Initializing... drag & drop files here
Supports: HEVC, MOV, MP4, M4V, MKV
You have an H.265 file and you want only the sound out of it. The container question answers itself — M4A is an audio-only MPEG-4 file (ISO/IEC 14496-14:2003) holding an AAC track, MP3 is the older MPEG-1 Audio Layer III — so the real question is which codec you want on the other side. The short version: if the audio inside your source was already AAC, which it usually is, M4A keeps you in the same codec family and costs you the least; if the file has to play on something old, MP3 is the safer bet. One thing decides this before either of them: whether your .hevc actually contains an audio track at all.
| Property | M4A (this page) | MP3 |
|---|---|---|
| Container | MPEG-4 Part 14, ISO/IEC 14496-14:2003 | MPEG-1 Audio, ISO/IEC 11172-3 |
| Codec inside | AAC | MPEG-1 Audio Layer III |
| Age | AAC arrived with MPEG-2 (ISO/IEC 13818-7) in 1997 and was extended in MPEG-4 | Standardised in 1993 |
| Efficiency | Higher — the gap widens as bitrate falls | Lower, but very well tuned after three decades |
| Bitrate ceiling on this page | 384 kbps | 320 kbps |
| Same codec as most HEVC soundtracks | Yes — containers carrying H.265 video normally pair it with AAC | No — always a codec change |
| Apple ecosystem | Native across iPhone, iPad, Mac, Apple Music and CarPlay | Plays, but is not the native library format |
| 2000s-era hardware | Some old players will not open an .m4a at all |
Plays on very nearly anything with a headphone socket |
| Metadata tagging | MPEG-4 atoms, handled by every current tagger | ID3v2, understood by essentially every tool ever written |
| Best for | Phones, Apple devices, modern editors, smallest file at a given quality | Car stereos, old DAPs, embedded players, handing a file to a stranger |
.m4a is the format the library expects and indexes cleanly..m4a is..hevc onto the page or click "+ Add Files" to browse. Several files can be queued and extracted with the same settings in one pass.HH:MM:SS.sss..m4a. Your file is uploaded over an encrypted connection, converted on our servers, and deleted automatically a few hours later. No sign-up, no watermark.| HEVC source | M4A output | |
|---|---|---|
| Media | Video, plus audio only if a container is wrapped around the stream | Audio only |
| Video track | H.265 | Discarded |
| Audio codec | Usually AAC when present; sometimes AC-3, Opus or LPCM | Always re-encoded to AAC |
Bare .hevc elementary stream |
Picture only — no audio exists to extract | Silent, correctly so |
| Chapters | Not present in an elementary stream | Not written |
| Tags | Elementary streams carry none; container metadata rarely maps to music tags | Expect an untagged file — tag it afterwards |
| Best use | Storage and delivery of the picture | Listening, music and podcast libraries |
Because the source had no audio track. HEVC is a video coding standard and defines nothing about sound, so a true .hevc file — an Annex B elementary stream from a screen recorder, a DVR, or an ffmpeg raw dump — has no soundtrack for the converter to find. No setting can create one. The exception is a .hevc written by our own HEVC converters, which is really a Matroska container that keeps the source audio; those extract normally. If yours is a genuine elementary stream, the audio you want is still in the .mp4, .mov or .mkv that the video was demuxed out of — convert that instead, for example with MOV to MP3.
Constant Bitrate, in most cases. It is the mode that gives you a number you chose and can predict the file size from — pick 128 kbps for speech, 192 or 256 kbps for music. Quality Preset is the mode the page opens on, but it is the automatic, hands-off option and the least predictable control here — and nothing is pre-selected in its dropdown, so leaving it alone means the server falls back to its own default rather than the "Highest" shown on screen. Use Custom Bitrate when you want a value the dropdown does not offer, and Specific file size when the size is the requirement and the bitrate is whatever falls out of it.
Some, because there are two lossy stages. The audio muxed alongside H.265 video was already compressed, usually with AAC, and our pipeline decodes that stream and re-encodes it rather than copying it across — so a second generation of loss is unavoidable. The way to make it inaudible is to encode at or a little above the source rate: if the original track was 128 kbps, exporting at 128–192 kbps loses nothing you can hear, while 320 kbps only grows the file without recovering anything.
Because the conversion always runs a decode-and-encode pass rather than lifting the compressed stream across untouched. That is what makes the trim, channel, sample-rate and bitrate controls work — every one of them needs decoded audio to operate on. The practical consequence is that an AAC-in-container source does not come out bit-identical, so pick a bitrate at least as high as the source if you want the second pass to be transparent.
Structurally nothing — both are MPEG-4 Part 14 files with an AAC track, and the extension is the whole difference. It is a behavioural cue rather than a technical one: Apple Books, Apple Podcasts and dedicated audiobook players treat .m4b as an audiobook and remember where you stopped between sessions, while .m4a is filed as music and restarts from the beginning. If you are extracting a lecture, a sermon or a long talk you will listen to across several sittings, use HEVC to M4B instead.
Rarely. AAC at 256 kbps is already transparent to most listeners on most material, and the returns above that are small enough that the extra 50% of file size buys very little. The one case where the top of the dropdown earns its place is when the extracted audio is an intermediate — something you will edit and re-encode again — because each generation of loss compounds and starting higher leaves more headroom. For anything you are simply going to listen to, 192 or 256 kbps is the sensible ceiling.
Generally not. A bare elementary stream carries no metadata at all, and the container-level metadata in a video file describes the video — duration, dimensions, frame rate — rather than the music-library fields a player looks for. Those do not map onto MPEG-4 audio metadata atoms, so expect the M4A to arrive untagged and to need tagging in your player or a dedicated tag editor.
Only for spoken word. Folding a stereo mix to mono roughly halves the data the encoder has to describe, which is a genuine saving on lectures, interviews and voice memos where the two channels are nearly identical anyway. On music it collapses the stereo image and can cancel out anything mixed out of phase, which is a real and audible loss. If size is the goal on a music track, lower the bitrate instead and leave Audio Channel on Original.
Yes. Set Trim to Time Range and enter a Start time and a Duration, in seconds or as HH:MM:SS.sss. Only that span is decoded and encoded, which also finishes noticeably faster on a long file. If you need several separate excerpts from one recording, export the full M4A first and then split it with the audio cutter, which takes M4A directly.
Your .hevc is uploaded over an encrypted connection, converted on our servers, and the .m4a is returned to you. Uploads and outputs are deleted automatically a few hours after the conversion — no sign-up, no watermark, and files are never shared or made public. The main practical constraint on a multi-hour source is upload size and connection speed rather than the encode, so trimming before uploading is worth the extra click.