using 2 microphones

[edf]

Hi
I am trying to use 2 usb microphones writting to a file (on windows 11 home 64 bits).
each of the microphones work one at a time, but when you try to use both it does not work.

Error is

as soon as use 2 usb mics   /not enough controller resources. usb
the record stop  seems ok.
on application halt
gives raised exception class 'External: ACCESS VIOLATION' with message Access violation reading from address $FFFFFFFFFFF.

The directory of the file is
20/11/2023  11:31 AM                46 rec23_11_20_11_31_03.wav
46 bytes (46 bytes)
it  wont play

The code is based on the example simple recorder. My appologies if I have not understood the use of uos.
Using 1 PlayerIndex and uos_AddIntoFile(PlayerIndex1, PChar(FileSaveOutoutName), -1, -1, -1, 4096, outformat);
code snip (InDev1 =1, InDev2=2, which are devices found by devinfo.) Both work 1 at a time.

   PlayerIndex1 := 0;
    // PlayerIndex : from 0 to what your computer can do ! (depends of ram, cpu, ...)
    // If PlayerIndex exists already, it will be overwritten...

    uos_CreatePlayer(PlayerIndex1);
    //// Create the player.
    //// PlayerIndex : from 0 to what your computer can do !
    //// If PlayerIndex exists already, it will be overwriten...

    //uos_AddIntoFileFromMem(PlayerIndex1, Pchar(edit3.Text));
    //// add Output into wav file (save record) from TMemoryStream  with default parameters
    bwav := true; // make it simple
    wantSpeakerOut := False ; // no speaker . make it simple
    if bwav then
     begin
    outformatst := '.wav';
    outformat := 0;
    end else
    begin
    outformatst := '.ogg';
    outformat := 3;
    end;
    FileSaveOutoutName := FileSaveBaseName + 'rec' +  UTF8Decode(formatdatetime('YY_MM_DD_HH_mm_ss', now)) + outformatst ;


     // saving in a file using a File-Stream:

     uos_AddIntoFile(PlayerIndex1, PChar(FileSaveOutoutName), -1, -1, -1, 4096, outformat);

 //   function uos_AddIntoFile(PlayerIndex: cint32; Filename: PChar; SampleRate: cint32;
 // Channels: cint32; SampleFormat: cint32 ; FramesCount: cint32 ; FileFormat: cint32): cint32;
// Add a Output into audio wav file with custom parameters from TFileStream
// PlayerIndex : Index of a existing Player
// FileName : filename of saved audio wav file
// SampleRate : delault : -1 (44100)
// Channels : delault : -1 (2:stereo) (1:mono, 2:stereo, ...)
// SampleFormat : default : -1 (2:Int16) (1:Int32, 2:Int16)
// FramesCount : default : -1 (= 65536)
// FileFormat : default : -1 (wav) (0:wav, 1:pcm, 2:custom, 3:ogg);

    //// add Output into wav or ogg file (save record) from TFileStream

    // saving in a Memory-Buffer:
    // SetLength(thebuffer, 0);
    // uos_AddIntoMemoryBuffer(PlayerIndex1, @thebuffer);

    // saving in a Memory-Stream:
    // if thememorystream = nil then thememorystream := tmemorystream.create;
    // uos_AddIntoMemoryStream(PlayerIndex1, (thememorystream),-1,-1,-1,-1);

    // saving in a file using a Menory-Stream:
    // uos_AddIntoFileFromMem(PlayerIndex1, Pchar(filenameEdit4.filename));
    //// add Output into wav file (save record)  with default parameters

      {$if defined(cpuarm) or defined(cpuaarch64)}  // need a lower latency
    out1index := uos_AddIntoDevOut(PlayerIndex1, -1, 0.8, -1, -1, -1, -1) ;
       {$else}
    out1index := uos_AddIntoDevOut(PlayerIndex1);
       {$endif}

    uos_outputsetenable(PlayerIndex1, out1index, wantSpeakerOut);

