Francois Gouget wrote:
Some Windows sound drivers allow wild frequencies such as the 2MHz one.
Is this a windows driver bug? Is the test really valid and it's just finding a real bug in the windows driver for that specific sound card?
I know it makes the test fail on bad windows drivers but is that bad?
On Sun, 25 Apr 2004, Robert Reif wrote:
Francois Gouget wrote:
Some Windows sound drivers allow wild frequencies such as the 2MHz one.
Is this a windows driver bug? Is the test really valid and it's just finding a real bug in the windows driver for that specific sound card?
I know it makes the test fail on bad windows drivers but is that bad?
It might be a buggy sound driver or it may siomply be that Windows does not specify the behavior. After all it's always possible to remap a 2MHz sampling rate to 44.1kHz...
So in doubt it seems best to just remove that test. Or currently just modify it to check for invalid bits per sample values in the hope it's a more robust test. If not we could also remove the test altogether.
Francois Gouget wrote:
On Sun, 25 Apr 2004, Robert Reif wrote:
Francois Gouget wrote:
Some Windows sound drivers allow wild frequencies such as the 2MHz one.
Is this a windows driver bug? Is the test really valid and it's just finding a real bug in the windows driver for that specific sound card?
I know it makes the test fail on bad windows drivers but is that bad?
It might be a buggy sound driver or it may siomply be that Windows does not specify the behavior. After all it's always possible to remap a 2MHz sampling rate to 44.1kHz...
So in doubt it seems best to just remove that test. Or currently just modify it to check for invalid bits per sample values in the hope it's a more robust test. If not we could also remove the test altogether.
How about something like this:
Adds test for illegal bits per sample. Fixes format printing.
Since we are opening the device in WAVE_FORMAT_DIRECT mode, the wave mapper is not involved so it should fail unless the hardware really supports 2MHz (which is highly unlikely).
Index: dlls/winmm/tests/wave.c =================================================================== RCS file: /home/wine/wine/dlls/winmm/tests/wave.c,v retrieving revision 1.16 diff -u -r1.16 wave.c --- dlls/winmm/tests/wave.c 27 Mar 2004 20:42:57 -0000 1.16 +++ dlls/winmm/tests/wave.c 26 Apr 2004 00:18:45 -0000 @@ -312,7 +312,7 @@ }
/* Try an invalid format to test error handling */ - trace("Testing invalid 2MHz format\n"); + trace("Testing invalid format: 2MHz sample rate\n"); format.wFormatTag=WAVE_FORMAT_PCM; format.nChannels=2; format.wBitsPerSample=16; @@ -322,7 +322,7 @@ format.cbSize=0; oformat=format; rc=waveOutOpen(&wout,d,&format,0,0,CALLBACK_NULL|WAVE_FORMAT_DIRECT); - ok(rc==WAVERR_BADFORMAT, + ok(rc==WAVERR_BADFORMAT || rc==MMSYSERR_INVALFLAG, "waveOutOpen: opening the device at 2MHz should fail %d: rc=%d\n",d,rc); if (rc==MMSYSERR_NOERROR) { trace(" got %ldx%2dx%d for %ldx%2dx%d\n", @@ -333,22 +333,25 @@ waveOutClose(wout); }
+ /* try an unsupported bit size */ + trace("Testing invalid format: 11 bits per sample\n"); format.wFormatTag=WAVE_FORMAT_PCM; format.nChannels=2; - format.wBitsPerSample=16; - format.nSamplesPerSec=2000000; /* 2MHz! */ + format.wBitsPerSample=11; + format.nSamplesPerSec=22050; format.nBlockAlign=format.nChannels*format.wBitsPerSample/8; format.nAvgBytesPerSec=format.nSamplesPerSec*format.nBlockAlign; format.cbSize=0; + oformat=format; rc=waveOutOpen(&wout,d,&format,0,0,CALLBACK_NULL|WAVE_FORMAT_DIRECT); ok(rc==WAVERR_BADFORMAT || rc==MMSYSERR_INVALFLAG, - "waveOutOpen: opening the device at 2MHz should fail %d: rc=%d\n",d,rc); + "waveOutOpen: opening the device in 11 bit mode should fail %d: rc=%d\n",d,rc); if (rc==MMSYSERR_NOERROR) { - trace(" got %ldx%2dx%d for %dx%2dx%d\n", + trace(" got %ldx%2dx%d for %ldx%2dx%d\n", format.nSamplesPerSec, format.wBitsPerSample, format.nChannels, - win_formats[f][1], win_formats[f][2], - win_formats[f][3]); + oformat.nSamplesPerSec, oformat.wBitsPerSample, + oformat.nChannels); waveOutClose(wout); } } @@ -519,7 +522,7 @@ }
