Loading ...
Sorry, an error occurred while loading the content.

RE: [nslu2-linux] HELP - MP3 playback using MPD on unslung 6.10

Expand Messages
  • Gregg C Levine
    Hello! If I remember correctly, the NSLU2 processor is one of those unlucky ones to be created without a FPU as part of its design. Technically the processor
    Message 1 of 5 , Mar 2, 2009
    • 0 Attachment
      Hello!
      If I remember correctly, the NSLU2 processor is one of those unlucky ones to
      be created without a FPU as part of its design.

      Technically the processor it uses is ideal for file management, and of
      course as it happens working in some routers, and of course the NSLU2.
      Because it does not have this FPU, Linux on it needs to emulate (some how)
      the floating point functions that we take for granted on desktops and full
      size servers, except for IBM's Z Series. So it spends its time physically
      doing that and causes the MP3 players such as that one to produce choppy
      sound. (And the other one to bollix the sampling actions.)

      If you have access to another system running Linux, (your choice of
      distributions), try your MP3s on the same player there.

      Next up: Is this a powered hub? I certainly hope so. You have your disk
      drive there that holds your collection, and the sound device dongle.

      The other issue is why did you abandon using the USB 2.0 hub? Supposedly the
      NSLU2 contains USB 2.0 connectors, so the hub should agree with them.

      The other thing that comes to mind is perhaps your choice of installed
      release, but I don't think it is at fault here.

      --
      Gregg C Levine hansolofalcon@...
      "The Force will be with you always." Obi-Wan Kenobi
       


      > -----Original Message-----
      > From: nslu2-linux@yahoogroups.com [mailto:nslu2-linux@yahoogroups.com] On
      Behalf
      > Of threeeyedtoad
      > Sent: Monday, March 02, 2009 10:47 PM
      > To: nslu2-linux@yahoogroups.com
      > Subject: [nslu2-linux] HELP - MP3 playback using MPD on unslung 6.10
      >
      > I'm sort of at the end of my rope on this one, and could really use
      > help from some of the seasoned experts...
      >
      > I've picked up a USB audio dongle (although I attempted to secure one
      > with a C-Media chip, I ended up with "3D SOUND" dongle with a JMTek
      > chip in it)... I've managed to configure the dongle with my NSLU2
      > running Unslung 6.10 Beta. I abandoned getting it to work with a USB
      > 2.0 hub, but have met with success using a Texas Instruments
      > TUSB2040/2070 USB 1.1 hub. I also have a Western Digital .5 Tb MyBook
      > sharing the USB hub, which has all my MP3s on it.
      >
      > When I use madplay (0.15.2) to playback an MP3, sound comes out of the
      > USB dongle (although it does sound like the MP3 is either downsampled,
      > or over-driven) and is delivered to my speakers.
      >
      > However, when I attempt to use MPD (0.14.2) with the wymypy web-based
      > front-end for playback, the playback is badly stuttered (like, a
      > half-second of audio is heard, then a second of silence, then another
      > half-second of audio, etc.). Top reveals MPD's CPU utilization is
      > pegged at 97%, leading me to believe the stuttering is a side affect
      > of the excessive utilization.
      >
      > I could go about getting another server-side playback utility running,
      > but I was hoping to avoid setting up PHP and/or a web server. Madplay
      > reports only using between 5% and 15% CPU utilization, so what's up
      > with MPD? Is it using some floating point calculation that can be
      > supressed through mpd.conf possibly?
      >
      > Also, any ideas on improving the fidelity of the sound sample coming
      > out of the USB dongle? I know the dongle is capable of much higher
      > quality, having used it with another system...
      >
      >
    • threeeyedtoad
      Thanks for the quick response. I m fully aware that the NSLU2 does not have a FPU. This is well documented in the wiki pages. Also well documented are the
      Message 2 of 5 , Mar 3, 2009
      • 0 Attachment
        Thanks for the quick response.

        I'm fully aware that the NSLU2 does not have a FPU. This is well
        documented in the wiki pages. Also well documented are the MP3
        players that only use integer math. Both madplay and MPD are both
        included in the list of supported MP3 players that use integer math.
        As I described, madplay works fine, but (to my knowledge) does not
        support remote server play via a web interface. MPD has many clients
        that support remote play via the web.

        I've read in several places online that replaygain (and possibly
        software volume control - whatever that is) MAY use FP math. My
        question for those who have gotten MPD successfully running on a Slug
        - is there a configuration that can be made in mpd.conf that would
        suppress any functionality that utilizes FPU?

