RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-07-27 11:47:46 +00:00
Code Issues Activity Actions

LibAudio: Add an adjustable buffer size to FlacLoader

Browse source 
This makes it easier to fine-tune the optimal input buffer size.
This commit is contained in:
kleines Filmröllchen 2021-12-16 23:18:49 +01:00 • committed by Brian Gianforcaro
parent eb1f00a940
commit 9fa3aa84e1
2 changed files with 16 additions and 13 deletions
Download patch file Download diff file Expand all files Collapse all files

15
Userland/Libraries/LibAudio/FlacLoader.h
View file

@ -19,15 +19,22 @@
namespace Audio {
class FlacInputStream : public Variant<Buffered<Core::InputFileStream>, InputMemoryStream> {
// Experimentally determined to be a decent buffer size on i686:
// 4K (the default) is slightly worse, and 64K is much worse.
// At sufficiently large buffer sizes, the advantage of infrequent read() calls is outweighed by the memmove() overhead.
// There was no intensive fine-tuning done to determine this value, so improvements may definitely be possible.
constexpr size_t FLAC_BUFFER_SIZE = 8 * KiB;
template<size_t Size = FLAC_BUFFER_SIZE>
class FlacInputStream : public Variant<Buffered<Core::InputFileStream, Size>, InputMemoryStream> {
public:
using Variant<Buffered<Core::InputFileStream>, InputMemoryStream>::Variant;
using Variant<Buffered<Core::InputFileStream, Size>, InputMemoryStream>::Variant;
bool seek(size_t pos)
{
return this->visit(
[&](Buffered<Core::InputFileStream>& buffered) {
[&](Buffered<Core::InputFileStream, Size>& buffered) {
// Discard the buffer, then seek normally.
if (!buffered.discard_or_error(buffered.buffered()))
return false;
@ -142,7 +149,7 @@ private:
// keep track of the start of the data in the FLAC stream to seek back more easily
u64 m_data_start_location { 0 };
OwnPtr<FlacInputStream> m_stream;
OwnPtr<FlacInputStream<FLAC_BUFFER_SIZE>> m_stream;
Optional<FlacFrameHeader> m_current_frame;
Vector<Sample> m_current_frame_data;
u64 m_current_sample_or_frame { 0 };