The claim made is within the realm of plausibility. A memcpy call in a playback loop was replaced with an optimized version, and this made some harsh audio artifact ("edge") go away.
The poster showed the code and what was done to it.
The code does strange things, though, like wait for a Win32 event on every iteration. And not only one time; there are multiple calls like this:
WaitForSingleObject(hNeedDataEvent, INFINITE);
Is that an auto-reset event that some other thread has to be constantly banging on to keep the loop going? its odd since the code has supposedly buffered the entire file of samples; so you'd think it would just be iterating over that buffer in some chunks and shoving them into some pipe/device. There is also a puzzling ZeroMemory call.
The poster showed the code and what was done to it.
The code does strange things, though, like wait for a Win32 event on every iteration. And not only one time; there are multiple calls like this:
WaitForSingleObject(hNeedDataEvent, INFINITE);
Is that an auto-reset event that some other thread has to be constantly banging on to keep the loop going? its odd since the code has supposedly buffered the entire file of samples; so you'd think it would just be iterating over that buffer in some chunks and shoving them into some pipe/device. There is also a puzzling ZeroMemory call.