Convert DSS to VOC — Free Online Tool
Convert DSS dictation recordings to VOC audio files directly in your browser, decoding Olympus/Philips ADPCM-compressed speech into uncompressed PCM audio compatible with vintage Sound Blaster applications. The conversion decodes the proprietary adpcm_ima_oki codec from your digital dictation device into 8-bit unsigned PCM (pcm_u8), the native audio format of Creative Labs VOC files.
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 use the adpcm_ima_oki codec — a form of Adaptive Delta Pulse Code Modulation developed specifically for digital dictation hardware by Olympus, Philips, and Grundig. During conversion, FFmpeg decodes this compressed ADPCM stream into raw PCM audio samples, then repackages them as 8-bit unsigned PCM (pcm_u8) inside a VOC container. Because DSS is speech-optimized and low-bitrate, the original audio is already somewhat limited in fidelity (typically narrow-band, 8–16 kHz sample rate), so encoding to 8-bit pcm_u8 in VOC is a natural fit — the VOC format's resolution ceiling is close to what the DSS source material can actually resolve. The output is a lossless representation of the decoded DSS audio in a format DOS-era sound hardware and software can natively read.
What Each Flag Does
| Flag | What it does |
|---|---|
ffmpeg
|
Invokes the FFmpeg multimedia processing tool. In the browser, this runs via FFmpeg.wasm (WebAssembly), executing the identical logic as the desktop binary but entirely within your browser session. |
-i input.dss
|
Specifies the input DSS file — a Digital Speech Standard recording containing audio encoded with the adpcm_ima_oki codec, typically produced by Olympus, Philips, or Grundig dictation hardware. |
-c:a pcm_u8
|
Sets the output audio codec to 8-bit unsigned PCM (pcm_u8), which is the native and most compatible audio encoding for the VOC container format. This decodes the adpcm_ima_oki ADPCM stream from the DSS file into raw, uncompressed audio samples suitable for Sound Blaster playback. |
output.voc
|
Defines the output file as a VOC file — Creative Labs' Sound Blaster audio container. FFmpeg infers the VOC format from the .voc extension and writes the uncompressed pcm_u8 audio into the VOC chunked structure. |
Common Use Cases
- Archiving digital dictation recordings from Olympus or Philips handheld recorders into a legacy-compatible format for DOS-based transcription software that expects VOC audio input.
- Integrating speech recordings from a DSS dictation device into a retro DOS game or multimedia CD-ROM project that uses VOC files for sampled voice clips.
- Preparing DSS voice memos for playback on vintage Sound Blaster hardware or emulators like DOSBox, which natively support the VOC format.
- Converting dictation audio captured in the field into VOC files for use with classic audio editing tools such as early Sound Blaster utilities that cannot open proprietary DSS files.
- Extracting and converting DSS recordings as raw PCM audio snapshots for forensic or archival analysis, where the VOC container's uncompressed format guarantees no further lossy encoding artifacts.
- Sampling speech segments from DSS recordings to use as sound effects or voice samples in retro-style game development projects targeting DOS compatibility.
Frequently Asked Questions
No — the conversion decodes the DSS file's adpcm_ima_oki compression into uncompressed PCM, but the VOC output cannot recover fidelity that was discarded when the original recording was made. DSS is a lossy, speech-optimized format with a narrow frequency range (typically below 8 kHz), so the VOC file will be an accurate but equally limited representation of that source audio. Think of it as unwrapping the compressed data into a lossless container, not upgrading the recording quality.
The VOC format was originally designed for Creative Labs Sound Blaster cards and supports both 8-bit unsigned (pcm_u8) and 16-bit signed (pcm_s16le) PCM. The default here is pcm_u8 because it matches the most common and broadly compatible VOC profile used in DOS-era software. Since DSS source audio is already narrow-band and low-fidelity, pcm_u8 captures essentially the same audible content with a smaller file size. If you need higher dynamic range, you can modify the command to use -c:a pcm_s16le instead.
Yes. On Linux or macOS you can loop over files with a shell command: for f in *.dss; do ffmpeg -i "$f" -c:a pcm_u8 "${f%.dss}.voc"; done. On Windows Command Prompt, use: for %f in (*.dss) do ffmpeg -i "%f" -c:a pcm_u8 "%~nf.voc". The browser tool processes one file at a time, so the FFmpeg command is especially useful when you have a folder full of dictation recordings to convert.
No. The VOC format has essentially no metadata support — it was designed purely as a raw audio container for Sound Blaster hardware and stores no tags for artist, date, or description. Any proprietary metadata embedded in the DSS file (such as the recorder model, recording timestamp, or author ID used by Olympus/Philips dictation systems) will be lost in the conversion. If preserving that metadata matters, consider exporting it separately before converting.
The output VOC file inherits the sample rate of the decoded DSS audio, which is typically 8000 Hz or 11025 Hz for standard DSS recordings, and up to 16000 Hz for DSS Pro files. FFmpeg does not resample the audio unless you explicitly add a -ar flag to the command (e.g., -ar 22050). Keep in mind that DOSBox and some legacy VOC players have limits on supported sample rates, so you may need to resample if targeting a specific vintage playback environment.
No — the conversion logic is identical. The browser tool uses FFmpeg.wasm, which is a WebAssembly port of the same FFmpeg binary you would run on your desktop. The exact command (ffmpeg -i input.dss -c:a pcm_u8 output.voc) is executed in both cases. The main practical difference is that the browser tool has a 1GB file size limit, while running FFmpeg locally has no such restriction — making the local command preferable for large batches of long dictation recordings.
Technical Notes
DSS (Digital Speech Standard) files encode audio with the adpcm_ima_oki codec, a variant of IMA ADPCM tuned for the low-bitrate, narrow-bandwidth characteristics of dictation hardware. Typical DSS files record at 8 kHz (standard quality) or up to 16 kHz (DSS Pro), with a mono channel, resulting in very small file sizes per minute of speech. When FFmpeg decodes this to pcm_u8 for the VOC container, the output is uncompressed 8-bit unsigned PCM, meaning file sizes will increase substantially — roughly 8x or more depending on the original compression ratio. The VOC format uses a chunked structure with a fixed header signature ('Creative Voice File') and supports basic looping metadata in some implementations, but FFmpeg writes a straightforward linear audio VOC with no loop or marker chunks. One known limitation is that some DSS files produced by certain Olympus recorder firmware versions use non-standard header variants that older FFmpeg builds may not parse correctly; if you encounter a decoding error, ensure you are using a recent FFmpeg version. The pcm_u8 codec has a theoretical dynamic range of about 48 dB, which is adequate for the speech content in DSS recordings but would be limiting for music.