FreeCalypso update: progress with DSP patches and voice calls

[Mychaela Falconia]

Hi DS,

> (btw note there is a slight error
> below, one should use gsmhandset.state not gsmheadset.state).

Oops, my apologies, didn't proofread before posting.

> I also
> tried setting the volume from 115 to 127 but on my particular GTA02
> it seems to not work as expected,

Hmm, it seems to work fine on mine.

> because only a loud piercing noise
> is made.

Hmm - what you are describing sounds like what I got when I tried
enabling the loudspeaker (gsmspeakerout.state) - it must not be
enabled at the same time as the debug serial channel, but I tried
making a voice call in the loudspeaker mode with the debug serial
channel disabled, using a trivial terminal program on the GTA02 AP to
issue the AT commands.  I was expecting to hear the call audio in the
loudspeaker, but I got a loud squeal instead.

> I also use AT+CLVL=255 before the call;

Yes, I do likewise.

> the sound was a bit muffled but otherwise ok.

One of the many many things that will get better when we get our own
FCDEV3B boards built and start using them instead of Openmoko devices
is the process for exercising voice calls.  There will be no Linux AP
and no ALSA scenarios to mess around with, instead there will be a
loudspeaker amplifier circuit on the board, connected directly to the
"earpiece" output from the Iota ABB.

When I exercised voice calls on the D-Sample board, using the 20020917
firmware it came flashed with, I could easily hear the downlink audio
while sitting in a chair next to the lab bench with the board, without
lifting any of the hardware off the bench and even without lifting the
D-Sample handset part out of its cradle - the D-Sample has an 8 ohm
speaker, physically located in the handset part which connects to the
main board via a round cable.

For our FCDEV3B I copied the loudspeaker circuit from TI's Leonardo
schematics; I assume that D-Sample and Leonardo must be similar in
this regard, so I expect the audio exercising capabilities on our
board to be no worse than on the D-Sample.  Our dev kit won't have the
fancy handset part like on TI's kits (too expensive), but putting the
loudspeaker directly on the board itself should be just as good, given
that there is no need to hold it up to one's ear.

M~