All times are UTC+01:00




Post new topic  Reply to topic  [ 707 posts ]  Go to page Previous 129 30 31 32 3371 Next
Author Message
PostPosted: Tue Mar 14, 2023 2:16 pm 

Joined: Fri May 27, 2022 9:33 pm
Posts: 34
Quote:
Quote:
The audio works now, but now I hear some periodic clicks after changing diversity delay, even at 1000ms, they don't seem to go away.

LE: Changing the value a few times seems to make ST crash.
I'm going to need to fix some driver issues; I have one system in which I can hear the output but it doesn't see any input (that's why I didn't notice the 90 dB issue), and another which does see input but I can't hear the output. That's a driver issue; the old version had the same behavior. Anyway, I need to fix it to hear those periodic clicks. I have a suspicion but I need to hear it (on my end, while looking at some logging) to verify.

Crashing: Which value do you mean?
Also the buffer size/diversity delay setting, sorry was late in evening and i was tired.
To my knowledge recent ubuntu/debian changed some ways of operation with jackd (scheduling and mixer) that may be problematic, but it's not on driver level, at least for most of it.
They say a setup from scratch is recommended.
I wish i could've helped more but now I use basic ALSA hardware on pipewire-jack, that attracted me cause it requires basically 0 setup at expense of 20ms more latency or so. (and of course, not utilizing pro gear to full potential)


Top
   
PostPosted: Tue Mar 14, 2023 5:44 pm 
Site Admin
User avatar

Joined: Mon Mar 17, 2008 1:40 am
Posts: 11187
Quote:
Quote:
Quote:
The audio works now, but now I hear some periodic clicks after changing diversity delay, even at 1000ms, they don't seem to go away.

LE: Changing the value a few times seems to make ST crash.
I'm going to need to fix some driver issues; I have one system in which I can hear the output but it doesn't see any input (that's why I didn't notice the 90 dB issue), and another which does see input but I can't hear the output. That's a driver issue; the old version had the same behavior. Anyway, I need to fix it to hear those periodic clicks. I have a suspicion but I need to hear it (on my end, while looking at some logging) to verify.

Crashing: Which value do you mean?
Also the buffer size/diversity delay setting, sorry was late in evening and i was tired.
To my knowledge recent ubuntu/debian changed some ways of operation with jackd (scheduling and mixer) that may be problematic, but it's not on driver level, at least for most of it.
They say a setup from scratch is recommended.
I wish i could've helped more but now I use basic ALSA hardware on pipewire-jack, that attracted me cause it requires basically 0 setup at expense of 20ms more latency or so. (and of course, not utilizing pro gear to full potential)
It's still half-way working here but I do see glitches in the GUI, and I just had a crash. I think I know where to look.

Edit: Whoops. There was a buffer that should be a multiple of the Jack period size, and it wasn't because I added a safety margin to it that wasn't a multiple. That caused both issues. And it could also cause issues if you change the Jack period - I've made it close and re-open the Jack devices now if you ever do that. I have also added checks elsewhere (for other device types) to catch this situation if it ever happens again.


Top
   
PostPosted: Wed Mar 15, 2023 11:59 am 
Site Admin
User avatar

Joined: Mon Mar 17, 2008 1:40 am
Posts: 11187
BETA053 is uploaded. I'll update the opening post of this thread after lunch, but if you want to test you can already download it.


Top
   
PostPosted: Wed Mar 15, 2023 4:44 pm 

Joined: Fri May 27, 2022 9:33 pm
Posts: 34
Oh yeah, i edited the link and could DL.
Glad you got it sorted, the audio issues are fixed now.
And very neat, livewire alsa and jack are now together supported, that's great!

LE: It seems that after closing ST, it's jack ports take a while to shut down, like a few seconds, if i restart it too quickly it's inputs get duplicated.

LE2: hmm a small issue: it doesn't like changing the sample rate to 96khz, it's inputs won't connect. Changing back to 48 works again.


Top
   
PostPosted: Wed Mar 15, 2023 4:54 pm 
Site Admin
User avatar

Joined: Mon Mar 17, 2008 1:40 am
Posts: 11187
Quote:
Oh yeah, i edited the link and could DL.
Glad you got it sorted, the audio issues are fixed now.
And very neat, livewire alsa and jack are now together supported, that's great!
Yes, I'm going to combine everything into a single binary and use dynamic linking for Jack, that will simplify things a lot (on our end as well).
Quote:
LE: It seems that after closing ST, it's jack ports take a while to shut down, like a few seconds, if i restart it too quickly it's inputs get duplicated.

LE2: hmm a small issue: it doesn't like changing the sample rate to 96khz, it's inputs won't connect. Changing back to 48 works again.
Hm, odd. So I assume that both of those issues did not occur with the old version?


Top
   
PostPosted: Wed Mar 15, 2023 5:09 pm 

Joined: Fri May 27, 2022 9:33 pm
Posts: 34
Quote:
Hm, odd. So I assume that both of those issues did not occur with the old version?
Yes, prior to b50 those didn't occur.
Correction: the first one doesn't occur on normal sample rate, only at 96 it gets stuck at exiting for 5ish seconds. My apologies. At 48 it closes normally.


