This bug results in the first period of samples being pitched at 1.0
regardless of whether or not ma_sound_set_pitch() had been called. What
happens is that the first period is processed at a rate of 1.0, and
then the pitch is applied after the fact. Then, the next period comes
along and resamples at the pitch set by ma_sound_set_pitch() which
results in a harsh sounding glitch.
This is useful for platforms that do not have support for threading or
programs that do not want to use an extra thread for resource
management and would rather process jobs manually.
When configuring the resource manager to not use threading, the
MA_RESOURCE_MANAGER_FLAG_NO_THREADING flag must be set in the config.
This implicitly enables the MA_RESOURCE_MANAGER_FLAG_NON_BLOCKING flag
because it requires programs to manually call
ma_resource_manager_process_next_job(), and since it's assumed that
won't ever be called from another thread, you'd never want that to be
blocking.
This sets up a framework for getting the resource manager working with
Emscripten.
This is an optional callback and is used by miniaudio as a hint to help
it determine how many input frames to read at a time. Without this,
miniaudio needs to guess how many frames to read, and in certain
situations may end up overestimating. This callback is only useful for
nodes that process input and output frames at different rates, i.e.
nodes that perform resampling.
This is a data source whose backing data is an application-controlled
pointer. No data is copied. It's a way of efficiently wrapping a raw
buffer and using it as a data source.
This is useful for retrieving information about some aspect of the
node. A good example is human readable names associated with the node
and it's input and output buses. This is useful for user interfaces
where a brief description of the node such as "Low Pass Filter" can be
drawn on the screen. It's also useful for buses to be named, such as
the source/carrier and excite/modulator on a vocoder effect which would
also need to be visible on a UI.
The following flags can now be associated with nodes via the vtable:
* MA_NODE_FLAG_PASSTHROUGH
* MA_NODE_FLAG_CONTINUOUS_PROCESSING
* MA_NODE_FLAG_ALLOW_NULL_INPUT
* MA_NODE_FLAG_DIFFERENT_PROCESSING_RATES
See commit changes for a description of these flags.
* The simplified callback has been removed.
* The `globalTime` parameter has been removed from the callback.
* The order of input and output frames and counts has been swapped to
be consistent with ma_data_converter_process_pcm_frames(), etc.
Variables are marked with this annotation to make it clear that access
to the variable should be done through atomics.
I've also review the use of volatile in this commit.
Public issue https://github.com/mackron/miniaudio/issues/259
This was setting the previous pointer of newly attached nodes to NULL
instead of a pointer to the dummy head node. This then results in the
dummy head node never being updated when the node is detached which in
turn results in an uninitialized node being dereferenced.
This would try reading from an uninitialized buffer thereby resulting
in a bad audio glitch. This does two things to fix the problem:
1) When there are no input nodes attached to an input bus, nothing is
read and 0 will be returned for the frames read variable.
2) The buffer is silenced by default.
This fixes a bug with ma_engine where it would glitch in the moment
just after the engine is initialized and before a sound or group is
attached.
David Reid