diff --git a/lib-src/portaudio-v19/src/hostapi/dsound/pa_win_ds.c b/lib-src/portaudio-v19/src/hostapi/dsound/pa_win_ds.c
index d5f463660..af18e884e 100644
--- a/lib-src/portaudio-v19/src/hostapi/dsound/pa_win_ds.c
+++ b/lib-src/portaudio-v19/src/hostapi/dsound/pa_win_ds.c
@@ -2046,9 +2046,18 @@ static PaError OpenStream( struct PaUtilHostApiRepresentation *hostApi,
     if( inputParameters )
     {
         /* IMPLEMENT ME - establish which  host formats are available */
+        /* JKC: CAN'T IMPLEMENT.  DirectSound does not have a way */
+        /* to interrogate for formats */
         PaSampleFormat nativeInputFormats = paInt16;
         /* PaSampleFormat nativeFormats = paUInt8 | paInt16 | paInt24 | paInt32 | paFloat32; */
 
+        /* July 2016 (Carsten and Uwe)
+         * http://bugzilla.audacityteam.org/show_bug.cgi?id=193
+         * Now we may over ride the paInt16.  
+         */
+        if (userData && *((int*)userData) == 24)
+            nativeInputFormats = paInt24;
+
         hostInputSampleFormat =
             PaUtil_SelectClosestAvailableFormat( nativeInputFormats, inputParameters->sampleFormat );
     }
diff --git a/src/AudioIO.cpp b/src/AudioIO.cpp
index ea88831be..b11efc8ff 100644
--- a/src/AudioIO.cpp
+++ b/src/AudioIO.cpp
@@ -1448,10 +1448,19 @@ bool AudioIO::StartPortAudioStream(double sampleRate,
    // rate is suggested, but we may get something else if it isn't supported
    mRate = GetBestRate(numCaptureChannels > 0, numPlaybackChannels > 0, sampleRate);
 
+   // July 2016 (Carsten and Uwe)
+   // BUG 193: Tell PortAudio sound card will handle 24 bit (under DirectSound) using 
+   // userData.
+   int captureFormat_saved = captureFormat;
    // Special case: Our 24-bit sample format is different from PortAudio's
    // 3-byte packed format. So just make PortAudio return float samples,
    // since we need float values anyway to apply the gain.
-   // ANSWER-ME: So we *never* actually handle 24-bit?! This causes mCapture to be set to floatSample below.
+   // ANSWER-ME: So we *never* actually handle 24-bit?! This causes mCapture to 
+   // be set to floatSample below.
+   // JKC: YES that's right.  Internally Audacity uses float, and float has space for
+   // 24 bits as well as exponent.  Actual 24 bit would require packing and
+   // unpacking unaligned bytes and would be inefficient.
+   // ANSWER ME: is floatSample 64 bit on 64 bit machines?
    if (captureFormat == int24Sample)
       captureFormat = floatSample;
 
@@ -1538,12 +1547,19 @@ bool AudioIO::StartPortAudioStream(double sampleRate,
    float oldRecordVolume = Px_GetInputVolume(mPortMixer);
 #endif
 #endif
+
+   // July 2016 (Carsten and Uwe)
+   // BUG 193: Possibly tell portAudio to use 24 bit with DirectSound. 
+   int  userData = 24;
+   int* lpUserData = (captureFormat_saved == int24Sample) ? &userData : NULL;
+
    mLastPaError = Pa_OpenStream( &mPortStreamV19,
                                  useCapture ? &captureParameters : NULL,
                                  usePlayback ? &playbackParameters : NULL,
                                  mRate, paFramesPerBufferUnspecified,
                                  paNoFlag,
-                                 audacityAudioCallback, NULL );
+                                 audacityAudioCallback, lpUserData );
+
 
 #if USE_PORTMIXER
 #ifdef __WXMSW__