glad to see PipeWire declaring itself production-ready, and I mostly agree:
Lack of S/PDIF was one of the many reasons I switched away from Fedora to Windows for a year (the others include power management, Nvidia, a few USB-C devices, MSVC, HDR, wireless game controllers, and audio equipment that doesn't run so well via Wine). I never noticed echo cancellation, but that's because I switched to using a USB lav mic years ago.
It's good to see rapid progress in these areas, although I wish there was a more intuitive and likely graphical setup for Pipewire/Wireplumber. (I haven't checked for the last year, maybe it's happened?)
Yesterday, I plugged in a single-sink, single-source USB audio device---the Shure MVX2U---using an up-to-date Fedora 39. It's an XLR/USB adapter that basically lets you plug in any standard professional microphone into a USB port. While there are two USB endpoints---one for config and the other for class-compliant audio---it's a really simple device.
On my Android phone? Works. On Windows? Works. On Linux? Fail.
No matter what audio driver I use, the sample frequency is off by 2x causing stuttering and playback that is an octave higher than recorded. Beyond that, the configuration software running in Wine doesn't detect the USB endpoint used to set gain, EQ, compression, monitor volume, etc. (I've already captured all of the config packets and hope to one day get around to writing a Python-based config client.)
So ... back to the regular grind of figuring out which group is actually responsible for unexpected device behavior, submitting bug reports, researching which workaround or conf file or udev rule or group membership is responsible for improperly matching sample rate and passing config data to the interface, and more complaining about how desktop Linux still falls short of expectations despite near-universal (albeit well deserved) praise of Pipewire, Wayland, Flatpak, Wine, and other modern improvements, and then getting chastised for not doing my part to contribute---despite the aforementioned USB capture and chipping away at each part in my spare time: https://gitlab.com/-/snippets/2595361#note_1620254493
P.S. while it would be nice to pin this on Shure, they're slow enough to release software and firmware upgrades for PC/Mac that would affect all end users (hanging software, audio distortion, high noise floor that's probably filterable, making the monitor sink disable-able, porting this adapter to their existing Android app). No one else makes an inline XLR-to-USB device. I guess it's on me or some other poor chap to get this device working on Linux until Shure gets caught by the competition (probably Behringer---yuck)
Interesting, FWIW I have the old model ("X2U") and it works flawlessly in Linux.
> No one else makes an inline XLR-to-USB device.
What about the mic mate? I've owned one in the past and it had some pretty awful white noise but maybe they've fixed themselves in the ~decade since I had one.
I tried again last night with just arecord, and that worked without issue on both 44100 and 48000 Hz. Something is amiss in Ardour, Reaper, and Audacity, so I suspect my sound server is just misconfigured and each DAW is using the misconfigured bit instead of direct device access. I'll try again later with dnf upgrade and a reboot.
Thanks for the recommendation. I'll put a watch out on that MXL. (It's not in my usual stores.) Noise, I can probably FetHead much of it away.
Lack of S/PDIF was one of the many reasons I switched away from Fedora to Windows for a year (the others include power management, Nvidia, a few USB-C devices, MSVC, HDR, wireless game controllers, and audio equipment that doesn't run so well via Wine). I never noticed echo cancellation, but that's because I switched to using a USB lav mic years ago.
It's good to see rapid progress in these areas, although I wish there was a more intuitive and likely graphical setup for Pipewire/Wireplumber. (I haven't checked for the last year, maybe it's happened?)
Yesterday, I plugged in a single-sink, single-source USB audio device---the Shure MVX2U---using an up-to-date Fedora 39. It's an XLR/USB adapter that basically lets you plug in any standard professional microphone into a USB port. While there are two USB endpoints---one for config and the other for class-compliant audio---it's a really simple device.
On my Android phone? Works. On Windows? Works. On Linux? Fail.
No matter what audio driver I use, the sample frequency is off by 2x causing stuttering and playback that is an octave higher than recorded. Beyond that, the configuration software running in Wine doesn't detect the USB endpoint used to set gain, EQ, compression, monitor volume, etc. (I've already captured all of the config packets and hope to one day get around to writing a Python-based config client.)
So ... back to the regular grind of figuring out which group is actually responsible for unexpected device behavior, submitting bug reports, researching which workaround or conf file or udev rule or group membership is responsible for improperly matching sample rate and passing config data to the interface, and more complaining about how desktop Linux still falls short of expectations despite near-universal (albeit well deserved) praise of Pipewire, Wayland, Flatpak, Wine, and other modern improvements, and then getting chastised for not doing my part to contribute---despite the aforementioned USB capture and chipping away at each part in my spare time: https://gitlab.com/-/snippets/2595361#note_1620254493
P.S. while it would be nice to pin this on Shure, they're slow enough to release software and firmware upgrades for PC/Mac that would affect all end users (hanging software, audio distortion, high noise floor that's probably filterable, making the monitor sink disable-able, porting this adapter to their existing Android app). No one else makes an inline XLR-to-USB device. I guess it's on me or some other poor chap to get this device working on Linux until Shure gets caught by the competition (probably Behringer---yuck)