     // uos_AddIntoDevOut(PlayerIndex1, -1, -1, 8000, -1, -1,65536, -1);   //// add a Output into device with custom parameters
    //////////// PlayerIndex : Index of a existing Player
    //////////// Device ( -1 is default Output device )
    //////////// Latency  ( -1 is latency suggested ) )
    //////////// SampleRate : delault : -1 (44100)
    //////////// Channels : delault : -1 (2:stereo) (0: no channels, 1:mono, 2:stereo, ...)
    //////////// SampleFormat : -1 default : Int16 : (0: Float32, 1:Int32, 2:Int16)
    //////////// FramesCount : -1 default : 65536
    // ChunkCount : default : -1 (= 512)
 // these are global variables setup in ini file
 // InDev1:               Integer;
 // InDev2:               Integer;

   useInDevDefault  := true; //default
   useInDev1 := (CheckBoxChan1.Checked) ;
   useInDev2 := (CheckBoxChan2.Checked) ;
   if ( useInDev1) then   useInDevDefault := false;
   if ( useInDev2) then   useInDevDefault := false;

   // we want useInDevDefault or one or both of   useInDev1  useInDev2
   if (useInDevDefault) then
   begin
        In1IndexD := uos_AddFromDevIn(PlayerIndex1);       // default
   end;

   if (useInDev1) then
   begin
         EditCh1.Text :=  InDev1.ToString;
         In1Index1 := uos_AddFromDevIn(PlayerIndex1, InDev1,-1,-1,-1,-1,-1,-1);
   end;
     if (useInDev2) then
   begin
         EditCh2.Text :=  InDev2.ToString;
         In1Index2 := uos_AddFromDevIn(PlayerIndex1, InDev2,-1,-1,-1,-1,-1,-1);
   end;

    //// add input from mic with custom parameters
    //////////// PlayerIndex : Index of a existing Player
    //////////// Device ( -1 is default Input device )
    //////////// Latency  ( -1 is latency suggested ) )
    //////////// SampleRate : delault : -1 (44100)
    //////////// OutputIndex : OutputIndex of existing Output // -1 : all output, -2: no output, other integer : existing output)
    //////////// SampleFormat : -1 default : Int16 : (0: Float32, 1:Int32, 2:Int16)
    //////////// FramesCount : -1 default : 4096   ( > = safer, < =  better latency )

    ////////// VolLeft : Left volume
    ////////// VolRight : Right volume
    volLeft  := 1.0;
    volRight := 1.0;

    if (useInDevDefault) then
   begin
        uos_InputAddDSPVolume(PlayerIndex1, In1IndexD, 1, 1);
         uos_InputSetDSPVolume(PlayerIndex1,  In1IndexD, volLeft,volRight, True); /// Set volume
   end;
   if (useInDev1) then
   begin
        uos_InputAddDSPVolume(PlayerIndex1, In1Index1, 1, 1);
         uos_InputSetDSPVolume(PlayerIndex1,  In1Index1, volLeft,volRight, True); /// Set volume
   end;
   if (useInDev2) then
   begin
        uos_InputAddDSPVolume(PlayerIndex1, In1Index2, 1, 1);
         uos_InputSetDSPVolume(PlayerIndex1,  In1Index2, volLeft,volRight, True); /// Set volume
   end;

    ///// DSP Volume changer
    //////////// PlayerIndex : Index of a existing Player
    ////////// In1Index : InputIndex of a existing input
    ////////// VolLeft : Left volume
    ////////// VolRight : Right volume



    uos_Play(PlayerIndex1);  /////// everything is ready to play...
    ButtonStart.Enabled:= false ;  // wait till stop
    ButtonStop.Enabled:= true ;  // allow stop
end;

[fredvs]

Hello edf.

Welcome to uos forum.

Hum, I am a bit lost with your code.

Are you able to make 1 usb microphone work with uos, following the uos simplerecorder demo?

What happens when you declare 2 microphones based on the simplerecorder demo?
For example:
OutputIndex1 := uos_AddFromDevIn(0,0,-1,-1,-1,-1,-1);
OutputIndex2 := uos_AddFromDevIn(0,1,-1,-1,-1,-1,-1);

What is the result of the demo deviceinfos when you connect both mic, what device is assigned for the mics?

