Conversation

Notices

  1. @eal Why would I need to do that if Firefox plays audio fine with my custom ~/.asoundrc? Chromium's the one that fails, remember?

    Friday, 14-Jul-17 11:24:49 UTC from gs.kawa-kun.com
    1. @eal Oh, now I see.

      I don't have apulse anyway, and I'd have to do an Arch update to get it. No time for that now.

      Friday, 14-Jul-17 11:25:44 UTC from gs.kawa-kun.com
      1. @eal Also, I really don't have it on my system. Chromium:

        ALSA lib pulse.c:243:(pulse_connect) PulseAudio: Unable to connect: Connection refused

        Friday, 14-Jul-17 11:33:39 UTC from gs.kawa-kun.com
        1. @takeFrankerZakenji @eal Check
          > drivers = alsa,null,wave
          That is, ‘alsa’ first, in ~/.alsoftrc. NB, not ~/.asoundrc!

          Friday, 14-Jul-17 11:40:00 UTC from gs.smuglo.li
          1. @tijagi @eal I'm just trying to get Chromium to use a different sound card from the default. Audio works fine on card 0, but my speakers there are dead. I have to use a different card (card 3) hooked up to my "home audio" speakers instead.

            The problem is that card 3 requires 32-bit audio and 48 khz audio, while card 0 accepts more typical parameters. Firefox and the ALSA utilities work fine with this, but Chromium gets that "Unable to get period size for PLAYBACK: Invalid argument" error. https://gs.kawa-kun.com/url/347077

            Since Chromium is the outlier, it seems that it's the problem here.

            Friday, 14-Jul-17 12:09:12 UTC from gs.kawa-kun.com
            1. @tijagi @eal The error doesn't yield any useful results when searching, either.

              Friday, 14-Jul-17 12:10:20 UTC from gs.kawa-kun.com
            2. @takemangoakenji @eal If I understood you correctly, then just make that 3rd card the default one. If line out on the first one (the internal card?) is dead, and you bought a new discrete as a replacement.

              Friday, 14-Jul-17 12:33:54 UTC from gs.smuglo.li
              1. @tijagi @eal Making card 3 the default one won't fix the issue at all. The issue is that Chromium won't output 32-bit audio at 48 khz. Switching devices is the easy part.

                The speakers themselves are dead, so I'm awaiting new ones. The line out on card 0 is fine, but I can't use them with the "home audio" speakers because they require optical input. Card 3 outputs optical, while card 0 only has normal 3.5mm audio jacks.

                Friday, 14-Jul-17 12:36:15 UTC from gs.kawa-kun.com
                1. @takekiwiakenji @eal Aah. But the requirement to pass the output exactly as 32 bit 48 kHz to a sound card looks odd to me. Usually it just resamples the stream as needed with libsamplerate (implying that the need to resample is to get richer output, and not provide the input desired by a sound card). https://gs.smuglo.li/attachment/652684

                  Friday, 14-Jul-17 12:45:49 UTC from gs.smuglo.li
                  1. @takegrapeakenji @eal What is that card, anyway?

                    Friday, 14-Jul-17 12:46:45 UTC from gs.smuglo.li
                    1. @tijagi @eal Some VIA PCIe thing.

                      Friday, 14-Jul-17 12:47:28 UTC from gs.kawa-kun.com
              2. @takeappleakenji @eal I’m afraid, that Chromium still tries to use the default card. Pure ALSA doesn’t have the ability to tell which app should use which card. You have to use dmix or pulsegrape for that.

                Though I remember doing something alike… but I’m not quite sure.

                Friday, 14-Jul-17 12:36:36 UTC from gs.smuglo.li
                1. @tijagi @eal If that were the case, I would get silence without errors. I'm getting an error that shows Chromium using the other card: "ALSA lib pcm.c:8473:(snd_pcm_set_params) Unable to get period size for PLAYBACK: Invalid argument"

                  When I tried using the card without the custom asoundrc, I was getting errors around either sampling rate or sound format. I get no errors at all with card 0. Also, Chromium has a command-line option to specify the ALSA device it uses.

