Package Details: noisetorch-git 0.12.2.r15.g3b04af2-1

Git Clone URL: https://aur.archlinux.org/noisetorch-git.git (read-only, click to copy)
Package Base: noisetorch-git
Description: Real-time microphone noise suppression on Linux.
Upstream URL: https://github.com/noisetorch/NoiseTorch
Licenses: GPL3
Conflicts: noisetorch
Provides: noisetorch
Submitter: erbrecht
Maintainer: gileri
Last Packager: gileri
Votes: 11
Popularity: 0.000015
First Submitted: 2020-07-18 04:21 (UTC)
Last Updated: 2023-08-06 18:55 (UTC)

Latest Comments

« First ‹ Previous 1 2 3 4 Next › Last »

gileri commented on 2022-05-19 16:15 (UTC)

Thank you for sharing that. I contemplated sabotaging temporarily this package with regard to this message, but decided against it, as the author only spoke about binaries.

But as you pointed out; commits could have been tampered too, even if first analysis seem to have not surfaced issues : https://github.com/lawl/NoiseTorch/issues/254

If multiple people here do not trust the git history I'll sabotage this package for a few days/weeks.

LiveLM commented on 2022-05-19 16:04 (UTC) (edited on 2022-05-19 16:07 (UTC) by LiveLM)

You guys likely already know this, but I'll say it just in case.
Yesterday, the author of NoiseTorch announced that he believes his key was compromised, and urged users to not trust neither the binaries, nor the source code for this project:

Release 0.116:

POTENTIAL COMPROMISE  
I HAVE REASON TO BELIEVE AT LEAST ONE OF MY SYSTEMS HOLDING KEY MATERIAL HAS BEEN COMPROMISED.  
DO. NOT. TRUST. THE. BINARIES.  
I'm deeply sorry about this, but it's better to be safe than sorry.

Readme.md:

At least one of my sytems has probably been compromised, dont use the source either. One could hide things in a large diff.

https://github.com/lawl/NoiseTorch/releases/tag/0.11.6
https://github.com/lawl/NoiseTorch/blob/master/README.md

gileri commented on 2022-01-16 09:54 (UTC)

Hello, thanks for the comments. On my system it worked without a hitch, I think pkexec did not work on your system [1} :

Anyway, I setcaped the binary in post_install/post_upgrade as you suggested. It should work on your end now.

[1] https://github.com/lawl/NoiseTorch/blob/245455d508f21e888b18e168b0a930a32ab5a24d/capability.go#L67

ronin commented on 2022-01-14 14:18 (UTC) (edited on 2022-01-14 14:24 (UTC) by ronin)

You should add sudo setcap 'CAP_SYS_RESOURCE=+ep' /usr/bin/noisetorch in your install file probably like something similar in voidlinux. im not sure if arch has post_install() function for PKGBUILD's but pls do add that if there is.

EDIT:

There is one after all in another AUR package https://aur.archlinux.org/cgit/aur.git/tree/noisetorch.install?h=noisetorch

montazar commented on 2021-05-02 02:35 (UTC)

I had issue with setting the setcap permission through the GUI. It kept exiting with a non-zero value so I had to manually set the setcap permission:

sudo setcap 'CAP_SYS_RESOURCE=+ep' $(which noisetorch)

I assume this is because NoiseTorch is looking for the installation in the ~/.local/bin directory when trying to set the said permission while this package installs it in the /usr/bin directory.

erbrecht commented on 2021-01-22 14:47 (UTC)

This is working again. Now that they are using their own implementation of the ladspa plugin, I was able to remove the dependency on noise-suppression-for-voice, and we no longer need the patches.

techge commented on 2021-01-20 11:23 (UTC)

With latest changes (https://github.com/lawl/NoiseTorch/commit/b747084bb0a033e10f55058b73b06238a666d8ae) the dependency of package noise-suppression-for-voice is no longer necessary as it is build-in the main package.

erbrecht commented on 2021-01-18 16:05 (UTC)

Thanks for the feedback. I'll be able to take a look at this tonight or tomorrow morning. Hopefully it's something I overlooked and we can get it working again.

NiklasM commented on 2021-01-18 15:58 (UTC)

Since the version 0.9.0.r5.ge7072b2 noisetorch doesn't work for me any more (But I am also a Manjaro [testing] user). If I load it on my microphone I always get a yellow warning Inconsistent state, please unload first. in the GUI. With downgrade I tested if previous installed versions (from 0.7.2.beta.r0.g02665f1 to 0.9.0.r3.g4f7ac32) did still work and all did. Is this a Manjaro specific bug (version of pulseaudio is 14.2) or does this also happen on Arch with the latest version 0.9.0.r5.ge7072b2?

I compared the logs (in the /tmp/ directory - https://gist.github.com/AnonymerNiklasistanonym/7a60bc422683cae61a5b6b32e18d4474) and in the latest version the biggest difference is after the line Loaded null sink as idx: 28 the next line says PulseAudio error: commandLoadModule -> Module initialization failed.

If it works on Arch you can see this issue as resolved for now - I'll just wait and check future versions while using the downgraded one that still works fine.

erbrecht commented on 2020-12-11 14:32 (UTC)

now depends on noise-suppression-for-voice (AUR) for the librnnoise_ladspa.so library. it was building the ladspa plugin from source previously, and embedding the .so file into the binary. should be more in line with packaging conventions/guidelines now.

new non-git release version coming soon!