XSB 2.0 Beta 3: supported input formats

[Timothy Walker 1294725]

XSB-AFV-Native: Sat Dec 14 20:06:46 2019: AudioDevice: device Built-in Microphone - can't handle sampling rate.

XSB-AFV-Native: Sat Dec 14 20:06:46 2019: AudioDevice: device SB Omni Surround 5.1 - doesn't support monaural audio

 

So I gather XSB 2 requires mono input, but what sample rate(s) and sample format(s) are supported? The log doesn't mention that.

 

In case of incompatible devices, it might be nice to log what XSB wants, as well as the detected capabilities of the device?

[Timothy Walker 1294725]

OK, so the plugin/AFV seems to expect 1-channel @ 48,000 Hz, and maybe doesn't care about the sample format.

 

I went back to XSB 1.3 for testing, and the mic setup dialog was able to use my 2-channel mic using the "setup mic" dialog to record and play my voice back without issues or complaints.

 

I understand the Opus codec and/or encoder only accepts 48 kHz input and software resampling is a PITA to implement, but requesting 1-channel input when 2-channel worked fine before seems like a pretty annoying regression. A proper downmix is not that hard, but if it's too tedious, surely discarding either of L or R channel should be trivial enough?

 

I got AFV working for a test using Rogue Amoeba's loopback version 1 (the latest loopback version is apparently unable to map 2-channel mics to 1-channel, and Audio MIDI Setup cannot do it either), but I'm somewhat unhappy about having to pay $32 for a legacy, no longer updated piece of software just to make XSB work (though at least I don't get hard-to-diagnose XSB crashes on my system like the swift client does).

 

Please implement support for Stereo microphones in XSB

[Craig Deitch]

I’m back to using v1.3.3 and AFV while waiting for a new beta. Also a PITA but better than the alternative. The latest version of xPilot causes terrible stutters on my system.

[Christopher Collins]

OK, so the plugin/AFV seems to expect 1-channel @ 48,000 Hz, and maybe doesn't care about the sample format.

 

Internally, AFV requires 48Khz, 1ch - sample format is left to the implementation. AFV-Native uses float internally and portaudio generally provides format conversion when it's necessary.

 

I went back to XSB 1.3 for testing, and the mic setup dialog was able to use my 2-channel mic using the "setup mic" dialog to record and play my voice back without issues or complaints.

 

Don't compare XSB 1.3's audio (VVL) to 2.x (AFV-Native). They're completely different implementations with completely different requirements - there's no common code at all, and we cannot take VVL code into AFV-Native due to the licensing differences.