[Released] DFVoice - VoIP Plugin

So I'm thinking, for now I've only got the means to compile Windows DLLs for Opus. So for now I'll likely just use those to get it working, then release the update with just those DLLs at first, and then:

1.) If you want to get it working on any platform of your choice, you can download Opus (it's free and open source), compile the source code for that platform, and drop in the library file and it should theoretically just work.
2.) Later I may drop in library files for other platforms as I go (if you figure out how to compile Opus for other platforms, you could also pester me over PM and send me the library files - just a suggestion )

 

So here's a test of the Opus codec. I used the song "The Corrupted", from Kingdom Hearts 2.5 HD Remix. The codec's bitrate is set to 64kbps.

The first second or so is silent, I think Opus throws away some amount of the first audio it receives or something. This first bit of silence seems to be a lot smaller with the MicrophoneInputDevice, for instance.

One change that had to be made to support this is that what frequency a given FrequencyMode maps to now depends on the codec - there's now an IFrequencyProvider interface for this. The Speex codec, for instance, maps frequency as follows:

Narrow - 8khz
Wide - 16khz
UltraWide - 32khz

Whereas the new Opus codec does not support 32khz, so it maps frequency as follows:

Narrow - 8khz
Wide - 16khz
UltraWide - 48khz

 

(So happy to see this moving along! Yay!)

 

Yep, pretty much finishing things up now. What I'll do is include the Opus stuff in a nested package. So if you want to use the experimental Opus codec, you'd import that package and then reboot Unity (the Opus wrapper, by necessity, uses unsafe methods, so the package will include smcs.rsp and gmcs.rsp files to enable unsafe code).
Also updating the included readme for two reasons:
1.) Clarify that the splash screen or other attribution isn't a requirement anymore.
2.) Explain how to import and use the experimental Opus codec.

Should be able to submit an update today

EDIT: Update submitted

 

Question about "threshhold" (when not using push-to-talk) is this value 0 to 1, or is it 0 to max float?
Also, is OPUS still Windows only? This project here:
https://github.com/martindevans/MumbleSharp

Looks like it is using an opus.dll (maybe it's multiplatform?)

-Carmine

 

Threshold is 0 to 1.

And as far as Windows only:

 

I'm having a problem with NSpeex.dll. When I try to build for iOS it gives me an AOT cross compile error. I tried .Net 2.0 and .net 2.0 subset (same issue with both).

(it worked fine for android and all other versions).

 

Has anybody been able to play with the Opus stuff yet? I see it was included in the latest update.

 

Is it possible to get Unity 4.6 compatible version of DFVoice if I purchase current version from Asset Store?

 

How is this on android?

 

I am having an issue setting this up with my project. When I import the asset I get this error.
Assets/DFVoice/Scripts/Codec/SpeexCodec.cs(29,39): error CS0433: The imported type `NSpeex.SpeexEncoder' is defined multiple times

Someone please help. I really need this ready to go before the launch of my beta.

 

Does this error happen if you import into an empty project?
If not, maybe you already had the NSpeex DLL in your project beforehand, so ensure your project only has one NSpeex DLL.

 

That fixed the issue. Thank you for the awesome support.

 

I am having trouble setting this up with UFPS multiplayer asset. I made the script that sends voice data over the network. But only 1 player is able to talk while others cannot.

 

If this is not abandon we love to give your product a try and reevaluate it since your last update.

Feel free to PM us.