Mumble voice suggestions (and improvements)

First of all, let me start this off by saying Mumble is underrated and quite amazing.

I currently use 3D audio and have been using it for 2 days now, it works great, but I do have some issues and/or suggestions that I’d like to put into one topic:

  1. Currently 3D audio heavily targets one side of the ear. For example, if a player is on the right side of you, it will only provide audio in your right ear. I believe that some audio should be provided to the left ear.

  2. When at high speeds with 3D voice and you are talking, the audio cuts in and out. This is a known issue, just figured I’d put it here though :blush:

  3. Please could we get a way to create voice effects. For example, when using the radio, I would love to apply a little distortion and static to voices. If this could become a feature, it would be great.

  4. Object collision voice. I’m not sure if I named it correctly, but whist on 3D audio audio is not suppressed when there is a door in between 2 people or a wall. It would be great if this would become a part of 3D audio, I’m aware it is a native audio feature.

That’s about it,
If you have any more suggestions, feel free to comment.

Thank you.

2 Likes

Use native audio.

Audio position calculation algorithms are hard :frowning:

Hm, I thought this was fixed already - doesn’t happen in nativeaudio I guess though.

Likely only for native audio, as it has a working DSP framework built in already.

Just use native audio. :confused:

Is there anything we can do to help test so you can improve the audio position calc algorithms?

D:

I would use Native audio but I don’t want people who are shouting or speaking normally at a distance from me sound like they’re on top of me. Main reason why I choose to go against using native audio.

Until there are more publicly available infos on how to tweak occlusion/reverb values I dont see native audio becoming the most commonly used audio mode, standing in central garage and hearing people as if you were in a cave is quite annoying. Voice filters are also a must, if you want people to use Mumble instead of Toko there needs to be 1:1 parity in terms of features, which we currently don’t have.

I’m probably gonna be crucified for saying this, but Toko has overall better audio quality, I know Mumble uses Opus too but bitrate or something else makes a lot of difference, Mumble is on par with Discord in terms of voice quality, but dont tell me it sounds as good as Toko.

Lower the output volume.

You shouldn’t even have to “tweak values”, wtf?

If there’s any issue, report it; the defaults should be fine, if it’s “unusable” without tweaking that is 100% unintentional, note that it’s only been tested with a fake audio source from a standalone Mumble client echoing a music file, as I don’t have a RP server or such myself to be using voice with in “real” contexts.

Really, native audio is meant to be “just enable it and it works” which one can’t even say about 3d audio, if anything is preventing it from being that it’s a bug and why did you not report it and assume you need “tweaking” which is kept “privately” (why would we keep any info privately? legit?) which makes no sense at all…

Also, does “Toko” whatever it is even have their audio code open sourced at all? Can you record a comparison and maybe do some testing of bitrates with a standalone Mumble client? I don’t have a “TeamSpeak Systems” license or friends to test voice with so I can’t test myself…

Also, sorry, but your two posts still nag at me… “go against” even, not just “decide not to”, but explicitly being against a feature that’s meant to be universally better than the old X3DA audio stuff… for reasons you’re not even sharing, and then wanting things improved in X3DA that come for free with native audio but you apparently decided not to use… no, “go against” even, like the people of Belarus went “against” their dictator, and for supposed issues that weren’t even known or mentioned in view anywhere at all, that you just assume are true and unchangeable, or maybe “so obvious if you use it that it’d be futile to report”… note that I don’t use FiveM except playing on RSM every so often and development work… I don’t hit all these issues you people get on normal RP servers that are all cookie-cutter as there’s no better free-to-use alternative as nobody is sharing and those who are get their stuff resold and hated enough to be instantly demotivated… seriously, I can’t do everything by myself but with everyone being “me me me me me” and those who aren’t easily being fucked over by “yeah this is an elitist culture” as I see any time I see RP stuff mentioned and just leaving… yeah… I get the whole point, but seriously, you folks should kinda have the freedom to report any issue, no matter how trivial, just as a “note”, as long as it’s clear it’s not a “ask for help” but just a “hey I hit this when trying that, is this expected?” as usually it isn’t…

1 Like

if it’s “unusable” without tweaking that is 100% unintentional

I wouldnt define it unusable, it comes down to personal preference and its very dependant on the location you’re in, being able to decide where and how much reverb/echo i want in each location is a must imo.

why would we keep any info privately?

