title says it all

apparently nokia had to implement software eq for speaker playback to allow higher volume without distortion "d case-resonance.

this puts 15-25 percent cpu load on while playing any audio.

this really hurts games performance, and i dont know if it hurts battery life on mp3 playback.

please fix speakers in hardware next time and allow disabling of eq now (even if that means a volume limit)

Feel free to file a bug report. The only at all related one at the moment is bug 5794.

What about disabling pulseaudio? Is it used by telephony apps?

What if you plugin a headset? It is my understanding that this will disable the speaker protection algorithms.

(As mentioned by Igor Stoppain in this discussion about N900 power usage with Ogg and MP3)

With headset usage drops a good deal.

I just discovered a very cool thing. It is possible to bypass pulse, but that would temporarily prevent any other app from sending sound.

Would you like the option for faster emulator, even if it blocks the phone, for e.g.?

Wouldn't it be possible to remove the relevant eq module?

Ok now I suddenly have a battery issue - didn't before. And I see that pulseaudio is using 16-18 % CPU, which must be the cause - it's constant. And I have not started the media player, or anything else that plays anything. So why is pulseaudio spending CPU?

If I plug in headphones it drops to 9-11%. There's no sound of course, because there's nothing playing! If I unplug the headphones the CPU goes back to 16-18% again. Something is wrong here.

No, there is nothing wrong. When sound goes over loudspeaker, an additional hardware protection algorithm chimes in. But if you use a headset, the algorithm/filter is not needed.

I also had older version of my mClock hogging PulseAudio, because PyGame audio is enabled in PyGame by default, and the mixer makes a mess. So if you use any PyGame-based app, take care.

he is seeing cpu use without telling device to play audio

i never saw that.

disabling eq wouldnt give much in terms of speed. it is the process itself that hurts emus with unnecessary context switching, imo

@twaelti: If you was replying to me - it's as Manuel says: Nothing playing, so no reason for pulseaudio to be active at all (well it's running all the time, but it should be _idle_, as it normally is when there's no application using sound).

In the meantime I have looked at all active processes and found nothing that should trigger this. So I killed pulseaudio, it then restarted (as it should), but now in idle mode, no CPU misbehaviour. There's something wrong here.