a way to fix it | 15:44 |
| zequence | The bridge is just an example utility really, so it could use more work anyway | 15:45 |
| OvenWerks | zequence: yes. including double buffering so increase latency as needed. | 15:46 |
| OvenWerks | (actually, I think it just needs buffering at all) | 15:46 |
| zequence | Maybe that would take some load off the CPU as well? (just guessing) | 15:47 |
| OvenWerks | zequence: yes. Right now latency forces time constrants on any sample rate conversion. | 15:48 |
| OvenWerks | So far as I know PA already does sliding latency with the audio devices (ALSA) | 15:49 |
| zequence | Only someone from the jack side of things will be interesting in working out a fix, and even in that bunch, the sum of people that includes becomes quite small | 15:56 |
| OvenWerks | The sum of people who are both interested and also have the time is even smaller. | 16:06 |
| OvenWerks | The origianl auther of the bridge has been responsive to change requests in the past... but those have been mostly passing paramerters from the jackdbus detector to the source/sink | 16:07 |
| zequence | Yeah, I remember David expressed he would have liked to do more, almost as if he was feeling guilty because he wasn't | 16:19 |
| zequence | He doesn't seem to be on IRC anymore | 16:43 |
| zequence | ..not as much, anyway | 16:44 |
| OvenWerks | IRC can waste a lot of time if we let it :) | 17:54 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!