Re: [Musicpd-dev-team] Best Sinc Interpolator
On 2009/06/16 19:49, Grant emailgr...@gmail.com wrote: Does this behavior make sense? Is an oprofile report necessary here? Without oprofile, we can only guess. You've been complaining about CPU usage for 5 weeks now, and still no oprofile, although that was the first thing I asked for. -- Crystal Reports - New Free Runtime and 30 Day Trial Check out the new simplified licensing option that enables unlimited royalty-free distribution of the report engine for externally facing server and web deployment. http://p.sf.net/sfu/businessobjects ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] Best Sinc Interpolator
I've got Best working after switching from oss to alsa in audio_output.
I cannot imagine how OSS or ALSA could have any impact on the
resampler. Can you give me an oprofile report of both configurations?
I'm curious who is really wasting all those CPU cycles.
I've installed oprofile. Can you tell me how you'd like it to be run
to get the info you need?
Start playing in MPD, start oprofile, wait a minute, then stop
oprofile. Generate a report with opreport -g --symbols. Clear the
oprofile data, and repeat with OSS.
If you can, specify a vmlinux image when starting oprofile, because
that allows oprofile to list details within the kernel.
Max
I've moved upsampling duties from mpd to dmix with the following in
/etc/asound.conf:
defaults.pcm.rate_converter samplerate_best
pcm.!default {
type plug
slave.pcm {
type dmix
ipc_key 1024
slave {
pcm hw:0,0
format S24_3LE
rate 96000
}
}
}
but the mpd process still uses almost 100% CPU according to top.
Should the mpd process be the one using all the CPU even though dmix
is the one resampling? Another confusing thing is I tried playing the
same file with mplayer and almost no CPU is used although it is also
going through dmix.
Does this behavior make sense? Is an oprofile report necessary here?
- Grant
--
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensing option that enables unlimited
royalty-free distribution of the report engine for externally facing
server and web deployment.
http://p.sf.net/sfu/businessobjects
___
Musicpd-dev-team mailing list
Musicpd-dev-team@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] Best Sinc Interpolator
how 'bout upsampling to 48KHz? It should sound significantly better than 44.1KHz. No, every resampling process reduces the (theoretical) quality of the file. Now if a 44.1 file upsampled to 96 sounds better than native 44.1 playback, then that must be a hardware quirk. Maybe the sound chip has a fixed sample rate, and plugs a poor hardware resampler in front of it. I think so too. The manufacturer actually suggested I upsample on the computer. - Grant -- OpenSolaris 2009.06 is a cutting edge operating system for enterprises looking to deploy the next generation of Solaris that includes the latest innovations from Sun and the OpenSource community. Download a copy and enjoy capabilities such as Networking, Storage and Virtualization. Go to: http://p.sf.net/sfu/opensolaris-get ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] Best Sinc Interpolator
Upsampling works, but only multiples of the native source - ie 44.1 to 96 the computer is not powerful enough. however 44.1 to 88.2 or 176.4 works also. Don't even try Secret Rabbit Code upsampling in Foobar, would stall right from the first second with the processor at 100%!! http://www.computeraudiophile.com/content/24192-USB-20-beer-budget-EMU-0404-mini-review?page=1 This was written about Windows of course, but it describes my same problem. 44.1-96 is too much for the CPU. Are you sure this isn't normal? That *might* be true for MPD, too, but we don't know currently. My point was: even if libsamplerate eats all your CPU, then switching between OSS and ALSA shouldn't make any difference. I'm curious what's the real cause of this. OK, I'll try to get that report together. Why are you resampling to 96 kHz anyway? My DAC sounds much better with libsamplerate upsampling than sending 16/44.1 to it. Upsampling to 88.1 is a lot easier on the CPU, but it doesn't sound quite as good. - Grant -- OpenSolaris 2009.06 is a cutting edge operating system for enterprises looking to deploy the next generation of Solaris that includes the latest innovations from Sun and the OpenSource community. Download a copy and enjoy capabilities such as Networking, Storage and Virtualization. Go to: http://p.sf.net/sfu/opensolaris-get ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] Best Sinc Interpolator