Trust me if I knew I’d gladly share how to do it, I’m doing research and talking to other devs, thats how I know some people have it figured out but dont share it publicly.

does “Toko” whatever it is even have their audio code open sourced at all?

I dont think it is, but there is no need to run any crazy tests, the difference is the same you can notice by switching from Toko to Discord, nowadays people are used to Discord sound quality, but for those of us who still use it its night and day, obviously Mumble/Discord arent that bad and you get used to them pretty quickly, but voice quality, the things I mentioned above and all those random minor Mumble bugs such as: [Help] Mumble voip needs to be turned on and off to be able to speak (Mumble-radio)
are all obstacles for Mumble to become THE Fivem voice system.

Better docs and more functionality would also be useful, I still dont know whether channel listeners can be used as replacement for channel voice targets and it would be cool if we had a MumbleGetVoiceChannelFromServerId alternative which returns the channel “name” instead of internal Mumble ID.

not able to write yet another very long reply on phone but a few points:

  • still stuff you experienced wasn’t reported :confused: how can we improve if people don’t mention stuff outright
  • i don’t have a teamspeak license or people to test with as said above so i can’t even use “toko” to try
  • “better docs” shouldn’t need to matter to most people since if you have to write a “custom” system there’s clearly something inherently wrong and that should be solved, not by writing “docs” of the triviality level you’d find on wikihow (also as it’s impossible to pre-empt every question that again it’s easier to answer when people ask for a specific thing, or just hit issues oneself but again I don’t have an RP server or people to test a voice resource with or I’d have written one myself and bundled it with server-data so none of you have to constantly suffer through researching stuff again and again as nobody who gets stuff working bothers sharing at all :confused: )

I can provide you with a Teamspeak server and people to test with if you’d like.

I’m your friend :tada:

  • still stuff you experienced wasn’t reported :confused: how can we improve if people don’t mention stuff outright

I can confirm you this still happens: [Help] Mumble voip needs to be turned on and off to be able to speak (Mumble-radio)

Today I switched to Mumble on my prod server and many players experienced it, they could hear people speaking but others couldnt hear them, it never happened while testing with 4/5 players. When it happened I tried to gather as many infos as possible while troubleshooting, resetting the voice target/voice channel didnt work, only way to fix it was to restart voice chat. One interesting thing I found out by looking at Mumble client is the player inactivity wasnt updated whenever he talked, so the client isnt sending voice packets for some reason.

This is very game breaking on RP servers and should be fixed asap, I know its hard af to repro and debug it with the tools at your disposal, if you want I can give you access to my prod server and test with 100+ players online. I really don’t want to go back to Toko, but we can’t keep going like this.

1 Like

I also have the above problem.

Yeah still happening and it makes a terrible experience for everybody.
VOIP is the what makes FiveM unique for me.
If there is any way to help, i am able to do with my all.

If i want to report some issues about VOIP, is there anything specific i need to do?

The understanding was that this was fixed when the whole connection loop was reimplemented… if it wasn’t, why did nobody report this in sight before? :frowning:

A full memory dump from Task Manager of a client that ‘pseudo-lost’ connection might help if anyone can upload one of these with their internet connection (as they do become 1.5 GB or so even when compressed) - also does anything of note show in the console/log with developer 1 set?

1 Like

I can probably get you a full memory dump. I’ll try for today <3

I’m not sure if this: Mumble constantly trying to connect bug is related but it’s something that I encounter and not 100% sure if it’s normal.

not likely to be related, sadly

Are there any plans to fix the 3D issues?
And if so, do you need anything from us?

I’ll try to provide you some dumps asap. Speaking of Nativeaudio I noticed 2 minor bugs: 1) if someone talks once other players will always see his mouth moving, 2) voice targets are directional so that makes radios, which already dont have a filter, even more confusing to use.

What I have noticed with NativeAudio is it turns interiors into Echo chambers