Force mpd to play at 41000 sample rate (high CPU usage issue)

shadyabhi · 2011-03-03 13:50:35

Playing a mp3 file with mpd causes high cpu usage but mplayer plays it without significant cpu usage.
Googling has brought be to the inference that its the issue of mpd playing with different settings (ie at higher sample_rate).
I added .asoundrc but then also mpd is not respecting it.

Basically, I want to play mpd with the same settings as mplayer plays.

My .asoundrc file:-

defaults.pcm.rate_converter "44100"

When I play a file with mplayer.

shadyabhi@archlinux-N210 ~ $ cat /proc/asound/card0/pcm0p/sub0/hw_params
access: RW_INTERLEAVED
format: S16_LE
subformat: STD
channels: 2
rate: 44100 (44100/1)
period_size: 1024
buffer_size: 16384
OSS format: S16_LE
OSS channels: 2
OSS rate: 44100
OSS period bytes: 4096
OSS periods: 16
OSS period frames: 1024
shadyabhi@archlinux-N210 ~ $

And when I have mpd running,

shadyabhi@archlinux-N210 ~ $ cat /proc/asound/card0/pcm0p/sub0/hw_params
access: MMAP_INTERLEAVED
format: S32_LE
subformat: STD
channels: 2
rate: 48000 (48000/1)
period_size: 1024
buffer_size: 8192
shadyabhi@archlinux-N210 ~ $

shadyabhi · 2011-03-03 14:09:29

Now, I have a real doubt. In the process of trying to solve the problem, I accidentally stopped alsa and then tried playing a mp3 from mplayer and I was able to play.

So, being able to play a mp3 without alsa DAEMON running, what does that mean? How do I have song enabled even when I have alsa stopped?

brebs · 2011-03-03 15:19:49

Note that you have "OSS ..." in your mplayer example. That's opensound, not ALSA, presumably.

A bit of quick debugging info:

zgrep SND /proc/config.gz | grep "="
lsmod | grep snd

skunktrader · 2011-03-03 15:34:20

The alsa "daemon" is simply a script that restores your mixer levels on startup and saves them on shutdown

