Original commit message from CVS:
* gst/filter/gstbpwsinc.c: (gst_bpwsinc_class_init),
(bpwsinc_set_property), (bpwsinc_get_property):
* gst/filter/gstbpwsinc.h:
* gst/filter/gstlpwsinc.c: (gst_lpwsinc_class_init),
(gst_lpwsinc_init), (lpwsinc_build_kernel), (lpwsinc_set_property),
(lpwsinc_get_property):
* gst/filter/gstlpwsinc.h:
* tests/check/elements/lpwsinc.c: (GST_START_TEST):
Use generator macros for the process functions for the different
sample types, add lower upper boundaries for the GObject properties
so automatically generated UIs can use sliders and change frequency
properties to floats to save a bit of memory, even ints would in
theory be enough. Also rename frequency to cutoff for consistency
reasons.
* docs/plugins/gst-plugins-bad-plugins.args:
* docs/plugins/gst-plugins-bad-plugins.signals:
* docs/plugins/inspect/plugin-gstrtpmanager.xml:
Regenerated for the above changes.
Original commit message from CVS:
* gst/filter/gstbpwsinc.c: (gst_bpwsinc_class_init),
(gst_bpwsinc_init), (process_32), (process_64),
(bpwsinc_build_kernel), (bpwsinc_push_residue),
(bpwsinc_transform), (bpwsinc_start), (bpwsinc_query),
(bpwsinc_query_type), (bpwsinc_event), (bpwsinc_set_property):
* gst/filter/gstbpwsinc.h:
Implement latency query and only forward those samples downstream
that actually contain the data we want, i.e. drop kernel_length/2
in the beginning and append kernel_length/2 (created by convolving
the filter kernel with zeroes) to the end.
* tests/check/elements/bpwsinc.c: (GST_START_TEST):
Adjust the unit test for this slightly changed behaviour.
* gst/filter/gstlpwsinc.c: (lpwsinc_build_kernel):
Reset residue length only when actually creating a residue.
Original commit message from CVS:
* gst/filter/gstlpwsinc.c: (gst_lpwsinc_class_init),
(gst_lpwsinc_init), (process_32), (process_64),
(lpwsinc_build_kernel), (lpwsinc_push_residue),
(lpwsinc_transform), (lpwsinc_start), (lpwsinc_query),
(lpwsinc_query_type), (lpwsinc_event), (lpwsinc_set_property):
* gst/filter/gstlpwsinc.h:
Implement latency query and only forward those samples downstream
that actually contain the data we want, i.e. drop kernel_length/2
in the beginning and append kernel_length/2 (created by convolving
the filter kernel with zeroes) to the end.
* tests/check/elements/lpwsinc.c: (GST_START_TEST):
Adjust the unit test for this slightly changed behaviour.
Original commit message from CVS:
* gst/filter/gstbpwsinc.c: (bpwsinc_build_kernel):
* gst/filter/gstlpwsinc.c: (lpwsinc_build_kernel):
Improve debugging a bit.
Original commit message from CVS:
* gst/filter/gstbpwsinc.c: (gst_bpwsinc_class_init),
(bpwsinc_start):
* gst/filter/gstlpwsinc.c: (gst_lpwsinc_class_init),
(lpwsinc_start):
Reset the residue in BaseTransform::start to get a clean residue
on stream changes.
Original commit message from CVS:
* gst/filter/gstbpwsinc.c: (process_32), (process_64):
* gst/filter/gstlpwsinc.c: (process_32), (process_64):
Fix processing with buffer sizes that are larger than the filter
kernel size.
Original commit message from CVS:
* gst/filter/gstbpwsinc.c: (bpwsinc_build_kernel):
Fix a segfault with more than one channel and don't rebuild
the kernel & residue with every buffer.
Original commit message from CVS:
* gst/filter/gstbpwsinc.c: (gst_bpwsinc_mode_get_type),
(gst_bpwsinc_window_get_type), (gst_bpwsinc_class_init),
(gst_bpwsinc_init), (bpwsinc_build_kernel), (bpwsinc_set_property),
(bpwsinc_get_property):
* gst/filter/gstbpwsinc.h:
Add support for a bandreject mode and allow specifying the window
function that should be used.
* gst/filter/gstlpwsinc.c:
And another small formatting fix.
