# SPHERE (.sph) to WAV Converter Convert a NIST SPHERE (.sph) speech-corpus file to a standard WAV in your browser. Parses the ASCII header, fixes big-endian byte order, expands mu-law and A-law. ## Run it - **CLI:** `gizza tool sphere-to-wav "TklTVF8xQQogICAgMjU2CnNhbXBsZV9yYXRlIC1pIDgwMDAKY2hhbm5lbF9jb3VudCAtaSAxCnNhbXBsZV9uX2J5dGVzIC1pIDIKc2FtcGxlX2J5dGVfZm9ybWF0IC1zMiAxMApzYW1wbGVfY29kaW5nIC1zMyBwY20Kc2FtcGxlX2NvdW50IC1pIDIwCmVuZF9oZWFkCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAAAD+Ad4ShYLgsuTCkTHv8RQQF48YTjRtht0kHRf9Y83+rtY/0PDRM="` - **Web:** https://gizza.ai/tools/sphere-to-wav/ - **Agents:** machine-readable descriptor (parameters JSON Schema) at https://gizza.ai/tools/sphere-to-wav/tool.json ## Inputs - `input` — SPHERE file bytes (base64 or hex) _(field)_ - `input_format` — Input encoding _(field)_ - `output` — Return _(field)_ - `encoding` — Sample encoding _(field)_ - `channel` — Channels _(field)_ - `container` — Container _(field)_ - `byte_order` — Byte order _(field)_ - `start_sample` — Start sample frame _(field)_ - `max_samples` — Max sample frames (0 = to the end) _(field)_ ## Output - Converted audio (text) ## Query parameters Open the tool pre-filled and auto-run via URL: - `input` — SPHERE file bytes (base64 or hex) - `input_format` — Input encoding - `output` — Return - `encoding` — Sample encoding - `channel` — Channels - `container` — Container - `byte_order` — Byte order - `start_sample` — Start sample frame - `max_samples` — Max sample frames (0 = to the end) Example: `https://gizza.ai/tools/sphere-to-wav/?input=TklTVF8xQQogICAgMjU2CnNhbXBsZV9yYXRlIC1pIDgwMDAKY2hhbm5lbF9jb3VudCAtaSAxCnNhbXBsZV9uX2J5dGVzIC1pIDIKc2FtcGxlX2J5dGVfZm9ybWF0IC1zMiAxMApzYW1wbGVfY29kaW5nIC1zMyBwY20Kc2FtcGxlX2NvdW50IC1pIDIwCmVuZF9oZWFkCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAAAD%2BAd4ShYLgsuTCkTHv8RQQF48YTjRtht0kHRf9Y83%2BrtY%2F0PDRM%3D&input_format=auto&output=data_url&encoding=pcm16&channel=all&container=wav&byte_order=auto&start_sample=0&max_samples=0` --- ## About this tool NIST SPHERE is the container almost every classic speech corpus ships in — TIMIT, Switchboard, Fisher, WSJ, CALLHOME, TEDLIUM. A `.sph` file is a fixed-size block of ASCII header text followed by raw interleaved samples, and nothing on a modern desktop plays it: the extension is often `.wav` even though the bytes are not RIFF, the samples are frequently big-endian, and telephone corpora store 8-bit mu-law or A-law instead of linear PCM. This converter reads the header field by field, applies what it says, and writes a standard RIFF/WAVE file you can open anywhere. Paste the `.sph` bytes as base64 or hex (or a `data:` URI), pick what you want back, and the conversion runs locally in your browser through WebAssembly — the recording never leaves the machine. What it does with the header: - `sample_byte_format` — `10` means big-endian, so every 16-bit sample is byte-swapped into the little-endian order WAV requires. `01` passes through untouched. Force it with **Byte order** when a corpus ships a wrong value and the result sounds like noise. - `sample_coding` — `pcm` is copied, `ulaw` / `mu-law` and `alaw` are expanded to 16-bit PCM by G.711 decoding so ordinary players handle them. - `channel_count` — two-channel conversation recordings can be split (`Channel 1 only` / `Channel 2 only`) or averaged into a mono downmix. - `sample_count`, `sample_n_bytes` — used to locate and bound the sample data; a file whose declared sample count exceeds the bytes actually present is reported as truncated rather than quietly half-converted. ### Worked example The demo file behind the “Convert to WAV” chip has this header, padded to 256 bytes and followed by 40 bytes of samples: ```text NIST_1A 256 sample_rate -i 8000 channel_count -i 1 sample_n_bytes -i 2 sample_byte_format -s2 10 sample_coding -s3 pcm sample_count -i 20 end_head ``` Its first sample pair on disk is `00 00 0f e0` — big-endian. With the defaults (16-bit PCM, all channels, WAV container) the output is an 84-byte file: a 44-byte RIFF header plus the 40 byte-swapped sample bytes, returned as ```text data:audio/wav;base64,UklGRkwAAABXQVZFZm10IBAAAAABAAEAQB8AAIA+AAACABAAZGF0YSgAAAAA… ``` Switch **Container** to `Raw, headerless samples` and **Return** to `Hex audio bytes` and you get the samples on their own, now little-endian: ```text 0000e00fe11d58280b2e4c2e1329ff1e4111780184f146e36dd841d27fd13cd6eadf63ed0ffd130d ``` Switch **Return** to `Header report` and you get the parsed field table instead of audio: every header line with its type token, then the derived sample rate, channel count, coding, byte order, frame count and duration, plus what the conversion would produce (container, encoding, size). For `Raw, headerless samples` the report ends with a ready-to-run re-import command such as `ffmpeg -f s16le -ar 8000 -ac 1 -i out.raw out.wav`, because headerless audio carries