Sorry but I am out of computer for the moment and can not try your code.
Also i never try to record 2 mic at the same time, it should work but I never try.

Fred

[edf]

thanks Fred.
re
Are you able to make 1 usb microphone work with uos, following the uos simplerecorder demo? YES

What happens when you declare 2 microphones based on the simplerecorder demo? GIVES ERROR
 (see first post)

if you look at following code
//<start of code>
 if (useInDev1) then
   begin
         
         In1Index1 := uos_AddFromDevIn(PlayerIndex1, InDev1,-1,-1,-1,-1,-1,-1);
   end;
 if (useInDev2) then
   begin
       
         In1Index2 := uos_AddFromDevIn(PlayerIndex1, InDev2,-1,-1,-1,-1,-1,-1);
   end;

//<end of code>

example
 where
InDev1 := 1; // a valid mic (actually a USB Mic)
InDev2 := 2; // another valid mic (actually another USB Mic)
one at a time works OK.

So if useInDev1 = true and useInDev2 = false then works fine

So if useInDev1 = false and useInDev2 = true then works fine

BUT if useInDev1 = true and useInDev2 = true then  BOMBs

Just in case I have not made it clear, BOTH mics are connected into the PC.
Just adding the 2 uos_AddFromDevIn makes it BOMB.
 It gives the error and the file produced is empty no matter how long the recorder runs.
 The file is not closed when you  uos_Stop(PlayerIndex1)

thanks for looking at it.

[fredvs]

Hello.

Hum, all what you have done is ok.

Are those 2 mic equal (same mark, same model) ?

 Did you try with different parameter for uos_AddFromDevIn() ?
( Device: cint32;
  Latency: CDouble;
  SampleRate: CDouble;
  OutputIndex: cint32;
  SampleFormat: cint32;
  FramesCount : cint32;
  ChunkCount: cint32): cint32;

For example with other Latency, SampleFormat, FramesCount and ChunkCount?

> Just adding the 2 uos_AddFromDevIn makes it BOMB.

Do you get any error message from the BOMB?
[EDIT] Ok, in first post you said:
>gives raised exception class 'External: ACCESS VIOLATION' with message Access violation reading from address $FFFFFFFFFFF.

Are you able to debug the program? (I know debug could not work with audio but maybe you can get something)

May I ask you the mark and model of your usb mic (so I can see their specification)?

Fred.

[fredvs]

Re-hello.

What happen if you use 2 players, one with the first mic and the second with the other mic and play the 2 players together?

[edf]

Thanks Fred
 I wonder if I am just asking too much to use 2 usb mics. (I tried this as it was the easest way to get 2 mics). But see on as well.

I dont want you to waste your time, so i am also trying different ways.

but in case you are interested.

Just going through the faults. It is clear I was a bit vague.
So trying to make it a bit clearer.
Assume you have 2 USB mics and code is
  In1Index1 := uos_AddFromDevIn(PlayerIndex1, InDev1,-1,-1,-1,-1,-1,-1);
  In1Index2 := uos_AddFromDevIn(PlayerIndex1, InDev2,-1,-1,-1,-1,-1,-1);

there is nothing appears wrong till

uos_Play(PlayerIndex1);  /////// everything is ready to play..

at which a message from windows comes up.
Not enough USB Controller resources.
(This is not from Lazarus, or debugger. It displays in the windows notification area on [in my case] very right hand side of windows )
 
not enough USB  controller resources

The  uos_Stop(PlayerIndex1);
  ClosePlayer1;
seems to be ok.

 But the file size is zero.

and when you close the application
the debugger gives the application raised exception class 'External: ACCESS VIOLATION' with message Access violation reading from address $FFFFFFFFFFF.

I put a breakpoint on form .Destroy
and it seemed the error was somewhere in uos_free (my guess)

The usb microphones are anko [from Kmart] keycode 42959540/jlr-74449

So I tried a different way.
1 USB mic and 1 using the 3.5 mm plug.
this gave no errors using 2 at once, but sound was a mess. (not just what you would expect with 2 audio at once. Cracks and etc. some real sound). intereresting that the code faults are not there. and that file has a length.
is this just me not understanding how 2 inputs writting to one file would be ?

