search⌘KtranslateENexpand_moreupload_fileConvert

home / audio / aac · m4a

Convert AAC Audio Free

Convert AAC audio for free, and see how the codec works: everything here runs in your browser.

AAC is the codec. .m4a and .aac are two ways to carry it.

An .m4a is an MP4 container: a two-byte AudioSpecificConfig describes the stream once, a sample table lists every frame, and iTunes-style atoms hold tags and cover art. A raw .aac is ADTS: the same frames, each with its own 7-byte header. Every AAC encoder adds silent priming samples at the start and padding at the end, and only some players cut them. The sample here is encoded by your browser, and the page measures that delay itself.

folder_openOpen .m4a / .aacSample .m4aSame frames as .aac (ADTS)downloadDownload

arpeggio.m4a: 259 AAC-LC frames from your browser, muxed here with edit list, iTunSMPB, tags and cover; encoder delay measured by decoding: 465 samples

AudioSpecificConfig of the loaded file (esds → DecSpecificInfo)12 10 → mp4a.40.2

00010AOT0100sf0010ch0fl0core0ext

audioObjectType2 = AAC LC

samplingFrequencyIndex4 = 44100 Hz

channelConfiguration2 = stereo (L R)

frameLengthFlag0 = 1024-sample frames

dependsOnCoreCoder0 = no

extensionFlag0 = 0

Where the real audio starts and stops

from iTunSMPB · 265,216 decoded samples in 259 frames

priming 465264,600 samples of audiopadding 151

who says how longsamplessecondswhere it comes from

all frames decoded265,2166.0140 sstts sample table: 259 frames

mdhd media duration265,2166.0140 strack header, in the media timescale

edit list (elst)264,6006.0000 sstarts at media time 465: the priming samples are skipped

iTunSMPB264,6006.0000 sApple gapless tag: delay 465, padding 151

your browser (decodeAudioData)264,6006.0000 strimmed to the real audio: gapless info honoured

bitrate per second · from stsz sample sizesavg 128 kb/s · frames 278-478 B · esds says avg 128, max 160

box tree · M4A [M4A mp42 isom] · 46 boxes

0ftypmajor M4A · compatible M4A mp42 isomfile type: which specs a reader needs to understand the filepayload 4D 34 41 20 00 00 00 00 4D 34 41 20 …28 B

28mooveverything except the audio bytes: tracks, timing, tags8.8 KB

36mvhdtimescale 1000 · duration 6000 = 6.000 s · next track 2movie header: overall timescale and lengthpayload 00 00 00 00 00 00 00 00 00 00 00 00 …108 B

144trakone track (here the only one, audio)1.5 KB

152tkhdtrack 1 · flags enabled in-movie · duration 6000 · volume 1track header: id, length, enabled flagpayload 00 00 00 03 00 00 00 00 00 00 00 00 …92 B

244edtsedit list container36 B

252elstplay 6000 (movie units) from media time 465 → skips 465 priming samplesedit list: which part of the decoded track to actually playpayload 00 00 00 00 00 00 00 01 00 00 17 70 …28 B

280mdiamedia container for this track1.4 KB

288mdhdtimescale 44100 · duration 265,216 samples · lang undmedia header: the track’s own clock and languagepayload 00 00 00 00 00 00 00 00 00 00 00 00 …32 B

320hdlrtype “soun” · name “SoundHandler”handler: what kind of data this track holdspayload 00 00 00 00 00 00 00 00 73 6F 75 6E …45 B

365minfmedia information1.3 KB

373smhdbalance 0sound media header: stereo balancepayload 00 00 00 00 00 00 00 0016 B

389dinfdata information36 B

397dref1 entrydata reference: where the samples are storedpayload 00 00 00 00 00 00 00 01 00 00 00 0C …28 B

413url flags 1: self-containedflag 1 = in this same filepayload 00 00 00 0112 B

425stblsample table: the index of every AAC frame1.2 KB

433stsd1 entrysample description: which codec, how many channels, what ratepayload 00 00 00 00 00 00 00 01 00 00 00 4B …91 B

449mp4a2 ch · 16-bit · 44100 HzMPEG-4 audio sample entrypayload 00 00 00 00 00 00 00 01 00 00 00 00 …75 B

485esdsoti 0x40 · avg 128 kb/s · max 160 · ASC 12 10 = mp4a.40.2elementary stream descriptor: carries the AudioSpecificConfigpayload 00 00 00 00 03 19 00 01 00 04 11 40 …39 B

524stts1 run(s): 259 × 1024 samplestime-to-sample: duration of each framepayload 00 00 00 00 00 00 00 01 00 00 01 03 …24 B

548stsc1 run(s): from chunk 1, 259 frames per chunksample-to-chunk: how frames are grouped into chunkspayload 00 00 00 00 00 00 00 01 00 00 00 01 …28 B

576stsz259 frames · 278-478 B eachsample sizes: bytes of each framepayload 00 00 00 00 00 00 00 00 00 00 01 03 …1.1 KB

