Python向Unity传输WAV音频文件出现高音调噪音问题求助

你的核心问题是音频格式解析不匹配：

• 测试用的WAV文件是16位PCM格式（2字节/采样），但你在Unity中错误地将接收的字节按32位浮点（4字节/采样）解析，导致采样数计算错误（实际采样数是接收字节数/2，但你按/4计算），播放速度直接翻倍，同时格式不匹配产生大量噪音。
• 硬编码跳过44字节WAV头不可靠，部分WAV文件包含额外元数据块，头长度可能超过44字节。

步骤1：修正Python发送端 #

先解析WAV头获取正确的音频参数，再发送参数+音频数据，避免硬编码假设格式：

import socket
import struct

HOST = '127.0.0.1'
PORT = 12345

def read_wav_info(file_path):
    with open(file_path, 'rb') as f:
        # 读取WAV头
        f.read(4)  # RIFF标识
        f.read(4)  # 文件总大小
        f.read(4)  # WAVE标识
        f.read(4)  # fmt块标识
        fmt_size = struct.unpack('<I', f.read(4))[0]
        audio_format = struct.unpack('<H', f.read(2))[0]
        channels = struct.unpack('<H', f.read(2))[0]
        sample_rate = struct.unpack('<I', f.read(4))[0]
        f.read(4)  # 字节率
        f.read(2)  # 块对齐值
        bits_per_sample = struct.unpack('<H', f.read(2))[0]
        
        # 跳过额外的fmt块数据（如果有的话）
        if fmt_size > 16:
            f.read(fmt_size - 16)
        
        # 定位到data块
        while True:
            chunk_id = f.read(4)
            if chunk_id == b'data':
                break
            chunk_size = struct.unpack('<I', f.read(4))[0]
            f.read(chunk_size)
        
        data_size = struct.unpack('<I', f.read(4))[0]
        audio_data = f.read(data_size)
        return sample_rate, channels, bits_per_sample, audio_data

# 读取音频信息和原始数据
sample_rate, channels, bits_per_sample, audio_data = read_wav_info('BabyElephantWalk60.wav')

# 创建Socket并发送数据
server_socket = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
server_socket.bind((HOST, PORT))
server_socket.listen()
print(f"Server listening on {HOST}:{PORT}")

client_socket, addr = server_socket.accept()
print(f"Connected to {addr}")

# 先发送音频参数（采样率、声道数、位深度、数据长度）
client_socket.send(struct.pack('<III', sample_rate, channels, bits_per_sample))
client_socket.send(struct.pack('<I', len(audio_data)))
# 发送音频原始数据
client_socket.sendall(audio_data)

print(f"Sent: {sample_rate}Hz, {channels} channels, {bits_per_sample}bit, {len(audio_data)} bytes")

client_socket.close()
server_socket.close()

步骤2：修正Unity接收端 #

根据接收的参数，正确解析16位PCM数据为AudioClip所需的float采样：

using System;
using System.Net.Sockets;
using System.IO;
using UnityEngine;

public class RealTimeAudioReceiver : MonoBehaviour
{
    public AudioSource audioSource;

    private TcpClient client;
    private NetworkStream stream;
    private BinaryReader reader;

    private byte[] receivedAudioData;
    private int sampleRate;
    private int channels;
    private int bitsPerSample;
    private bool play = false;

    private void Start()
    {
        Debug.Log("Time.timeScale: " + Time.timeScale);
        ConnectToServer();
    }

    private void Update()
    {
        if (play)
        {
            audioSource.Play();
            play = false;
        }
    }

    private void ConnectToServer()
    {
        try
        {
            client = new TcpClient("127.0.0.1", 12345);
            stream = client.GetStream();
            reader = new BinaryReader(stream);
            StartCoroutine(ReceiveAudioData());
        }
        catch (Exception e)
        {
            Debug.LogError("Error connecting to server: " + e.Message);
        }
    }

    private System.Collections.IEnumerator ReceiveAudioData()
    {
        // 先接收音频参数
        sampleRate = reader.ReadInt32();
        channels = reader.ReadInt32();
        bitsPerSample = reader.ReadInt32();
        int dataLength = reader.ReadInt32();
        Debug.Log($"Received params: {sampleRate}Hz, {channels} channels, {bitsPerSample}bit, {dataLength} bytes");

        // 按指定长度接收音频数据
        receivedAudioData = new byte[dataLength];
        int bytesRead = 0;
        while (bytesRead < dataLength)
        {
            int read = reader.Read(receivedAudioData, bytesRead, dataLength - bytesRead);
            if (read == 0) break;
            bytesRead += read;
        }

        CreateAndPlayAudioClip();

        // 清理资源
        reader.Close();
        stream.Close();
        client.Close();

        yield return null;
    }

    private void CreateAndPlayAudioClip()
    {
        if (receivedAudioData == null || receivedAudioData.Length == 0) return;

        int sampleCount = receivedAudioData.Length / (bitsPerSample / 8);
        AudioClip clip = AudioClip.Create("ReceivedAudio", sampleCount, channels, sampleRate, false);
        float[] samples = new float[sampleCount];

        if (bitsPerSample == 16)
        {
            // 16位PCM整数转float（映射到[-1,1]范围）
            for (int i = 0; i < sampleCount; i++)
            {
                short rawSample = BitConverter.ToInt16(receivedAudioData, i * 2);
                samples[i] = rawSample / 32768f;
            }
        }
        else if (bitsPerSample == 32)
        {
            // 兼容32位浮点格式
            for (int i = 0; i < sampleCount; i++)
            {
                samples[i] = BitConverter.ToSingle(receivedAudioData, i * 4);
            }
        }

        clip.SetData(samples, 0);
        audioSource.clip = clip;
        play = true;
    }
}

1. 格式匹配：将16位PCM整数转换为float时，除以32768（16位有符号整数的最大值），将数值映射到Unity AudioClip要求的[-1,1]范围。
2. 动态参数同步：从WAV头解析真实参数并发送给Unity，避免硬编码导致的参数不匹配。
3. 可靠数据接收：先发送数据长度，再按长度接收，解决TCP粘包或数据截断问题。
4. 正确采样数计算：采样数=总字节数/(位深度/8)，16位格式下为字节数/2，修正了之前采样数减半导致的播放加速问题。

内容的提问来源于stack exchange，提问作者Fantuan