Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(94)

Unified Diff: media/base/audio_hardware_config.cc

Issue 235723003: Use larger buffer sizes for lower power on Linux. (Closed) Base URL: svn://svn.chromium.org/chrome/trunk/src
Patch Set: Created 6 years, 8 months ago
Use n/p to move between diff chunks; N/P to move between comments. Draft comments are only viewable by you.
Jump to:
View side-by-side diff with in-line comments
Download patch
Index: media/base/audio_hardware_config.cc
diff --git a/media/base/audio_hardware_config.cc b/media/base/audio_hardware_config.cc
index d72fce7b4e234c2f225e28b528c15900dd2582d2..7f1ebf7a13ede7ed749194139d4b76948fb2b5fb 100644
--- a/media/base/audio_hardware_config.cc
+++ b/media/base/audio_hardware_config.cc
@@ -4,16 +4,36 @@
#include "media/base/audio_hardware_config.h"
+#include <algorithm>
+
+#include "base/logging.h"
+#include "build/build_config.h"
+
using base::AutoLock;
using media::AudioParameters;
namespace media {
+// Taken from "Bit Twiddling Hacks"
+// http://graphics.stanford.edu/~seander/bithacks.html#RoundUpPowerOf2
+static uint32_t RoundUpToPowerOfTwo(uint32_t v) {
+ v--;
+ v |= v >> 1;
+ v |= v >> 2;
+ v |= v >> 4;
+ v |= v >> 8;
+ v |= v >> 16;
+ v++;
+ return v;
+}
+
AudioHardwareConfig::AudioHardwareConfig(
const AudioParameters& input_params,
const AudioParameters& output_params)
: input_params_(input_params),
output_params_(output_params) {
+ DCHECK(input_params_.IsValid());
+ DCHECK(output_params_.IsValid());
}
AudioHardwareConfig::~AudioHardwareConfig() {}
@@ -77,4 +97,29 @@ void AudioHardwareConfig::UpdateOutputConfig(
output_params_ = output_params;
}
+int AudioHardwareConfig::GetHighLatencyBufferSize() const {
+ AutoLock auto_lock(config_lock_);
+#if defined(OS_LINUX)
+ // Empirically, use the nearest higher power of two buffer size corresponding
+ // to 20 ms worth of samples. For a given sample rate, this works out to:
no longer working on chromium 2014/04/14 08:51:50 curiously, 20ms does not sound "high" latency, why
DaleCurtis 2014/04/14 18:08:18 As discussed in the meeting this morning, I choose
+ //
+ // <= 3200 : 64
+ // <= 6400 : 128
+ // <= 12800 : 256
+ // <= 25600 : 512
+ // <= 51200 : 1024
+ // <= 102400 : 2048
+ // <= 204800 : 4096
+ //
+ // On Linux, the minimum hardware buffer size is 512, so the lower calculated
+ // values are unused. OSX may have a value as low as 128. Windows is device
henrika (OOO until Aug 14) 2014/04/14 08:03:24 Should comments about OSX and Windows be within de
DaleCurtis 2014/04/14 18:08:18 I hope eventually this section will have all OS, s
+ // dependent but will generally be sample_rate() / 100.
+ const int high_latency_buffer_size =
+ RoundUpToPowerOfTwo(2 * output_params_.sample_rate() / 100);
+ return std::max(output_params_.frames_per_buffer(), high_latency_buffer_size);
+#else
+ return output_params_.frames_per_buffer();
+#endif
+}
+
} // namespace media

Powered by Google App Engine
This is Rietveld 408576698