shadyabhi · 2011-03-03 22:18:10

shadyabhi@archlinux-N210 ~ $ zgrep SND /proc/config.gz | grep "="
CONFIG_SND=m
CONFIG_SND_TIMER=m
CONFIG_SND_PCM=m
CONFIG_SND_HWDEP=m
CONFIG_SND_RAWMIDI=m
CONFIG_SND_JACK=y
CONFIG_SND_SEQUENCER=m
CONFIG_SND_SEQ_DUMMY=m
CONFIG_SND_OSSEMUL=y
CONFIG_SND_MIXER_OSS=m
CONFIG_SND_PCM_OSS=m
CONFIG_SND_PCM_OSS_PLUGINS=y
CONFIG_SND_SEQUENCER_OSS=y
CONFIG_SND_HRTIMER=m
CONFIG_SND_SEQ_HRTIMER_DEFAULT=y
CONFIG_SND_DYNAMIC_MINORS=y
CONFIG_SND_SUPPORT_OLD_API=y
CONFIG_SND_VERBOSE_PROCFS=y
CONFIG_SND_VMASTER=y
CONFIG_SND_DMA_SGBUF=y
CONFIG_SND_RAWMIDI_SEQ=m
CONFIG_SND_OPL3_LIB_SEQ=m
CONFIG_SND_EMU10K1_SEQ=m
CONFIG_SND_MPU401_UART=m
CONFIG_SND_OPL3_LIB=m
CONFIG_SND_VX_LIB=m
CONFIG_SND_AC97_CODEC=m
CONFIG_SND_DRIVERS=y
CONFIG_SND_DUMMY=m
CONFIG_SND_ALOOP=m
CONFIG_SND_VIRMIDI=m
CONFIG_SND_MTPAV=m
CONFIG_SND_MTS64=m
CONFIG_SND_SERIAL_U16550=m
CONFIG_SND_MPU401=m
CONFIG_SND_PORTMAN2X4=m
CONFIG_SND_AC97_POWER_SAVE=y
CONFIG_SND_AC97_POWER_SAVE_DEFAULT=0
CONFIG_SND_SB_COMMON=m
CONFIG_SND_SB16_DSP=m
CONFIG_SND_PCI=y
CONFIG_SND_AD1889=m
CONFIG_SND_ALS300=m
CONFIG_SND_ALS4000=m
CONFIG_SND_ALI5451=m
CONFIG_SND_ASIHPI=m
CONFIG_SND_ATIIXP=m
CONFIG_SND_ATIIXP_MODEM=m
CONFIG_SND_AU8810=m
CONFIG_SND_AU8820=m
CONFIG_SND_AU8830=m
CONFIG_SND_AW2=m
CONFIG_SND_AZT3328=m
CONFIG_SND_BT87X=m
CONFIG_SND_CA0106=m
CONFIG_SND_CMIPCI=m
CONFIG_SND_OXYGEN_LIB=m
CONFIG_SND_OXYGEN=m
CONFIG_SND_CS4281=m
CONFIG_SND_CS46XX=m
CONFIG_SND_CS46XX_NEW_DSP=y
CONFIG_SND_CS5530=m
CONFIG_SND_CS5535AUDIO=m
CONFIG_SND_CTXFI=m
CONFIG_SND_DARLA20=m
CONFIG_SND_GINA20=m
CONFIG_SND_LAYLA20=m
CONFIG_SND_DARLA24=m
CONFIG_SND_GINA24=m
CONFIG_SND_LAYLA24=m
CONFIG_SND_MONA=m
CONFIG_SND_MIA=m
CONFIG_SND_ECHO3G=m
CONFIG_SND_INDIGO=m
CONFIG_SND_INDIGOIO=m
CONFIG_SND_INDIGODJ=m
CONFIG_SND_INDIGOIOX=m
CONFIG_SND_INDIGODJX=m
CONFIG_SND_EMU10K1=m
CONFIG_SND_EMU10K1X=m
CONFIG_SND_ENS1370=m
CONFIG_SND_ENS1371=m
CONFIG_SND_ES1938=m
CONFIG_SND_ES1968=m
CONFIG_SND_ES1968_INPUT=y
CONFIG_SND_FM801=m
CONFIG_SND_FM801_TEA575X_BOOL=y
CONFIG_SND_FM801_TEA575X=m
CONFIG_SND_HDA_INTEL=m
CONFIG_SND_HDA_HWDEP=y
CONFIG_SND_HDA_RECONFIG=y
CONFIG_SND_HDA_INPUT_BEEP=y
CONFIG_SND_HDA_INPUT_BEEP_MODE=1
CONFIG_SND_HDA_INPUT_JACK=y
