[opus] Opus Tools -- low bitrates

[Jan Stary]

On Aug 07 08:30:31, hans at stare.cz wrote:
> On Aug 07 00:41:52, petrparizek2000 at yahoo.com wrote:
> >     #1. To test encoding at low bitrates, I encoded a sine sweep at 12 kbps
> > with Opusenc and then decoded the resulting file with Opusdec.

> 1) Opusenc --bitrate 12 --downmix-mono Sweep50.wav Sweep50.opus

Why are you using a stereo file
containing the same sweep in both channels
and then downmixing to mono?

Also, why are you repeating the sweep twice in every channel?

Also also, the file you linked to is Sweep50FlatMidSide.wav
as opposed to Sweep50.wav - is it really the same file?

> 2) Opusdec --float Sweep50.opus Sweep50out.wav

Why are you using floats for the decoded file?

> > The strange
> > thing was that even though the output wave file was at 48 kHz, it showed
> > clear marks of being resampled from 16 kHz

The encoded opus file is 48kHz,
so how would the output wav be resampled from 16kHz?
What are those "clear signs" exactly?

> and where my original frequency
> > reached 16 kHz, the frequency in the output file practically fell down to
> > zero.

The intermediate opus file already contains
only the first of the two repeating sweeps.
After around 50s, there is silence.
The resulting wav file too then, obviously.

As far as I can tell (see the attached spectrograms),
the sound dies out around the point the freq reaches 12kHz.

Also see the transcript below.

> >     Unfortunately, it seems that the lowpass filtering is somehow a part of
> > the actual resampling algorithm.

What resampling?

> > Why do I think so? Opus normally doesn't
> > encode any frequencies above 20 kHz, which would suggest that 20 kHz should
> > be aliased down to 4 kHz followed by silence.

What makes you think a sampling frequency of 16 kHz
is involved anywhere in the signal path?

> However, right at the spot
> > where my original sine sweep went all the way up to almost 24 kHz, the
> > aliased frequency wen to almost 8 instead of being absent altogether.
> >     Now I can hear someone say that natural sounds are not composed solely
> > of some very high frequencies. But I still think that aliasing is an
> > undesirable effect. After all, when the final sample rate is supposed to be
> > 48 kHz anyway, why should low bitrates make it necessary to first downsample
> > the audio and then upsample again?

What makes you think there is any downsampling involved?

	Jan


$ soxi sweep.wav

Input File     : 'sweep.wav'
Channels       : 2
Sample Rate    : 48000
Precision      : 24-bit
Duration       : 00:01:51.32 = 5343360 samples ~ 8349 CDDA sectors
File Size      : 32.1M
Bit Rate       : 2.30M
Sample Encoding: 24-bit Signed Integer PCM

$ opusenc --bitrate 12 --downmix-mono sweep.wav sweep.opus

Encoding0using_libopusi1.3.1m(audio)av1B20De_A12C8Dopus4me4mn24mc 24m--bitrate 12 --downmix-mono sweep.wav sweep.opus1B61Dn_A7C1C4mn4mc24m1B11Dc_A6C24me24mn24mc1B
-----------------------------------------------------
   Input: 48 kHz, 1 channel
  Output: 1 channel (1 uncoupled)
	  20ms packets, 12 kbit/s VBR
 Preskip: 312

Encoding0complete06 1440x realtime, 9.854 kbit/s
-----------------------------------------------------
       Encoded: 1 minute and 51.34 seconds
       Runtime: 2 seconds
		(55.67x realtime)
	 Wrote: 127667 bytes, 5567 packets, 114 pages
       Bitrate: 8.49542 kbit/s (without overhead)
 Instant rates: 3.2 to 23.2 kbit/s
		(8 to 58 bytes per packet)
      Overhead: 7.39% (container+metadata)

$ opusdec sweep.opus out-pcm.wav
$ opusdec --float sweep.opus out-float.wav

Processing file "sweep.opus"...

New logical stream (#1, serial: 2dc0e27f): type opus
Encoded with libopus 1.3.1, libopusenc 0.2.1
User comments section follows...
	ENCODER=opusenc from opus-tools 0.2
	ENCODER_OPTIONS=--bitrate 12 --downmix-mono
Opus stream 1:
	Pre-skip: 312
	Playback gain: 0 dB
	Channels: 1
	Original sample rate: 48000 Hz
	Packet duration:   20.0ms (max),   20.0ms (avg),   20.0ms (min)
	Page duration:	 1000.0ms (max),  994.1ms (avg),  340.0ms (min)
	Total data length: 127667 bytes (overhead: 7.39%)
	Playback length: 1m:51.319s
	Average bitrate: 9.175 kbit/s, w/o overhead: 8.497 kbit/s
Logical stream 1 ended

$ soxi out-pcm.wav

Input File     : 'out-pcm.wav'
Channels       : 1
Sample Rate    : 48000
Precision      : 16-bit
Duration       : 00:01:51.32 = 5343360 samples ~ 8349 CDDA sectors
File Size      : 10.7M
Bit Rate       : 768k
Sample Encoding: 16-bit Signed Integer PCM

$ soxi out-float.wav

Input File     : 'out-float.wav'
Channels       : 1
Sample Rate    : 48000
Precision      : 25-bit
Duration       : 00:01:51.32 = 5343360 samples ~ 8349 CDDA sectors
File Size      : 21.4M
Bit Rate       : 1.54M
Sample Encoding: 32-bit Floating Point PCM

-------------- next part --------------
A non-text attachment was scrubbed...
Name: spectrogram-input.png
Type: image/png
Size: 71852 bytes
Desc: not available
URL: <http://lists.xiph.org/pipermail/opus/attachments/20240807/d2e9b26b/attachment-0002.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: spectrogram-output.png
Type: image/png
Size: 190760 bytes
Desc: not available
URL: <http://lists.xiph.org/pipermail/opus/attachments/20240807/d2e9b26b/attachment-0003.png>