What my actual new approch will be.
use 2 3.5 mm connected mics.
use 3.5mm plug to 2 rca plug cables. connect 1 mic to LHS RCA plug, and the other Mic to RHS RCA plug.
So record 1 mic to LHS channel, and other to RHS Channel.

getting back to what you suggested.
a) using 2 different PlayerIndex1 and PlayerIndex2. i tried this first.
producing 2 files. one  mic would work and produced a valid file. other produced zero length file.
(it was a while ago, and i cant remember if it produced access faults.)

b) using different settings. If you could suggest some to try , happy to. (I dont understand the UOS well enough to know what to use.)

But if you think you have wasted too much time on this, donot waste any more.

thanks for your help (and patience)
Ed

[fredvs]

Hello Ed.

>So I tried a different way.
>1 USB mic and 1 using the 3.5 mm plug.
>this gave no errors using 2 at once,but sound was a mess.

Ha, that is already a excellent news.
If you get something but the sound is mess it is because of wrong parameters used.
For the input it could be from those:
  Latency: CDouble;
  SampleFormat: cint32;
  FramesCount : cint32;

Also check parameters of output uos_AddIntoDevOut() and uos_AddIntofile():
 Channels : delault : -1 (2:stereo) (1:mono, 2:stereo, ...)
 SampleFormat : default : -1 (2:Int16) (1:Int32, 2:Int16)
 FramesCount : default : -1 (= 65536)

If you get some result (even with bad noise) you will win, it is only a problem of parameters.
But I cannot say witch one (at first view it is about SampleFormat or/and Channels or/and Latency).

>But if you think you have wasted too much time on this, donot waste any more.

No, not at all!  It interest me a lot because I was on plan to create a multi track recorder.
Sadly I am outside for a few weeks without my audio-computer.

> ACCESS VIOLATION' with message Access violation reading from address $FFFFFFFFFFF

Hum, strange, I need to buy a usb mic to try, I only have try with using the 3.5 mm plug or the build in of the computer and it was ok.

Fred

[trustfm]

Hello i have the same problems using two inputs

this problem goes away only if 1 stream is allowed or by volume like this
    uos_InputAddDSPVolume(PlayerIndex1, In1Index1, 1, 1);
    uos_InputAddDSPVolume(PlayerIndex1, In1Index2, 0, 0);

or by InputSetEnable

procedure TForm1.CheckBoxSpeakersChange(Sender: TObject);
begin
  uos_InputSetEnable(PlayerIndex1, In1Index1, CheckBoxSpeakers.Checked);
end;

procedure TForm1.CheckBoxMicChange(Sender: TObject);
begin
  uos_InputSetEnable(PlayerIndex1, In1Index2, CheckBoxMic.Checked);
end;

When i try to enable both the sound oversaturates i have tried different cards same results ...

adding uos_AddFromEndlessMuted makes things worse ...
    //mut := uos_AddFromEndlessMuted(PlayerIndex1, 2,1024);
    //uos_InputSetEnable(PlayerIndex1, mut, True);  

[fredvs]

Hello.

Did you try with the uos/examples/simplerecorder demo, is it ok?

[trustfm]

yes for that demo i am talking about.
If you try to add a second inpout you get the intermittent problems

Full code below nothing special (simplerecorder with second input)

procedure TForm1.ButtonRecordClick(Sender: TObject);
var
  outformatst: string;
  outformat, numchan: integer;
begin
  if (checkbox1.Checked = True) or (checkbox2.Checked = True) then
  begin

    PlayerIndex1 := 0;
    // PlayerIndex : from 0 to what your computer can do ! (depends of ram, cpu, ...)
    // If PlayerIndex exists already, it will be overwritten...

    uos_CreatePlayer(PlayerIndex1);
    // Create the player.
    // PlayerIndex : from 0 to what your computer can do !
    // If PlayerIndex exists already, it will be overwriten...