CONFIG_SND_HDA_PATCH_LOADER=y
CONFIG_SND_HDA_CODEC_REALTEK=y
CONFIG_SND_HDA_CODEC_ANALOG=y
CONFIG_SND_HDA_CODEC_SIGMATEL=y
CONFIG_SND_HDA_CODEC_VIA=y
CONFIG_SND_HDA_CODEC_HDMI=y
CONFIG_SND_HDA_CODEC_CIRRUS=y
CONFIG_SND_HDA_CODEC_CONEXANT=y
CONFIG_SND_HDA_CODEC_CA0110=y
CONFIG_SND_HDA_CODEC_CMEDIA=y
CONFIG_SND_HDA_CODEC_SI3054=y
CONFIG_SND_HDA_GENERIC=y
CONFIG_SND_HDA_POWER_SAVE=y
CONFIG_SND_HDA_POWER_SAVE_DEFAULT=0
CONFIG_SND_HDSP=m
CONFIG_SND_HDSPM=m
CONFIG_SND_HIFIER=m
CONFIG_SND_ICE1712=m
CONFIG_SND_ICE1724=m
CONFIG_SND_INTEL8X0=m
CONFIG_SND_INTEL8X0M=m
CONFIG_SND_KORG1212=m
CONFIG_SND_LX6464ES=m
CONFIG_SND_MAESTRO3=m
CONFIG_SND_MAESTRO3_INPUT=y
CONFIG_SND_MIXART=m
CONFIG_SND_NM256=m
CONFIG_SND_PCXHR=m
CONFIG_SND_RIPTIDE=m
CONFIG_SND_RME32=m
CONFIG_SND_RME96=m
CONFIG_SND_RME9652=m
CONFIG_SND_SONICVIBES=m
CONFIG_SND_TRIDENT=m
CONFIG_SND_VIA82XX=m
CONFIG_SND_VIA82XX_MODEM=m
CONFIG_SND_VIRTUOSO=m
CONFIG_SND_VX222=m
CONFIG_SND_YMFPCI=m
CONFIG_SND_SPI=y
CONFIG_SND_USB=y
CONFIG_SND_USB_AUDIO=m
CONFIG_SND_USB_UA101=m
CONFIG_SND_USB_USX2Y=m
CONFIG_SND_USB_CAIAQ=m
CONFIG_SND_USB_CAIAQ_INPUT=y
CONFIG_SND_USB_US122L=m
CONFIG_SND_PCMCIA=y
CONFIG_SND_VXPOCKET=m
CONFIG_SND_PDAUDIOCF=m
CONFIG_SND_SOC=m
CONFIG_SND_SOC_I2C_AND_SPI=m
shadyabhi@archlinux-N210 ~ $ lsmod | grep snd
snd_seq_dummy           1479  0 
snd_seq_oss            29240  0 
snd_seq_midi_event      5516  1 snd_seq_oss
snd_seq                50562  5 snd_seq_dummy,snd_seq_oss,snd_seq_midi_event
snd_seq_device          5281  3 snd_seq_dummy,snd_seq_oss,snd_seq
snd_pcm_oss            39509  1 
snd_mixer_oss          17730  1 snd_pcm_oss
snd_hda_codec_realtek   295589  1 
snd_hda_intel          21837  1 
snd_hda_codec          74609  2 snd_hda_codec_realtek,snd_hda_intel
snd_hwdep               6222  1 snd_hda_codec
snd_pcm                72321  3 snd_pcm_oss,snd_hda_intel,snd_hda_codec
snd_timer              19537  2 snd_seq,snd_pcm
snd                    58906  11 snd_seq_oss,snd_seq,snd_seq_device,snd_pcm_oss,snd_mixer_oss,snd_hda_codec_realtek,snd_hda_intel,snd_hda_codec,snd_hwdep,snd_pcm,snd_timer
soundcore               6161  2 snd
snd_page_alloc          7361  2 snd_hda_intel,snd_pcm
shadyabhi@archlinux-N210 ~ $

