The resampler of libspeexdsp1 is used by PulseAudio and ALSA (if libasound2-plugins is installed) by default. It has serious distortion when converting loud audio from 44100 to 48000. There is a bug report at 2015 (https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=799176). Now (2018), it is still a unsolved issue. It's a fixed point overflow problem.
Solutions
fix the libspeexdsp1 library in debian repository. The one in the repository is built from speex source, but now speexdsp is separated from speex source as a standalone library (https://github.com/xiph/speexdsp)
or just don't use speex's resampler. For ALSA, add defaults.pcm.rate_converter "samplerate" to /etc/asound.conf or ~/.asoundrc. There are other resamplers, use ls /usr/lib/arm-linux-gnueabihf/alsa-lib to find more resamplers. For PulseAudio, set resample-method = ffmpeg in /etc/pulse/daemon.conf, use pulseaudio --dump-resample-methods to find more resampler.
If you wonder which resampler of PulseAudio is fit for you, there is a resampler quality evaluation from Alexander E. Patrakov. See
The resampler of
libspeexdsp1
is used by PulseAudio and ALSA (if libasound2-plugins is installed) by default. It has serious distortion when converting loud audio from 44100 to 48000. There is a bug report at 2015 (https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=799176). Now (2018), it is still a unsolved issue. It's a fixed point overflow problem.Solutions
fix the libspeexdsp1 library in debian repository. The one in the repository is built from speex source, but now speexdsp is separated from speex source as a standalone library (https://github.com/xiph/speexdsp)
or just don't use speex's resampler. For ALSA, add
defaults.pcm.rate_converter "samplerate"
to/etc/asound.conf
or~/.asoundrc
. There are other resamplers, usels /usr/lib/arm-linux-gnueabihf/alsa-lib
to find more resamplers. For PulseAudio, setresample-method = ffmpeg
in/etc/pulse/daemon.conf
, usepulseaudio --dump-resample-methods
to find more resampler.If you wonder which resampler of PulseAudio is fit for you, there is a resampler quality evaluation from Alexander E. Patrakov. See
http://comments.gmane.org/gmane.comp.audio.pulseaudio.general/21051 https://imgur.com/a/jtIEj
It compares resampler of ffmpeg, sox and speex.