    //uos_AddIntoFileFromMem(PlayerIndex1, Pchar(edit3.Text));
    // add Output into wav file (save record) from TMemoryStream  with default parameters

    if bwav.Checked then
    begin
      outformatst := '.wav';
      outformat := 0;
    end
    else
    begin
      outformatst := '.ogg'; //ogg opus
      outformat := 3; //     //opus 4
    end;

    edit3.Text := 'rec_' + UTF8Decode(formatdatetime('YY_MM_DD_HH_mm_ss', now)) +
      outformatst;


    //In1Index1 := uos_AddFromDevIn(PlayerIndex1,-1,0.8,44100,-1,0,-1,-1); //-1 is speaker - 18 pulse works
    In1Index1 := uos_AddFromDevIn(PlayerIndex1,  -1,0.02,-1   ,-1,2,1024,-1); //-1 is speaker - 18 pulse works
    numchan := uos_InputGetChannels(PlayerIndex1, In1Index1);
    label3.Caption:=inttostr(numchan);



    In1Index2 := uos_AddFromDevIn(PlayerIndex1,  0,0.02, -1,   -1,2,1024,-1); //0 is mic
    numchan := uos_InputGetChannels(PlayerIndex1, In1Index2);
    label8.Caption:=inttostr(numchan);

    //uos_AddFromFile(PlayerIndex0, PChar(Edit11.Text), -1, 0,

    // This for a dummy endless input, set enable to true, must be last input
    // Needed to make all input reach the end.

    //mut := uos_AddFromEndlessMuted(PlayerIndex1, 2,1024);
    //uos_InputSetEnable(PlayerIndex1, mut, True);


    // add input from mic with custom parameters
    // PlayerIndex : Index of a existing Player
    // Device ( -1 is default Input device )
    // Latency  ( -1 is latency suggested ) )
    // SampleRate : delault : -1 (44100)
    // OutputIndex : OutputIndex of existing Output // -1 : all output, -2: no output, other integer : existing output)
    // SampleFormat : -1 default : Int16 : (0: Float32, 1:Int32, 2:Int16)
    // FramesCount : -1 default : 4096   ( > = safer, < =  better latency )



    // saving in a file using a File-Stream:
    if CheckBox2.Checked then
      uos_AddIntoFile(PlayerIndex1, PChar(edit3.Text), -1, 2, -1,1024, outformat); //-1

    //   function uos_AddIntoFile(PlayerIndex: cint32; Filename: PChar; SampleRate: cint32;
    // Channels: cint32; SampleFormat: cint32 ; FramesCount: cint32 ; FileFormat: cint32): cint32;
    // Add a Output into audio wav file with custom parameters from TFileStream
    // PlayerIndex : Index of a existing Player
    // FileName : filename of saved audio wav file
    // SampleRate : delault : -1 (44100)
    // Channels : delault : -1 (2:stereo) (1:mono, 2:stereo, ...) if 2 it will force to stereo if input is mono
    // SampleFormat : default : -1 (2:Int16) (1:Int32, 2:Int16)
    // FramesCount : default : -1 (= 65536)
    // FileFormat : default : -1 (wav) (0:wav, 1:pcm, 2:custom, 3:ogg);

    //  saving in a Memory-Buffer:
    //  SetLength(thebuffer, 0);
    //  uos_AddIntoMemoryBuffer(PlayerIndex1, @thebuffer);

    // saving in a Memory-Stream:  
    // if thememorystream = nil then thememorystream := tmemorystream.create;
    // uos_AddIntoMemoryStream(PlayerIndex1, (thememorystream),-1,-1,-1,-1);

    // saving in a file using a Menory-Stream:  
    // uos_AddIntoFileFromMem(PlayerIndex1, Pchar(filenameEdit4.filename));
    // add Output into wav file (save record)  with default parameters

