Convert DSS to AU — Free Online Tool
Convert DSS dictation recordings to AU format by decoding Olympus/Philips ADPCM-IMA audio into uncompressed 16-bit big-endian PCM. This is ideal for making speech recordings created on digital dictation hardware accessible on Unix/Linux systems and legacy audio pipelines that expect the simple Sun AU container.
to
FFmpeg Command
Copy this command to run the same conversion locally with FFmpeg on your desktop. Download FFmpeg
Drop your DSS file here
or click to browse
Free — no uploads, no signups. Your files never leave your browser.
Settings
Note: Browser-based encoding uses approximate quality targets. For precise CRF compression, copy the FFmpeg command above and run it on your desktop.
Estimated output:
Conversion Complete!
DownloadHow It Works
DSS files store audio using the adpcm_ima_oki codec — a proprietary variant of IMA ADPCM compression developed specifically for low-bitrate speech recording on Olympus and Philips digital dictation devices. During conversion, FFmpeg fully decodes this compressed audio stream into raw PCM samples, then re-encodes them as 16-bit signed big-endian PCM (pcm_s16be) wrapped in the Sun AU container. This is a full decode-and-re-encode process, not a remux, because the AU format does not support any ADPCM codec — it only stores uncompressed or G.711-style PCM variants. The result is a significantly larger file with no lossy re-compression, since PCM is lossless storage of the decoded audio.
What Each Flag Does
| Flag | What it does |
|---|---|
ffmpeg
|
Invokes the FFmpeg tool. In the browser-based version of this tool, FFmpeg runs as a WebAssembly binary (ffmpeg.wasm) entirely within your browser — no files leave your machine. |
-i input.dss
|
Specifies the input DSS file. FFmpeg identifies the DSS container and locates the adpcm_ima_oki compressed audio stream inside it, which is the proprietary codec used by Olympus and Philips digital dictation recorders. |
-c:a pcm_s16be
|
Decodes the adpcm_ima_oki audio from the DSS file and re-encodes it as 16-bit signed big-endian PCM — the standard uncompressed audio encoding for the Sun AU container format, required by its big-endian Unix origins. |
output.au
|
Specifies the output file with the .au extension, which tells FFmpeg to wrap the pcm_s16be audio stream in the Sun AU container format — a simple binary format with a minimal 24-byte header followed by raw PCM data. |
Common Use Cases
- Archiving digital dictation recordings from Olympus or Philips recorders into a format readable by Unix/Linux audio tools like SoX or Audacity without proprietary plugin dependencies.
- Feeding transcribed speech recordings into legacy legal or medical transcription pipelines running on Unix workstations that accept AU as their standard audio input format.
- Preserving DSS recordings in a non-proprietary format — since DSS is a closed standard, converting to AU with uncompressed PCM ensures the audio remains accessible regardless of future software availability.
- Pre-processing dictation audio for speech recognition or NLP pipelines on Unix-based servers that expect simple headerless-adjacent PCM formats like AU.
- Preparing dictation recordings for playback or analysis in classic Sun Solaris or early Java-based multimedia applications that natively support the AU format.
- Converting batches of DSS interview or meeting recordings to AU for ingestion into audio forensics or voice analysis software running on Unix infrastructure.
Frequently Asked Questions
The DSS format is already lossy — audio was compressed using the adpcm_ima_oki codec when it was recorded on the dictation device. Converting to AU with pcm_s16be stores the decoded output as uncompressed PCM, which introduces no additional lossy compression. However, the original quality lost during DSS recording cannot be recovered, so the AU file will reflect the same speech quality as the source DSS file.
DSS uses the adpcm_ima_oki codec, which compresses speech audio at very low bitrates (typically 13–16 kbps) specifically to fit long recordings on small storage devices. The AU output uses uncompressed 16-bit PCM at the full decoded sample rate, which typically produces files 6–10 times larger. This is normal and expected — the AU format is not designed for compression.
Yes. The AU container supports several PCM variants. You can substitute '-c:a pcm_mulaw' for 8-bit mu-law encoding (used in telephony, smaller file size but slightly more lossy), '-c:a pcm_alaw' for A-law encoding, or '-c:a pcm_s8' for 8-bit signed PCM. For example: ffmpeg -i input.dss -c:a pcm_mulaw output.au. For dictation speech, mu-law or A-law are often good enough and roughly halve the file size compared to pcm_s16be.
No. The Sun AU format has a very minimal header structure that supports only a basic text annotation field, and DSS metadata fields (recording date, author name, work type codes used by Olympus and Philips devices) have no standardized mapping to AU. FFmpeg will not transfer this metadata during conversion. If preserving metadata matters, consider copying the DSS metadata to a sidecar file or choosing a richer output format like FLAC or WAV instead.
On Linux or macOS, you can loop over all DSS files in a directory with: for f in *.dss; do ffmpeg -i "$f" -c:a pcm_s16be "${f%.dss}.au"; done. On Windows Command Prompt, use: for %f in (*.dss) do ffmpeg -i "%f" -c:a pcm_s16be "%~nf.au". This applies the same decode and re-encode process to each file, producing one AU output per DSS input.
The Sun AU format was designed by Sun Microsystems for SPARC-based Unix workstations, which use big-endian byte order. The format specification defines its PCM encoding as big-endian by convention, so pcm_s16be (16-bit signed big-endian) is the correct and most compatible default. Using little-endian PCM in an AU container would produce a file that sounds like noise on most AU-compatible players.
Technical Notes
The adpcm_ima_oki codec inside DSS files is a proprietary variant of IMA ADPCM tuned for narrow-band speech, typically recorded at 8 kHz or 11.025 kHz mono. FFmpeg's decoder handles the most common DSS variants (DSS v1 and v2) but may struggle with DSS Pro or newer Olympus DS-series proprietary extensions. The resulting AU file will inherit the source sample rate — most commonly 8000 Hz for standard dictation — so the AU output will sound telephone-quality, which is accurate to the source. The Sun AU container has a 24-byte minimum header and no support for multi-channel audio or sample rates above 65535 Hz, neither of which is relevant for DSS dictation content. Because AU files are headerless beyond that minimal descriptor and store raw PCM, they are well-suited for streaming and piping in Unix shell pipelines. One practical limitation: some older AU players cap file size at 2GB (32-bit offset field), but this is unlikely to matter for dictation recordings even after expanding to uncompressed PCM.