Audio playback lag and brief interruptions when streaming WAV to call

[PrasannaLakshmiD]

AlloyAIS.wav

Hi team,

I'm using SIPSorcery to play an audio prompt into a live call session. The playback works, but sometimes I hear small breaks or lags in the middle of the audio—it doesn’t play smoothly every time.

Observed behavior:
The audio starts correctly but sometimes lags or briefly cuts out midway.
The WAV file itself is clean (verified by local playback).
The issue occurs intermittently—not on every playback.
I am making multiple simultaneous calls to the same SIP extension and playing different requested audio files concurrently (each call uses its own AudioExtrasSource).

Environment details:
.NET version: Core 9
SIPSorcery version: 8.0.23
Platform: Linux and Windows both
Audio file: 8kHz, 16-bit mono WAV(3CX Supported wav file format)

Here’s a simplified version of the code I’m using:

public async void AttachToTransport(SIPTransport transport)
{
var registrationAgent = new SIPUserAgent(transport, null, true);

registrationAgent.OnIncomingCall += async (ua, req) =>
{
    string callId = req.Header?.CallId ?? "(none)";
    try
    {
        var user = req.Header.To.ToURI.ToString();
        var start = user.IndexOf(':') + 1;
        var end = user.IndexOf('@');
        string bargeinExtNum = user.Substring(start, end - start);
        _log.Information("Call [INVITE] from {From} Call-ID={CallId}", req.Header?.From?.ToString(), callId);
        string audioPath = "Audio/SupportedWavFile.wav";
        var audioSession = new AudioSendOnlyMediaSession { AcceptRtpFromAny = true };
        audioSession.OnAudioFormatsNegotiated += formats =>
        {
            foreach (var f in formats)
                _log.Information("🎚️ [Negotiated Codec] {Codec} {Clock}Hz", f.Codec, (int)f.ClockRate);
        };

        audioSession.AudioExtrasSource.RestrictFormats(fmt =>
            fmt.Codec == AudioCodecsEnum.PCMU || fmt.Codec == AudioCodecsEnum.PCMA);
        // Each incoming call gets its own SIPUserAgent
        var callAgent = new SIPUserAgent(transport, null, true);
        var uas = callAgent.AcceptCall(req);

        bool answered = await callAgent.Answer(uas, audioSession);
        audioSession.AudioExtrasSource.SetSource(AudioSourcesEnum.Silence);
        if (!answered)
        {
            _log.Warning("Failed to answer barge-in call.");
            try { ua.Hangup(); } catch { }
            return;
        }

        // Play prompt or beep
        if (string.IsNullOrWhiteSpace(audioPath) || !File.Exists(audioPath) || new FileInfo(audioPath).Length == 0)
        {
            _log.Warning("No valid audio file({audioPath}) found.", audioPath);
        }            
        else
        {
            try
            {
                byte[] audioBytes = await File.ReadAllBytesAsync(audioPath);
                using var ms = new MemoryStream(audioBytes);
                await audioSession.AudioExtrasSource.SendAudioFromStream(ms, AudioSamplingRatesEnum.Rate8KHz);
            }
            catch (Exception ex)
            {
                _log.Error(ex, "Error sending audio from file {AudioPath}: {Error}", audioPath, ex.Message);
            }
        }
        audioSession.AudioExtrasSource.SetSource(AudioSourcesEnum.Silence);
        // Add small delay before hangup to flush RTP
        await Task.Delay(300);
        callAgent.Hangup();
    }
    catch (Exception ex)
    {
        _log.Error(ex, "Error handling Call-ID={CallId} exception is : {Error}", callId, ex.Message );
        try { ua?.Hangup(); } catch { }
    }
    finally
    {
        if (!string.IsNullOrEmpty(hCallID))
            BargeHistoryDictionaryCache.Remove(hCallID);
    }
};

}

Can you please check if there’s a known issue with audio streaming consistency or any recommended pattern (like using timers, fixed frame chunking, or buffered streaming) to avoid these small interruptions?

Thanks!

[PrasannaLakshmiD]

Hello team,
Can anyone respond to this? I am able to play the audio into the call. But in the Linux environment, while playing audio, small gaps are coming. The audio is continuing after the gap.

[PrasannaLakshmiD]

Hello Team,

I’m facing an issue with audio playback. When the extension is registered on a Windows softphone, the audio prompt plays smoothly with no issues. However, when the same extension is registered on a Linux softphone, I experience noticeable audio gaps and breaks during playback.

Could you please help me understand why this happens or how to resolve it?

[sipsorcery]

It may be due to the timer mechanism being used in this library to trigger the audio sampling. The AudioExtrasSource class uses a standard timer. It was only ever intended as a demo to provide a handy way to get some audio on test calls. There are more reliable ways available that could be used instead of the Timer approach. If you point CoPilot or ChatGPT at that class it will probably come up with some good suggestions.

[PrasannaLakshmiD]

SendAudio.txt
Hi, I tried in different ways. With this attached code, I am able to play the audio, but sometimes tiny gaps are still coming. Once, can you help me with this?