[RaspberryPi] audio clicks

[Dmole]

Raspberrypi
The audio clicks pops on AV start stop
The sound is perfect for blowing a speaker if the volume is cranked

[olivier8509]

I would also add that sound by HDMI is also clicking sometimes. It is not smooth.

[mdjong]

I'm also hearing clicks while playing sound by HDMI. The clicks are on the same time when I rewind and play again.

[Kampie]

Using: 12.0-ALPHA3 Git:39c680e
I am also hearing click/pops when playing audio only (mp3/flac). A bit like playing a scratched record.
Also i think ReplayGain is not working or somehow RPi hdmi sound is way louder then from my ION hdmi.

[Joshfindit]

When the Raspberry 'clicks', to me this sounds like resetting the hardware before sending audio data to it.
Sound due to voltage spike during this reset.

Since I have no way to measure and confirm, this is only an impression based on years of previous experience with audio gear.

[Dmole]

yah that's exactly what is sounds like, only most systems I have used do that only on boot or shutdown. So why is the rpi doing it before and after every video, and seemingly at random during the navigation.
need to find that line of code and comment it out.

[Dmole]

grep -R -i reset . | grep -i rpi

found this:

./packages/linux/patches/linux-3.2.23-601-RPi_support-f23da7f.patch

which among other things patches this:

linux-3.2.23/sound/arm/bcm2835.c

which contains this:

+/* HACKY global pointers needed for successive probes to work : ssp

• * But compared against the changes we will have to do in VC audio_ipc code
• * to export 8 audio_ipc devices as a single IPC device and then monitor all
• * four devices in a thread, this gets things done quickly and should be easier
• * to debug if we run into issues
• */

so maybe that's a place to start

[2226cc]

Dmole, this does seem like a classic DC problem on the audio output. I was thinking of popping some DC blocking cap on the analogue audio output. If the issue stops then to me that says it is definitely a DC issue most likely because audio hardware/amp is being toggled during pause/resume. Don't know why that would be getting done at the driver level.

[Joshfindit]

As a semi-wild guess, possibly it's opening the driver before checking if it should be playing navigation sounds, resulting in the pop.

I'm very curious to see the the differences in this issue and the downsampling issue when Raspbmc RC4 release comes out: there may be a good amount of work that can be shared :)
Edit: Just ran across this: http://forum.stmlabs.com/showthread.php?tid=202 "Fixed in rc4"

[2226cc]

Would be nice if USB audio devices were supported.

[Dmole]

I like the Raspbmc builds more because it's not as stripped down (still has apt) and it does not use 100% cpu all the time, but at the same time it's a bit backwards using private svn hosting instead of github, Raspbmc claims that this bug is fixed but there is no build to prove it yet... neither project has had a nightly build in 3 weeks, so good things might come with the next one.

[sraue]

@Dmole if OpenELEC is to "stripped down" then no one stops you from using other distros. thats no issue, OpenELEC is a embedded OS and not a desktop distro!

[sraue]

If this issue is still present with our latest builds please report XBMC related bugs to http://trac.xbmc.org because RPi support is now merged to XBMC officially.

[Dmole]

I would rather not use trac. This is still an issue. Maybe I'll add it to the rpi kernel~

[tohojo]

There's a reference to an updated firmware version at the issue linked in the previous comment, which seems to fix the clicking sound issue. Any chance of having that update folded into openelec? Or is updating the firmware something that needs to be done outside of openelec?

Edit: Seems the clicking issue is gone in the RC5 build. Yay! :)