I think because I use an MC cartridge this hasn’t been a problem for me. The cartridge maker states a minimum input impedance of 100ohms and the mic pre inputs are far higher than this. I thought I might need to solder resistors across the xlr pins, but I’ve not heard anything unusual, so didn’t bother in the end.
I have a problem on my raspberry Pi 3B+ pop and crackle on the playback device.
I've found it occurs with 6 channels and IIR filters. Does not happen with only 4 channels and filters. All filters are IIR.
Top shows 32% user CPU for camilla running 6 channels and around only 11% for 4 channels. The last 2 channels don't have anything too complex.
I am running the pre-compiled Camilla DSP v1.0.3 on Archlinux
I've tried to:
increase chunk size
reduce queue length
enable rate adjust
reduce the adjust period.
All to no avail - the crackle / pop is immediately available. there no signs of buffer depletion.
Is the Pi just not grunty enough? I'd appreciate any help / advice please or please let me know if this should be a new thread. thank you
Edit: I'm just testing by playing WAV files via aplay. Whether I go direct (to loopback aka hw:1 in my system) or via an ALSA plugin to resample. no difference (still crackles)
I've found it occurs with 6 channels and IIR filters. Does not happen with only 4 channels and filters. All filters are IIR.
Top shows 32% user CPU for camilla running 6 channels and around only 11% for 4 channels. The last 2 channels don't have anything too complex.
I am running the pre-compiled Camilla DSP v1.0.3 on Archlinux
I've tried to:
increase chunk size
reduce queue length
enable rate adjust
reduce the adjust period.
All to no avail - the crackle / pop is immediately available. there no signs of buffer depletion.
Is the Pi just not grunty enough? I'd appreciate any help / advice please or please let me know if this should be a new thread. thank you
Edit: I'm just testing by playing WAV files via aplay. Whether I go direct (to loopback aka hw:1 in my system) or via an ALSA plugin to resample. no difference (still crackles)
Last edited:
USB bus throughput might be the culprit? Not sure about the 3+ revision but the former RPi3 is not actually capable of true 480Mbps USB2 speed whith onboard network card connectet to the onboard USB bus. Not sure how the later 3+ revision with updated ethernet adapter make use of the USB bus?
Possible starting point to research this https://elinux.org/RPi_HardwareHistory
Fixed the crackle issue. using alsa plug rather than direct hw playback device. I presume the plug handles rate conversion mismatches?. here's the revised devices section:
CPU with 44.1/16 throughout is ~ 16% so hopefully not risk of underruns. Can't hear any. This is 6 channels and IIR filters
CPU with 44.1/16 throughout is ~ 16% so hopefully not risk of underruns. Can't hear any. This is 6 channels and IIR filters
Code:
devices:
samplerate: 44100
chunksize: 4096
queuelimit: 1
enable_rate_adjust: true
capture:
type: Alsa
channels: 2
device: "hw:Loopback,1,0"
format: S16LE
playback:
type: Alsa
channels: 6
#device: "hw:0,1"
# no crackle going via plug!
device: "plughw:0"
format: S16LEYou are correct. My ALSA knowledge is dismal sadly.
I had assumed the plug will be handling the rate/sample conversion which fixes my "crackle" problem before routing to hw:0,1 (the HDMI output on the raspberry pi)
However plughw:0 is a 2 channel device and I need 6 channels.
What I want is "whatever" the plughw:0 does to resample/rate AND setup 6 channel output. I'd really appreciate a couple of pointers to try in my asound.conf if you have the time please! thank you.
I added support for most CamillaDSP IIR filters with YAML import/export in the latest REW V5.20.14 early access build.
Thanks for your help Henrik,
I've uploaded the file that plays clean here:
https://drive.google.com/file/d/1pujm-CMVjAILjSCdz3GrIAjo7UCrargD/view?usp=share_link
it is an REW 10-20000Hz 44.1/16 stereo sweep.
I can play it without crackle via:
aplay -D hw:0,1 256k_ntr2.wavI've uploaded both crackle and no crackle Camilla DSP configs. As you can see, all I am doing is switching mixers for 4 (no crackle) and 6 (crackle) channels.
I can't work out how to record the crackle itself. Any pointers on how I configure CamillaDSP please? (or maybe I create some sort of ALSA device to route the output somehow). My alsa knowledge is very low..
I will try and figure out how to record the crackled output.
What I do know is if I replace the direct HDMI hw output device with the plug version (which must do resampling?) then I have no issues... the problem is , the plug is 2 channel. I am trying to work out how I create a multi-channel plug
#device: "hw:0,1"
# no crackle going via plug!
device: "plughw:0"The problem seems to be setting 6 channels of output on this 7.1 device. ALSA is remapping the channels as well as causing the crackling.
Taking Camilla DSP out of the loop I can reproduce the error using this ALSA config:
If I change channels to 2,4,8 = no crackle
Only 6 causes a problem.
Not an issue as I will just set to 8 channels for now (even though I don't need 2). I was just wanting to keep CPU usage as low as possible (may not increase anyway).
Taking Camilla DSP out of the loop I can reproduce the error using this ALSA config:
Code:
pcm.hdmi_hw {
type hw
card 0
device 1
}
pcm.simple {
type plug
slave sosimple
}
pcm_slave.sosimple {
pcm hdmi_hw
format S16_LE
channels 6
rate 48000
}If I change channels to 2,4,8 = no crackle
Only 6 causes a problem.
Not an issue as I will just set to 8 channels for now (even though I don't need 2). I was just wanting to keep CPU usage as low as possible (may not increase anyway).
If it's a stereo file, then it's not testing the device hw:0,1 in the same setting, i.e. in 6 channels. You can try e.g. sox which will generate any number of channels for you and play that directly to alsa.
You would need plughw:0,1 to test the same device.
I have not seen people complaining about broken 5.1 and working 7.1 HDMI PCM. IMO the problem with your setup is somewhere else.
Totally agree. I went straight to using Camilla to apply the crossover, just after doing a speaker-test to ensure each channel was working. I realised I skipped some important checking / configuration. Camilla DSP was never the issue (never meant to imply it was either sorry).
If it refuses to work properly with 6 channels, then it's an ok workaround to use 8 channels. It won't really affect CPU usage.
Don't spend s lot of time on recording. Recording a short sample of the sound from a speaker with the mic of a mobile phone will do fine. It's just to get an idea of what kind of glitches you are getting.
Don't spend s lot of time on recording. Recording a short sample of the sound from a speaker with the mic of a mobile phone will do fine. It's just to get an idea of what kind of glitches you are getting.
I am not planning to add VST plugin support.
It's possible to implement crossfeed just with the built-in filters and mixers. I know several people have done it and published the configs. Maybe someone remembers where to find them?