    {$if defined(cpuarm) or defined(cpuaarch64)}
    // need a lower latency
    //out1index := uos_AddIntoDevOut(PlayerIndex1, -1, 0.8, -1, numchan, -1, -1, -1) ;
    {$else}
    //out1index := uos_AddIntoDevOut(PlayerIndex1, -1, -1, -1, numchan, -1, -1, -1);
    //out1index := uos_AddIntoDevOut(PlayerIndex1, 2, -1, -1, -1, -1, -1, -1);
    {$endif}
   
    // function uos_AddIntoDevOut(PlayerIndex: cint32; Device: cint32; Latency: CDouble;
    // SampleRate: CDouble; Channels: cint32; SampleFormat: cint32 ;
    // FramesCount: cint32 ; ChunkCount: cint32): cint32;

    // Add a Output into Device Output
    // Device ( -1 is default device )
    // Latency  ( -1 is latency suggested )
    // SampleRate : delault : -1 (44100)
    // Channels : delault : -1 (2:stereo) (0: no channels, 1:mono, 2:stereo, ...)
    // SampleFormat : default : -1 (2:Int16) (0: Float32, 1:Int32, 2:Int16)
    // FramesCount : default : -1 (= 65536)
    // ChunkCount : default : -1 (= 512)
    //  result :  Output Index in array  -1 = error

    //uos_OutputSetEnable(PlayerIndex1, out1index, checkbox1.Checked);

    uos_InputAddDSPVolume(PlayerIndex1, In1Index1, 1, 1);
    //uos_InputSetDSPVolume(PlayerIndex1, In1Index1, TrackBar1.position / 100, TrackBar3.position / 100, True); // Set volume


    uos_InputAddDSPVolume(PlayerIndex1, In1Index2, 0, 0);
    //uos_InputSetDSPVolume(PlayerIndex1,In1Index2, 0.01,0.01, True);
    //uos_InputSetDSPVolume(PlayerIndex1, In1Index2, TrackBar1.position / 100, TrackBar3.position / 100, True); // Set volume

    // procedure to execute when stream is terminated
    //   uos_EndProc(PlayerIndex1, @ClosePlayer1);
    // Assign the procedure of object to execute at end
    // PlayerIndex : Index of a existing Player
    // ClosePlayer1 : procedure of object to execute inside the loop

    uos_Play(PlayerIndex1);  // everything is ready to play...

    ButtonRecord.Enabled := False;
    Button3.Enabled := True;
    ButtonPlay.Enabled := False;
    //CheckBox1.Enabled := False;
    CheckBox2.Enabled := False;
  end;
end;    

[fredvs]

> If you try to add a second inpout you get the intermittent problems

Not sure to understand, you have then 2 sound cards and want to add a second input in your player after adding the fist input from your first sound card, is it that?
I never try this but it should work.

[fredvs]

>    In1Index1 := uos_AddFromDevIn(PlayerIndex1,  -1,0.02,-1   ,-1,2,1024,-1); //-1 is speaker - 18 pulse works
>    numchan := uos_InputGetChannels(PlayerIndex1, In1Index1);
>    label3.Caption:=inttostr(numchan);

>    In1Index2 := uos_AddFromDevIn(PlayerIndex1,  0,0.02, -1,   -1,2,1024,-1); //0 is mic
>    numchan := uos_InputGetChannels(PlayerIndex1, In1Index2);
 >   label8.Caption:=inttostr(numchan);


Hum, the second uos_AddFromDevIn seems strange and the first should target a input device.
There is a demo devicesinfos to see what input device you have.

So it should be something like:

    In1Index1 := uos_AddFromDevIn(PlayerIndex1,  -1,0.02,-1   ,-1,2,1024,-1); //-1 should be the default device in
    numchan := uos_InputGetChannels(PlayerIndex1, In1Index1);
    label3.Caption:=inttostr(numchan);

    In1Index2 := uos_AddFromDevIn(PlayerIndex1,  deivicein2, 0.02,-1   ,-1,2,1024,-1); //devicein2  should be the default device in of second card.
    numchan := uos_InputGetChannels(PlayerIndex1, In1Index2);
   label8.Caption:=inttostr(numchan);

[trustfm]

Hello thanks for the prompt repliues!
Yes i use the device demo to find my devices

