[Bug 37890] New: MCI 24bit .wav playback on top of pulseaudio (both with winealsa.drv,winepulse.drv) broken
https://bugs.winehq.org/show_bug.cgi?id=37890 Bug ID: 37890 Summary: MCI 24bit .wav playback on top of pulseaudio (both with winealsa.drv,winepulse.drv) broken Product: Wine Version: unspecified Hardware: x86 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: winmm&mci Assignee: wine-bugs(a)winehq.org Reporter: jonas.th(a)web.de Distribution: --- Created attachment 50462 --> https://bugs.winehq.org/attachment.cgi?id=50462 output of: pacmd list-sinks MCI 24bit .wav playback on top of pulseaudio (both with winealsa.drv,winepulse.drv) appears to be broken. A minimal test case with the necessary C code (<30 lines), the resulting .exe (compiled with MinGW) and a test .wav file is attached. The resulting test case: * works on all Windows 7 64bit machines tested * breaks with winealsa.drv on top of pulseaudio * breaks with winepulse.drv on top of pulseaudio * works with winealsa.drv on top of alsa (without pulseaudio) * works if the file is converted to a 16bit .wav file Furthermore, the attached output of pacmd list-sinks appears to show that pulseaudio doesn't advertise 24bit capabilities - so wine probably mistakenly tries to use 24bit audio with pulseaudio without support for that being present. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=37890 --- Comment #1 from jonas.th(a)web.de --- Created attachment 50463 --> https://bugs.winehq.org/attachment.cgi?id=50463 minimal test case to reproduce issue -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=37890 Sebastian Lackner <sebastian(a)fds-team.de> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |sebastian(a)fds-team.de -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=37890 --- Comment #2 from Sebastian Lackner <sebastian(a)fds-team.de> --- Created attachment 50464 --> https://bugs.winehq.org/attachment.cgi?id=50464 Log with WINEDEBUG=+winmm,+alsa WARNING: Do not wear headphones while running this test app! ;) I can confirm this issue with vanilla wine 1.7.34. In other wine apps (and also in winecfg) sound is working fine, so its not a general issue with my sound configuration. I have attached a log with WINEDEBUG=+winmm,+alsa. $ pulseaudio --version pulseaudio 5.0 $ git describe wine-1.7.34 -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=37890 --- Comment #3 from debuser <debuser1(a)mt2014.com> --- Created attachment 50468 --> https://bugs.winehq.org/attachment.cgi?id=50468 Check wBitsPerSample in PCM_StreamOpen() This is a bug in wine: wav-file contains 24-bit audio. Pulse-alsa plugin only advertises S16 and S32, no S24 format (see `aplay -v --dump-hw-params -d1 -Ddefault /dev/zero`). When attempting to set it snd_pcm_hw_params_set_format() fails to set S24_3LE, failing winealsa:AudioClient_Initialize. WINMM_OpenDevice catches the fail and calls WINMM_MapDevice to try other rates/formats. 16-bit one succeeds, calling acmStreamOpen, which calls PCM_StreamOpen, which sets cvtKeepRate to one of PCM_ConvertKeepRate[] values. This where things break. PCM_StreamOpen assumes that wBitsPerSample is either 8 or 16, so it sets conversion to 8->16 bits instead of 24->16 bits that it does not support. When sound plays acmStreamConvert->PCM_StreamConvert->cvtKeepRate->cvtSS816K gets called and turns 24-bit sound into noise. Attached is a tiny patch for PCM_StreamOpen to fail if format is not 8 or 16-bit. It's not a workaround but it stops wine from producing noise. The bug never happens with alsa because alsa provides a default=plug->softvol->dmix->hw chain, and its "plug" autoconversion module supports any possible input format, so snd_pcm_hw_params_set_format() never fails and wine's conversion code never gets called. To reproduce the bug and hear the noise without pulseaudio, using just alsa, you need a "broken" ~/.asoundrc bypassing plug autoconversion and limited supported formats, for example: pcm.!default { type hw card 0 format S16_LE rate 44100 channels 2 } The workaround for pulseaudio is to fix its alsa-plugin config (usually /usr/share/alsa/alsa.conf.d/pulse.conf or /usr/share/alsa/pulse-alsa.conf or similar) adding "plug" autoconversion there replacing: pcm.!default { type pulse } with pcm.!default { type plug slave.pcm { type pulse } } -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=37890 Ken Sharp <imwellcushtymelike(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Version|unspecified |1.7.34 -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=37890 Andrew Eikum <aeikum(a)codeweavers.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |aeikum(a)codeweavers.com --- Comment #4 from Andrew Eikum <aeikum(a)codeweavers.com> --- Thanks for looking so far into this. Sounds like the correct fix would be to support 24-bit formats in msacm32. We have a separate bug, where msacm32 fails to error on formats it doesn't support. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=37890 --- Comment #5 from Andrew Eikum <aeikum(a)codeweavers.com> --- Just curious, is there an actual game or application that uses this 24-bit format? I don't know if I've ever encountered one in the wild, so it would be nice to add it to my lits of audio test programs. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=37890 winetest(a)luukku.com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |winetest(a)luukku.com --- Comment #6 from winetest(a)luukku.com --- Still valid wine-staging 1.9.23. Cant test this bug against pure wine, since my build is for some reason missing development files. But did recent changes improve or fix this bug? http://source.winehq.org/git/wine.git/commit/6cc673477dd4a0749a292a6d6f60474... -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=37890 --- Comment #7 from Andrew Eikum <aeikum(a)codeweavers.com> --- (In reply to winetest from comment #6)
Still valid wine-staging 1.9.23. Cant test this bug against pure wine, since my build is for some reason missing development files.
But did recent changes improve or fix this bug?
http://source.winehq.org/git/wine.git/commit/ 6cc673477dd4a0749a292a6d6f604749d548fb6b
It's likely fixed, yes. Jonas, could you re-run your test case in wine-git (or Wine 1.9.25, when it is released) to confirm that it's fixed? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=37890 Bruno Jesus <00cpxxx(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Keywords| |download, testcase --- Comment #8 from Bruno Jesus <00cpxxx(a)gmail.com> --- I just tested but as expected on ALSA it works with or without the patch. Needs real pulseaudio user. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=37890 --- Comment #9 from winetest(a)luukku.com --- wine 2.7-git winecfg says that I got winealsa.drv. wine 2.7 winecfg says I got pulseaudio.drv Works for me. Tried with fresh wineprefix each. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=37890 Andrew Nguyen <arethusa26@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Resolution|--- |DUPLICATE Status|UNCONFIRMED |RESOLVED --- Comment #10 from Andrew Nguyen <arethusa26@gmail.com> --- The test case plays back the test audio file correctly on wine-11.0. I confirmed that the originally reported issue was indeed fixed by commit 6cc673477dd4a0749a292a6d6f604749d548fb6b, which is covered by bug 41599, so I'm marking this bug as a duplicate. *** This bug has been marked as a duplicate of bug 41599 *** -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=37890 Alistair Leslie-Hughes <leslie_alistair@hotmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED --- Comment #11 from Alistair Leslie-Hughes <leslie_alistair@hotmail.com> --- Closing duplicate. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
participants (2)
-
wine-bugs＠winehq.org -
WineHQ Bugzilla