werman/noise-suppression-for-voice

noise-suppression-for-voice | Arch Package does not work out of the box | carla

hugeblank opened this issue · 13 comments

hugeblank commented

No matter what I do I can't seem to get any output when putting audio through the plugin:
image
I also tried lowering and increasing values to their mins/maxes with no result.

I saw that this could be because librnnoise is missing dependencies ( #138 ), so I checked using ldd librnnoise_ladspa.so and got:

        linux-vdso.so.1 (0x00007ffc5457f000)
        libatomic.so.1 => /usr/lib/libatomic.so.1 (0x00007fa48a518000)
        libstdc++.so.6 => /usr/lib/libstdc++.so.6 (0x00007fa48a200000)
        libm.so.6 => /usr/lib/libm.so.6 (0x00007fa48a118000)
        libgcc_s.so.1 => /usr/lib/libgcc_s.so.1 (0x00007fa48a4f8000)
        libc.so.6 => /usr/lib/libc.so.6 (0x00007fa489f31000)
        /usr/lib64/ld-linux-x86-64.so.2 (0x00007fa48a57d000)

All dependencies satisfied, not the problem here.

I see that in the README there's some configuration that needs to be done in order for the plugin to work with pipewire, but I assume that was either not necessary for folks that are using a rack/patchbay, or was necessary but had already been handled by the arch package on installation. Am I doing something wrong?

werman commented

When developing the plugin Carla is one of the main apps I use for testing.

So, I installed aur/noise-suppression-for-voice-git, added plugin in Carla:

image

Set correct sample rate:
image

Routed mic through Carla:

image

And it is correctly denoised:

image

I saw that this could be because librnnoise is missing dependencies ( #138 ), so I checked using ldd librnnoise_ladspa.so and got:

Since you have built it yourself the dependencies should be good.

I see that in the README there's some configuration that needs to be done in order for the plugin to work with pipewire,

For Carla you also do the configuration by adding plugin and routing the inputs/outputs.

hugeblank commented

I did all of those things and it's not working for me. Is the buffer size important?
image
Note that without the noise suppression, and just routing vout_2 to the Calf Reverb mic output is fine.
Full patchbay in case it's relevant:
image

edit: oh wait, you added -git. Let me try doing that.
edit 2: didn't work but I should probably try restarting my PC.

werman commented

And you should try with the most basic config like I shown.

hugeblank commented

image
mic -> plugin -> headset
image
params set to around the same you have them set to

still nothing. are there any other diagnostic steps I could take to maybe isolate the problem?

edit: tested with -git, and did restart my PC before testing.

werman commented

You could look at "Logs" tab.

Also you could try VST3 or VST2 plugins which have GUI with some debug information.

hugeblank commented

image
I only see the lv2 plugin in carla. I see that there's VST3/2 plugins in /usr/lib/vst3 and /usr/lib/lxvst, but they don't show up. Carla is configured properly:
image
image

werman commented

Have you refreshed VST2 and VST3 plugins:
image

hugeblank commented

ahhhhh yep! that's what I needed to do. The ladspa plugin works perfectly. Thanks for the help!

  • 👍1
hugeblank commented

Not sure if I should leave this open or not, since the lv2 plugin still technically doesn't work

werman commented

Checked LV2 and it does work...

hugeblank commented

doesn't for me, and the vst3 plugin has 2 inputs and 2 outputs, but only one of them seems to function. The rest of them work though, so that's good!

werman commented

the vst3 plugin has 2 inputs and 2 outputs, but only one of them seems to function

Well, that's the only thing I can reproduce.

  • 👍1
digitalsignalperson commented

ahhhhh yep! that's what I needed to do. The ladspa plugin works perfectly. Thanks for the help!

same here. arch linux, no output with LV2, but LADSPA worked!

  • 👍1