| Summary: | Regression: Bad/Crackling audio with pcsx2 (spu2x) using pulseaudio since 2.0.4 | ||
|---|---|---|---|
| Product: | SDL | Reporter: | zilla |
| Component: | audio | Assignee: | Ryan C. Gordon <icculus> |
| Status: | RESOLVED ABANDONED | QA Contact: | Ryan C. Gordon <icculus> |
| Severity: | blocker | ||
| Priority: | P2 | CC: | gregory.hainaut, philipp.wiesemann, slouken |
| Version: | 2.0.4 | Keywords: | target-2.0.6 |
| Hardware: | x86_64 | ||
| OS: | Linux | ||
| Attachments: | screenshot of spu2-x settings | ||
Additional info: Here is the relevant code in PCSX2: https://github.com/PCSX2/pcsx2/blob/master/plugins/spu2-x/src/SndOut_SDL.cpp We are trying to debug this, with regards to behaviour change in 2.0.3 vs 2.0.4 and a question has come up by one of the PCSX2 devs: "User requests a 1024 samples of 16 bits on 2 channel. Therefore a sample frame is sizeof(u16) * 2 = 4. But the len of callback buffer is only 2048. And therefore it could only contain 2048/4=512 samples instead of the 1024 requested. I just want to know if it is done on purpose or if it's a not-expected side effect of another change?" Could you give us some insight in this? Thank you. More on this @ https://github.com/PCSX2/pcsx2/issues/1095 Specifically: "So the bug (probably) is (2.0.4): id = open_audio_device(NULL, 0, desired, NULL, 0, 1); and 2.0.3: id = open_audio_device(NULL, 0, desired, desired, 0, 1); So it is better to use obtained in PCSX2, because 2.0.4 has a bug (or 2.0.4 doc has a bug)..." Hello, I understand this change was done to correct a bug but it introduces a change on the API. Quoting the wiki (I don't know if it is the official doc) https://wiki.libsdl.org/SDL_OpenAudio?highlight=%28\bCategoryAPI\b%29|%28SDLFunctionTemplate%29 Section: Return Value If obtained is NULL, ...... If obtained is NULL, *desired will have fields modified*. I think, we will rewrite the code to avoid this pitfall in the future. But it is annoying for already released version. And it might impact others project too.
> If obtained is NULL, *desired will have fields modified*.
This phrase is bad in the wiki, but it does explain later on the page that it will only update fields like "silence" to be correct, and not actually change anything about the requested format. The intention was and always will be that if you specify "obtained" as NULL, you're saying "I _must_ get the format requested in 'desired' because I can't handle anything but that format, so if you can't give me that format, pretend you did and have SDL convert for me transparently."
I have to look at the code to see if we're failing to do it this way, but that's the idea.
--ryan.
I'm not marking it as a duplicate yet, but Bug #3749 is probably also a victim of this confusion. --ryan. > I'm not marking it as a duplicate yet, but Bug #3749 is probably also a > victim of this confusion. Also Bug #3710. --ryan. Hello,
> you're saying "I _must_ get the format requested in 'desired' because
> I can't handle anything but that format, so if you can't give me that
> format, pretend you did and have SDL convert for me transparently."
Yes it makes sense. It would be nice to update the wiki to state clearly the behavior of "NULL obtained" and which field is updated.
I think the confusion also came from SDL 2.0.3 which updated all desired field value.
Hello, and sorry if you're getting dozens of copies of this message by email. We are closing out bugs that appear to be abandoned in some form. This can happen for lots of reasons: we couldn't reproduce it, conversation faded out, the bug was noted as fixed in a comment but we forgot to mark it resolved, the report is good but the fix is impractical, we fixed it a long time ago without realizing there was an associated report, etc. Individually, any of these bugs might have a better resolution (such as WONTFIX or WORKSFORME or INVALID) but we've added a new resolution of ABANDONED to make this easily searchable and make it clear that it's not necessarily unreasonable to revive a given bug report. So if this bug is still a going concern and you feel it should still be open: please feel free to reopen it! But unless you respond, we'd like to consider these bugs closed, as many of them are several years old and overwhelming our ability to prioritize recent issues. (please note that hundred of bug reports were sorted through here, so we apologize for any human error. Just reopen the bug in that case!) Thanks, --ryan. |
