Experiment: better sound on remote end

Al Johnson openmoko at mazikeen.demon.co.uk
Wed Dec 2 19:35:21 CET 2009


On Wednesday 02 December 2009, rms at 1407.org wrote:
> Experiment: better sound on remote end
> http://blog.1407.org/2009/12/02/experiment-better-sound-on-remote-end/
> 
> One of the problems people complain about my calls from the Freerunner is
>  *bad*sound*. Lot's of background noise, static, etc. I'm not talking about
>  the familiar buzz problem, as I've had the buzz fix applied to my phone.
> 
> http://picasaweb.google.com/RuiMSeabra/DebConf9OpenMokoBuzzFixParty#
> 
> However what I found to fix this problem on
> my phone model (GTA02v5 + buzz fix) goes totally against the official
>  information.
> 
> http://wiki.openmoko.org/wiki/Neo_Freerunner_audio_subsystem#Alsamixer_chan
> nel_controls
> 
> So much so that there's at least one person who, from his very good
>  knowledge about this subject (Really! This is *not* sarcasm!), thinks this
>  fix is as good as putting some peanut butter over the microphone :)

The suggested default values on the wiki (48 on max) will cause distortion for 
a lot of people, but the procedure described below isn't bad if you set 
control 12 to 5. 

> http://wiki.openmoko.org/index.php?title=Talk:Neo_Freerunner_audio_subsyste
> m&oldid=77886
> 
> I know of at least one other person who claims to have better sound on
>  remote end from this fix, but I think I need to drop the challenge. Can
>  you try the current SHR Unstable, apply this fix and then report on the
>  mailing lists your results?

I still don't believe there's one mixer setting that suits everyone. If we had 
a working AGC we might have a better chance though.

> You can? Good! So here's the fix:
> 
> 	1. call someone who noticed your bad sound and check whether the sound is
>  still bad, then finish the call
> 	2. edit /etc/freesmartphone/alsa/default/gsmhandset
> 	3. change control 63 (it's also on line 63), so that it ends in 3 rather
>  than in 2: * Old value: 63:'Mic Sidetone Mux':1:2
> 	 * New value: 63:'Mic Sidetone Mux':1:3
> 	4. call that person again and check the result

Unless you've stumbled upon some settings that make the AGC do something 
useful, there's nothing to be gained by routing audio through the PGA that 
can't be achieved just by setting the gain controls correctly on the existing 
route. You could try fso-simplemixer.py though I've not checked to see if it 
still works with the changes in the new SHR-U yet. Those that tried it said it 
made adjustment much easier, and improved the audio quality in calls.
http://thread.gmane.org/gmane.comp.handhelds.openmoko.community/50771/focus=50802

> Alternatively, in one call, you can ssh into your phone and use alsamixer
>  to change the same control, but this change will not be persistent until
>  you do the steps above.
> 
> http://blog.1407.org/wp-content/uploads/2009/12/oldValue.png
> 
> http://blog.1407.org/wp-content/uploads/2009/12/newValue.png
> 
> So let's hear your results, nothing like a good empirical experiment.
> 
> In the following recordings (complete set here),
> 
> http://files.1407.org/openmoko/mic_sidemux_test/
> 
> I said «The quick brown fox jumps over the lazy dog» starting with Mic2,
>  then Right PGA, then Left PGA.
> 
> 	1. in perfect conditions;
>           
>  http://files.1407.org/openmoko/mic_sidemux_test/test1-perfect-conditions.w
> av 2. with tv making background noise (news);
>           
>  http://files.1407.org/openmoko/mic_sidemux_test/test2-with-tv.wav 3. 2. +
>  computer playing loud music («Te Quiero Puta» from Rammstein) at the same
>  time.
>  http://files.1407.org/openmoko/mic_sidemux_test/test3-with-tv-and-music.wa
> v

An interesting result. It suggests your max setting on control 48 is ok, and 
the distortion is introduced by having control 12 set to 7, boosting the 
signal by 6dB. Try dropping it to 5, and increasing control 5 if needed. 



More information about the community mailing list