SPA_MEMBER is misleading, all we're doing here is pointer+offset and a
type-casting the result. Rename to SPA_PTROFF which is more expressive (and
has the same number of characters so we don't need to re-indent).
When allocating memory, just export the mmaped memory and mark the
memory as MemPtr, not MemFd. Set the fd to -1.
Otherwise we might send this fd to the client, making the complete
device accessible to the client.
PipeWire does not give DMABUF anymore to clients unless they
explicitly ask for it. This results in v4l2 sending an fd and
the mmapoffset + size fields in the buffer.
GStreamer does not handle the mmapoffset in the fd allocator and
thus runs into mmap problems when using the memory.
Enable DMABuf explicitly to work around the problem until
GStreamer is fixed.
Fixes#1115
When a device as added in dbus but we already knew about the
device, don't assert but reuse the device we already have.
Seems to happen when resuming from suspend.
See rhbz#1948776
When we have a soft Mute or Volume, use the soft volume.
When we get a volume update with only a channel Mute/Volume, use
the channel volumes.
See #1140
Only run the lowpass filter on the LFE channel when we are upmixing
and there is a valid cutoff frequency defined.
Otherwise we might filter away a valid LFE channel.
Fixes: rhbz#1941366
Set source user data for all dbus sources and set a destroy notify
when removed.
Remove the dbus user data to remove the source user data.
Clean up remaining sources when destoying a connection
Clean up remaining connections when freeing the dbus plugins.
Fixes#1114
Support creating Struct as well. When an object property is marked as
Struct, use [] to start the struct and copy each item with its type into
the struct:
pw-cli s 0 Route '{ info = [ "foo": 0.1 "bar": null [ 0.1 10 "hat" ]] }'
Make a new softVolume property that contains only the soft volume
to apply.
In the case of HW/SW volume, we pass the real volume in the
channelVolume and the leftover volume in softVolume. We don't
use the monitorVolume for this anymore because it is a completely
separate volume handled by the merger node.
This way, channelVolume always represents the effective volume
set on routes, channelmix and merger and only the softVolume (when
available) is applied as software volume by channelmix.
This makes things map a bit better to what is actually happening with
the real volume and leftover software volumes after applying the
hardware volumes in the device.
With this change, the volume on the monitor is not affected by the
sink volume anymore and we can use the monitorVolume for this later.
This also means that the monitor volume in pavucontrol of the sinks
does not change when the sink volume changes. PulseAudio is inconsistent
here: If the volume is HW, the monitor volume is not affected, if the
volume is SW, it is. In PipeWire there is an option in merger to
let the volume affect the monitor with monitor.channel-volumes = true.
To set the io on the mixer ports, we need to use the same id that
was used to add the port, not the id we use to identify the mixer
structure.
Fixes#759
FreeBSD ports tree does not have alsa-lib >=1.1.7 yet,
build the newer one locally in CI until a newer version
of alsa-lib lands to the FreeBSD ports tree.
This reverts commit 09106151d3.
We need the actual name and description of the mapping, not the
profile or else we end up with the full profile description in the
node description instead of the part of the profile that applies to
the device.
Only implement the null echo cancel implementation for now.
And skeleton webrtc echo cancel implementation
It uses 4 streams arranged as:
input ---+---> source
^
|
sink ---+---> output
The output of the source is the filtered input of the input stream
(linked a master source) based on the data going from sink to
the output (linked to a master sink).
All streams are arranged in the same group so that the echo canceler
does not have to deal with clock drift. For echo cancelers that can
handle clock drift we might want place the source and sink chains
in different groups.