/* Try an invalid format to test error handling */ - trace("Testing invalid 2MHz format\n"); + trace("Testing invalid format: 2MHz sample rate\n"); format.wFormatTag=WAVE_FORMAT_PCM; format.nChannels=2; format.wBitsPerSample=16; @@ -529,7 +532,7 @@ format.cbSize=0; oformat=format; rc=waveInOpen(&win,d,&format,0,0,CALLBACK_NULL|WAVE_FORMAT_DIRECT); - ok(rc==WAVERR_BADFORMAT, + ok(rc==WAVERR_BADFORMAT || rc==MMSYSERR_INVALFLAG, "waveInOpen: opening the device at 2MHz should fail %d: rc=%d\n",d,rc); if (rc==MMSYSERR_NOERROR) { trace(" got %ldx%2dx%d for %ldx%2dx%d\n", @@ -540,22 +543,25 @@ waveInClose(win); }
+ /* try an unsupported bit size */ + trace("Testing invalid format: 11 bits per sample\n"); format.wFormatTag=WAVE_FORMAT_PCM; format.nChannels=2; - format.wBitsPerSample=16; - format.nSamplesPerSec=2000000; /* 2MHz! */ + format.wBitsPerSample=11; + format.nSamplesPerSec=22050; format.nBlockAlign=format.nChannels*format.wBitsPerSample/8; format.nAvgBytesPerSec=format.nSamplesPerSec*format.nBlockAlign; format.cbSize=0; + oformat=format; rc=waveInOpen(&win,d,&format,0,0,CALLBACK_NULL|WAVE_FORMAT_DIRECT); ok(rc==WAVERR_BADFORMAT || rc==MMSYSERR_INVALFLAG, - "waveInOpen: opening the device at 2MHz should fail %d: rc=%d\n",d,rc); + "waveInOpen: opening the device in 11 bit mode should fail %d: rc=%d\n",d,rc); if (rc==MMSYSERR_NOERROR) { - trace(" got %ldx%2dx%d for %dx%2dx%d\n", + trace(" got %ldx%2dx%d for %ldx%2dx%d\n", format.nSamplesPerSec, format.wBitsPerSample, format.nChannels, - win_formats[f][1], win_formats[f][2], - win_formats[f][3]); + oformat.nSamplesPerSec, oformat.wBitsPerSample, + oformat.nChannels); waveInClose(win); } } }
On Sun, 25 Apr 2004, Robert Reif wrote: [...]
How about something like this:
Adds test for illegal bits per sample. Fixes format printing.
Looks ok though if we keep the 2MHz etst then we should investigate what's goign on on Windows in more details.
Since we are opening the device in WAVE_FORMAT_DIRECT mode, the wave mapper is not involved so it should fail unless the hardware really supports 2MHz (which is highly unlikely).
True. But I believe Alsa will happily resample (or claim to resample) a 2MHz signal which makes it impossible for us to report an error. Unless we check for any frequency >96KHz in our driver but this test could be wrong on high-end sound cards: I believe some support 192KHz sampling rates.
IOW I'm not sure it's worth trying to be 100% compatible with Windows there. Especially since there's some doubt about what Windows actually does.
Francois Gouget wrote:
On Sun, 25 Apr 2004, Robert Reif wrote: [...]
How about something like this:
Adds test for illegal bits per sample. Fixes format printing.
Looks ok though if we keep the 2MHz etst then we should investigate what's goign on on Windows in more details.
Since we are opening the device in WAVE_FORMAT_DIRECT mode, the wave mapper is not involved so it should fail unless the hardware really supports 2MHz (which is highly unlikely).
True. But I believe Alsa will happily resample (or claim to resample) a 2MHz signal which makes it impossible for us to report an error. Unless we check for any frequency >96KHz in our driver but this test could be wrong on high-end sound cards: I believe some support 192KHz sampling rates.
IOW I'm not sure it's worth trying to be 100% compatible with Windows there. Especially since there's some doubt about what Windows actually does.
I know for a fact that DirectSound used to have a 100kHz limit which was recently raised to 200kHz but I don't know if that limit was imposed in Direct Sound or was inherited from the driver itself. We could also check the wave api against the Direct Sound limits when ALSA is used with plugins.
-
Francois Gouget -
Robert Reif