Skip to content

Sample Audio Files for Transcription Testing

Download Sample File

Free test file, safe content, instant download.

SHA25630785d72dfafc11ff8754c33...

Testing a speech-to-text pipeline needs actual speech — tones and music do not exercise a model the way a human voice does. This page ships two purpose-built public-domain speech files: a 16kHz mono WAV whose exact transcript is known ("speech prosody"), and a genuine 1921 studio recording of Marcus Garvey for noisy, real-world hard-mode testing.

The short WAV is deliberately in the exact format most ASR systems natively consume: 16 kHz sample rate, 16-bit PCM, single channel — the same input Whisper uses internally. Because the ground-truth transcript is documented, you can compute exact word error rate (WER) on a known reference instead of eyeballing output.

Both files are public domain (CC0 / pre-1923 US recording), so you can ship them in test fixtures, CI pipelines, and demos without attribution obligations.

File Information

File 1
sample-speech-short.wav — 1.46s, 16kHz mono PCM, transcript: "speech prosody"
File 2
sample-speech-archive.ogg — Marcus Garvey 1921 speech, real-world noise
License
Public domain / CC0 — use freely in fixtures
Best for
Whisper, Deepgram, AssemblyAI, Google STT, WER evaluation

How to Use This Audio (Speech) Test File

  1. 1Download sample-speech-short.wav (the known-transcript reference).
  2. 2Run it through your ASR pipeline or Whisper CLI: whisper sample-speech-short.wav --model tiny.
  3. 3Compare output against the known transcript to measure accuracy.
  4. 4Re-run with sample-speech-archive.ogg to test noise robustness.

Frequently Asked Questions

What makes a good audio file for transcription testing?
Real human speech (not tones or music), a documented transcript for WER measurement, 16 kHz or higher sample rate, and ideally both a clean and a noisy variant. Our two files cover the clean and noisy cases; both are public domain so you can embed them in test suites.
Is 1.5 seconds of audio enough for STT testing?
For unit tests and CI smoke tests, yes — a short utterance with a known transcript verifies the whole pipeline (upload → decode → transcribe → output) in milliseconds. For accuracy benchmarking across models, use the longer archive recording or concatenate multiple clips.
What is the transcript of the short WAV file?
A male voice says "speech prosody" (a linguistics recording released as CC0). That documented ground truth is what makes WER computation trivial: one reference, no ambiguity.
Can I use these files with the Whisper API?
Yes. The WAV is 16kHz mono PCM — Whisper’s native format. Via the OpenAI API: POST it as a file to /v1/audio/transcriptions with model whisper-1. Via the local CLI: whisper sample-speech-short.wav --model base. Both files also work with Deepgram, AssemblyAI, and Google Speech-to-Text.
Are these speech files free to use in commercial test suites?
Yes. The short WAV is CC0 and the 1921 recording predates US sound-recording copyright — both are public domain. No attribution required, though crediting Wikimedia Commons (where they are hosted) is a kind gesture.

Related Sample Files

Browse Related Formats

Need a different size or format?

Use our free Generator to create custom test files in any format and size — right in your browser, no upload needed.

Try Our Generator