The setup is easy (this is with only one sound card)
I want to be able to record the main soundcard output + the mic
The soundcard is at -1 and the mic is at 0
Individually work perfectly . simutaneusly not

    In1Index1 := uos_AddFromDevIn(PlayerIndex1,  -1 ,0.02,-1   ,-1,2,1024,-1); //-1 is speaker
    In1Index2 := uos_AddFromDevIn(PlayerIndex1,  0 ,0.02, -1,   -1,2,1024,-1); //0 is mic

The current solution to this problem was to
1) Go to volume control
Set : analog stereo dublex
 

2)  From command line make an audio loop  
pactl load-module module-loopback latency_msec=1
To stop it running, run:
pactl unload-module module-loopback

3) Set only one input  
    In1Index1 := uos_AddFromDevIn(PlayerIndex1,  -1,0.02,-1   ,-1,2,1024,-1); //-1 is speaker

This works but the mic loopback at the speakers (obviusly)


**************

Adding a second usb card did not solved any problem (i changed the ids ) again same problem
Single input work double input not

****
The other solution could be to record 2 tracks one for the mic one for the speakers and mix them :(

[fredvs]

Nice that you find a workaround.
As already said, I am out of the game for minimum 2 weeks, sorry.

[trustfm]

yes no worries ! I just writing down the problems because i will loose track me too :)


If i find something i will let you know !

[fredvs]

Perfect!
Good luck and have fun.

[trustfm]

@edf
i have found another way to fix the clipping of 2 inputs
I had to set
sampleformat:=0; // 0: Float32
and not to default value
sampleformat:=-1; //2:Int16


***********************
procedure TForm1.ButtonRecordSpeakersClick(Sender: TObject);
var
  outformatst: string;
  sampleformat,outformat, numchan: integer;
begin

    PlayerIndex1 := 0;
    // PlayerIndex : from 0 to what your computer can do ! (depends of ram, cpu, ...)
    // If PlayerIndex exists already, it will be overwritten...

    uos_CreatePlayer(PlayerIndex1);

    sampleformat:=0; //0works ! 1 does not work !  2 does not work!


    if RadioButtonWav.Checked then
    begin
      outformatst := '.wav';
      outformat := 0;
    end
    else
    begin
      outformatst := '.ogg';
      outformat := 3;
    end;

    EditSpeakersPath.Text := 'rec_' + UTF8Decode(formatdatetime('YY_MM_DD_HH_mm_ss', now)) + outformatst;


    //In1Index1 := uos_AddFromDevIn(PlayerIndex1,-1,0.8,44100,-1,0,-1,-1); //-1 is speaker - 18 pulse works
    In1Index1 := uos_AddFromDevIn(PlayerIndex1,  -1, -1 ,-1   ,-1,sampleformat,1024,-1); //-1 is speaker - 18 pulse works
    numchan := uos_InputGetChannels(PlayerIndex1, In1Index1);
    LabelSpeakersChannel.Caption:=inttostr(numchan);

    In1Index2 := uos_AddFromDevIn(PlayerIndex2,  0, -1 ,  -1   ,-1,sampleformat,1024,-1); //0 is mic - 18 pulse works
    numchan := uos_InputGetChannels(PlayerIndex2, In1Index2);
    LabelMicChannel.Caption:=inttostr(numchan);

    //mut := uos_AddFromEndlessMuted(PlayerIndex1,2,1024);
    //uos_InputSetEnable(PlayerIndex1, mut, True);


    uos_AddIntoFile(PlayerIndex1, PChar(EditSpeakersPath.Text), -1, 2,sampleformat,1024, outformat); //-1



    uos_InputAddDSPVolume(PlayerIndex1, In1Index1, 1, 1);
    uos_InputAddDSPVolume(PlayerIndex1, In1Index2, 1, 1);

    //uos_InputSetDSPVolume(PlayerIndex1, In1Index1, TrackBar1.position / 100, TrackBar3.position / 100, True); // Set volume



    uos_Play(PlayerIndex1);  // everything is ready to play...

