HE-AAC : how to turn an album into 22 MB of data

[ron999]

Hi
There's a stand-alone fdkaac encoder available from here ---> https://github.com/nu774/fdkaac

Use fdkaac -h to show the options.

Examples of use.

CBR
For AAC-LC, like this:-
(Format profile : LC)

fdkaac -b 128k -o foo.m4a foo.wav

For AAC-HE, like this:-
(Format profile : HE-AAC / LC)

fdkaac -p 5 -b 64k -o foo.m4a foo.wav

For AAC-HEv2, like this:-
(Format profile : HE-AACv2 / HE-AAC / LC)

fdkaac -p 29 -b 32k -o foo.m4a foo.wav

VBR[1-5]

fdkaac -m 3 -o foo.m4a foo.wav

(afterburner is enabled by default, use -a 0 to disable it)

[evilsoup]

Having just played around with it a little bit, I have to mostly concur with Shantiq about the high quality of HE-AAC. I think there were a few audio artefacts, but they were so hard to notice that they could just as easily be my imagination, and I certainly wouldn't pick them up with regular listening, where I'm not trying to find those things. An audiophile with top-end gear might disagree, of course, and I haven't tested this on speakers yet (which is where most MP3s fall down for me), but so far it's sounding good.

It's nice to have a good-quality AAC encoder within ffmpeg at last, I've always found it annoying having to use neroaacenc. Testing out libfdk_aac's -vbr setting, even the lowest setting of -vbr 1 puts out (to me) audio indistinguishable from the source FLAC. I'll probably be sticking to regular AAC for my DVD rips, since not all hardware devices support HE-AAC yet - assuming it sounds good with surround-sound (and I see no reason why it shouldn't).

ron999, what does the -signaling flag do in the ffmpeg examples you gave? I like to know what all the options do when I'm using the CLI, and the ffmpeg documentation is typically unhelpful. I would guess that it puts some kind of metadata flag on the file, to tell the player what AAC profile is being used, or something like that?

EDIT: huh, weird. I used mediainfo on one of the files I encoded with -vbr, and it says

Bit rate mode                            : Constant

...so the -vbr tag actually doesn't output a variable bit rate file? I'll have to check out that standalone fdk_aac encoder, to test if it is just an ffmpeg problem.

[shantiq]

even the lowest setting of -vbr 1 puts out (to me) audio indistinguishable from the source FLAC

so true and that for me is the whole point ... it sounds [to my ears] as good altho 22000kHz and 38kbps as opposed to 48000khz
and say 700kbps

for more detailed info

as Ron showed me

ffmpeg --help full

gives you

libfdk_aac AVOptions:
-afterburner <int> E...A. Afterburner (improved quality) (from 0 to 1)
-eld_sbr <int> E...A. Enable SBR for ELD (for SBR in other configurations, use the -profile parameter) (from 0 to 1)
-signaling <int> E...A. SBR/PS signaling style (from -1 to 2)
default E...A. Choose signaling implicitly (explicit hierarchical by default, implicit if global header is disabled)
implicit E...A. Implicit backwards compatible signaling
explicit_sbr E...A. Explicit SBR, implicit PS signaling
explicit_hierarchical E...A. Explicit hierarchical signaling
-latm <int> E...A. Output LATM/LOAS encapsulated data (from 0 to 1)
-header_period <int> E...A. StreamMuxConfig and PCE repetition period (in frames) (from 0 to 65535)
-vbr <int> E...A. VBR mode (1-5) (from 0 to 5)

and

fdkaac -h
fdkaac 0.0.8
Usage: fdkaac [options] input_file
Options:
-h, --help Print this help message
-p, --profile <n> Profile (audio object type)
2: MPEG-4 AAC LC (default)
5: MPEG-4 HE-AAC (SBR)
29: MPEG-4 HE-AAC v2 (SBR+PS)
23: MPEG-4 AAC LD
39: MPEG-4 AAC ELD
129: MPEG-2 AAC LC
132: MPEG-2 HE-AAC (SBR)
156: MPEG-2 HE-AAC v2 (SBR+PS)
-b, --bitrate <n> Bitrate in bits per seconds (for CBR)
-m, --bitrate-mode <n> Bitrate configuration
0: CBR (default)
1-5: VBR
(VBR mode is not officially supported, and
works only on a certain combination of
parameter settings, sample rate, and
channel configuration)
-w, --bandwidth <n> Frequency bandwidth in Hz (AAC LC only)
-a, --afterburner <n> Afterburner
0: Off
1: On(default)
-L, --lowdelay-sbr Enable ELD-SBR (AAC ELD only)
-s, --sbr-signaling <n> SBR signaling mode
0: Implicit, backward compatible(default)
1: Explicit SBR and implicit PS
2: Explicit hierarchical signaling
-f, --transport-format <n> Transport format
0: RAW (default, muxed into M4A)
1: ADIF
2: ADTS
6: LATM MCP=1
7: LATM MCP=0
10: LOAS/LATM (LATM within LOAS)
-c, --adts-crc-check Add CRC protection on ADTS header
-h, --header-period <n> StreamMuxConfig/PCE repetition period in
transport layer

-o <filename> Output filename
--ignore-length Ignore length of WAV header

Tagging options:
--title <string>
--artist <string>
--album <string>
--genre <string>
--date <string>
--composer <string>
--grouping <string>
--comment <string>
--album-artist <string>
--track <number[/total]>
--disk <number[/total]>
--tempo <n>

[ron999]

... what does the -signaling flag do ... I would guess that it puts some kind of metadata flag on the file, to tell the player what AAC profile is being used...

This is known as "SBR signalling"

...so the -vbr tag actually doesn't output a variable bit rate file

Maybe MediaInfo gets confused.

Files converted @CBR with fdkaac encoder show as Variable with MediaInfo.
This is probably true, I think AAC CBR is just very constrained VBR.

[evilsoup]

Well at the end of the day I'll just have to trust that the people who made it knew what they were doing. It passes the listening test anyway, and that's the only one that really matters.