1,632stco1 chunk(s) · first at byte 8,863 = start of mdat data ✓chunk offsets: where each chunk starts in the file (32-bit)payload 00 00 00 00 00 00 00 01 00 00 22 9F20 B

1,652udtauser data container7.2 KB

1,660metahandler mdir/applmetadata container (iTunes style)payload 00 00 00 00 00 00 00 21 68 64 6C 72 …7.2 KB

1,672hdlrtype “mdir” · name “”handler: what kind of data this track holdspayload 00 00 00 00 00 00 00 00 6D 64 69 72 …33 B

1,705ilst8 tagsiTunes item list: one child per tag7.2 KB

1,713©namtitle37 B

1,721datatype 1 UTF-8: “Arpeggio in G”tag value with a type codepayload 00 00 00 01 00 00 00 00 41 72 70 65 …29 B

1,750©ARTartist34 B

1,758datatype 1 UTF-8: “FileFormer”tag value with a type codepayload 00 00 00 01 00 00 00 00 46 69 6C 65 …26 B

1,784©albalbum38 B

1,792datatype 1 UTF-8: “Format samples”tag value with a type codepayload 00 00 00 01 00 00 00 00 46 6F 72 6D …30 B

1,822©dayyear28 B

1,830datatype 1 UTF-8: “2026”tag value with a type codepayload 00 00 00 01 00 00 00 00 32 30 32 3620 B

1,850©tooencoding tool60 B

1,858datatype 1 UTF-8: “WebCodecs AAC encoder via FileFormer”tag value with a type codepayload 00 00 00 01 00 00 00 00 57 65 62 43 …52 B

1,910trkntrack number of total32 B

1,918datatype 0 binary: 00 00 00 03 00 0C 00 00tag value with a type codepayload 00 00 00 00 00 00 00 00 00 00 00 03 …24 B

1,942covrcover art6.7 KB

1,950datatype 14 PNG: 6.7 KB imagetag value with a type codepayload 00 00 00 0E 00 00 00 00 89 50 4E 47 …6.7 KB

8,667----freeform tag188 B

8,675mean“com.apple.iTunes”namespace of a freeform tag28 B

8,703name“iTunSMPB”name of a freeform tag20 B

8,723datatype 1 UTF-8: “ 00000000 000001D1 00000097 0000000000040998 00000000 00000000 0000000”tag value with a type codepayload 00 00 00 01 00 00 00 00 20 30 30 30 …132 B

8,855mdat96,535 bytes of AAC frames, starting at byte 8,863media data: the raw AAC frames, back to back96.5 KB

©namArpeggio in G

©ARTFileFormer

©albFormat samples

©day2026

©tooWebCodecs AAC encoder via FileFormer

trkn3 of 12

----:iTunSMPB 00000000 000001D1 00000097 0000000000040998 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000

check_circleAAC in MP4.Sample entry mp4a, objectTypeIndication 0x40 (MPEG-4 Audio), codec string mp4a.40.2.

check_circlemoov before mdat.Playback can start as soon as the first bytes arrive.

check_circleCover art embedded.covr atom, 6.7 KB.

AudioSpecificConfig lab · tested against your browser

Two bytes decide whether a file plays

Pick a profile, rate and layout to build the config a muxer would write. HE-AAC can be signalled three ways: explicitly (object type 5 first), backward-compatibly (an LC config with a sync extension 0x2B7 that old decoders skip), or not at all, leaving the decoder to discover SBR in the bitstream and play at half the rate if it doesn’t. Each result is checked live with MediaSource and WebCodecs.

profile

AAC LCHE-AACHE-AAC v2MainLD

output sample rate

96 kHz48 kHz44.1 kHz32 kHz24 kHz22.05 kHz

channels

monostereo5.17.1

SBR signalling

12 10mp4a.40.2

00010AOT0100sf0010ch0fl0core0ext

MediaSource.isTypeSupported(“audio/mp4; codecs=mp4a.40.2”)supported

audio.canPlayType(same)probably

AudioDecoder.isConfigSupported(with these bytes)supported

ADTS could carry it?yes, profile 1

Hover a bit group for its meaning.

ADTS header lab

The same facts, repeated in front of every frame

ADTS was built for broadcast: a receiver can tune in anywhere, find the 0xFFF sync word and start decoding. That costs 7 bytes per frame (9 with CRC), about 2.3 kb/s at 44.1 kHz, and it can only describe profiles 1-4, so HE-AAC in ADTS is always signalled implicitly. Below is the header that goes with the config built above, for a frame of the size you set.

raw frame bytesor decode a header (hex)no CRC (7 bytes)

FF F1 50 80 2F 5F FC

111111111111syncword0ID00layer1prot01profile0100sf0private010ch0orig0home0cid bit0cid start0000101111010length11111111111buffer00blocks

syncword0xFFF ✓

IDMPEG-4

layer0

protection_absentno CRC

profileAAC LC (AOT − 1)

sf_index44100 Hz

channel_configstereo (L R)

frame_length378 bytes incl. header

buffer_fullness0x7FF = variable bitrate

raw_blocks1 raw data block(s)

