Since Spotify audio is always 16-bit depth, optimize the conversion pipeline:
- Always dither at 16-bit level regardless of output format
- Preserve fractional precision until final rounding for better requantization
- Replace floating-point multiplication with compile-time bit shifts
- Add comprehensive inlining to eliminate function call overhead
- Specialize 24-bit clamping to remove runtime branching
This maintains proper dithering of the original 16-bit quantization artifacts
while maximizing performance through bit-shift operations and eliminating
unnecessary runtime calculations.
- Update many dependencies to latest versions across all crates
- Switch from `once_cell::OnceCell` to `std::sync::OnceLock` where appropriate
- Update OAuth to use stateful `reqwest` for HTTP requests
- Fix Rodio backend to honor the requested sample format
This saves up to 1-2% CPU useage on a PI 4 depending on how much normalisation is actually being done.
* We don't need to test against EPSILON. The factor will never be over 1.0 in basic normalisation mode.
* Don't check the normalisation mode EVERY sample.
* Do as little math as possible by simplfiying all equations as much as possible (while still retaining the textbook equations in comments).
* Misc cleanup
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.
* Remove deprecated use of std::u16::MAX
* Use `FromStr` for fallible `&str` conversions
* DRY up strings into constants
* Change `as_ref().map()` into `as_deref()`
* Use `Duration` for time constants and functions
* Optimize `Vec` with response times
* Move comments for `rustdoc` to parse
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