                  Friday, 14-Jul-17 12:38:36 UTC from gs.kawa-kun.com
                  1. @takebatcaveakenji @eal
                    > If that were the case, I would get silence without errors. I'm getting an error that shows Chromium using the other card: "ALSA lib pcm.c:8473:(snd_pcm_set_params) Unable to get period size for PLAYBACK: Invalid argument"
                    Ah, right. That probably tells, that Chromium uses the new card.

                    As for the period size, set it. This page seem to be explaining period time and period size https://www.alsa-project.org/main/index.php/FramesPeriods

                    Friday, 14-Jul-17 12:56:07 UTC from gs.smuglo.li
                    1. @takePotato Knishesakenji @eal I’d also check pcm.dmix section here as an example. https://www.alsa-project.org/main/index.php/Asoundrc#dmix

                      Friday, 14-Jul-17 12:57:10 UTC from gs.smuglo.li
                      1. @tijagi @eal Unfortunately, I already tried all this before ranting.

                        Friday, 14-Jul-17 12:59:13 UTC from gs.kawa-kun.com
                        1. @takemangoakenji @eal Latest alsa-lib 1.4.1.1?

                          Friday, 14-Jul-17 13:01:11 UTC from gs.smuglo.li
                          1. @tijagi @eal Can't update my OS now because I'm telecommuting using this computer.

                            Friday, 14-Jul-17 13:02:45 UTC from gs.kawa-kun.com
                            1. @takepapayaakenji @eal Should install Gentoo. We know no such problems.

                              That’s only alsa-lib, well, probably alsa-plugins, gst-plugins and alsa-utils, but none of them should require a large system update. Hmm, alsa-utils can, if they pull new major ncurses, but you can just not update it for the time being. Other packets shouldn’t require a new kernel or glib. Restarting alsasound won’t break your X session.
                              It’s good to have # and #
                              https://gs.smuglo.li/attachment/86227

                              Friday, 14-Jul-17 13:10:39 UTC from gs.smuglo.li
                              1. @tijagi @eal I used to use Gentoo, but I remember 36-hour straight compile times. I don't have the energy for that mango anymore.

                                Besides, changing OSes just because a set of speakers died is honestly pretty stupid.

                                Friday, 14-Jul-17 13:11:55 UTC from gs.kawa-kun.com
                                1. @takeFluffle Puffakenji @eal You tell about it like you have to be bound to your PC for 36 hours straight and can’t even open a browser while something compiles. I take it slowly: resolve deps, prepare for upgrade, then compile. It distracts from mangoposting for 15 mins fours times a day or so.

                                  Just the other day I could solve problems with steam only thanks to seeing the dependencies clearer than in some ubuntu or fedora.

                                  Friday, 14-Jul-17 13:19:14 UTC from gs.smuglo.li
                                  1. @tijagi @eal I used Gentoo for ten years and left for several reasons. I don't need to argue with you about them.

                                    Friday, 14-Jul-17 13:20:17 UTC from gs.kawa-kun.com
                                    1. @takeFrankerZakenji @eal As you wish. https://gs.smuglo.li/attachment/79976

                                      Friday, 14-Jul-17 13:21:30 UTC from gs.smuglo.li
                            2. @takepapayaakenji @tijagi @eal That's why you need Windows 10. Little things like "I'm telecommuting", "I actually wanted to use my computer today", "I am on a limited connection and will be charged per MB" or "Good job my webcam isn't on to broadcast me Fluffle Puffing" are things of the past.
                              It'll be up to date and everything.

                              Friday, 14-Jul-17 15:33:16 UTC from gs.kawa-kun.com
                              1. @daggsy @takeFluffle Puffakenji @eal Uh-huh. https://gs.smuglo.li/notice/1326683 https://gs.smuglo.li/attachment/180844

                                Friday, 14-Jul-17 15:57:49 UTC from gs.smuglo.li
                    2. @tijagi @eal Unfortunately, period_size isn't supported with the 'rate' device type.

                      Friday, 14-Jul-17 12:57:20 UTC from gs.kawa-kun.com
                      1. @takePotato Knishesakenji @eal How about making another PCM that pre-coverts the rate then passes it to pcm.chrome via dmix?  https://gs.smuglo.li/attachment/53248

                        Friday, 14-Jul-17 13:03:35 UTC from gs.smuglo.li