Convert WAV to AMR

Some, and it is permanent — this is the one direction where information is genuinely thrown away. WAV holds every sample exactly; AMR discards what a psychoacoustic model predicts you will not hear. At 192 kbps and above that prediction is good enough to be transparent on most material and most equipment. Keep the WAV as your master, because you cannot get back to it from the AMR. WAV normally stores raw uncompressed PCM — every sample kept exactly, at whatever bit depth and sample rate was captured. AMR is a narrowband speech codec running between 4.75 and 12.2 kbps, modelling the human vocal tract rather than the full audio spectrum.

It is folded down to stereo, because AMR has no practical multichannel mode. The downmix uses the standard coefficients so centre-channel dialogue stays clear and the surrounds are mixed in at reduced level. Anything that depended on discrete channel placement is gone once that happens. AMR destroys music: its 8 kHz bandwidth removes everything above telephone frequencies, and no amount of re-encoding brings it back.

Because one side of it is a speech codec rather than a music codec. It models the human vocal tract and discards everything outside telephone bandwidth, which is fine for a voice memo and destructive for anything with instruments in it. Converting away from it cannot restore the missing frequencies — they were never encoded in the first place. WAV is enormous — roughly 10 MB per stereo minute at CD quality — and the classic format has a 4 GB ceiling and no room for proper tags.

Free accounts cap at 100 MB per file, and WAV sources run large because the format carries substantial per-chunk overhead on top of the video data. Premium lifts the cap considerably. If the file is split into numbered parts because of the format ceiling, upload them together and they are treated as one job.

Default is 192 kbps for lossy AMR; pass-through for lossless AMR. Override to 320 kbps for audiophile or 64-96 kbps for spoken-word audiobooks (transparent for voice, halves the file size).

If WAV is lossy and AMR is lossless (e.g. MP3 → WAV), the AMR file is no better than the WAV — you can't recover information that's already been thrown away. If WAV is lossless and AMR is lossy, expect the AMR codec to recompress; at 192 kbps this is transparent for most content.

Yes — title, artist, album, year, track number, album art, and chapter markers (important for audiobooks) are read from WAV and written into the AMR container where the AMR format supports tags. M4B audiobook structure round-trips into M4B and AAX → M4B output.

Yes — drop a folder of WAV files in and we process them in parallel. Premium has more parallel workers and no per-file size cap, so a 500-file batch finishes in minutes rather than tens of minutes.

By default yes (48 kHz WAV → 48 kHz AMR). If you need to downsample for compatibility (e.g. 96 kHz → 44.1 kHz for CD burning) the advanced sample-rate option does this with high-quality resampling.

Yes — the loudness-normalize option applies ITU-R BS.1770 / EBU R128 normalization to the AMR output, targeting -14 LUFS (streaming standard), -16 LUFS (podcast standard), or -23 LUFS (audiobook standard). Useful when batch-converting tracks with varying mastering levels.

MP3 plays universally. AAC plays on Apple, most Android, Sonos. FLAC plays on Sonos and Android, less well on older iPods. M4B audiobooks play with chapter navigation on iOS Books, Audiobookshelf, and Smart Audiobook Player. The advanced options include device presets.