Practise / Cover Workflow

I am looking for some advice on improving my cover workflow.

I want to practise / do takes of my vocals over another cover online before I do my own guitar. JACK stops non-jack audio.
Currently, I must YouTube to mp3 the cover, convert it to WAV in the Linux terminal and import it in to Ardour.
I understand this is seen as heresy but don’t see the problem as the vocals and guitar that make the final cover will be of the highest quality.

How do I do this or why is my use of Ardour / JACK, unreasonable / wrong?

I would do it the other way around by first recording in the guitar with the metronome and then practice with that. The advantages for this workflow are: you can adjust the tempo to your liking and also transpose the music so that it better fits your vocal range. Also when you record with the metronome you can copy and paste patterns of music and Ardour will take care of it staying in sync. You just need to record one successful take of each part of the song and then you can copy and paste the whole song together.

Jack can be configured to work with audio from youtube, either by running pulse audio (see https://www.linuxmusicians.com/viewtopic.php?f=19&t=13234), which is configured by default in, for example, AVLinux2017 or if your version of firefox supports alsa audio using an alsa loopback device which is done in kxstudio. With these methods you can record youtube directly into Ardour.

Epiphany web browser has Jack support (at least here, I’m using Elementary OS 0.4.1). When you open a tab with youtube it’ll create a jack client that you can route to Ardour and record the incoming audio there.

Good luck!