shadyabhi · 2011-03-03 22:19:06

skunktrader wrote:

The alsa "daemon" is simply a script that restores your mixer levels on startup and saves them on shutdown

I never really looked the init script of alsa. Now, I just opened it and figured out thats its just about saving and restoring alsa levels. Thanks.

Square · 2011-03-04 09:33:10

In your mpd.conf, add/edit your audio_output{} section to use a specific format and set it to what you want.

This should help you:
http://mpd.wikia.com/wiki/Configuration

shadyabhi · 2011-03-04 13:05:12

Square wrote:

In your mpd.conf, add/edit your audio_output{} section to use a specific format and set it to what you want.
This should help you:
http://mpd.wikia.com/wiki/Configuration

I already did it by un-commenting the OSS part in /etc/mpd.conf but how can I do the same for alsa?

brebs · 2011-03-04 13:20:53

Sounds like you're doing it the wrong way around. People should usually be setting things up for *ALSA*, because ALSA's what's in the kernel, and ALSA's emulation of OSS bypasses dmix.

As an example, here's my ~/.asoundrc

shadyabhi · 2011-03-04 13:30:28

brebs wrote:

Sounds like you're doing it the wrong way around. People should usually be setting things up for *ALSA*, because ALSA's what's in the kernel, and ALSA's emulation of OSS bypasses dmix.
As an example, here's my ~/.asoundrc

I used your .asoundrc and still when I play mpd, my output is still at 48000.
is it that mpd bypasses those settings. (Also, do I need to restart or something?)

Last edited by shadyabhi (2011-03-04 13:34:14)

Rasi · 2011-03-04 13:35:57

https://bbs.archlinux.org/viewtopic.php … 79#p897379

brebs · 2011-03-04 13:46:38

shadyabhi wrote:

I used your .asoundrc

Well, *look* at it first

Notice that it deliberately contains 48000.

I'm still not sure whether 44100 or 48000 is best. Listening with decent headphones, I've not been able to notice a difference between stereo music outputted at 44100, versus upmixed to 48000.

Rasi · 2011-03-04 14:05:46

brebs wrote:

shadyabhi wrote:
I used your .asoundrc
Well, *look* at it first
Notice that it deliberately contains 48000.
I'm still not sure whether 44100 or 48000 is best. Listening with decent headphones, I've not been able to notice a difference between stereo music outputted at 44100, versus upmixed to 48000.

thats easy to answer. if you listen mostly to music, you should use 44100. If you mainly watch movies use 48000 - its all about resampling, which uses quite some CPU

Last edited by Rasi (2011-03-04 14:06:09)

shadyabhi · 2011-03-04 14:14:29

Rasi wrote:

https://bbs.archlinux.org/viewtopic.php … 79#p897379

After following your configs, my cpu usage is down to between 3% and 5%. And also, now I can see 41000 in "/proc/asound/card0/pcm0p/sub0/hw_params".

But, you said in the post that yours is 1% so I think there is still some issue.

Last edited by shadyabhi (2011-03-04 14:18:26)

Rasi · 2011-03-04 14:19:27

shadyabhi wrote:

Rasi wrote:
https://bbs.archlinux.org/viewtopic.php … 79#p897379
After following your configs, my cpu usage is down to between 3% and 5%. And also, now I can see 41000 in "/proc/asound/card0/pcm0p/sub0/hw_params".
But, you said in the post that yours is 1% so I think there is still some issue.

i dont think so... 3-5% sounds perfectly fine. decoding needs some cpu too. And i have an core-i7

Last edited by Rasi (2011-03-04 14:20:05)

brebs · 2011-03-04 14:24:09

Rasi wrote:

resampling, which uses quite some CPU

Only about 1%, compared to my rough test. 4-5% CPU usage according to top, listening to audacious with my pcm.headphones at 44100, versus 5-6% CPU when at 48000 (using converter "samplerate_medium").

This rises to 9-10% with samplerate_best, but medium (or less) is perfectly fine.

What I'm more interested in, is something that just works, with music, games, movies, and even Skype.

shadyabhi · 2011-03-04 14:26:14

Rasi wrote:

i dont think so... 3-5% sounds perfectly fine. decoding needs some cpu too. And i have an core-i7

Ya. May be.. Things are more noticeable on Intel Atoms than on Core-i7.

Rasi · 2011-03-04 15:22:12

brebs wrote:

Rasi wrote:
resampling, which uses quite some CPU
Only about 1%, compared to my rough test. 4-5% CPU usage according to top, listening to audacious with my pcm.headphones at 44100, versus 5-6% CPU when at 48000 (using converter "samplerate_medium").
This rises to 9-10% with samplerate_best, but medium (or less) is perfectly fine.
What I'm more interested in, is something that just works, with music, games, movies, and even Skype.

interesting. i consider my hardware more or less state of the art and resampling still takes around 20-30%

considering your 2nd question: my asound.conf works with all stuff i tried by now. it includes a plug for skype to make it work.

