Grandorgue audio crackles when recording

[bwv582]

I'm running Grandorgue with Ubuntu Linux 12.04 on a fairly new laptop. I have it hooked up to the MIDI outputs from a Hammond 825 console. All is well until I try to record to a wav file. When recording to wav, I get periodic crackling in the sound which affects the recording. Currently I have an external USB SB1090 plugged in, but I had the same problem when using the onboard audio, so the sound card does not seem to make a difference. Again, it works fine when not recording, but as soon as record is enabled there is trouble. Anybody else run into this problem?

[L.Palo]

Hi!

Your computer is under too high workload when recording. Change the buffer size if possible to a higher number, increase desired latency, make sure nothing else makes your computer work at the same time. If nothing else helps - upgrade the computer to higher performance!

Kind regards

Lars P

[bwv582]

Thanks. I will try experimenting with some of those settings. I'm running a Core i5 with 4GB RAM. That should be enough, right?

[e9925248]

Are the cracks on the recorded wav too?

[bwv582]

Are the cracks on the recorded wav too?

The cracks are in the recording too, but in a slightly different way. When I play back the wav, at places where the crackling wasn't too bad, I don't notice it much, but in other places the sound "hitches." That is, it will speed up or sound like it is skipping audio frames at the points where it was crackling.

I set the audio output buffers as high as they would go (1024) and I disabled all the extra features (interpolation, randomized pipe speaking, etc.) It has helped marginally, but basically the trouble is still there. I needed to record some hymns yesterday and I finally got them well enough, but it took several tries before I got clean enough recordings to use.

I imagine to really troubleshoot I will have to start trying different combinations of computers, operating systems, and sound cards. It's just so much effort to do that.

[e9925248]

The cracks are in the recording too, but in a slightly different way. When I play back the wav, at places where the crackling wasn't too bad, I don't notice it much, but in other places the sound "hitches." That is, it will speed up or sound like it is skipping audio frames at the points where it was crackling.

I took a look at the code: With reverb disabled, the recoding should not skip any samples. At the point of cracks, input events are processed a little bit faster (as GO manages to calculate less frames). So the audio quality of the recoding should not be affected, while length of some played notes might be a little bit different.

[bwv582]

So the audio quality of the recoding should not be affected, while length of some played notes might be a little bit different.

I think you're right and this seems to match what I've observed. The audio quality of the recording is not worse, but note lengths are wrong. Particularly, notes appear to be played with shorter duration at the point of the crackling. The effect in the end is that certain measures or notes "speed up" when the recording is played back.

[L.Palo]

However, the strange thing is that you have those crackles at all when you're recording. I've got less capable processors on my computers but have been able to record without problems. Have you tried different audio backends (both the RtAudio versions and Portaudio ones)? If you record with other softwares do you have issues too?

Kind regards

Lars P

[bwv582]

Hello again. I'm sorry if bumping such an old thread is bad form, but nearly three years later I still have the same problem recording with GO. I'm using the current version (2093) and I've tried everything I can think of. I've tried different operating systems (Linux and Windows), I've added memory, I've switched between internal and external sound cards, I've tried different speaker systems... I still get very loud popping at random intervals when I record. For a time I thought it was related to disc access as it seemed to happen more when the hard disc was working harder in the background, but lately I've observed that theory doesn't always hold.

Here's the thing-- I tried Hauptwerk for awhile and the problem disappeared. I really want GO to work because HW is so expensive and I prefer free open source software, yet I'm reaching the end of my patience.

I'm currently using a Hammond 825 console to play to a Windows 7 AMD dual-core with 8GB RAM running GO 3.1.2093. Has anybody else had this problem lately and can offer any further insight?

[e9925248]

Although the situation has improved in recent GOs, recording with a larger latency is safer (= large samples per buffer + desired latency). I'm having troubles reproducing that [well, if run a program in the background, which generates CPU spikes, I get audio dropouts].

If you load something small [eg. the demo set] and use only a few stops, does it happen too?

[bwv582]

Yes, I have the problem no matter how large or small the sample set and no matter what stop combinations are pulled. I have the organs load in mono and have tried to make other settings as easy as possible on the processor. But really, I should have plenty of computing power to run this. Right now the buffer setting is at 1024 (as high as it will go) and the latency is 50ms. Are those the settings you meant? Would you suggest increasing the audio latency?

I even thought it was maybe a grounding problem with the AC wiring in the wall or the speakers, but then when I didn't have the trouble using Hauptwerk, I figured it couldn't be that.