Created attachment 89612 [details] [review]
patch against 0.10.3

First attempt at writing GStreamer code.

Ok, I suck. The 24-bit code is completely wrong.

Created attachment 89643 [details] [review]
patch with working 24-bit support

The caps in the pad templates have a small bug, you only want width==depth caps in there. In your patch width==32 and depth==16 would be allowed for example.
Other than that this looks fine from a fast look...

Oh, and the negative minimum values are always in absolute one larger than the positive maximums, i.e. 
+#define VOLUME_MAX_INT32             2147483647
+#define VOLUME_MIN_INT32             -2147483646
+#define VOLUME_MAX_INT24             8388607
+#define VOLUME_MIN_INT24             -8388606

should be

+#define VOLUME_MAX_INT32             2147483647
+#define VOLUME_MIN_INT32             -2147483648
+#define VOLUME_MAX_INT24             8388607
+#define VOLUME_MIN_INT24             -8388608

Also I'm not sure about the bitshift stuff in the very beginning of the patch, would be better if someone who understands it looks at this too :)

#define VOLUME_UNITY_INT24 2097152   /* (2^(24-3)) */
#define VOLUME_UNITY_INT32 536870912 /* (2^(32-3)) */

Committed with those changes and int8 support.

2007-08-23  Sebastian Dröge  <slomo@circular-chaos.org>

	Based on a patch by: Davyd <davyd at madeley dot id dot au>

	* gst/volume/gstvolume.c: (volume_choose_func),
	(volume_update_real_volume), (gst_volume_set_volume),
	(gst_volume_init), (volume_process_int32),
	(volume_process_int32_clamp), (volume_process_int24),
	(volume_process_int24_clamp), (volume_process_int16),
	(volume_process_int16_clamp), (volume_process_int8),
	(volume_process_int8_clamp), (volume_update_volume), (plugin_init):
	* gst/volume/gstvolume.h:
	Add support for int32, int24 and int8 to the volume element.
	Fixes #445529.