none of those parameters with it. ### Limits and edge cases - Decoded input is capped at 6 MiB and produced audio at 12 MiB; the hex rendering is capped at 4 MiB of audio because it doubles again as text. Use **Start sample frame** / **Max sample frames** to excerpt a long recording — one frame is one sample per channel, so seconds × `sample_rate` gives the frame index. - Shorten-compressed payloads (`sample_coding: pcm,embedded-shorten-v1.09` and friends) are detected and reported, not decoded. Decompress those with a desktop converter that bundles a shorten decoder, then convert the uncompressed file here. - The sample rate is never changed — this tool re-containers audio, it does not resample. - 8-bit output is written unsigned, because that is what WAV requires of 8-bit PCM; 16-, 24- and 32-bit output stays signed little-endian. - A header with no `sample_byte_format` and samples wider than one byte is an error rather than a guess; set **Byte order** explicitly. - Files with a header size other than the usual 1024 bytes are fine — the size on line 2 is what is honoured. ## FAQ
Why does my .sph file already end in .wav, and why won't it play? Several corpora (TIMIT most famously) name their SPHERE files `.WAV`. The extension is a lie: the bytes begin with the ASCII magic `NIST_1A`, not `RIFF`. Players read the first four bytes, fail to find a RIFF chunk, and refuse the file. Converting it here rewrites the container so the extension and the contents finally agree.
Do I have to upload the recording anywhere? No. The converter is a WebAssembly module that runs in the page, so the bytes you paste stay in the browser tab. That is also why the input is base64 or hex text rather than a file picker, and why the size caps are lower than a desktop tool's — everything happens in one sandboxed process.
What is the difference between "16-bit PCM" and "Keep the file's own encoding"? `16-bit PCM` always produces linear signed 16-bit samples: mu-law and A-law corpora are expanded by G.711 decoding, and 8-bit PCM is scaled up. That is the widest-compatibility option and the default. `Keep the file's own encoding` preserves the original bit depth and companding, fixing only what WAV strictly requires — byte order, and the unsigned convention for 8-bit PCM. A mu-law file converted that way stays mu-law inside the WAV (format tag 7, with the `fact` chunk WAVE requires), which halves the size but is not understood by every editor.
The audio came out as loud static. What went wrong? Almost always the byte order. If a header claims `sample_byte_format 01` but the samples are actually big-endian (or the field is missing and you guessed), every 16-bit sample is read with its halves swapped, which sounds like harsh noise at the right duration. Set **Byte order** to the opposite value and convert again — the duration and the header report stay identical, only the samples change.
How do I convert just one side of a telephone conversation? Two-channel corpora such as Switchboard and Fisher interleave the two speakers. Set **Channels** to `Channel 1 only` or `Channel 2 only` to keep one side, which also halves the output size. Pick `Mono downmix (average)` instead if you want both speakers mixed into a single track. Combine either with **Start sample frame** and **Max sample frames** to pull out one turn of the conversation.
Can it read a headerless file, or write AU/AIFF/SPHERE output? No on both counts. A detached header (a separate file describing raw samples) needs a second input, and this page takes one payload. For output, WAV and raw samples are supported; AU, AIFF and re-wrapping back into SPHERE are not — the raw option plus the re-import command in the header report covers the same ground for pipelines that need something else.
## Related tools - [Convert AIFF to FLAC](https://gizza.ai/tools/aiff-to-flac/): Convert AIFF or AIF audio to lossless FLAC in your browser — preserve samples and textual tags, choose FLAC compression level 0-12, and download the result. - [Bleep or Mute Regions of an Audio File](https://gizza.ai/tools/audio-bleep-censor/): Censor an audio file in your browser — bleep, mute, or duck the exact time regions you pick. Set a tone frequency and output format. Nothing is uploaded, free. - [Audio Channel Tool](https://gizza.ai/tools/audio-channel/): Swap left/right channels, downmix stereo to mono, up-mix mono to stereo, or copy one side onto both — right in your browser. Nothing is uploaded, free. - [Compress an Audio File](https://gizza.ai/tools/audio-compress/): Shrink MP3, WAV, M4A or any audio file right in your browser — pick a target bitrate (96 kbps default) and format. Nothing is uploaded, free. - [Compress Audio Dynamic Range](https://gizza.ai/tools/audio-compressor/): Even out loud and quiet parts of an audio file in your browser with threshold, ratio, attack, release and make-up gain. Local and private.