Bug #1267, any news ?
Mike Montour
mail at mmontour.net
Sat Jul 12 06:25:27 CEST 2008
Julien Cassignol wrote:
> 1) My voice was too low, and there was a buzzing
> 2) There was a big echo
>
> I saw on trac that this bug was notified some time ago (#1267),
These are two different (although related) issues, and #1267 is only
about the echo. The buzzing is covered by #883, #1352, and possibly #938.
To recap, there are two "intended" audio paths (ignoring sidetone, which
is not relevant here):
1. your mouth -> Neo microphone -> GSM -> remote earpiece -> remote ear
2. remote mouth -> remote microphone -> GSM -> Neo earpiece -> your ear
The buzzing affects path 1:
1a. your mouth -> Neo microphone [+ RFI] -> GSM -> remote earpiece ->
remote ear
That is, interference from the GSM radio transmitter in the Neo is being
picked up on the same wire that is coming from the microphone. The level
of this interference is proportional to the strength of the transmitted
radio signal; if you are far from a tower (fewer "bars" on the received
signal) then your transmit power will be higher and the interference
will be louder.
The echo heard by the remote end follows the path:
3. remote mouth -> remote mic -> GSM -> Neo earpiece -> Neo mic -> GSM
-> remote earpiece -> remote ear
It seems that most of the coupling between the Neo earpiece and the Neo
mic is through the inside of the case, because if you cover the earpiece
hole with your finger (blocking the sound transmitted out of the case)
then the coupled audio becomes louder.
A complication with the "echo" case is that some component in the phone
network, or the phone on the other end, might be performing active echo
cancellation. I don't know enough about GSM telephony to be able to
comment on this point.
The mixer controls on the Neo (the only elements in this whole picture
that can be affected by user-controlled software) basically give you two
knobs to turn:
A. Gain on the "Neo mic -> GSM" path
B. Gain on the "GSM -> Neo earpiece" path
Adjusting "A" will affect the buzz. Adjusting "A" or "B" will affect the
echo. However, "A" will also affect the intended audio path 1 and "B"
will also affect the intended audio path 2. Therefore, the choice of the
"best" mixer settings will have to be a trade-off between making the
intended paths loud enough and the unintended paths quiet enough.
Whether or not it is possible to come up with a "good enough" set of
mixer settings is still an open question at this time.
More information about the openmoko-devel
mailing list