Audio! both AudioSink and AudioSource dont deliver any data / bytes.

[UweR70]

Hello forum,

Abstract:
My app registration works fine,
incoming calls are fetched, the body says HasAudio = true,
Caller id (aka his tel.number) is received, etc.

Assumption: useragent.AcceptCall(...) and useragent.Answer(...) seems to work fine;
The assumption base on the fact that the ringing stops;
assumption correct?

Issue:
Did not get any audio. "AudioSink" nor "AudioSource" are not delivering any data (bytes).
The SipExamples "RecordCall" and all the other examples are well known.
I adjusted "RecordCall" to my needs.
Aka: The adjusted "RecordCall" app registers successful with my Fritzbox; incoming call is fetched, but
private static void OnRtpPacketReceived(...)
is never been called!

My setup

AVM Fritzbox 7490.
Two devices are in the Fritzbox defined and configured:
"Softphone_52" and "Softphone_53"
"52" and "53" because two of my three land line tel.numbers are ending with 52 and 53.
So the "device"/ "Softphone_52" is mapped to the land line tel.number ending with "52";
the other "device"/ "Softphone_53" is mapped to the land line tel.number ending with "53".
Their settings: both LAN/WLAN, both with unique Username and Passwords.

To avoid any issue/minimize potential issues I am using "Softphone_53" only in my test app.
Because there is no other (real or software) device handling the land line tel.number ending with 53.

Using:
Windows 11, Visual Studio Prof. V17.13.2,
C# project setup
(
a) Target framework: .NET 8.0
b) Target OS: Windows
c) Target OS version: 10.0.17763.0
d) Supported OS version: 10.0.17763.0
), WPF.
Latest SIPSocery nuget packages SIPSorcery (8.0.11) and SIPSorceryMedia.Windows (8.0.7).
NAudio (2.2.1)

Used landline telefon and mobile phone to make the call.
As said, the incomming call is receive and can behandled, but always no audio.

Any idea?
is there a working expample that
fetches incoming calls and stores the incoming (AudioSink) and(!) the outgoing (AudioSource); keyword NAudio mixer?

Thanks in advance!

[UweR70]

Errata:
It was not working with SIPSorcery version 8.0.10 (ten).
Updated to new version 8.0.11 (eleven).

Now my extended RecordCall test app with regard to authentication is capable of:
a) record anything the caller say.
b) the caller hears a wave file when I fire
await using var waveFileStream = new FileStream(pathAndFilename, FileMode.Open, FileAccess.Read);
await voipMediaSession.AudioExtrasSource.SendAudioFromStream(waveFileStream, AudioSamplingRatesEnum.Rate16KHz);

BUT now I am totally confused.
What does
await winAudioEp.PauseAudio();
and
await winAudioEp.ResumeAudio();
do?

In case we are talking about a incoming call:
Are these commands turning off/on caller microfons?
So the app didnt get an audio data (= AudioSink = incoming stream) from the caller?
OR
Are they used to pause/resume the AudioSource (= outgoing / datastream) to the caller?

I beleave the last is true.
So using await winAudioEp.PauseAudio();
sets _isAudioSourcePause to true.
So we are talking about AudioSource, so it is the outgoing audio stream, means from my app to the caller.
BUT
It seems to have no effect!
The caller hears a played/streamed wav file and also a SineWave, Noise etc.
regardles whether _isAudioSource is equal to true or false.

So, PLEASE give us all more details and working examples WITH comments why this or that command is used and what the command does!
PLEASE!

[sipsorcery]

I know us maintainers need to have thick skins but comments like the below are very frustrating. Lots of effort has gone into the demos and to a lesser extent docs. BUT this is NOT a commercial library. If you don't want to invest your time you can go and find an alternative that you'll most likely have to pay for. Ideally you'd even contribute something back so other devs facing your challenge would have an easier time of it.

So, PLEASE give us all more details and working examples WITH comments why this or that command is used and what the command does! PLEASE!

Against my better judgement I'll put it down to a language thing and will try and help. Normally devs that come with the attitude above never end up happy and just want to whinge.

Are these commands turning off/on caller microfons?

If you're using the WindowsAudioEndPoint the answer is yes, the microphone is being turned on and off.

The caller hears a played/streamed wav file and also a SineWave, Noise etc

For the helper AudioExtrasSource class the pause and resume simply sets a flag so you can use it on your app. Something that would take 30s to verify by looking at the source. If you want your app to stop sending RTP packets which MAY or MAY NOT be what the app wants to do you have a few options.

• A typical reason to pause one audio source is when a call gets put on hold. In that case you typically want to stop the microphone source and swtich to music on hold or something. It's diconcerting for the other end of a VoIP call to get dead air.

• If you really do want to completely stop sending RTP packets you can go so far as to unhook the audio encoder from the RTP session:

Media.AudioSource.OnAudioSourceEncodedSample -= SendAudio;

[UweR70]

Sorry, but I must have hit a nerve. Nothing could be further from my mind than (just) moaning and complaining. I've simply been looking for a solution for days to save the incoming and outgoing audio channel in an audio file (wav or mp3) if possible.
So everything that comes from the caller side (= audio and dmtf tones) and really everything that my application transmits to the caller, i.e. an (audio) announcement and if the caller transmits e.g. the DMTF tone for key 5, what the called person speaks / delivers as audio.
I think that the OnAudioSourceEncodedSample event is what I am looking for.
It would just be nice if you could read which audio stream this event takes care of. AudioSource, i.e. outgoing, or AudioSink (i.e. incoming) or both or everything except when an app streams audio to the caller or only when noise is sent or or or.
This is not whining, but a request for reasonably up-to-date documentation.

[sipsorcery]

The RecordCall demo shows how to save the incoming audio to a file.

If you're using the WindowsAudioEndPoint (which is for microphone & speakrs on Windows) there is no way to get the raw audio samples before they are sent. The OnAudioSourceRawSample event could be wired up easily enough in LocalAudioSampleAvailable if you're so inclined.

After that you'll still have challenges:

• If you want a single file with incoming and outgoing streams you'll need to use an audio mixer, NAudio has one.
• DTMF tones often come as RTP events which aren't part of the audio stream.

[UweR70]

As said:
I adjuseted "RecordCall" already and works since I use latest SIPSorcery version 8.0.11 in such a manner that my app (aka adjusted "RecordCall") is capable to fetch the complete AudioSink / incoming audio.
After fetching the call, an announcement (wav file) is played and than waited for the DMTF tone 5.
Instead of using a deprecated/obsolete marked event I follow now the approach to copy the announcement (wave file) and merge both files in the hungup event via an NAudio mixer. (both = the wav file containing the incoming audio (AudioSink) and the file containing the announcement (wav) file.

Another approach would be to implement all members of IAudioSource and IAudioSource inclucding a NAudio mixer... but may be later / in the future.

In case the dmtf tone 5 is pressed I am going to play a "Key5KeyWasPressed" wave or when not a "Key5WasNotPressed" wave file and follow the above approach: Copy one or the other, and mix the finally all three wave file into on "CompletINcomingAndOutgoingAudio" wave file (on hungup event).

One issue with this approach is the fact that the caller can be hang up during the (a) announcement, (b) the "Key5WasPressed" or (c) the "Key5WasNotPressesd" wave file is played but not finished.
While the finally resulting (mixed) wave file would containing the complete wav file (a), (b) or (c).
So for sure. For version 0.0.1 is the approach ok but will be changed later!

But may be you have a better Idea how to get the complete outgoing audio (AudioSource) on byte[] level ...