Make dither noise as a value between -0.5 and 0.5 and add this
to the scaled samples.
For this, we first need to do the scaling and then the CLAMP to
the target depth. This optimizes to the same code but allows us
to avoid under and overflows when we add the dither noise.
Add more dithering methods.
Expose a dither.method property on audioconvert. Disable dither when
the target depth > 16.
We need to do dithering and noise when converting f32 to the
target format. This is more natural because we can work in 32 bits
integers instead of floats.
This will also make it possible to actually calculate the error between
source and target values and implement some sort of feedback and
noise shaping later.
Move the noise setting in the dither struct so that it can be
handled separately.
Setup dither separately.
Set used cpu_flags in structures after setup.
The quantize is the amount of bits we want to keep from the original
signal, subtract the amount of bits for noise. Clamp this to 0 (all
noise).
Calculate the scale factor better with powf() and avoid overflows.
Fixes#2479
Rename empty.noise -> dither.noise and always add this amount of noise
when > 0. This also adds the noise to silent sounds, not only when
nothing is connected because that would also be a problem when an amp
needs to be kept alive with an non-0 signal.
Rename noise -> dither because we can use this also for dithering later.
See #705
Using `int` results in UndefinedBehaviorSanitizer errors
when `noise::intensity` is 31 as that would shift the 1 into
the sign bit of a signed integer type.
Add an empty.noise option that specifies the number of bits to
use for noise when the input signal is pure silence.
Some amplifiers can go into suspend mode pretty easily when they
get pure silence. With empty.noise = 1, audioconvert will now generate
a bitpattern that can keep those amplifiers alive, together with
disabling suspend in the session manager.
Fixes#705
Add an EMPTY chunk flag to mark a piece of memory as 'empty'. For audio
this means silence.
Use the empty flag to avoid mixing 0 samples.
Set the empty flag in output buffers on audioconvert.
this->monitor enabled adds an additional port in reconfigure_mode. If
there was already the maximum 64, this will crash.
Make maximum number of ports one larger than max channels to avoid
problems.
Use the NEAREST flag when setting a format. This only works for raw
formats and will update the format with the nearest accepted rate
or channels. We can then query the real configured format and use that
for the converter.
This makes things work when a driver tells us it can do 44100Hz but then
refuses and changes the rate to 48000.
See #2197, #2457, #2455, rhbz#2096193
Only advance the in_offset with the number of samples that were consumed
by the resampler. In case when the resampler is filling up an old
buffer, this can be less than n_samples.
Fixes a2dp source and possibly others.
Use the offset to skip entries in the sequence array.
Use one loop to handle intermediate and trailing samples.
Fixes an issue where the last chunk of a sequence would be ignored.
