(CA) Fixed issue where you can not stream audio in Windows 8. There a…

…re still some issues however. We do not generate data on silence... so your filter graph will not receive data and thus will pause A/V until more data comes down the line. A silence generator for the buffer is still needed

source_code/acam/loopback-capture.cpp

@@ -381,8 +381,11 @@ HRESULT propagateBufferOnce() {
         }
 
         if (0 == nNextPacketSize) {
-            // no data yet, we're waiting, between incoming chunks of audio. [occurs even with silence on the line--it just means no new data yet]
+            // (CA) - this condition appears on Win8 when there is "silence" in the audio stream 
+			// (CA) - we don't appear to hit this condition in Win7 .. so I have eliminated the logic in here for right now...
 
+			/*
+			// no data yet, we're waiting, between incoming chunks of audio. [occurs even with silence on the line--it just means no new data yet]
 			DWORD millis_to_fill = (DWORD) (1.0/SECOND_FRACTIONS_TO_GRAB*1000); // truncate is ok :) -- 1s
 			assert(millis_to_fill > 1); // sanity
 			DWORD current_time = timeGetTime();
@@ -399,6 +402,8 @@ HRESULT propagateBufferOnce() {
 			  Sleep(1); // doesn't seem to hurt cpu--"sleep x ms"
 			  continue;
 			}
+			*/
+			continue;
         } else {
 		  gotAnyAtAll = TRUE;
 		  totalSuccessFullyread++;
@@ -437,7 +442,11 @@ HRESULT propagateBufferOnce() {
 			  // we'll let fillbuffer set bFirstPacket = false; since it uses it to know if the next packet should restart, etc.
 			} else if (bFirstPacket && AUDCLNT_BUFFERFLAGS_DATA_DISCONTINUITY == dwFlags) {
 				ShowOutput("Probably spurious glitch reported on first packet, or two discontinuity errors occurred before it read from the cached buffer\n");
-				bFirstPacket = true; // won't hurt, even if it is a real first packet :)
+
+				// (CA) - we take this out as it causes problems in Win8 and does not seem to cause issues for win7
+				//bFirstPacket = true; // won't hurt, even if it is a real first packet :)
+
+
 				// LODO it should probably clear the buffers if it ever gets discontinuity
 				// or "let" it clear the buffers then send the new data on
 				// as we have any left-over data that will be assigned a wrong timestamp
@@ -451,26 +460,34 @@ HRESULT propagateBufferOnce() {
 				  pAudioCaptureClient->Release();		  
 				  pAudioClient->Release();            
 				  return E_UNEXPECTED;*/
-				  bFirstPacket = true;
+
+				  // (CA) - we take this out as it causes problems in Win8 and does not seem to cause issues for win7
+				  //bFirstPacket = true;
 			} else if (AUDCLNT_BUFFERFLAGS_SILENT == dwFlags) {
      		  // ShowOutput("IAudioCaptureClient::silence (just) from GetBuffer after %u frames\n", pnFrames);
 			  // expected if there's silence (i.e. nothing playing), since we now include the "silence generator" work-around...
 			} else {
 			  // probably silence + discontinuity
      		  ShowOutput("IAudioCaptureClient::unknown discontinuity GetBuffer set flags to 0x%08x after %u frames\n", dwFlags, pnFrames);
-			  bFirstPacket = true; // probably is some type of discontinuity :P
+
+			   // (CA) - we take this out as it causes problems in Win8 and does not seem to cause issues for win7
+			  //bFirstPacket = true; // probably is some type of discontinuity :P
 			}
 
 			if(bFirstPacket)
 				totalBlips++;
 
 			if (0 == nNumFramesToRead) {
+
+				// (CA) - This condition happens in Win8 with silence on the line... Again we don't seem to fall into this condition on Win7
+				/*
 				ShowOutput("death failure: IAudioCaptureClient::GetBuffer said to read 0 frames after %u frames\n", pnFrames);
 				pAudioClient->Stop();
 				AvRevertMmThreadCharacteristics(hTask);
 				pAudioCaptureClient->Release();
 				pAudioClient->Release();
-				return E_UNEXPECTED;            
+				return E_UNEXPECTED;  
+				*/
 			}
 
 			pnFrames += nNumFramesToRead; // increment total count...