Hi there,

I´m still hoping to reach my goal getting studio-quality-sound in vmix-calls :-)

The germany-based coder Sebastian Reimers has developed a vst-/au-Plugin for digital audio workstations, which brings a studio-level audio-quality for calls over the internet. Maybe it´s worth to check out his technology for you nice guys at vmix to integrate this easy to use software in vmix / vmix-call?!

All you need to know can be found under www.studio-link.de .

Thanks for checking this out and have a great time,

Jonas

wich version? there is studio link and studio link [on air]

All the various services mentioned leverage WebRTC, including vMix Call. WebRTC standards support only two audio codecs; G.711 and Opus.

G.711 is old skool narrowband telephony, sampled at 8 KHz.

Opus defaults to sampling at 48KHz, which is truly studio quality. However, it's quite variable as it's adaptive to the network conditions. If the network connection struggles the codec can vary its operation to ensure continuous audio, event at the expense of quality.

Curiously, when you say ISDN-quality that implies G.722, which was the codec typically associated with an ISDN BRI circuit.

G.722 is characterised by sample rate of 16 KHz, with a usable passband of 50 Hz - 7 KHz (aka HDVoice.) It's is still widely used for voice. Most IP telephones (soft & hard) support G.722.

Chrome and Firefox both support G.722 audio, even though it's not part of the formal WebRTC standard.

I suspect that vMix call uses nominal audio constraints when asking the browser to deliver audio.

In my experience, when a WebRTC link is not satisfactory you have a problem with network performance or the end-users situation. It's not usually a problem with the codec or sample rate.

Telephony grade headsets (Logitech, Plantronics, Jabra, VXi and the like) have microphones that roll-off high frequencies. Some at 4 KHz, most at 7 KHz. Few headset mics reach beyond 10 KHz, which is fine for voice, but may seem a little dull compared to a proper studio microphone.

Further, the browsers audio handling is less than stellar. In particular, we've found that echo cancellers in browsers quite often degrade the sound from the mic while there is also sound flowing to the speakers. They're not entirely full duplex. This leads to mangled sounding speech.

We have found it useful to be able to turn off echo cancellation, noise reduction and auto gain control in the browser. In the WebRTC world the "constraints" used when calling getusermedia() allow this to happen.

Of course, turning off EC implies that your guest us using a headset, or at least headphones to avoid an echo loop.

P.S. - My employer was the very first HDVoice-capable conference service. We still support audio conferencing using PSTN and SIP, supporting G.711, G.722 and Opus.