meeting-assistant/MeetingAssistant.Tests/AudioMixingTests.cs

using MeetingAssistant.Recording;
using Microsoft.Extensions.Logging.Abstractions;

namespace MeetingAssistant.Tests;

public sealed class AudioMixingTests
{
    [Fact]
    public async Task CompositeAudioSourceCleansMicrophoneBeforeAddingSystemAudio()
    {
        var microphone = new FixedAudioSource(Pcm16(12_000));
        var system = new FixedAudioSource(Pcm16(10_000));
        var source = CreateSource(microphone, system, new FixedEchoCanceller(Pcm16(2_000)));

        var chunks = await ReadChunks(source);

        Assert.Single(chunks);
        Assert.Equal(12_000, BitConverter.ToInt16(chunks[0].Pcm));
    }

    [Fact]
    public async Task CompositeAudioSourceClampsAfterConfiguredGain()
    {
        var microphone = new FixedAudioSource(Pcm16(30_000));
        var system = new FixedAudioSource(Pcm16(10_000));
        var source = CreateSource(microphone, system);

        var chunks = await ReadChunks(source);

        Assert.Equal(short.MaxValue, BitConverter.ToInt16(chunks[0].Pcm));
    }

    [Fact]
    public async Task CompositeAudioSourceUsesRunOptionsForChildSourcesAndGains()
    {
        var microphone = new OptionsCapturingAudioSource(Pcm16(10_000));
        var system = new OptionsCapturingAudioSource(Pcm16(10_000));
        var source = CreateSource(microphone, system);
        var options = new MeetingAssistantOptions
        {
            Recording =
            {
                SampleRate = 24000,
                Channels = 1,
                MicrophoneMixGain = 0.5,
                SystemAudioMixGain = 0.25
            }
        };

        var chunks = await ReadChunks(source, options);

        var chunk = Assert.Single(chunks);
        Assert.Equal(7_500, BitConverter.ToInt16(chunk.Pcm));
        Assert.Same(options, microphone.CapturedOptions);
        Assert.Same(options, system.CapturedOptions);
    }

    [Fact]
    public async Task CompositeAudioSourceWaitsForMatchingStreamsBeforeMixing()
    {
        var microphone = new DelayedAudioSource(TimeSpan.Zero, Pcm16(2_000));
        var system = new DelayedAudioSource(TimeSpan.FromMilliseconds(100), Pcm16(10_000));
        var source = CreateSource(microphone, system);

        var chunks = await ReadChunks(source);

        var chunk = Assert.Single(chunks);
        Assert.Equal(12_000, BitConverter.ToInt16(chunk.Pcm));
    }

    [Fact]
    public async Task CompositeAudioSourceEmitsMicrophoneWithSilentSystemWhenSystemAudioIsQuiet()
    {
        var microphone = new DelayedAudioSource(TimeSpan.Zero, Pcm16(2_000));
        var system = new DelayedAudioSource(TimeSpan.FromMilliseconds(500), Pcm16(10_000));
        var source = CreateSource(microphone, system);

        var chunks = await ReadChunks(source);

        Assert.Collection(
            chunks,
            first =>
            {
                Assert.Equal(2_000, BitConverter.ToInt16(first.Pcm));
            },
            second =>
            {
                Assert.Equal(10_000, BitConverter.ToInt16(second.Pcm));
            });
    }

    [Fact]
    public async Task CompositeAudioSourceFlushesContinuouslyArrivingMicrophoneBeforeLateSystemAudio()
    {
        var microphone = new RepeatedAudioSource(50, TimeSpan.FromMilliseconds(10), Pcm16(2_000));
        var system = new DelayedAudioSource(TimeSpan.FromMilliseconds(750), Pcm16(10_000));
        var source = CreateSource(microphone, system);

        var chunks = await ReadChunks(source);

        Assert.NotEmpty(chunks);
        Assert.Equal(2_000, BitConverter.ToInt16(chunks[0].Pcm));
    }

    [Fact]
    public void AdaptiveEchoCancellerReducesEchoFromMicrophoneSignal()
    {
        var canceller = new AdaptiveFilterAcousticEchoCanceller();
        var system = Enumerable.Repeat(Pcm16(8_000), 80).SelectMany(bytes => bytes).ToArray();
        var microphone = Enumerable.Repeat(Pcm16(8_000), 80).SelectMany(bytes => bytes).ToArray();

