* Update GStreamer backend to 0.18
* Don't manually go through all intermediate states when shutting down the GStreamer backend; that happens automatically
* Don't initialize GStreamer twice
* Use less stringly-typed API for configuring the appsrc
* Create our own main context instead of stealing the default one; if the application somewhere else uses the default main context this would otherwise fail in interesting ways
* Create GStreamer pipeline more explicitly instead of going via strings for everything
* Add an audioresample element before the sink in case the sink doesn't support the sample rate
* Remove unnecessary `as_bytes()` call
* Use a GStreamer bus sync handler instead of spawning a new thread with a mainloop; it's only used for printing errors or when the end of the stream is reached, which can also be done as well when synchronously handling messages.
* Change `expect()` calls to proper error returns wherever possible in GStreamer backend
* Store asynchronously reported error in GStreamer backend and return them on next write
* Update MSRV to 1.56
- Switch from `lewton` to `Symphonia`. This is a pure Rust demuxer
and decoder in active development that supports a wide range of
formats, including Ogg Vorbis, MP3, AAC and FLAC for future HiFi
support. At the moment only Ogg Vorbis and MP3 are enabled; all
AAC files are DRM-protected.
- Bump MSRV to 1.51, required for `Symphonia`.
- Filter out all files whose format is not specified.
- Not all episodes seem to be encrypted. If we can't get an audio
key, try and see if we can play the file without decryption.
- After seeking, report the actual position instead of the target.
- Remove the 0xa7 bytes offset from `Subfile`, `Symphonia` does
not balk at Spotify's custom Ogg packet before it. This also
simplifies handling of formats other than Ogg Vorbis.
- When there is no next track to load, signal the UI that the
player has stopped. Before, the player would get stuck in an
infinite reloading loop when there was only one track in the
queue and that track could not be loaded.
While `Xoshiro256+` is faster on 64-bit, it has low linear complexity in the
lower three bits, which *are* used when generating dither.
Also, while `Xoshiro128StarStar` access one less variable from the heap,
multiplication is generally slower than addition in hardware.
* Improve error handling
* Harmonize `Seek`: Make the decoders and player use the same math for converting between samples and milliseconds
* Reduce duplicate calls: Make decoder seek in PCM, not ms
* Simplify decoder errors with `thiserror`
* More meaningful error messages
* Use F32 if a user requests F64 (F64 is not supported by PulseAudio)
* Move all code that can fail to `start` where errors can be returned to prevent panics
* Use drain in `stop`
Better error handling in Alsa backend
* More consistent error messages
* Bail on fatal errors in player
* Capture and log the original error as a warning when trying to write to PCM before trying to recover
This caused quite a bump in CPU usage, which be acceptable if this
actually improved sound quality. However, it turns out that this
function only has one decimal precision, i.e. it would consider
all values from `0.50..0.60` (exclusive) as `0.5` which is in
error for our purposes.
Dithering lowers digital-to-analog conversion ("requantization") error, linearizing output, lowering distortion and replacing it with a constant, fixed noise level, which is more pleasant to the ear than the distortion.
Guidance:
- On S24, S24_3 and S24, the default is to use triangular dithering. Depending on personal preference you may use Gaussian dithering instead; it's not as good objectively, but it may be preferred subjectively if you are looking for a more "analog" sound akin to tape hiss.
- Advanced users who know that they have a DAC without noise shaping have a third option: high-passed dithering, which is like triangular dithering except that it moves dithering noise up in frequency where it is less audible. Note: 99% of DACs are of delta-sigma design with noise shaping, so unless you have a multibit / R2R DAC, or otherwise know what you are doing, this is not for you.
- Don't dither or shape noise on S32 or F32. On F32 it's not supported anyway (there are no integer conversions and so no rounding errors) and on S32 the noise level is so far down that it is simply inaudible even after volume normalisation and control.
New command line option:
--dither DITHER Specify the dither algorithm to use - [none, gpdf,
tpdf, tpdf_hp]. Defaults to 'tpdf' for formats S16
S24, S24_3 and 'none' for other formats.
Notes:
This PR also features some opportunistic improvements. Worthy of mention are:
- matching reference Vorbis sample conversion techniques for lower noise
- a cleanup of the convert API
Usage: `--format {F32|S16}`. Default is F32.
- Implemented for all backends, except for JACK audio which itself
only supports 32-bit output at this time. Setting JACK audio to S16
will panic and instruct the user to set output to F32.
- The F32 default works fine for Rodio on macOS, but not on Raspian 10
with Alsa as host. Therefore users on Linux systems are warned to set
output to S16 in case of garbled sound with Rodio. This seems an issue
with cpal incorrectly detecting the output stream format.
- While at it, DRY up lots of code in the backends and by that virtue,
also enable OggData passthrough on the subprocess backend.
- I tested Rodio, ALSA, pipe and subprocess quite a bit, and call on
others to join in and test the other backends.
Probably more simple than the previous approach which
doubles the code: Instead of implementing the `Open` trait,
we simply use custom SinkBuilder, one for the default host,
and one for the "jack" host.
