My first post here; I think this is the right group to ask this
question; if not, or if there are other relevant groups, please feel
free to direct me to them.

Background:  I have a C++ WinXP speech recognition application that
takes spoken input from a mike.   It must interact with different people
at different  times, and these people's voices may have very different
levels of amplitude.  One person's speech might cause clipping, for
example, while another's may be perfectly normal, or even too soft.  It
is thus not possible to select a fixed recording volume that is suitable
for all.

My query: How can my application change the machine's recording volume
at will?  For example, I'd like the app to be able to turn the volume
down if it detects clipping.   I know there is a setWaveOutVolume API,
but I have also learned from various searches that there is no
corresponding setWaveInVolume API.   There *must* be a solution somehow.
   I'd prefer to have it be a simple one, if possible.  ;)

Just to make matters worse, we also use an Andrea USB SoundPod adapter,
as well as direct connection the standard mike jack.   Is there a
solution that would work for both? (probably this is a dumb question)

    thanks,
      Dave

On Jan 29, 3:38 am, Dave Stallard <stall...@nospam.net> wrote:

I think the only way is to get involved with windows Mixer API. I
found this a seriously brain-tiring activity <g>.

My first step was (and possibly yours would be) to get a document
entitled "Mixing it up with the Mixer" by Steven Roman. This _was_
with MSDN (Microsoft Developers Network) but I think that later MSDN
issues have omitted it.

Then having understood mixers, you would search for a source mixer-
line of type SRC_MICROPHONE and then look for a control of type VOLUME
on that line.

I have an MSWord copy of "Mixing it up with the Mixer" and also a
small app which enumerates all mixer lines and displays their
controls. I'll send you a copy if you email me. However while the app
would tell you what is there, the source is in Delphi, but might be
helpful to you.

Alan Lloyd

On 29 Jan 2007 00:05:07 -0800, "alangll...@aol.com"

I'll second that suggestion to get "Mixing it up with the mixer".
I also found it very useful to download Mixer Browser (you'll
have to Google for it), which shows you all the controls and
their line IDs, etc in a neat, logical window layout.

The only other way out of this would be to use a sleazy
trick:  Set the recording volume manually using the
mixer sliders such that nobody ever goes into clipping.
That will mean that many people will come through
too soft, and you will need to boost their signals in
software.  That's a bad idea in general since it will
also boost internal preamp noise, and may run into
quantization problems if carried too far.  But for
many applications the acoustic background noise
will be higher than the internal noise anyway, so
this trick won't be a problem.  

Consider that if you have a 16-bit system you can
probably afford to have the normal "soft" voices
48 dB down, where they will be equivalent to an 8-bit
system (which is usually OK for voice).  You can then
adjust the soft ones up (or the loud ones down) as
needed for the final mix.

Of course, whichever approach you choose, you
will need to be able to determine the peak and
average levels of the voices in software, either to
change the mixer or to scale the signals.

Best regards,

Bob Masta
dqatechATdaqartaDOTcom

            D A Q A R T A
Data AcQuisition And Real-Time Analysis
           www.daqarta.com
Scope, Spectrum, Spectrogram, Signal Generator
    Science with your sound card!

Reading what you described it all depends on how you design your
application.

If you receive the audio streams independently and play them independently
(and simultaneously) then the Mixer API will not work for you. Because
though you can pass the waveout of which waveout created still the volume
change will affect all the streams being played.
I would probably choose this one
- if you are having separate input streams from eachone then convert each
stream to a PCM format.
- Process the PCM stream to determine the maximum/minimum levels and adjust
them accordingly.
Now that each stream is OK play them separately and simultaneously or use
some algorithm to mix them and play them.

Generally however if an input stream is too low or high then you'd better
adjust the microphone level (wavein) of the producer. Then you can filter or
do whatever processing needed.

To set the microphone volume use this

function SetMicrophoneVolume(ADeviceID: Integer; bValue: Word; AOpenFlag:
Cardinal): Boolean;
var
  hMix: HMIXER;
  mxlc: MIXERLINECONTROLS;
  mxcd: TMIXERCONTROLDETAILS;
  vol: TMIXERCONTROLDETAILS_UNSIGNED;
  mxc: MIXERCONTROL;
  mxl: TMixerLine;
  intRet: Integer;
  nMixerDevs: Integer;
  cConnections: DWord;
  j: DWORD;
begin
  // Check if Mixer is available
  Result := True;
  nMixerDevs := mixerGetNumDevs();
  if (nMixerDevs < 1) then
  begin
    Result := False;
    exit;
  end;

  // open the mixer
  intRet := mixerOpen(@hMix, ADeviceID, 0, 0, AOpenFlag);
  if intRet = MMSYSERR_NOERROR then
  begin
    mxl.dwComponentType := MIXERLINE_COMPONENTTYPE_DST_WAVEIN;
    mxl.cbStruct := SizeOf(mxl);

    // get line info
    intRet := mixerGetLineInfo(hMix, @mxl,
MIXER_GETLINEINFOF_COMPONENTTYPE);

    if intRet = MMSYSERR_NOERROR then
    begin

      cConnections := mxl.cConnections;

      for j := 0 to cConnections - 1 do
      begin
        mxl.dwSource := j;
        mixerGetLineInfo(hMix, @mxl, MIXER_GETLINEINFOF_SOURCE);
        if (MIXERLINE_COMPONENTTYPE_SRC_MICROPHONE = mxl.dwComponentType)
        then break;
      end;

      ZeroMemory(@mxlc, SizeOf(mxlc));
      mxlc.cbStruct := SizeOf(mxlc);
      mxlc.dwLineID := mxl.dwLineID;
      mxlc.dwControlType := MIXERCONTROL_CONTROLTYPE_VOLUME;
      mxlc.cControls := 1;
      mxlc.cbmxctrl := SizeOf(mxc);

      mxlc.pamxctrl := @mxc;
      intRet := mixerGetLineControls(hMix, @mxlc,
MIXER_GETLINECONTROLSF_ONEBYTYPE);

      if intRet = MMSYSERR_NOERROR then
      begin
        ZeroMemory(@mxcd, SizeOf(mxcd));
        mxcd.dwControlID := mxc.dwControlID;
        mxcd.cbStruct := SizeOf(mxcd);
        mxcd.cMultipleItems := 0;
        mxcd.cbDetails := SizeOf(Vol);
        mxcd.paDetails := @vol;
        mxcd.cChannels := 1;

        vol.dwValue := bValue;

        intRet := mixerSetControlDetails(hMix, @mxcd,
MIXER_SETCONTROLDETAILSF_VALUE);
        if intRet <> MMSYSERR_NOERROR then
          Result := False;
      end
      else
        Result := False;
    end
    else Result := False;
    mixerClose(hMix);
  end;
end;

Where if you have the wavein handle call the function thisway:
SetMicrophoneVolume(waveinhadle, micvolume; MIXER_OBJECTF_HWAVEIN)

If you don't have the wavein handle you could call the function this way
SetMicrophoneVolume(SoundDeviceHandle, micvolume; MIXER_OBJECTF_WAVEIN)

The function returns True if there is a Microphone level control else
returns False

Hope this helps

news:8Pednc1ca8CB8SDYnZ2dnUVZ_o6gnZ2d@comcast.com...