Since all algorithms will need to specify a block size (the amount of
data to be processed together), we make this a common parameter and have
the implementation set it at initialisation time.
This adds Andre Adrian's AEC implementation from his intercom project
(http://andreadrian.de/intercom/) as an alternative to the speex echo
cancellation routines. Since the implementation was in C++ and not in
the form of a library, I have converted the code to C and made a local
copy of the implementation.
The implementation actually works on floating point data, so we can
tweak it to work with both integer and floating point samples (currently
we just use S16LE).
Since the source and sink specification will need to be determined by
the AEC algorithm (can it handle multi-channel audio, does it work with
a fixed sample rate, etc.), we negotiate these using inout parameters at
initialisation time.
There is opportunity to make the sink-handling more elegant. Since the
sink data isn't used for playback (just processing), we could pass
through the data as-is and resample to the required spec before using in
the cancellation algorithm. This isn't too important immediately, but
would be nice to have.
The speex echo canceler prefers a power of 2 for the frame size. Round down the
ideal frame_size to the nearest power of two. This makes sure we don't create
more than the requested frame_size_ms latency while still providing a power of 2
to the speex echo canceller.
Try to keep the drift between source and sink within 4ms now that we have more
accurate timings.
Don't force a resync on latency changes but let the drift code handle it.
Tag the source and sink with the phone media roles so that they automatially
connect to phone streams such as Empathy when using the intended-rols module.
Make a new echo-cancel module that exposes a new sink and source. All data sent
to the sink is matched against the data captured from the source and
echo-canceled using the speex echo canceler.
The same logic is applied to the sample rate adjustments in module-rtp-recv,
module-loopback and module-combine:
- Each time an adjustment is made, the new rate can differ at most 2‰ from the
old rate. Such a step is equal to 3.5 cents (a cent is 1/100th of a
semitone) and as 5 cents is generally considered the smallest observable
difference in pitch, this results in inaudible adjustments.
- The sample rate of the stream can only differ from the rate of the
corresponding sink by 25%. As these adjustments are meant to account for
very small clock drifts, any large deviation from the base rate suggests
something is seriously wrong.
- If the calculated rate is within 20Hz of the base rate, set it to the base
rate. This saves CPU because no resampling is necessary.
This is needed to better support out of tree builds (including
distcheck) and to ensure the necessary folders are created in the
build tree on configure and also works around an intl-tools bug
(https://bugs.launchpad.net/intltool/+bug/605826)
The Makefile.am's used are minimal (and in some cases completely
blank). At present they do not include anything interesting
with the majority of the real work still done by the monolitic
src/Makefile.am
It may make sense to start splitting out src/Makefile.am into
smaller chunks but this commit makes the minimum changes to address
the issues that result from using make distcheck and other out of
tree builds.
Note: This 'breaks' the ability to type make in e.g. the src/modules
folder and have all of PA rebuilt accordingly (this is because the
static Makefiles previously present just did a "make -C ..") which
was purportedly for use in emacs. But I'm sure there will be a better
and more robust way to configure emacs to do your builds properly if
this behaviour is still desirable.
There is a call to this function where 'skip' variable is NULL. Looks
like this code doesn't get hit very often, probably because a suitable
default sink can be found to move the stream to. However, if we can't
move to the default sink and skip is NULL, there will be a segfault.
BugLink: https://launchpad.net/bugs/680810
Some laptops have 'Internal Mic 1' exposed as an 'Input Source', e.g., Dell
XPSM 1530, so handle these, too.
Signed-off-by: Daniel T Chen <crimsun@ubuntu.com>
Earlier, if slave sinks were unlinked in non-automatic mode, their
re-appearance was disregarded. Now they are added back to the list of outputs.
Signed-off-by: Antti-Ville Jansson <antti-ville.jansson@digia.com>
Reviewed-by: Tanu Kaskinen <tanu.kaskinen@digia.com>
The PCM handle is already opened with the SND_PCM_NONBLOCK flag.
This additional call is useless.
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@intel.com>
If the user specifically removes the device element from the stream
restore rule, we have to clear the save_sink/save_source flag of the
stream. This means that other stream routing systems
(e.g. module-device-manager) can take over routing for this
stream. In order to facilitate the reapplication of other routing
rules, we fire a stream change event. Arguably the stream itself
has not changed, but the rules governing its routing have, so
I feel this is justified.
Make new defines for the smoother window size and adjust time constants instead
of reusing some unrelated constant.
Increase the smoother window size even more because the bigger it is, the
better. Since we have a 200ms max update interval and the max smoother history
is 64 entries, 10seconds is a good default.
Decrease the smoother adjust time to 1 second. The previous value of 4 seconds
was too much to adapt quickly after a resume.
Use snd_pcm_avail_delay() in pa_alsa_safe_delay() so that we can check the delay
value against the avail value and patch it up when it looks invalid. Only do
this for capture.
Move the code to start the capture and the smoother closer together to improve
smoother accuracy.
Rework things to look more like the alsa sink where the device is started in
only one place.
Currently when rewinding alsa, a fixed value of 256 bytes is used,
which represents 1.33ms @ 48kHz (2ch, 16bit). This is typically fine
and due to DMA constraints we would not want to rewind less than this.
However with more demanding sample specs, (e.g. 8ch 192kHz 32bit)
256 bytes is likely not sufficient, so calculate what 1.33ms would
be and use which ever value is bigger.
Discussed with David Henningsson and Pierre-Louis Bossart here:
http://thread.gmane.org/gmane.comp.audio.pulseaudio.general/7286