On 2009/06/03 22:08, Stefan Monnier monn...@iro.umontreal.ca wrote: how 'bout upsampling to 48KHz? It should sound significantly better than 44.1KHz. No, every resampling process reduces the (theoretical) quality of the file. Now if a 44.1 file upsampled to 96 sounds better than native 44.1 playback, then that must be a hardware quirk. Maybe the sound chip has a fixed sample rate, and plugs a poor hardware resampler in front of it. Max -- OpenSolaris 2009.06 is a cutting edge operating system for enterprises looking to deploy the next generation of Solaris that includes the latest innovations from Sun and the OpenSource community. Download a copy and enjoy capabilities such as Networking, Storage and Virtualization. Go to: http://p.sf.net/sfu/opensolaris-get ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] Best Sinc Interpolator
Maybe the sound chip has a fixed sample rate, and plugs a poor hardware resampler in front of it. That's what I assumed, yes, otherwise his report makes no sense. Stefan -- OpenSolaris 2009.06 is a cutting edge operating system for enterprises looking to deploy the next generation of Solaris that includes the latest innovations from Sun and the OpenSource community. Download a copy and enjoy capabilities such as Networking, Storage and Virtualization. Go to: http://p.sf.net/sfu/opensolaris-get ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] Best Sinc Interpolator
On 2009/05/29 17:36, Grant emailgr...@gmail.com wrote: Upsampling works, but only multiples of the native source - ie 44.1 to 96 the computer is not powerful enough. however 44.1 to 88.2 or 176.4 works also. Don't even try Secret Rabbit Code upsampling in Foobar, would stall right from the first second with the processor at 100%!! http://www.computeraudiophile.com/content/24192-USB-20-beer-budget-EMU-0404-mini-review?page=1 This was written about Windows of course, but it describes my same problem. 44.1-96 is too much for the CPU. Are you sure this isn't normal? That *might* be true for MPD, too, but we don't know currently. My point was: even if libsamplerate eats all your CPU, then switching between OSS and ALSA shouldn't make any difference. I'm curious what's the real cause of this. Why are you resampling to 96 kHz anyway? Max -- OpenSolaris 2009.06 is a cutting edge operating system for enterprises looking to deploy the next generation of Solaris that includes the latest innovations from Sun and the OpenSource community. Download a copy and enjoy capabilities such as Networking, Storage and Virtualization. Go to: http://p.sf.net/sfu/opensolaris-get ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] Best Sinc Interpolator
On Wed, May 13, 2009 at 9:30 AM, Max Kellermann m...@duempel.org wrote: On 2009/05/12 21:20, Grant emailgr...@gmail.com wrote: I've got Best working after switching from oss to alsa in audio_output. I cannot imagine how OSS or ALSA could have any impact on the resampler. Can you give me an oprofile report of both configurations? I'm curious who is really wasting all those CPU cycles. I still need to figure out oprofile, but I read this: Upsampling works, but only multiples of the native source - ie 44.1 to 96 the computer is not powerful enough. however 44.1 to 88.2 or 176.4 works also. Don't even try Secret Rabbit Code upsampling in Foobar, would stall right from the first second with the processor at 100%!! http://www.computeraudiophile.com/content/24192-USB-20-beer-budget-EMU-0404-mini-review?page=1 This was written about Windows of course, but it describes my same problem. 44.1-96 is too much for the CPU. Are you sure this isn't normal? - Grant -- Register Now for Creativity and Technology (CaT), June 3rd, NYC. CaT is a gathering of tech-side developers brand creativity professionals. Meet the minds behind Google Creative Lab, Visual Complexity, Processing, iPhoneDevCamp as they present alongside digital heavyweights like Barbarian Group, R/GA, Big Spaceship. http://p.sf.net/sfu/creativitycat-com ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] Best Sinc Interpolator
On 2009/05/14 17:36, Grant emailgr...@gmail.com wrote: I've got Best working after switching from oss to alsa in audio_output. I cannot imagine how OSS or ALSA could have any impact on the resampler. Can you give me an oprofile report of both configurations? I'm curious who is really wasting all those CPU cycles. I've installed oprofile. Can you tell me how you'd like it to be run to get the info you need? Start playing in MPD, start oprofile, wait a minute, then stop oprofile. Generate a report with opreport -g --symbols. Clear the oprofile data, and repeat with OSS. If you can, specify a vmlinux image when starting oprofile, because that allows oprofile to list details within the kernel. Max -- Crystal Reports - New Free Runtime and 30 Day Trial Check out the new simplified licensing option that enables unlimited royalty-free distribution of the report engine for externally facing server and web deployment. http://p.sf.net/sfu/businessobjects ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
