Sudhakar,
Will this fix channel swapping that occurs *many* times during a
recording or just initially when the recording starts? Either way it
needs to be fixed.
I also found out why I wasn't getting stereo - I'm using a different ASP
than the EVM and it doesn't have a CLKS input and that is the default
clk setup. Once I changed it to use CLKR, I got my stereo. I'm actually
amazed that I was getting anything at all with the way it was configured.
Thanks, I'll take a look at this and see if it helps.
Steve
Rajashekhara, Sudhakar wrote:
> Steve,
>
> Along with this mail, I am sending following 2 patches:
>
> tx_swap.patch: This patch fixes the channel swap issue on the TX side
> (play).
> rx_swap.patch: This patch fixes the channel swap issue on RX side
> (record).
>
> In both the patches, the frame sync pins are being configured as GPIO
> and the first frame sync is being ignored. Only after the first frame
> sync, the pins are being configured as Audio pins.
>
> Please note that, this patch is for Linux 2.6.10 MV Kernel.
>
> Regards, Sudhakar
> ------------------------------------------------------------------------
> *From:* [email protected]
> [[email protected]] On Behalf Of
> Stephen Berry [[email protected]]
> *Sent:* Wednesday, February 11, 2009 2:16 AM
> *To:* [email protected]
> *Subject:* stereo audio from aic33 (ASP help)
>
>
> I'm having trouble getting stereo data from the AIC33 - well to be
> more correct, the codec seems to be supplying stereo data, but the ASP
> seems to be generating mono audio. And just to top it off, it looks
> like the channel that is active seems to swap randomly.
>
> MVL 2.6.10 - custom hardware (using ASP0 not 1)
>
> Here is what I've seen so far:
>
> The ASP setup seems to be looking for two phases of data per frame
> - but the bit that enables phases wasn't set. Setting this didn't help.
> The second channel of audio is generally midline (0).
> DMA seems to be setup to grab a 32bit word from the interface (DRR
> register for received data)
> With the codec in master mode I've observed on the scope that
> there is truly stereo data being fed to the AIC33
> The encoder demo uses a nice averaging trick that hides the
> channel swap beautifully.
>
> I'm unsure of the logic the ASP uses to determine when to tell the DMA
> engine it is ready with data. I also don't know what happens with more
> than one "phase" of data. Are there two DMA transfers or one packed 32
> bit word?
>
> Any help from some ASP experts would be appreciated.
>
> Thanks,
>
> Steve Berry
>
_______________________________________________
Davinci-linux-open-source mailing list
[email protected]
http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source