        As for the hub - yes, I'm using a powered USB 1.1 hub. Again, using
        maadplay, everything is hunky dory playing MP3s off a HD plugged into
        the same 1.1 hub. I bailed on a USB 2.0 hub after reading multiple
        places that the USB audio dongle may not play nice with a 2.0 hub - at
        least, not using an unslung distro. The USB 2.0 hub is fine with the
        unslung NSLU to hang HD drives off, but not for the audio.

        So - thoughts on how to get MPD working without FPU math under unslung?


        --- In nslu2-linux@yahoogroups.com, "Gregg C Levine"
        <hansolofalcon@...> wrote:
        >
        > Hello!
        > If I remember correctly, the NSLU2 processor is one of those unlucky
        ones to
        > be created without a FPU as part of its design.
        >
        > Technically the processor it uses is ideal for file management, and of
        > course as it happens working in some routers, and of course the NSLU2.
        > Because it does not have this FPU, Linux on it needs to emulate
        (some how)
        > the floating point functions that we take for granted on desktops
        and full
        > size servers, except for IBM's Z Series. So it spends its time
        physically
        > doing that and causes the MP3 players such as that one to produce choppy
        > sound. (And the other one to bollix the sampling actions.)
        >
        > If you have access to another system running Linux, (your choice of
        > distributions), try your MP3s on the same player there.
        >
        > Next up: Is this a powered hub? I certainly hope so. You have your disk
        > drive there that holds your collection, and the sound device dongle.
        >
        > The other issue is why did you abandon using the USB 2.0 hub?
        Supposedly the
        > NSLU2 contains USB 2.0 connectors, so the hub should agree with them.
        >
        > The other thing that comes to mind is perhaps your choice of installed
        > release, but I don't think it is at fault here.
        >
        > --
        > Gregg C Levine hansolofalcon@...
        > "The Force will be with you always." Obi-Wan Kenobi
        >
        >
        >
        > > -----Original Message-----
        > > From: nslu2-linux@yahoogroups.com
        [mailto:nslu2-linux@yahoogroups.com] On
        > Behalf
        > > Of threeeyedtoad
        > > Sent: Monday, March 02, 2009 10:47 PM
        > > To: nslu2-linux@yahoogroups.com
        > > Subject: [nslu2-linux] HELP - MP3 playback using MPD on unslung 6.10
        > >
        > > I'm sort of at the end of my rope on this one, and could really use
        > > help from some of the seasoned experts...
        > >
        > > I've picked up a USB audio dongle (although I attempted to secure one
        > > with a C-Media chip, I ended up with "3D SOUND" dongle with a JMTek
        > > chip in it)... I've managed to configure the dongle with my NSLU2
        > > running Unslung 6.10 Beta. I abandoned getting it to work with a USB
        > > 2.0 hub, but have met with success using a Texas Instruments
        > > TUSB2040/2070 USB 1.1 hub. I also have a Western Digital .5 Tb MyBook
        > > sharing the USB hub, which has all my MP3s on it.
        > >
        > > When I use madplay (0.15.2) to playback an MP3, sound comes out of the
        > > USB dongle (although it does sound like the MP3 is either downsampled,
        > > or over-driven) and is delivered to my speakers.
        > >
        > > However, when I attempt to use MPD (0.14.2) with the wymypy web-based
        > > front-end for playback, the playback is badly stuttered (like, a
        > > half-second of audio is heard, then a second of silence, then another
        > > half-second of audio, etc.). Top reveals MPD's CPU utilization is
        > > pegged at 97%, leading me to believe the stuttering is a side affect
        > > of the excessive utilization.
        > >
        > > I could go about getting another server-side playback utility running,
        > > but I was hoping to avoid setting up PHP and/or a web server. Madplay
        > > reports only using between 5% and 15% CPU utilization, so what's up
        > > with MPD? Is it using some floating point calculation that can be
        > > supressed through mpd.conf possibly?
        > >
        > > Also, any ideas on improving the fidelity of the sound sample coming
        > > out of the USB dongle? I know the dongle is capable of much higher
        > > quality, having used it with another system...
        > >
        > >
        >
      • Brian Zhou
        I just disabled libsamplerate in mpd configure except on x86 optware target. When libsamplerate is enabled, it does quite a lot of FP calculation. Please give
        Message 3 of 5 , Mar 3, 2009
        • 0 Attachment
          I just disabled libsamplerate in mpd configure except on x86 optware target. When libsamplerate is enabled, it does quite a lot of FP calculation. Please give mpd_0.14.2-3 a try and see if it makes any difference.

