Category: music

Posted on

Some tips for musicians working with Windows

I haven’t done much on Windows the last 20 years or so, but through a thread post on recording.de I stumbled upon Pete Brown’s “Unofficial Windows 10 Audio Workstation build and tweak guide” which I found nice, and which could be helpful for those who make music and try to record themselves using machines with Windows 10 or 11.

I knew for instance that the Cubase support staff recommends to switch off the so-called “E” (for “efficiency”) cores on newer (12th gen and up) Intel processors, but that’s true only for Windows 10, not 11, as the commenter on the thread also wrote:

“Steinberg empfiehlt an anderer Stelle auch, die E-Cores zu deaktivieren, mit dem Argument, der Audio-Thread würde als “Hintergrund-Thread” sonst am Ende dort landen. Für Windows 10, dessen Task-Scheduler noch nichts von der P-/E-Core-Architektur von Intel weiß, mag das sogar einen gewissen Sinn haben. Bei Windows 11 ist das aber auch auch Bullshit, denn der Audio-Thread ist eben KEIN Hintergrund-Thread und als Vordergrund Task mit hoher Priority landet der IMMER auf einem P-Core. Unter Windows 11 die E-Cores zu deaktivieren bringt nur Einbußen, weil man einen großen Teil der CPU-Ressourcen schlicht lahmlegt.”

Pete from Microsoft explains that in more detail, so have a look at his 3-part guide which started on my birthday in 2021 and was finished in December. Pete’s history, of which he spoke a bit in the first embedded video there, ran parallel to my own one for a while – like him, I also started with a Commodore C64, and later had an Intel 286 processor in my first “real” PC, first with DOS, and later with Windows 286, a predecessor of Windows 3 (and the better known 3.11). But I gave up on all that while he joined Microsoft to work on their audio-related components, which is much more than the “pro audio” stuff we musicians care and think about.

Interesting guy, and I hope that these links are helpful for some of you.

And like always, thanks for reading.

Posted on

Some really good tutorials and tools for musicians

If you make music and even record yourself and care about sound, you should have a look at the Youtube channel from Dan Worrall, a senior live mixer and sound engineer from the UK.

Dan is using the really cost-insensitive (read: cheap but worth it) DAW Reaper, and he also works with the guys from FabFilter. These are professional tools, although he also often checks the “stock plugins” from DAWs like Reaper, Ableton, or Cubase, and compares them to the more expensive ones.

On the FabFilter channel he has some great tutorials about “The Philosophy of Bass“, and also one of the best guides about compressors I’ve ever seen, the “Beginner’s Guide to Compression (part 1)“. Really recommended stuff if you care about making a singer or a voice-over heard. And even if you’re using just the “stock plugins” of your DAW, you can achieve great results if you know what to do…

I myself very often use the Linux Studio Plugins, which are free and open source plugins for – as the name says – Linux. On Arch, I have the newest ones as the listing shows:

unfa (from Poland I think) shows the LSP 16-band stereo eq here, and the LSP Project has their Youtube channel as well of course. So go and have a look at Dan’s and unfa’s, and enjoy your mixing.

Like always, thanks for reading.

Posted on

How to avoid double conversions

My computer normally runs with a 48000Hz sampling rate for audio, that’s the one you would also use for most video productions like when producing something for the ‘tubes and such.

But CDs had 44100Hz which is also perfectly fine, and which saves some space if you record with that frequency – and so some (or even most?) of my friends over at Wikiloops use that sampling rate for their music.

No problem; Ardour checks when importing, and would normally automatically convert the imported 44.1kHz files to 48kHz ones. But that would mean that I’d make it harder for others who would probably like to add my single tracks to the rest (with 44.1Khz). And also, each conversion diminishes the quality just a tiny bit, so it’s always best if/when you can avoid these and use the material as it comes. Even Ardour says so:

But how to temporarily set Ardour to 44.1kHz? Easy in case you’re using the new pipewire! I just wrote the following short shell script which I named ‘ardour44k.sh’:

pw-metadata -n settings 0 clock.force-rate 44100
PIPEWIRE_LATENCY=128/44100 pw-jack ardour
wait
pw-metadata -n settings 0 clock.force-rate 0

