Mumble voice suggestions (and improvements)

[nta]

@nta

Thanks, I guess we should change the output volume scaling by default then.

This part is done now.

Some other native audio changes (making volume override non-directional, fixing some game allocator leaks) are being worked on still and are pending for a future release, as is submix support.

[Licious]

What about the mouth anim bug?

[nta]

Fixed, too, and deployed on canary client.

Also the bitrate was set to 24k by default which is a fair bit lower than most defaults so it’s configurable now (on canary too until next push) and defaults to 48k as well.

[Licious]

Looking great, if you manage to fix the voice restart bug and maybe sneak in submix support Mumble will really start to shine.

[nta]

@Licious

if you manage to fix the voice restart bug

still waiting for a full dump for a broken client

[manzarek]

Also connecting to an external server can lead to the mumble client crashing:

Mumble voip crashing

GTA V version: Steam Up to date: Yes Legit or Pirate copy: Legit Windows version: Windows 10 Did you try to delete caches.xml and try again?: Yes Error screenshot (if any): [image] System specifications: 1650 super, i5 4670k, 12GB What did you do to get this issue? Connect to a server using an external mumble server (MumbleSetServerAddress) What server did you get this issue on: gta.bay.life CitizenFX crash zip fileCfxCrashDump_2020_10_01_20_27_50.zip (1.5 MB) CitizenFX.log file Ci…

I think it is due to the disconnect/reconnect by the MumbleSetServerAddress

[Flawws]

I don’t think this post was made for people to dump other issues here … still waiting for the actual thread to get a yes or no if it’s going to be worked on

[darksaid98]

@nta

Just use native audio.

I’d say that’s a no. Either way, I think we should apply our energy to polishing native audio rather than the old system since that would lead to native audio becoming the new standard. Bettering VOIP for all of the community.

[Flawws]

But Native audio just isn’t RP immersive. This is why people choose to use Toko over Mumble. No one wants to hear someone who is far away sound like they’re in your head.

[darksaid98]

Not sure what setup you were running, this wasn’t my recent experience with native audio. On foot having a conversation sounded exactly like it did when I was using Toko.

I just need to tune reverb inside interiors since it is somewhat overkill. The only edge Toko still has is it comes with the RadioFX plugin, however, reading bubbles reply it seems we’ll be able to accomplish this in Mumble soon anyways.

Because of audio occlusion, I’d say native audio is much better suited for RP than the clip through wall nightmare that Toko and the old Mumble was.

[Flawws]

All right, I’ll give Native audio another go.
Can you tell me which options you have enabled?

I assume:

voice_useNativeAudio, voice_useSendingRangeOnly, and maybe voice_use2dAudio?

Thanks.

[Flawws]

All right, I switched to:

setr voice_useSendingRangeOnly true
setr voice_useNativeAudio true

You are correct, it is not bad, but high speeds still causes voices to cut out.
I’m on version: 3161.

[manzarek]

@darksaid98

The only edge Toko still has is it comes with the RadioFX plugin, however, reading bubbles reply it seems we’ll be able to accomplish this in Mumble soon anyways.

Also you can talk outside of the game and use PTT in NUI without having to play with SetNuiFocusKeepInput and also some issue sometimes with external servers on Mumble

[Flawws]

Refer to: Allow mumble voice whilst NUI enabled
I’m sure it’s being worked on already.

[darksaid98]

Running Canary & server artifact 3160, was running ~315* yesterday but the new artifacts should do fine.

+setr voice_use2dAudio false
+setr voice_use3dAudio false 
+setr voice_useSendingRangeOnly false # Had no effect with my testing or I'm pepega
+setr voice_useNativeAudio true

Again, there was still way too much reverb inside interiors for my liking, however outside and some vanilla interiors seemed fine.

Use these natives instead of NetworkSetTalkerProximity, since NetworkSetTalkerProximity sets your input and output distance. I found that dividing my old talking ranges by about half worked pretty nicely.

MumbleSetAudioInputDistance(talkingRange) -- You talk up to this distance (between players).
MumbleSetAudioOutputDistance(hearingRange) -- You can hear up to this distance (between players).

I did manage to brick Frazzles radio completely but I’m going to have a look at that later today I believe.

The high-speed cut-outs are related to crossing between grid zones. I assume the whole clear voice targets and reassign part is what is causing this.

[Flawws]

@darksaid98

The high-speed cut-outs are related to crossing between grid zones. I assume the whole clear voice targets and reassign part is what is causing this.

So it’s a script related problem not a out of the box mumble related issue?

[manzarek]

No its not because of the grid when using persistent channels.
it happen when your NetworkSetTalkerProximity is too low. So most likely position sync issue/delay.

[Licious]

Try enabling 2d audio for car passengers

[Flawws]

That’s a workaround, not a fix

[Flawws]

Would I replace

NetworkSetTalkerProximity(mumbleConfig.voiceModes[playerData.mode][1] + 0.0)

with:

MumbleSetAudioInputDistance(mumbleConfig.voiceModes[playerData.mode][1] + 0.0)
MumbleSetAudioOutputDistance(mumbleConfig.voiceModes[playerData.mode][1] + 0.0

not sure if I’m using it correctly.