Audio to WAV / FLAC Converter

WAV and FLAC are the two formats that store audio without losing a single sample — both are lossless. They do it in very different ways. WAV is the raw container that stores uncompressed PCM samples exactly as the microphone or synthesizer produced them: every amplitude snapshot lands verbatim in the file, which is why WAV is large but perfectly editable and the universal intermediate format for studios. FLAC wraps the same PCM in a lossless compressor that typically shrinks the file by 50–60% while still allowing a bit-exact decode back to the original bytes — decode(encode(x)) equals x, every time. Both formats keep every sample, unlike lossy formats like MP3, AAC and Ogg Vorbis (covered on their own pages); the trade-off is file size.

PCM (Pulse Code Modulation) is the universal raw format for digital audio: the analog waveform is sliced into equal time intervals (the sample rate) and at each interval the amplitude is rounded to the nearest of a fixed number of discrete levels (the bit depth). A 44.1 kHz / 16-bit / stereo recording stores 44,100 amplitude measurements per second per channel, each rounded to one of 2¹⁶ = 65,536 levels (24-bit: one of 2²⁴ = 16,777,216). The Nyquist–Shannon sampling theorem says a sampled signal can be perfectly reconstructed as long as its highest frequency is below sampleRate / 2. Human hearing tops out around 20 kHz, so 44.1 kHz (capturing up to 22.05 kHz) is mathematically sufficient for perfect reproduction — which is why the CD standard adopted exactly that rate.

The bit depth controls dynamic range (the ratio between loudest and quietest signal), not frequency response. Each extra bit adds about 6.02 dB of theoretical signal-to-noise ratio: 16-bit gives 98.1 dB, 24-bit gives 146.2 dB, 32-bit float (used internally in DAWs) gives 194.4 dB. Real hardware never reaches the theoretical figure — a well-engineered 24-bit ADC delivers about 120 dB of usable range (vinyl: ~70 dB, cassette: ~60 dB).

WAV (or WAVE, formally WAVEform Audio File Format) is the Microsoft/IBM RIFF container defined in 1991. Its on-disk layout is three chunks back-to-back: a 12-byte RIFF header that names the file as WAVE, a fmt chunk describing the PCM format (sample rate, channels, bit depth), and a data chunk containing the raw samples themselves. Bytes are little-endian; the format has not changed in 35 years; every operating system and DAW reads it natively.

Because WAV stores raw PCM with no compression and no overhead beyond the header, the file size is exactly predictable: bytes = sampleRate × bitDepth × channels × duration_seconds / 8. A CD-quality track (44.1 kHz / 16-bit / stereo, bitrate = 1,411,200 bps = 1411.2 kbps) chews through 1,411,200 ÷ 8 = 176,400 bytes per second, which works out to 10.584 MB per minute or about 42.336 MB for a 4-minute song. The 48 kHz / 24-bit / stereo studio format (bitrate 2304 kbps) climbs to 17.280 MB per minute. This predictability is why WAV remains the industry's exchange format.

FLAC (Free Lossless Audio Codec) was designed by Josh Coalson in 2001 as an open, royalty-free answer to Apple's then-proprietary ALAC. It works on blocks of samples (typically 4,096 at level 5) in three steps: inter-channel decorrelation (making the two stereo channels more similar), linear-predictive coding (predict the next sample from previous ones, store only the small residual), and Rice entropy coding (a near-optimal way to pack small integers into bits). The crucial property: decoding returns the original PCM byte-for-byte. FLACdecode(FLACencode(x)) === x. That identity is what makes FLAC lossless — not "high quality" but bit-identical.

FLAC exposes 9 compression levels (0–8) that trade encoding speed against file size. Level 0 is fastest and only achieves about 30% reduction; level 5 is the default and reaches about 55%; level 8 squeezes out about 62% but takes 4–5 times longer to encode. Real-world ratios depend on the music: solo piano or vocal tracks reduce by 45–50% because the signal is sparse; dense orchestral or electronic music reduces by 55–65% because the noise floor is more predictable. A typical 4-minute CD-quality track at level 5 lands around 19.05 MB — almost exactly half of the 42.336 MB WAV.