Original commit message from CVS:
* gst/filter/gstbpwsinc.c: (gst_bpwsinc_class_init),
(gst_bpwsinc_init), (process_32), (process_64),
(bpwsinc_build_kernel), (bpwsinc_setup), (bpwsinc_get_unit_size),
(bpwsinc_transform), (bpwsinc_set_property),
(bpwsinc_get_property):
* gst/filter/gstbpwsinc.h:
Apply the same changes to the bandpass filter:
- Support double input
- Fix processing for input with >1 channels
- Specify frequency in Hz
- Specify actual filter kernel length
- Use transform instead of transform_ip as we're working
out of place anyway
- Factor out filter kernel generation and update the filter
kernel when the properties are set
Fix bandpass filter kernel generation to actually generate
a bandpass filter by creating a highpass instead of a second
lowpass.
* gst/filter/gstlpwsinc.c: (gst_lpwsinc_class_init):
Small formatting fix.
Original commit message from CVS:
* gst/filter/gstlpwsinc.c: (gst_lpwsinc_class_init),
(gst_lpwsinc_init), (process_32), (process_64),
(lpwsinc_build_kernel), (lpwsinc_set_property),
(lpwsinc_get_property):
* gst/filter/gstlpwsinc.h:
Specify the actual filter length instead of a weird
2N+1. Setting the property will round to the next odd number.
Also remove now obsolete FIXMEs.
Original commit message from CVS:
* gst/filter/gstlpwsinc.c: (gst_lpwsinc_window_get_type),
(gst_lpwsinc_class_init), (gst_lpwsinc_init),
(lpwsinc_build_kernel), (lpwsinc_set_property),
(lpwsinc_get_property):
* gst/filter/gstlpwsinc.h:
Allow choosing between hamming and blackman window. The blackman
window provides a better stopband attenuation but a bit slower
rolloff.
Original commit message from CVS:
* gst/filter/gstlpwsinc.c: (process_32), (process_64),
(lpwsinc_build_kernel):
Fix processing if the input has more than one channel.
Original commit message from CVS:
* gst/filter/gstbpwsinc.c: (gst_bpwsinc_dispose),
(gst_bpwsinc_init), (bpwsinc_setup), (bpwsinc_transform_ip),
(bpwsinc_set_property), (bpwsinc_get_property):
"this" is a C++ keyword, use "self" instead.
Add TODOs and FIXMEs and remove two wrong FIXMEs.
* gst/filter/gstlpwsinc.c:
Add FIXMEs and a new TODO.
Original commit message from CVS:
* gst/filter/gstlpwsinc.c: (gst_lpwsinc_dispose),
(gst_lpwsinc_class_init), (gst_lpwsinc_init), (process_32),
(process_64), (lpwsinc_build_kernel), (lpwsinc_setup),
(lpwsinc_get_unit_size), (lpwsinc_transform),
(lpwsinc_set_property), (lpwsinc_get_property):
* gst/filter/gstlpwsinc.h:
Add double support, replace "this" with "self" as the former
is a C++ keyword.
Implement the frequency property in Hz instead of fraction
of sampling frequency.
Remove some unecessary FIXMEs and add some TODOs, add some
required locking and refactor the kernel generation into a
separate function that is also called when the properties
change now.
And use BaseTransform::transform instead of transform_ip
as the convolution is done out of place anyway. Should
be done in place later.
Original commit message from CVS:
* gst/filter/Makefile.am:
* gst/filter/gstbpwsinc.c: (gst_bpwsinc_dispose),
(gst_bpwsinc_base_init), (gst_bpwsinc_class_init),
(gst_bpwsinc_init), (bpwsinc_setup):
* gst/filter/gstbpwsinc.h:
* gst/filter/gstlpwsinc.c: (gst_lpwsinc_dispose),
(gst_lpwsinc_base_init), (gst_lpwsinc_class_init),
(gst_lpwsinc_init), (lpwsinc_setup):
* gst/filter/gstlpwsinc.h:
Use GstAudioFilter as base class and don't leak the memory
of the filter kernel and residue.
Original commit message from CVS:
Remove all config.h includes from header files, add it to each source file and remove duplicate config.h includes from several source files
Original commit message from CVS:
New mimetypes gone into effect today - this commit changes all old mimetypes over to the new mimetypes spec as described in the previous commit's document. Note: some plugins will break, some pipelines will break, expect HEAD to be broken or at least not 100% working for a few days, but don't forget to report bugs
Original commit message from CVS:
compatibility fix for new GST_DEBUG stuff.
Includes fixes for missing includes for config.h and unistd.h
I only ensured for plugins I can build that they work, so if some of them are still broken, you gotta fix them yourselves unfortunately.
Original commit message from CVS:
another batch of connect->link fixes
please let me know about issues
and please refrain of making them yourself, so that I don't spend double
the time resolving conflicts
