Increase default audio quality and reduce latency (#6936)#7164
Increase default audio quality and reduce latency (#6936)#7164ViniciusMassuda wants to merge 3 commits intomumble-voip:masterfrom
Conversation
- Increased default bitrate (iQuality) from 40000 to 96000. - Reduced default frames per packet from 2 (20ms) to 1 (10ms).
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (1)
WalkthroughAudio quality default settings were adjusted in the Settings structure. The 🚥 Pre-merge checks | ✅ 3 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
| float fVADmin = 0.80f; | ||
| float fVADmax = 0.98f; | ||
| int iFramesPerPacket = 2; | ||
| int iFramesPerPacket = 1; |
There was a problem hiding this comment.
omg
- 28% overhead from current (and already pretty high) value of 14%!
- 2x packet rate (sic!)
NO, thanks! it would introduce unnecessary packet loss in periods of higher load (read: prime time) or technical issues on networks
you should fix your audio setup first after making proper measurements! mumble itself is already blazing fast!
FYI: 20ms is already 2/10th of average human reaction, way MORE than enough for SPEECH!
p.s.
i've been using murmur + mumble for the whole 16 years already (since ~2009 or so) and heard many devices in different scenarios, short and long distances (up to 9000km) and can state next:
10ms: only for wired LANs, or very specific scenarios like real-time singing or musical instruments20ms: good default, balanced overhead vs packet rate40ms: much better for mobile/radio/Wi-Fi devices or really looong RTT (physical distance) to server
| /// backend. | ||
| QString qsTTSLanguage = {}; | ||
| int iQuality = 40000; | ||
| int iQuality = 96000; |
There was a problem hiding this comment.
agreed
this default could be increased about 2x:
40000with overhead totals:<= 64kbps80000with overhead totals:<= 96kbps- server configuration can limit it back -- nothing will break, yay!
please test these values as DEFAULTs and report your findings!
thank you!
There was a problem hiding this comment.
Thanks for the technical breakdown @winex!
I've updated the code to 80000 and 2. These default values are ideal because:
Setting iQuality to 80000 hits the "sweet spot" for high-fidelity audio, while staying safely under the 96 kbps server-side limit once network overhead is factored in.
Maintaining iFramesPerPacket = 2 ensures there is no regression in network performance, as this is already the proven stable value for Mumble’s current architecture.
Revert default latency to 20ms based on feedback
This PR updates the default audio settings as discussed in #6936:
Fixes #6936