The bit rate of any PCM stream is given by one short formula: bps = sampleRate × bitDepth × channels. Common reference points:

• 1411.2 kbps — CD audio (44.1 kHz / 16-bit / stereo)
• 1536 kbps — 48 kHz / 16-bit / stereo (DAT, early DVD)
• 2304 kbps — 48 kHz / 24-bit / stereo (studio master)
• 4608 kbps — 96 kHz / 24-bit / stereo (Hi-Res)
• 9216 kbps — 192 kHz / 24-bit / stereo (ultra-Hi-Res)

To turn bit rate into file size, multiply by duration and divide by 8 (bytes per bit). For the CD-quality 4-minute example: 1411200 × 240 / 8 = 42,336,000 bytes = 42.336 MB. For FLAC, multiply the WAV size by 1 - compression_ratio: at the level-5 default of ~55%, the same track is 42,336,000 × 0.45 ≈ 19,051,200 bytes ≈ 19.05 MB. A 48 kHz / 24-bit / stereo track weighs 17.280 MB per minute in WAV and roughly 7.78 MB per minute in FLAC level 5. Lossy codecs (MP3, AAC, Opus) hit the same audio at 128–256 kbps, ten times smaller again — but they discard samples to do it. One unit trap worth knowing: bit rates are quoted in kilobits per second (kbps) while file managers count kilobytes — divide by 8 to convert (1411.2 kbps ÷ 8 = 176.4 kB/s).

Both formats are lossless, so the question is workflow and bandwidth, not quality. Use WAV when an editor or DAW needs raw PCM input, when you're rendering an intermediate bounce that will be processed further, or when you want a master file that every program on the planet can read without libraries. Use FLAC when you want lossless audio at half the size for archival, music libraries, network streaming (one lossless CD fits in about 19 MB instead of 42 MB), and backups of physical media. FLAC also supports rich metadata via Vorbis Comment which is more capable than WAV's RIFF LIST INFO (which lacks an album artist field — this is why the tag editor above hides the album artist input when WAV is selected).

One size tip: voice recordings are mono by nature — one mouth, one microphone. Recording speech in stereo doubles the bitrate with no information gain (a 1-hour 44.1 kHz / 16-bit voice recording is 317.5 MB in mono, 635.0 MB in stereo), so switch the Channels option above to Mono for voice.

• FLAC is "better quality" than WAV. No — both are bit-perfect lossless. FLAC simply stores the same audio in less space. Decoding FLAC back to PCM gives the identical byte sequence WAV would have stored.
• Higher bit depth always sounds better. 24-bit gives more headroom for editing (low-level signals sit well above the noise floor), but the final consumer delivery is usually 16-bit. The extra 8 bits are working precision in the DAW, not audible quality in the speakers.
• Hi-Res (96 kHz / 192 kHz) sounds better than CD (44.1 kHz). For perfectly dithered music, humans cannot distinguish frequencies above about 20 kHz. 96 kHz captures ultrasonic content that no human hears and no earphone reproduces. The arguable benefit of Hi-Res is fewer aliasing artefacts in the recording chain, not playback quality.
• Converting MP3 to WAV restores the original quality. No. Once lossy compression discards samples, those samples are gone forever. The resulting WAV file contains MP3 artefacts — it is not bit-identical to the pre-MP3 master, no matter the file extension.

Related tools: Audio to MP3 / ID3 Tags for lossy compression, Audio to OGG / Opus for open lossy codecs, Audio to AIFF / ALAC / WavPack for other lossless formats, Audio Spectrum to visualise what frequencies are in your audio, and MIDI to WAV (MIDI is instructions, not audio — a different conversion pipeline).