    ButtonRecordSpeakers.Enabled := False;
    ButtonStopRecordSpeakers.Enabled := True;


end;

[fredvs]

@trustfm:  
Thanks to find the trick!
But I have to check, asap, why int16 makes problem.
Also it seems that int16 is not anymore the "default" format from device-in sound cards.
Maybe change the default (-1) format to float32 (0) ?

[trustfm]

I dont know it is clearly not your fault . BUT I HAVE GREAT NEWS
Here is a solution recording sound and mic at the same time on linux-pulseaudio
using your simplerecorder demo unmodifided
Just use one input chanel the default one at -1

Now

Place your headphones at your back headphone jack of the PC
PulseAudio volume control > Configuration tab  > Analog Stereo Duplex
 

PulseAudio clients can send audio to "sinks" and receive audio from "sources".
So sinks = outputs (audio goes there speakers headphones)
   sources = inputs (audio comes from there mic)
   

pactl info

a) virtual speaker "recording"
pacmd load-module module-null-sink sink_name=recording sink_properties=device.description=recording
[recording]

b) loopback mirrors microphone to virtual speaker "recording"
   module-loopback: routes your input(source microphone) to your output(sink speakers)
find your sources using
pacmd list-sources
*alsa_output.pci-0000_00_1b.0.analog-stereo.monitor
*alsa_input.pci-0000_00_1b.0.analog-stereo  ---> my mic
pactl load-module module-loopback source=alsa_input.pci-0000_00_1b.0.analog-stereo sink=recording latency_msec=1
you get at PulseAudio volume control
tab Playback  [Loopback from Build-In Audio Analog Stereo on [recording]]
tab Recording [Loopback to recording from [Build-In Audio Analog Stereo]]

c) mirror your speakers to virtual speaker "recording"
find your sinks using
pacmd list-sinks
*alsa_output.pci-0000_00_1b.0.analog-stereo
pacmd load-module module-combine-sink sink_name=combined sink_properties=device.description=combined slaves=recording,alsa_output.pci-0000_00_1b.0.analog-stereo
this sets output of alsa_output.pci-0000_00_1b.0.analog-stereo=recording
you get at PulseAudio volume control
tab Playback  [Simultaneus output on recording on [recording] ]
tab Playback  [Simultaneus output on Built-In Audio Analog Stereo on [Built-In Audio Analog Stereo] ]


Now from PulseAudio volume control select
playback set your player with [combined]
recording set your recorder with [Alsa projectmy simplerecorder [Monitor of recording]]
This records mic and speakers [The user does not hear himself]

use
pulseaudio -k
to clear all modifications and restart pulseaudio  

[trustfm]

Recapping ...
Here is a simple recorder demo that can record Sound and Mic at the same time
https://trim.uk.to/my-simple-recorder.zip
(Please ignore the threaded test included at the same demo)

********************************************

Linux Speakers Only :
sample format 0: Float32 works , 1:Int32 works, 2:Int16 works
[use "Record speakers" only demo]

Linux Mic Only :
sample format 0: Float32 works , 1:Int32 works, 2:Int16 works
[use "Record Mic" only demo]

Linux Speakers + Mic (pulseaudio)
Using "Record Both" i get
sample format 0: Float32 works but desynchronizes after a while
              1:Int32 does not work (hiss noise)
              2:Int16 does not work (hiss noise)

Solution :
[use "Record speakers" only demo]

AND

use the solution above this post


***********************************
***********************************
***********************************


Windows Speakers Only :
sample format 0: Float32 works , 1:Int32 works, 2:Int16 works
[use "Record speakers" only demo]

Windows Mic Only :
sample format 0: Float32 works , 1:Int32 works, 2:Int16 works
[use "Record Mic" only demo]


Windows Speakers + Mic at the same time
[use "Record Both" only demo]
sample format 0: Float32 works !
              1:Int32 does not work (hiss noise)
              2:Int16 does not work (hiss noise)


So i can confirm that even at windows (another machine) we have problems with dual input when sampleformat is 1 or 2 BUT if 0 selected then works nicely without synch issues!