Last edited by Rasi (2011-03-04 15:23:24)

TigTex · 2011-03-04 16:59:40

I've disabled dmix on my system to reduce cpu usage with this .asoundrc

pcm.!default {
 type plug
 slave.pcm hw
 }

Works, it disables resampling, but you can only play 1 sample rate at the same time. (you can't play a 44.1khz mp3 with a 48khz video on background for example).
You can use this if you really want the lowest possible cpu usage.

shadyabhi · 2011-03-04 17:06:47

TigTex wrote:

I've disabled dmix on my system to reduce cpu usage with this .asoundrc
Works, it disables resampling, but you can only play 1 sample rate at the same time. (you can't play a 44.1khz mp3 with a 48khz video on background for example).
You can use this if you really want the lowest possible cpu usage.

Thats not really possible for me because if I disable dmix, I wont hear two sounds at once. For ex, if I am listening to songs, I wont be able to hear any gtalk notifications which I am getting in browser window.

TigTex · 2011-03-04 17:21:10

that's what i've said... but you could use it just to see if you can get your cpu usage lower than 4-5%, just for testing. I get 0-1% on a pentium4...

Arch Linux

#1 2011-03-03 13:50:35

Force mpd to play at 41000 sample rate (high CPU usage issue)

#2 2011-03-03 14:09:29

Re: Force mpd to play at 41000 sample rate (high CPU usage issue)

#3 2011-03-03 15:19:49

Re: Force mpd to play at 41000 sample rate (high CPU usage issue)

#4 2011-03-03 15:34:20

Re: Force mpd to play at 41000 sample rate (high CPU usage issue)

#5 2011-03-03 22:18:10

Re: Force mpd to play at 41000 sample rate (high CPU usage issue)

#6 2011-03-03 22:19:06

Re: Force mpd to play at 41000 sample rate (high CPU usage issue)

#7 2011-03-04 09:33:10

Re: Force mpd to play at 41000 sample rate (high CPU usage issue)

#8 2011-03-04 13:05:12

Re: Force mpd to play at 41000 sample rate (high CPU usage issue)

#9 2011-03-04 13:20:53

Re: Force mpd to play at 41000 sample rate (high CPU usage issue)

#10 2011-03-04 13:30:28

Re: Force mpd to play at 41000 sample rate (high CPU usage issue)

#11 2011-03-04 13:35:57

Re: Force mpd to play at 41000 sample rate (high CPU usage issue)

#12 2011-03-04 13:46:38

Re: Force mpd to play at 41000 sample rate (high CPU usage issue)

#13 2011-03-04 14:05:46

Re: Force mpd to play at 41000 sample rate (high CPU usage issue)

#14 2011-03-04 14:14:29

Re: Force mpd to play at 41000 sample rate (high CPU usage issue)

#15 2011-03-04 14:19:27

Re: Force mpd to play at 41000 sample rate (high CPU usage issue)

#16 2011-03-04 14:24:09

Re: Force mpd to play at 41000 sample rate (high CPU usage issue)

#17 2011-03-04 14:26:14

Re: Force mpd to play at 41000 sample rate (high CPU usage issue)

#18 2011-03-04 15:22:12

Re: Force mpd to play at 41000 sample rate (high CPU usage issue)

#19 2011-03-04 16:59:40

Re: Force mpd to play at 41000 sample rate (high CPU usage issue)

#20 2011-03-04 17:06:47

Re: Force mpd to play at 41000 sample rate (high CPU usage issue)

#21 2011-03-04 17:21:10

Re: Force mpd to play at 41000 sample rate (high CPU usage issue)

Board footer