little audio glitches

Hello Everyone,

I’m having a problem with the latest release of ardour where every now and then I’ll hear little audio glitches during playback. The sound is a brief kind crackling noise. Sort of like the background noise from a record player. I don’t think this is an xrun because qjackctl doesn’t show any messages about xruns (nor does ardour, and I seem to recall that ardour normally marks xruns when they occur). At any rate, the sound is more subtle than I remember xruns being, thought the quality is kind-of similar.

And it occurs quite inconsistently. At first I figured I’d made some mistakes while editing my tracks (there’s a lot of cut/paste/duplicate work I’ve had to do with my project), but actually when you play a given time in a different region sometimes I hear a click or two and sometimes I don’t. In fact, I’ve just noticed that if I loop a certain region, then I’m more likely to hear the clicks then if I put the playhead in the location and do normal playback.

Ardour logs a few messages to console when I’m looping:

drop flushed event on the floor, time 0 to early for 282 + 0
drop flushed event on the floor, time 0 to early for 282 + 0
drop flushed event on the floor, time 0 to early for 282 + 0
drop flushed event on the floor, time 0 to early for 564 + 0
drop flushed event on the floor, time 0 to early for 846 + 0
drop flushed event on the floor, time 0 to early for 282 + 0
drop flushed event on the floor, time 0 to early for 564 + 0
drop flushed event on the floor, time 0 to early for 846 + 0

which seems odd, but maybe it’s not relevant because the glitches will occasionally also occur during normal playback in which case there doesn’t seem to be anything logged to the console.

I suppose it’s possible that it’s a hardware issue, like a bad connector or bad cables, however, the glitches seem to effect certain audio tracks (there’s two audio tracks that I’ve done extensive copy/paste/duplicate editing on) and not others. They also seem to be present if I do a stem export of the tracks and/or route a track output to another track’s input and record in, which suggest to me that it’s not an issue with either my headphones or output jack. Since I suspect the problem might be related to me having so many overlapping regions, I tried doing a stem export and then finding sections of the playback where glitches occurred and then cover this up with a region from one of the original tracks. However, it seems that if I do this and then export again, I still pick up a couple new glitches.

I’m using an onboard soundcard, by the way:
00:03.0 Audio device: Intel Corporation Xeon E3-1200 v3/4th Gen Core Processor HD Audio Controller (rev 06)
00:1b.0 Audio device: Intel Corporation 8 Series/C220 Series Chipset High Definition Audio Controller (rev 05)

This is driving me crazy; I’m going to try working with the tracks on another computer with a different sound card and see if that makes any difference. But if anybody has any idea what might be going on, please comment!

Thanks!

Strangely enough I’ve been hearing similar things. I’ve always been in too much of a rush to investigate further. However, it only seems to be on playback, the exported tracks/mixes don’t seem to have the glitches (perhaps I’ve just not noticed though). My main Ardour work is all pre production demo recordings so I’ve not been too bothered by it so far, but if they are finding their way into stem exports or mixes, that is more worrying.

I’ve thought for me it was occouring after cross fades, like if I have about a short cross fade I’ll get a playback glitch occuring shortly after that. I’ll see if I can get some time to investigate further.

We will be moving onto final tracking soon, so I would like to find out what’s going on before I reach this stage.

the messages are not relevant.

Maybe this is related to http://tracker.ardour.org/view.php?id=5953
Can you try the attached patch? (ardour-click_2.patch)
It worked for me, but it needs further testing.

tartina: we continued discussing your patch on IRC today. still up in the air about which parts of it to use.

I’ll try the patch, though it might not be until next weekend that I have some time to work on this again. I still haven’t gotten the chance to try playback on another computer either.

It seems possible that the issue tartina has encountered may be related, but I’m really not sure. The clicks, quite frustratingly, are not consistently reproducible, though they do tend to occur in certain places. I’ll post as soon as I get a chance to compile and test the patched version, but (as mentioned) it might be awhile before I get the chance.

The clicks depend on the position you start playback, so they are not easily reproducibile.
@paul: I understand, that is a core part of ardour, so you should think twice before modifying it, I can try try to join the chat, but the different time zone doesn’t help…

After testing Tartina’s patch, and listening for about a half an hour I can’t pick out any of these little glitches anymore using the patched build. It’s tricky, because as tartina notes, the click depends on where the playback starts. But there are a few sections in my project where I tend to hear clicks in the un-patched version of ardour, but haven’t heard them in the patched version, despite moving the playback head around a few times. So I think Tartina has probably managed to fix this problem (thanks tartina!). Hopefully nothing else breaks because of the fix :slight_smile:

to follow up:

tartina’s patch was merged in Ardour 3.5-4331-gc6e71a6 on 2015-Jan-19.
see also http://tracker.ardour.org/view.php?id=5953

This is similar to the problem I had with 3.5 and 4.0rc2-dbg. I do hear little clicks sometimes but the random skips are even worse. Could this patch be adapted to version 4?

@mashworth

Pretty sure it already is in 4 actually.

Seablade

it IS in version 4, which will be officially released in a few days.

Different issue. (The bug tartina fixed was always present, Mashworth and don3 only have this during export).
It is still present in 4.0:

http://tracker.ardour.org/view.php?id=6096
http://tracker.ardour.org/view.php?id=6238

I meant that Tartina’s patch is in version 4 :slight_smile:

My second patch was about a very rare case when the last frame of a region was at the beginning of the next audio buffer and you got a very loud sample, the problem was in Curve::_get_vector.
The click noise were corrected by colinf, reworking coverage in some other way than mine.
That problems were always reproducible if you started playback from a particular position.

Ciao

@x42 For the record, I’m also experiencing glitch/pop/dropout-like phenomena when recording and playing, also with no reported xruns. In bug ID 6096 I focused on exports, mainly because they’re easiest for me to instrument. Also, at the moment, most of the recordings I work on are being made on a machine with a PCI Firewire interface, and that machine has a previous history with PCI bus contention issues (when using a PCI audio card) – I wanted to avoid complicating the discussion with all that. Naturally I’m hoping that whatever cause is eventually found for the export issue will also apply to recording and playing, but at this point I can’t say whether they’re even related.