procsiab/denoiseit-lc repository overview

⁠DenoiseIt Linux Container

A Linux Container minimal distribution based on Alpine, to make available the compiled binary of the DenoiseIt project.

An AUR package maintained by the author is publicly available, however a Linux Container is best suited to integrate this handy tool into a pipeline.

Thank You DragoonAethis, for putting together DenoiseIt!

Find out more on the author's repo⁠.

⁠Build the Container Image

Run the following command in this repo's directory:

podman build -f Containerfile.amd64 -t mycompany/denoiseit-lc .

NOTE: You may want to use the Containerfile.aarch64 to produce an image for an ARM64v8 target platform. In that case, the QEMU AARCH64 static binary (qemu-aarch64-static) must be in the same folder of the Containerfile.

NOTE: You can choose which commit of the author's repo use as a codebase for the build process, passing the argument --build-arg=commit=<SHA> to the podamn build command. Replace with the hash of the designed commit.

⁠Run the Image

From the folder you want to share inside the container, run the following command (replace podman with docker if using the latter):

podman run --rm -it -v:$(pwd):/workdir:Z docker.io/procsiab/denoiseit-lc:v1.0-amd64

You will be presented with a Bash prompt, inside the shared folder you chose with the -v argument.

⁠Prepare the audio for DenoiseIt

At the time of writing this, I find myself using a lot the following FFMPEG command to prepare a video file to be consumed by the denoiseit command: its input must be a WAV file, with a sample rate of 48KHz; therefore it seems you cannot feed it directly a container file (e.g. MP4) with audio and video streams in it.

In short, run the following command on your input.mp4 file to extract the audio and make it readable from denoiseit:

ffmpeg -i input.mp4 -acodec pcm_s16le -ac 2 -af "aresample=48000" denoiseit-input.wav

Afterwards, I stick together the output of denoiseit and the video stream from the original input file.