          -Brian Zhou

          --- In nslu2-linux@yahoogroups.com, "threeeyedtoad" <threeeyedtoad@...> wrote:
          >
          > I'm sort of at the end of my rope on this one, and could really use
          > help from some of the seasoned experts...
          >
          > I've picked up a USB audio dongle (although I attempted to secure one
          > with a C-Media chip, I ended up with "3D SOUND" dongle with a JMTek
          > chip in it)... I've managed to configure the dongle with my NSLU2
          > running Unslung 6.10 Beta. I abandoned getting it to work with a USB
          > 2.0 hub, but have met with success using a Texas Instruments
          > TUSB2040/2070 USB 1.1 hub. I also have a Western Digital .5 Tb MyBook
          > sharing the USB hub, which has all my MP3s on it.
          >
          > When I use madplay (0.15.2) to playback an MP3, sound comes out of the
          > USB dongle (although it does sound like the MP3 is either downsampled,
          > or over-driven) and is delivered to my speakers.
          >
          > However, when I attempt to use MPD (0.14.2) with the wymypy web-based
          > front-end for playback, the playback is badly stuttered (like, a
          > half-second of audio is heard, then a second of silence, then another
          > half-second of audio, etc.). Top reveals MPD's CPU utilization is
          > pegged at 97%, leading me to believe the stuttering is a side affect
          > of the excessive utilization.
          >
          > I could go about getting another server-side playback utility running,
          > but I was hoping to avoid setting up PHP and/or a web server. Madplay
          > reports only using between 5% and 15% CPU utilization, so what's up
          > with MPD? Is it using some floating point calculation that can be
          > supressed through mpd.conf possibly?
          >
          > Also, any ideas on improving the fidelity of the sound sample coming
          > out of the USB dongle? I know the dongle is capable of much higher
          > quality, having used it with another system...
          >
        • threeeyedtoad
          Rock on. Disabling libsamplerate did the trick for me in mpd_0.14.2-3. Now that I know what to search for, there are some threads that have recently spun up
          Message 4 of 5 , Mar 4, 2009
          • 0 Attachment
            Rock on. Disabling libsamplerate did the trick for me in mpd_0.14.2-3. Now that I know what to search for, there are some threads that have recently spun up on the linux.debian.bugs.dist Google Group (specifically Bug#517976), complaining about libsamplerate as a CPU hog for machines without FPU. I guess there have been changes to MPD recently that changed how this was implemented.

            Brian, thanks for the quick response and attention in this matter. I'm amazed and grateful for how well the *nix community continues to support the little Slug, five years after it was first introduced to the market. Amazing.


            --- In nslu2-linux@yahoogroups.com, "Brian Zhou" <b88zhou@...> wrote:
            >
            > I just disabled libsamplerate in mpd configure except on x86 optware target. When libsamplerate is enabled, it does quite a lot of FP calculation. Please give mpd_0.14.2-3 a try and see if it makes any difference.
            >
            > -Brian Zhou
            >
            > --- In nslu2-linux@yahoogroups.com, "threeeyedtoad" <threeeyedtoad@> wrote:
            > >
            > > I'm sort of at the end of my rope on this one, and could really use
            > > help from some of the seasoned experts...
            > >
            > > I've picked up a USB audio dongle (although I attempted to secure one
            > > with a C-Media chip, I ended up with "3D SOUND" dongle with a JMTek
            > > chip in it)... I've managed to configure the dongle with my NSLU2
            > > running Unslung 6.10 Beta. I abandoned getting it to work with a USB
            > > 2.0 hub, but have met with success using a Texas Instruments
            > > TUSB2040/2070 USB 1.1 hub. I also have a Western Digital .5 Tb MyBook
            > > sharing the USB hub, which has all my MP3s on it.
            > >
            > > When I use madplay (0.15.2) to playback an MP3, sound comes out of the
            > > USB dongle (although it does sound like the MP3 is either downsampled,
            > > or over-driven) and is delivered to my speakers.
            > >
            > > However, when I attempt to use MPD (0.14.2) with the wymypy web-based
            > > front-end for playback, the playback is badly stuttered (like, a
            > > half-second of audio is heard, then a second of silence, then another
            > > half-second of audio, etc.). Top reveals MPD's CPU utilization is
            > > pegged at 97%, leading me to believe the stuttering is a side affect
            > > of the excessive utilization.
            > >
            > > I could go about getting another server-side playback utility running,
            > > but I was hoping to avoid setting up PHP and/or a web server. Madplay
            > > reports only using between 5% and 15% CPU utilization, so what's up
            > > with MPD? Is it using some floating point calculation that can be
            > > supressed through mpd.conf possibly?
            > >
            > > Also, any ideas on improving the fidelity of the sound sample coming
            > > out of the USB dongle? I know the dongle is capable of much higher
            > > quality, having used it with another system...
            > >
            >
          Your message has been successfully submitted and would be delivered to recipients shortly.