一款面向 Windows 的本地音频分析工具，通过频谱截止特征与启发式码率映射，辅助识别疑似由有损音频转制而来的“伪无损”文件。
A local Windows audio analysis tool that uses spectral cutoff characteristics and heuristic bitrate mapping to help identify lossless files that may have been transcoded from lossy sources.
Important
本工具提供的是启发式分析结果，不是实验室或司法级鉴定。老录音、磁带、母带低通、自然高频衰减及特殊制作手法都可能影响结果。
This tool provides heuristic results, not laboratory- or forensic-grade proof. Old recordings, tape sources, mastering filters, natural high-frequency roll-off, and unusual production choices may affect the result.
- 支持 MP3、FLAC、WAV、AIFF、AAC、M4A、WMA 和 OGG。
- 支持拖放、批量添加和大规模文件队列。
- 使用 4096 点 FFT、多片段平均频谱和自适应截止频率检测。
- 使用顺序解码采样，避免压缩音频随机定位造成结果不一致。
- 提供“通过”“无法确定”“疑似伪无损”“有损格式”等结果。
- 可调节鉴定灵敏度；最宽松端对 192 kbps 档更宽容。
- 可疑文件可以批量添加
可疑_前缀，或直接清理。 - 暗黑主题与清新主题，可自由切换。
- 所有分析均在本地完成，音频不会上传网络。
- Release 提供无需安装 .NET 的 Windows x64 单文件便携版。
- 前往 Releases 下载
无损音频真伪鉴定器.exe。 - 双击运行，无需安装。
- 拖入音频文件或点击“添加文件”。
- 根据需要调整鉴定灵敏度，然后点击“开始鉴定”。
- 结合“有效截止频率”“推测来源码率”和说明综合判断。
“一键清理”会直接删除可疑列表中的源文件，请谨慎操作。
程序从音频主体范围内顺序取得最多 100 个频谱片段，对每个声道执行 4096 点 FFT，并对结果进行平均和平滑。随后根据高频能量下降、局部与宽频落差、采样率和容器编码信息估计有效截止频率及可能的有损来源码率。
灵敏度只影响启发式阈值，不会把分析变成确定性证明。最宽松端的大致分级为：
| 推测来源码率 | 结果 |
|---|---|
| ≤ 128 kbps | 疑似伪无损 |
| 160 kbps | 无法确定 |
| ≥ 192 kbps | 通过 |
实际结果还会受到截止线是否足够明显、采样率和频谱完整度影响。
要求：
- Windows 10 1809 或更高版本
- .NET 8 SDK
dotnet restore
dotnet build -c Release
dotnet publish src/LiteAudioVerifier/LiteAudioVerifier.csproj `
-c Release -r win-x64 --self-contained true `
-p:PublishSingleFile=true `
-p:IncludeNativeLibrariesForSelfExtract=true算法思路改编自 Angel2mp3/AudioAuditor。本项目采用 Apache License 2.0，第三方依赖遵循各自许可证。
- Supports MP3, FLAC, WAV, AIFF, AAC, M4A, WMA, and OGG.
- Drag-and-drop input, batch import, and large file queues.
- 4096-point FFT, multi-segment spectrum averaging, and adaptive cutoff detection.
- Sequential decoding prevents inconsistent samples caused by repeated seeking in compressed audio.
- Results include Passed, Inconclusive, Suspected Fake Lossless, and Lossy Format.
- Adjustable sensitivity with a more permissive 192 kbps tier at the most lenient setting.
- Batch-prefix suspicious files with
可疑_, or delete them directly. - Switchable Dark and Fresh themes.
- Analysis runs entirely on the local machine; audio is never uploaded.
- Releases include a self-contained, single-file Windows x64 portable executable.
- Open the latest release and download
无损音频真伪鉴定器.exe. - Run the executable; installation is not required.
- Drop audio files into the window or select Add Files.
- Adjust sensitivity if needed, then start the analysis.
- Review the effective cutoff frequency, estimated source bitrate, and explanation together.
One-click cleanup permanently deletes the source files listed as suspicious. Use it carefully.
The analyzer sequentially samples up to 100 spectral windows from the main body of each file. It performs a 4096-point FFT per channel, averages and smooths the spectra, then estimates an effective cutoff frequency and a likely lossy-source bitrate from high-frequency energy loss, local and broad-band drops, sample rate, and codec metadata.
Sensitivity changes heuristic thresholds only; it does not turn the result into definitive proof. At the most lenient setting, the approximate tiers are:
| Estimated source | Result |
|---|---|
| ≤ 128 kbps | Suspected fake lossless |
| 160 kbps | Inconclusive |
| ≥ 192 kbps | Passed |
The final result also depends on cutoff confidence, sample rate, and overall spectral completeness.
Requirements:
- Windows 10 version 1809 or later
- .NET 8 SDK
dotnet restore
dotnet build -c Release
dotnet publish src/LiteAudioVerifier/LiteAudioVerifier.csproj `
-c Release -r win-x64 --self-contained true `
-p:PublishSingleFile=true `
-p:IncludeNativeLibrariesForSelfExtract=trueThe algorithmic approach is adapted from Angel2mp3/AudioAuditor. This project is licensed under the Apache License 2.0. Third-party dependencies remain under their respective licenses.