So if I start Ardour using that, I can use 44100Hz just perfectly fine – and when Ardour ends, the system will be set back to 48000Hz; just what I wanted. Here are some screenshots from Ardour’s Edit and Mixer windows while it ran with 44100Hz:

And when I stop Ardour, the script ends with:

set property: id:0 key:clock.force-rate value:0 type:(null)

Just what I always wanted, as Tigger would say 🙂 Thanks to the pipewire crew, and thanks also to my friends over at Wikiloops 🙂

Oh, and what I’m also using with pipewire (which is now the standard audio “engine” on Debian and most other Linux distributions) is a program called qpwgraph, and that is a graphical patchbay like the older tools (qjackctl, Carla, Catia & Co). Looks like this:

Here you see three inputs for my upright bass on the left, which go into Ardour. The right side shows Ardour’s monitor section and its metronome going out into my sound interface, and from there, into my headphones. The outputs of individual tracks go back into Ardour’s master track, which gives you this figure 8 shape. Easy peasy, isn’t it? Virtual cabling, so to say…

Thanks to you for reading.

Posted on

… and back to a triple boot system

I had installed the new and upcoming Debian 12 (aka “Bookworm”) on my machine, parallel to the stable version (Debian 11 aka “Bullseye”) and Windows 11 – so I had a triple boot operating system again since a while.

The Windows part is a bit controversial – since I have this new self-built machine with the AMD Ryzen 7 5700G processor, my Windows 10 offered to upgrade itself to Win11 which I did. But in recent times, more and more reports arrive saying that Microsoft is forcing ads upon its clients all over the place – I’m running it with a local account and haven’t seen them yet. But the day I will, it’ll be a “goner” as they say.

Anyway, I was also looking at Arch Linux again since that is always the latest and greatest (like Debian unstable aka “Sid”, it’s what they call a “rolling release”). First I tried some things in virt-manager and QVM/KEMU, but then I decided to overwrite my old stable Debian 11 with Arch. Went fine, except that both Arch and Debian have different ideas about where their respective /boot folders are mounted. They’re both of the EFI partition alongside Windows, but still – anyway, maybe that’s a good thing; at least they won’t overwrite each others’ kernels and/or firmware. But both run fine, even if at the moment I can’t start Arch from Debian’s grub or vice versa; doesn’t matter.

Once I damaged my Debian 12 part, accidentally deleted the firmware, so it wouldn’t boot. Didn’t matter the slightest bit since for Debian I’ll always have my /home and system parts on different partitions – so wipe it with the latest (RC2 at this time) installer – and I just saw that since today there’s even an RC3 installer – and all is well. Except of course a bit of manual labour with reinstalling Ardour and all, but even that could be remembered and more or less automated when using Debian; have done so in the past with saving and later restoring its list of installed packages…

Anyway, here’s a screenshot where I newly registered the only commercial program I’m using on Debian, it’s Sonarworks’ Reference 4 headphone correction which I use in the monitoring bus in Ardour:

Haven’t installed Ardour in Arch (yet) since at this moment they’re close – with version 7.3 in Debian’s “unstable” and 7.4.1 (or so) in Arch.

The only programs which I still use in Windows from time to time are the OM Workspace from the former Olympus guys, and Nik’s Silver Efex Pro2 which you could get for free from Google for the time they’ve owned it (sold by now to DXO, not sure what they’re going to do with it…). So it’s kind of a jump-through-the-hoops for photography, but for music I’m on Debian alone since long, like for everything else as well.

And now, from time to time, I’ll have a look at/into Arch again. Normally when you read about new program versions with new features somewhere, looking into Arch means that you’ll have that newest version already. And Debian will stay my main and stable machine once that Debian 12 will be made official on June 10th.

Oh, by the way: Arch is slim, as they say on their homepage. Unlike Debian or other distributions, it doesn’t come with LibreOffice or any other programs pre-installed, so it’s *you* who has to decide what’s needed. Together with the Gnome desktop plus Firefox, Thunderbird and a few goodies, even with all my Wikiloops albums copied onto it, it’s still less than 10GB as you can see here – one third of that is my data so far:

That blue and purple stuff is all music (with the purple bits being published albums, and the outer blue one being raw and unpublished songs in .wav form)…

Like always, thanks for reading.