Created attachment 2359 [details] screenshot of spu2-x settings Probably this will be for Ryan :) Description/context: pcsx2 uses a plugin system, the plugin for sound is spu2x. The plugin uses SDL2. When using pulseaudio backend in SDL2, the sound is very bad (sound artefacts / crackling) since 2.0.4. If I switch to alsa backend, sound is fine. If I reinstall 2.0.3 lib32-sdl2 then sound is fine, as before, even when using the pulseaudio backend. Reproduce: Not sure if it's system specific, but (assuming one is familiar with pcsx2) all you have to do is configure spu2x to use SDL audio & pulseaudio backend. I'll add a screenshot of the spu2x settings menu. The crackling is noticeable even from the bootup of the PS2 bios, so it's not game specific. I'm using PCSX2 built from git (1.3.1-20160105063026). I tried to narrow it down a bit, reverse patching from 2.0.4 source and recompiling SDL2 but alas I didn't get very far, probably because I'm not familiar with Mercurial and didn't know exactly what patches or patchsets to use. I'm no programmer but if you have patches for me to test, I'll be glad to try them out! --- System info --- $ uname -r ; glxinfo | grep 'OpenGL version' 4.3.3-2-ARCH OpenGL version string: 4.5.0 NVIDIA 358.16 $ sudo lshw -short [sudo] password for x: H/W path Device Class Description ================================================ system P55-USB3 /0 bus P55-USB3 /0/0 memory 128KiB BIOS /0/4 processor Intel(R) Core(TM) i5 CPU 750 @ 2.67GHz /0/4/a memory 64KiB L1 cache /0/4/b memory 8MiB L2 cache /0/1a memory 4GiB System Memory /0/1a/0 memory 2GiB DIMM 1520 MHz (0.7 ns) /0/1a/1 memory DIMM [empty] /0/1a/2 memory 2GiB DIMM 1520 MHz (0.7 ns) /0/1a/3 memory DIMM [empty] /0/100 bridge Core Processor DMI /0/100/3 bridge Core Processor PCI Express Root Port 1 /0/100/3/0 display GF104 [GeForce GTX 460] /0/100/3/0.1 multimedia GF104 High Definition Audio Controller /0/100/8 generic Core Processor System Management Registers /0/100/8.1 generic Core Processor Semaphore and Scratchpad Registers /0/100/8.2 generic Core Processor System Control and Status Registers /0/100/8.3 generic Core Processor Miscellaneous Registers /0/100/10 generic Core Processor QPI Link /0/100/10.1 generic Core Processor QPI Routing and Protocol Registers /0/100/1a bus 5 Series/3400 Series Chipset USB Universal Host Controller /0/100/1a/1 usb1 bus UHCI Host Controller /0/100/1a/1/2 input USB Keyboard /0/100/1a.1 bus 5 Series/3400 Series Chipset USB Universal Host Controller /0/100/1a.1/1 usb2 bus UHCI Host Controller /0/100/1a.2 bus 5 Series/3400 Series Chipset USB Universal Host Controller /0/100/1a.2/1 usb3 bus UHCI Host Controller /0/100/1a.2/1/1 input Xornet gaming mouse /0/100/1a.7 bus 5 Series/3400 Series Chipset USB2 Enhanced Host Controller /0/100/1a.7/1 usb7 bus EHCI Host Controller /0/100/1a.7/1/4 multimedia Webcam C270 /0/100/1b multimedia 5 Series/3400 Series Chipset High Definition Audio /0/100/1c bridge 5 Series/3400 Series Chipset PCI Express Root Port 1 /0/100/1c/0 storage JMB363 SATA/IDE Controller /0/100/1c/0.1 storage JMB363 SATA/IDE Controller /0/100/1c.1 bridge 5 Series/3400 Series Chipset PCI Express Root Port 2 /0/100/1c.1/0 enp3s0 network RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller /0/100/1c.2 bridge 5 Series/3400 Series Chipset PCI Express Root Port 3 /0/100/1c.2/0 bus uPD720200 USB 3.0 Host Controller /0/100/1c.2/0/0 usb6 bus xHCI Host Controller /0/100/1c.2/0/1 usb5 bus xHCI Host Controller /0/100/1d bus 5 Series/3400 Series Chipset USB Universal Host Controller /0/100/1d/1 usb4 bus UHCI Host Controller /0/100/1d/1/2 generic Xbox 360 Wireless Receiver for Windows /0/100/1d.1 bus 5 Series/3400 Series Chipset USB Universal Host Controller /0/100/1d.1/1 usb9 bus UHCI Host Controller /0/100/1d.2 bus 5 Series/3400 Series Chipset USB Universal Host Controller /0/100/1d.2/1 usb10 bus UHCI Host Controller /0/100/1d.3 bus 5 Series/3400 Series Chipset USB Universal Host Controller /0/100/1d.3/1 usb11 bus UHCI Host Controller /0/100/1d.7 bus 5 Series/3400 Series Chipset USB2 Enhanced Host Controller /0/100/1d.7/1 usb8 bus EHCI Host Controller /0/100/1e bridge 82801 PCI Bridge /0/100/1f bridge 5 Series Chipset LPC Interface Controller /0/100/1f.2 storage 5 Series/3400 Series Chipset 6 port SATA AHCI Controller /0/100/1f.3 bus 5 Series/3400 Series Chipset SMBus Controller /0/101 bridge Core Processor QuickPath Architecture Generic Non-Core Registers /0/102 bridge Core Processor QuickPath Architecture System Address Decoder /0/103 bridge Core Processor QPI Link 0 /0/104 bridge Core Processor QPI Physical 0 /0/105 bridge Core Processor Integrated Memory Controller /0/106 bridge Core Processor Integrated Memory Controller Target Address Decoder /0/107 bridge Core Processor Integrated Memory Controller Test Registers /0/108 bridge Core Processor Integrated Memory Controller Channel 0 Control Registers /0/109 bridge Core Processor Integrated Memory Controller Channel 0 Address Registers /0/10a bridge Core Processor Integrated Memory Controller Channel 0 Rank Registers /0/10b bridge Core Processor Integrated Memory Controller Channel 0 Thermal Control Registers /0/10c bridge Core Processor Integrated Memory Controller Channel 1 Control Registers /0/10d bridge Core Processor Integrated Memory Controller Channel 1 Address Registers /0/10e bridge Core Processor Integrated Memory Controller Channel 1 Rank Registers /0/10f bridge Core Processor Integrated Memory Controller Channel 1 Thermal Control Registers