HydrogenAudio

Hydrogenaudio Forum => Listening Tests => Topic started by: hyaudio on 2021-04-02 04:39:57

Title: 24 kbps stereo music listening test
Post by: hyaudio on 2021-04-02 04:39:57
1. Hardware: Samsung headset SHS-270V
2. Software: foobar2000 v1.6.4
3. Samples: 37 samples from Microsoft story remix app
4. Encoders and Settings
- 48 kHz and stereo
- AAC: Winamp FhG v03.02.16, CBR, HE-AAC v2 profile
- Opus: libopus 1.3.1, CVBR, music mode
- USAC was excluded because no free encoders are available
5. Results
- Both were rated for low sound quality that could be detected even with cheap($12) audio devices.
- Averagely Opus(2.31) was slightly inferior than AAC(2.59).
Title: Re: 24 kbps stereo music listening test
Post by: kode54 on 2021-04-02 05:20:24
- USAC was excluded because no free encoders are available
Were you that concerned with using Exhale without securing a patent license?
Title: Re: 24 kbps stereo music listening test
Post by: hyaudio on 2021-04-02 06:09:05
- USAC was excluded because no free encoders are available
Were you that concerned with using Exhale without securing a patent license?

Exhale supports only down to 36 kbps for stereo contents.
Title: Re: 24 kbps stereo music listening test
Post by: kode54 on 2021-04-02 07:29:35
Oh, right.
Title: Re: 24 kbps stereo music listening test
Post by: fabiorug on 2021-04-02 10:00:58
no it doesn't work the bitrate is still quite high average 42 kbps.
Title: Re: 24 kbps stereo music listening test
Post by: C.R.Helmrich on 2021-04-02 11:42:29
True. Very interesting test, though, and the 24-kbps stereo operating point definitely is one of the sweetspots for xHE-AAC (along with 16 and 32 kbps stereo, see also http://www.ecodis.de/audio.htm#mpeg, Figure 2).

Is there a particular reason why you chose the Winamp HE-AACv2 encoder over the FDK-AAC encoder?

Chris
Title: Re: 24 kbps stereo music listening test
Post by: Kamedo2 on 2021-04-02 12:36:58
Great test, so Winamp FhG is one of the strong HE-AACv2 encoder at 24kbps.
As expected, Opus struggles at this low rate, in the absence of the patented technology.
Title: Re: 24 kbps stereo music listening test
Post by: Porcus on 2021-04-03 11:04:34
only down to 36 kbps
... begging the question. A 36 kbit/s comparison, anyone? Not only exhale, but it would also be possible to include mp3. If nothing else, as an anchor and for a reminder that it is a notch off the minimum supported by that format.

(I have always taken for granted that the reason for that 32 kbit/s "limitation" is the obvious one. I am surely open to hear a better explanation.) 
Title: Re: 24 kbps stereo music listening test
Post by: Kamedo2 on 2021-04-04 17:18:01
Machine-readable results here.
Code: [Select]
HE-AACv2	Opus
%feature 10 fhgaac v03.02.16 libopus 1.3.1
%feature 11 --cbr 24 --profile hev2 --cvbr 24 --music
2.7 1.6
2.9 2.0
2.4 2.1
1.8 2.7
3.5 1.2
2.5 2.5
1.9 2.2
1.8 2.4
2.7 2.4
2.0 2.0
2.2 2.3
2.2 2.8
3.7 1.4
2.5 1.7
3.4 2.4
2.2 1.6
2.3 2.0
3.1 3.4
4.0 2.5
2.4 2.0
3.2 2.3
2.5 3.0
2.0 2.6
2.2 2.7
2.4 2.4
1.7 2.1
3.3 1.9
2.5 2.5
2.8 2.8
3.7 1.3
2.4 2.4
2.1 1.8
2.6 3.0
3.8 3.4
3.1 2.7
1.5 3.7
1.7 1.7
Title: Re: 24 kbps stereo music listening test
Post by: pr0m3th3u5 on 2021-04-04 22:55:26
All these listening tests are done using CBR... but how accurate are these especially given that most codecs default to VBR and nobody bothers to change mode to CBR in practice...?
Title: Re: 24 kbps stereo music listening test
Post by: celona on 2021-04-05 02:18:36
Exhale supports only down to 36 kbps for stereo contents.

I don't understand this sentence, not even the files obtained from the tests are stereo, they contain a central mono channel with the sum of the left and right channels sampled at 24kHz and compressed with a general purpose encoder, plus SBR and the PS information necessary to reconstruct the decoded signal.

I did a little stereo test with xHE-AAC (https://celona.altervista.org/exhale/24kbps/stereo-test.mp4) (USAC) to compare with Exhale (https://celona.altervista.org/exhale/24kbps/stereo-test.m4a), HE-AACv2 (http://celona.altervista.org/exhale/24kbps/stereo-test.m4b), and Opus (http://celona.altervista.org/exhale/24kbps/stereo-test.caf) (only CELT).

To me they are all stereo in the same way and except the first one they are all at 32kbps (which is less than 36kbps).

The output of ffprobe
Code: [Select]
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '001 AAC 24kbps.m4a':
  Metadata:
    major_brand     : mp42
    minor_version   : 0
    compatible_brands: mp42isom
    creation_time   : 2021-04-02T02:30:00.000000Z
    encoder         : fhgaac v03.02.16;CBR=24000
    iTunSMPB        :  00000000 000017C1 000001EF 00000000000DB650 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
  Duration: 00:00:18.86, start: 0.000000, bitrate: 27 kb/s
    Chapter #0:0: start 0.126687, end 18.858667
    Metadata:
      title           :
    Stream #0:0(eng): Audio: aac (HE-AACv2) (mp4a / 0x6134706D), 48000 Hz, stereo, fltp, 24 kb/s (default)
    Metadata:
      creation_time   : 2021-04-02T02:30:00.000000Z
    Stream #0:1(eng): Data: none (mp4s / 0x7334706D), 0 kb/s (default)
    Metadata:
      creation_time   : 2021-04-02T02:30:01.000000Z
    Stream #0:2(eng): Data: none (mp4s / 0x7334706D), 0 kb/s (default)
    Metadata:
      creation_time   : 2021-04-02T02:30:01.000000Z
Title: Re: 24 kbps stereo music listening test
Post by: celona on 2021-04-05 05:50:26
Now I understand, the "A" option starts at a bitrate of 36kbps. It's time to go to sleep.  :))
SimplePortal 1.0.0 RC1 © 2008-2021