        AudioChunk cleaned = new(microphone, 16000, 1);
        for (var i = 0; i < 40; i++)
        {
            cleaned = canceller.CleanMicrophone(
                new AudioChunk(microphone, 16000, 1),
                new AudioChunk(system, 16000, 1));
        }

        var average = Enumerable.Range(0, cleaned.Pcm.Length / sizeof(short))
            .Select(index => (double)Math.Abs(BitConverter.ToInt16(cleaned.Pcm, index * sizeof(short))))
            .Average();
        Assert.True(average < 1_000);
    }

    private static async Task<IReadOnlyList<AudioChunk>> ReadChunks(
        IMeetingAudioSource source,
        MeetingAssistantOptions? options = null)
    {
        var chunks = new List<AudioChunk>();
        await foreach (var chunk in source.CaptureAsync(options ?? new MeetingAssistantOptions(), CancellationToken.None))
        {
            chunks.Add(chunk);
        }

        return chunks;
    }

    private static byte[] Pcm16(short sample)
    {
        return BitConverter.GetBytes(sample);
    }

    private static CompositeMeetingAudioSource CreateSource(
        IMeetingAudioSource microphone,
        IMeetingAudioSource system,
        IAcousticEchoCanceller? echoCanceller = null)
    {
        return new CompositeMeetingAudioSource(
            microphone,
            system,
            new FixedEchoCancellerFactory(echoCanceller ?? new NoopAcousticEchoCanceller()),
            NullLogger<CompositeMeetingAudioSource>.Instance);
    }

    private sealed class FixedAudioSource : IMeetingAudioSource
    {
        private readonly AudioChunk chunk;

        public FixedAudioSource(byte[] pcm)
        {
            chunk = new AudioChunk(pcm, 16000, 1);
        }

        public async IAsyncEnumerable<AudioChunk> CaptureAsync(
            [System.Runtime.CompilerServices.EnumeratorCancellation] CancellationToken cancellationToken)
        {
            await Task.Yield();
            yield return chunk;
        }
    }

    private sealed class DelayedAudioSource : IMeetingAudioSource
    {
        private readonly TimeSpan delay;
        private readonly AudioChunk chunk;

        public DelayedAudioSource(TimeSpan delay, byte[] pcm)
        {
            this.delay = delay;
            chunk = new AudioChunk(pcm, 16000, 1);
        }

        public async IAsyncEnumerable<AudioChunk> CaptureAsync(
            [System.Runtime.CompilerServices.EnumeratorCancellation] CancellationToken cancellationToken)
        {
            if (delay > TimeSpan.Zero)
            {
                await Task.Delay(delay, cancellationToken);
            }

            yield return chunk;
        }
    }

    private sealed class RepeatedAudioSource : IMeetingAudioSource
    {
        private readonly int count;
        private readonly TimeSpan interval;
        private readonly AudioChunk chunk;

        public RepeatedAudioSource(int count, TimeSpan interval, byte[] pcm)
        {
            this.count = count;
            this.interval = interval;
            chunk = new AudioChunk(pcm, 16000, 1);
        }

        public async IAsyncEnumerable<AudioChunk> CaptureAsync(
            [System.Runtime.CompilerServices.EnumeratorCancellation] CancellationToken cancellationToken)
        {
            for (var index = 0; index < count; index++)
            {
                if (interval > TimeSpan.Zero)
                {
                    await Task.Delay(interval, cancellationToken);
                }

                yield return chunk;
            }
        }
    }

    private sealed class OptionsCapturingAudioSource : IMeetingAudioSource
    {
        private readonly AudioChunk chunk;

        public OptionsCapturingAudioSource(byte[] pcm)
        {
            chunk = new AudioChunk(pcm, 16000, 1);
        }

        public MeetingAssistantOptions? CapturedOptions { get; private set; }

        public IAsyncEnumerable<AudioChunk> CaptureAsync(CancellationToken cancellationToken)
        {
            return CaptureAsync(new MeetingAssistantOptions(), cancellationToken);
        }

        public async IAsyncEnumerable<AudioChunk> CaptureAsync(
            MeetingAssistantOptions options,
            [System.Runtime.CompilerServices.EnumeratorCancellation] CancellationToken cancellationToken)
        {
            await Task.Yield();
            CapturedOptions = options;
            yield return chunk;
        }
    }

    private sealed class FixedEchoCanceller : IAcousticEchoCanceller
    {
        private readonly byte[] cleanedPcm;

        public FixedEchoCanceller(byte[] cleanedPcm)
        {
            this.cleanedPcm = cleanedPcm;
        }

        public void Reset()
        {
        }

        public AudioChunk CleanMicrophone(AudioChunk microphone, AudioChunk systemAudio)
        {
            return microphone with { Pcm = cleanedPcm };
        }
    }

    private sealed class FixedEchoCancellerFactory : IAcousticEchoCancellerFactory
    {
        private readonly IAcousticEchoCanceller echoCanceller;

        public FixedEchoCancellerFactory(IAcousticEchoCanceller echoCanceller)
        {
            this.echoCanceller = echoCanceller;
        }

        public IAcousticEchoCanceller Create()
        {
            return echoCanceller;
        }
    }
}