Reference

AOTprofilewhat it is forbrowser

1AAC Mainoriginal MPEG-2 profile with prediction; rarely usedMSE no

2AAC LCLow Complexity: the AAC in almost every .m4aMSE yes

5HE-AACLC core at half rate + Spectral Band Replication; streaming at 48-64 kb/sMSE yes

29HE-AAC v2adds Parametric Stereo; down to about 24-32 kb/sMSE yes

23AAC LDlow delay, 480/512-sample frames for callsMSE no

39AAC ELDenhanced low delay with SBR; FaceTime-class voiceMSE no

42xHE-AACUSAC: speech and music at any bitrate, adaptive streamingMSE yes

34MP3 in MP4mp4a.40.34 or oti 0x6B: MP3 inside an MP4 container-

samplingFrequencyIndex (4 bits; 15 = explicit 24-bit rate)

096,000188,200264,000348,000444,100532,000624,000722,050816,000912,0001011,025118,000127,350

MIMEaudio/mp4 (.m4a), audio/aac (ADTS), audio/x-m4a (Apple, unregistered)

Extensions.m4a music, .m4b audiobook with chapters, .m4r ringtone, .aac / .adts raw

Magic bytesMP4: “ftyp” at offset 4 with brand M4A, M4B or mp42 · ADTS: FF F1 (MPEG-4) or FF F9 (MPEG-2)

Frame1024 samples (960 in DAB+); HE-AAC outputs 2048 per frame

DelayApple encoders 2112 samples, FFmpeg AAC 1024, Fraunhofer FDK 2048 for HE-AAC

BitratesLC transparent for most music near 192-256 kb/s stereo; HE-AAC aimed at 32-64 kb/s

StandardsISO/IEC 13818-7 (MPEG-2 AAC), ISO/IEC 14496-3 (MPEG-4 Audio), 14496-14 (MP4 file), 14496-12 (ISOBMFF)

1997MPEG-2 AACFraunhofer, AT&T, Dolby, Sony and Nokia standardise AAC as the successor to MP3.

1999MPEG-4 AudioAAC becomes the core of MPEG-4, with object types and the AudioSpecificConfig.

2003iTunes StoreApple sells music as 128 kb/s AAC in .m4a/.m4p.

2003HE-AACSpectral Band Replication added; used by DAB+ and internet radio.

2006HE-AAC v2Parametric Stereo standardised for very low bitrates.

2012xHE-AACUSAC (ISO/IEC 23003-3) unifies speech and music coding.

2020Android & ApplexHE-AAC playback ships widely for adaptive streaming.

Doing something with AAC

M4A to MP3Decodes AAC and writes MP3 for players that only take MP3.song.m4a · AAC 256arrow_forwardsong.mp3 · 320 kb/s AAC to MP3Turns raw ADTS streams and radio rips into tagged MP3s.radio.aac · ADTSarrow_forwardradio.mp3 · ID3 MP3 to M4ARe-encodes into AAC with proper gapless edit list.album.mp3arrow_forwardalbum.m4a · elst M4A to WAVLossless PCM for editing: the priming should be trimmed here.take.m4a · 2112 delayarrow_forwardtake.wav · sample-exact Audio CompressorEncodes AAC or Opus at a chosen bitrate and measures the cutoff.podcast.wav · 50 MBarrow_forwardpodcast.m4a · 64 kb/s MP4 formatThe same box structure, explored for video tracks and fast start.moov · trak · stblarrow_forwardcodec string per track

sourcesISO/IEC 14496-3 Apple: gapless (iTunSMPB, elst)MP4 registration authority RFC 6381 codecs parameter WebCodecs AAC registration MultimediaWiki: ADTS

Frequently Asked Questions

How do I convert AAC to MP3?

Add your .aac file to the converter, choose MP3 as the output, and download it. Both AAC and MP3 are lossy formats, so the audio is re-encoded; picking a reasonable bitrate keeps quality close. MP3 plays on nearly every device, which is the main reason people convert from AAC.

Can Audacity or VLC convert AAC to MP3?

Both Audacity and VLC can convert AAC to MP3, though each needs a few settings changed first. You can skip that setup and convert AAC to MP3 right here in your browser instantly, with no install and no signup, then download the finished MP3.

Can I convert AAC to FLAC or WAV?

Yes. You can convert AAC to WAV for uncompressed editing or to FLAC for lossless compression. Because AAC is already lossy, these targets preserve the current quality but cannot restore detail AAC removed. Both conversions are free and run in your browser.

Can you convert AAC to ALAC?

Yes. You can convert AAC to ALAC (Apple Lossless), which is convenient for an Apple-centric library. Keep in mind that wrapping lossy AAC in a lossless ALAC container does not add back quality; it simply stores the existing audio without further loss. Conversion here is free.