WMA is the format you'll find on professional dictaphones bought in the 2000s, in archival folders of meeting recordings from the Windows-XP era, and in any institution that standardised on Windows Media tooling before MP3 became the universal default. Most modern transcription tools quietly refuse WMA and demand a transcode to MP3 first. MDisBetter accepts WMA directly.
WMA in 2026 — who still uses it
WMA (Windows Media Audio) was Microsoft's answer to MP3 in the late 1990s and 2000s. It saw heavy adoption in three places: professional dictaphones (Olympus, Philips, Sony) that captured directly to WMA for desktop transfer to a Windows machine; institutional archives (legal firms, healthcare practices, insurance investigators) that built workflows around Windows Media Player and never migrated; and pre-2010 Windows Voice Recorder recordings made by anyone using the built-in Windows app. None of those archives have disappeared — they've just become harder to process because most modern audio software has dropped WMA support.
Why direct WMA support matters for archival projects
The standard advice — "convert to MP3 first using ffmpeg" — works but adds friction at scale. For an archival project digitising a thousand WMA dictaphone files, the transcode step alone adds hours of compute time and (since WMA is already lossy) a small but non-zero accuracy loss vs decoding the WMA directly into the transcription engine. MDisBetter skips that step.
What you get back
The same structured Markdown output as for any other format: a top heading, sub-sections at topic shifts, properly punctuated paragraphs, and timestamps ([HH:MM:SS]) at section boundaries. We do not identify speakers, so the transcript reads as continuous prose. For dictaphone recordings (typically single-speaker, professional dictation) that is exactly the right shape: a clean paragraph-structured document.
Before / After
Before (PDF):
[WMA audio file]
Windows Media Audio v9, 3.4 MB, 28 minutes mono
Metadata: WM/Author, WM/Composer (legacy dictaphone defaults)
(no extractable text — WMA-encoded waveform)
After (Markdown):
# Dictation — Case File 2087
## Background
Client interview conducted on May 9, 2026 at the firm's main office. The client, Mr. Henderson, presented for an initial consultation regarding a contract dispute with his former employer.
## Facts as Presented
Mr. Henderson states he was employed as a senior engineer at Acme Corp from January 2019 through March 2026. The dispute centres on the calculation of his accrued bonus for fiscal year 2025…
Frequently asked questions
Is WMA still actually used in 2026?
Less than it was, but yes — significantly in legacy institutional archives (legal, medical, insurance), in professional dictaphone workflows that haven't migrated, and in any archive of recordings made on Windows machines pre-2010. For modern recordings on modern devices, MP3 or M4A has displaced it; for archives you can't reach for, WMA support remains useful.
Are dictaphone recordings handled correctly?
Yes — professional dictaphone WMA files (typically mono, 22 kHz or 44.1 kHz, single speaker, often with custom field markers) transcribe cleanly. The output is structured paragraphs with no speaker attribution, which suits single-voice dictation exactly. Custom dictaphone field markers (e.g., the Olympus DSS-style "patient", "case number" prompts) appear as inline text in the output.
Does this work for both WMA Standard and WMA Lossless?
Both. WMA Standard (the common lossy format) and WMA Lossless are decoded internally before transcription. The lossless variant gives marginally better accuracy on noisy or low-volume source audio, identical accuracy on clean recordings.
Can I batch-process a folder of legacy WMA files?
The MDisBetter web tool processes one file at a time — for a 1000-file archival project, you'd want a self-hosted batch pipeline. The OSS path: ffmpeg-decode each WMA to PCM in a loop, feed each to faster-whisper running locally, save each transcript as a .md file mirroring the input filename. About 50 lines of Python; zero per-file interaction once it's running. The web tool is for one-off files where setting up the pipeline isn't worth it.
What about WMA files with DRM (Windows Media DRM)?
DRM-protected WMA files (sold by some music stores in the 2000s, used by some institutional document-management systems) can't be decoded outside the Windows Media DRM ecosystem. Most archival material is DRM-free; if your WMA refuses to play in standalone players (VLC, etc.), it's likely DRM-locked and not transcribable by any tool that doesn't bypass the protection.