Top
   
PostPosted: Wed Mar 15, 2023 5:18 pm 
Site Admin
User avatar

Joined: Mon Mar 17, 2008 1:40 am
Posts: 11187
Quote:
Quote:
Hm, odd. So I assume that both of those issues did not occur with the old version?
Yes, prior to b50 those didn't occur.
Correction: the first one doesn't occur on normal sample rate, only at 96 it gets stuck at exiting for 5ish seconds. My apologies. At 48 it closes normally.
Weird. I found a bug fix in Jack itself from a year ago where apparently jack_deactivate when used with PipeWire didn't disconnect the ports. So that might be what you're seeing - but then I don't understand why it worked with the old code. I'll add an explicit removal just in case. [done, next build]

When you say "changing the sample rate", do you mean while Stereo Tool is running? I see that that fails here indeed. But the reason appears to be that the Jack instance is gone - I cannot even close it anymore, that returns an error. So I don't think that that can be solved, or can have worked in the old version. I see no issues at different sample rates if I don't touch them while Stereo Tool is running.

I have everything loading dynamically now, so I'm going to merge the Jack and non-Jack builds now. No more _alsa and _jack binaries!


Top
   
PostPosted: Wed Mar 15, 2023 6:43 pm 

Joined: Fri May 27, 2022 9:33 pm
Posts: 34
Quote:
Weird. I found a bug fix in Jack itself from a year ago where apparently jack_deactivate when used with PipeWire didn't disconnect the ports. So that might be what you're seeing - but then I don't understand why it worked with the old code.
It can be that the new audio subsystem simply uses parts of the Jack standard that PipeWire breaks (i'm sure it does at least some things to get the device integration features), that can explain why it didn't happen in the old one. It's certainly PW's fault from here on.
I know the native jackd is to be recommended (and it's certainly doable on Pi where it's specialized use only for that), but on a general purpose system it's too difficult to set up, you get it :)
Quote:
I'll add an explicit removal just in case. [done, next build]
Indeed that is worth it, even at 48khz the devices seem to linger on, e.g. when switching to alsa (on both in and out) the jack sink was still showing.
Image
Quote:
When you say "changing the sample rate", do you mean while Stereo Tool is running? I see that that fails here indeed. But the reason appears to be that the Jack instance is gone - I cannot even close it anymore, that returns an error.
Yeah i tried while running (what i usually did) and while closed. It seems that the reported sample rate by ST is 44100, the input seems disconnected, the output sems connected, but both are dead. Switching back to 48khz and they come back alive, just with 'Restart sound cards' option, with sample rate reported correctly 48000.

Again this is no biggie, the audio now works fine at 48, it's enough for now. I know you guys have more important stuff to do.
Quote:
I have everything loading dynamically now, so I'm going to merge the Jack and non-Jack builds now. No more _alsa and _jack binaries!
Nice, certainly useful and simpler.


Top
   
PostPosted: Wed Mar 15, 2023 7:00 pm 
Site Admin
User avatar

Joined: Mon Mar 17, 2008 1:40 am
Posts: 11187
Quote:
Quote:
Weird. I found a bug fix in Jack itself from a year ago where apparently jack_deactivate when used with PipeWire didn't disconnect the ports. So that might be what you're seeing - but then I don't understand why it worked with the old code.
It can be that the new audio subsystem simply uses parts of the Jack standard that PipeWire breaks (i'm sure it does at least some things to get the device integration features), that can explain why it didn't happen in the old one. It's certainly PW's fault from here on.
I know the native jackd is to be recommended (and it's certainly doable on Pi where it's specialized use only for that), but on a general purpose system it's too difficult to set up, you get it :)
Quote:
I'll add an explicit removal just in case. [done, next build]
Indeed that is worth it, even at 48khz the devices seem to linger on, e.g. when switching to alsa (on both in and out) the jack sink was still showing.
Image
Quote:
When you say "changing the sample rate", do you mean while Stereo Tool is running? I see that that fails here indeed. But the reason appears to be that the Jack instance is gone - I cannot even close it anymore, that returns an error.
Yeah i tried while running (what i usually did) and while closed. It seems that the reported sample rate by ST is 44100, the input seems disconnected, the output sems connected, but both are dead. Switching back to 48khz and they come back alive, just with 'Restart sound cards' option, with sample rate reported correctly 48000.

Again this is no biggie, the audio now works fine at 48, it's enough for now. I know you guys have more important stuff to do.
Quote:
I have everything loading dynamically now, so I'm going to merge the Jack and non-Jack builds now. No more _alsa and _jack binaries!
Nice, certainly useful and simpler.
What we do in the Jack code should be 100% identical to what we did before, except for 2 things:
- Instead of the (deprecated) _jack_client_new we use _jack_client_open. Should do basically the same thing.
- The restart behavior when an error occurs (which appears to happen when you change the sample rate) might be somewhat different, because it's using the new I/O framework. But it should still try to close and re-open Jack, and that code hasn't changed.

So I can't explain why the behavior would be different.

Anyway, I'm going to start a new build in a moment, with the explicit closing etc. And without separate Jack versions :)


Top
   
PostPosted: Wed Mar 15, 2023 11:49 pm 

Joined: Mon Oct 19, 2020 7:29 pm
Posts: 4
I've been running various v10 builds on Windows Server and it's been solid. Trying to switch over to Linux (Ubuntu 22.04) and having lots of issues. Are there known issues with the latest Linux (build 52), or should I keep take better notes and try to post about it? I'm pretty consistently getting a seg fault and then it won't start. I have to rename it to play around again. Also, I can't seem to enable the webserver. I tried 9.92 for fun, and everything works fine.


Top
   
Display posts from previous:  Sort by  
Post new topic  Reply to topic  [ 707 posts ]  Go to page Previous 129 30 31 32 3371 Next

All times are UTC+01:00


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB® Forum Software © phpBB Limited