How to passthrough mic to headphone?

[OevreFlataeker]

Hi,

thanks for your great work which is really a kickstart in working with the nRF5340 DK. I spent the last couple of hours playing with the various settings and understand how it all interrelates.

I know want to simply passthrough the audio from the board's PDM microphone right to the headphone out.

What I have for line-in passthrough is:

	// Enable line-in
	{ CS47L63_INPUT2_CONTROL1, 0x00050020 },/* MODE=analog */ 
	{ CS47L63_IN2L_CONTROL1, 0x10000000 },  /* SRC=IN2LP */
	{ CS47L63_IN2R_CONTROL1, 0x10000000 },  /* SRC=IN2RP */
	{ CS47L63_INPUT_CONTROL, 0x0000000C },  /* IN2_EN=1 */
	// Set volume for line-in
	{ CS47L63_IN2L_CONTROL2, 0x00800080 },  /* VOL=0dB, MUTE=0 */
	{ CS47L63_IN2R_CONTROL2, 0x00800080 },  /* VOL=0dB, MUTE=0 */
	{ CS47L63_INPUT_CONTROL3, 0x20000000 }, /* VU=1 */
        // Map to output
        { CS47L63_OUT1L_INPUT3, (0x2B  << CS47L63_OUT1LMIX_VOL3_SHIFT) | (0x012 << CS47L63_OUT1L_SRC3_SHIFT)    // 0x12=IN2L },
	{ CS47L63_OUT1L_INPUT4, (0x2B  << CS47L63_OUT1LMIX_VOL4_SHIFT) | (0x013 << CS47L63_OUT1L_SRC4_SHIFT)    // 0x13=IN2R },

and from the Nordic's audio demo app I borrowed this for enabling the onboard PDM and route it to the headphone as well
(I've commented the line-in code above so it's no longer used)

        // Enable digital MIC
	/* Set MICBIASes */
	{ CS47L63_LDO2_CTRL1, 0x0005 },
	{ CS47L63_MICBIAS_CTRL1, 0x00EC },
	{ CS47L63_MICBIAS_CTRL5, 0x0272 },
	/* Enable IN1L */
	{ CS47L63_INPUT_CONTROL, 0x000F },
	/* Enable PDM mic as digital input */
	{ CS47L63_INPUT1_CONTROL1, 0x50021 },
	/* Un-mute and set gain to 0dB */
	{ CS47L63_IN1L_CONTROL2, 0x800080 },
	{ CS47L63_IN1R_CONTROL2, 0x800080 },
	/* Volume Update */
	{ CS47L63_INPUT_CONTROL3, 0x20000000 },
        // Map to output
        { CS47L63_OUT1L_INPUT3, (0x2B  << CS47L63_OUT1LMIX_VOL3_SHIFT) | (0x010 << CS47L63_OUT1L_SRC3_SHIFT)    // 0x10=IN1L },
	{ CS47L63_OUT1L_INPUT4, (0x2B  << CS47L63_OUT1LMIX_VOL4_SHIFT) | (0x011 << CS47L63_OUT1L_SRC4_SHIFT)    // 0x11=IN1R },

However I don't hear anything when I talk close to the microphone which appears to be in between of the two 3.5mm jacks?
Maybe that's a stupid question but I do NOT have to connect anything, the mic is right there on the board, right?

I've also mixed in the sine tone you generate with i2s and it is also audible in both configurations (code not included above), so I know the output itself is working.

Did I miss anything for enabling the PDM output? I've literally only added the lines above to an already working config of a noise/tone generator + line-in app. Maybe some more clocks needed?

I also have the WISCE software from Cirrus installed and am able to live manipulate the codec. Maybe you could also tell me which option to activate where so it should work? Independent from what the application does I mean...

[OevreFlataeker]

What's interesting is that if I blow gently against the area where the PDM is located I can hear the noise in the headset. But when I just "talk to the mic" nothing can be heard?

[ace-johnny]

I'm glad to hear that you found my example useful. I originally uploaded this repo to assist another beginner user on Nordic's DevZone forum, and they quickly made much more progress with the Audio DK than I have so far. They've even posted several repos of their own, including one that seemingly exactly satisfies your question: nrf5340_i2s_echo.

Additionally, I'm curious how you're able to control the codec in real-time with WISCE. I've installed that software and its requisite CS47L63 package, and am able to conveniently generate register settings, but was under the impression that additional firmware was required to directly liaison with the PC. I'd appreciate learning how you got that working.

Sorry that I can't provide a more straightforward definitive answer, but I'm still studying the details as well. I hope this information at least helps.

[OevreFlataeker]

Well it was straight forward:
Install the WISCE app from there: https://www.cirrus.com/support/wisce/
Install the "driver" for the CS47L63 from here: https://www.cirrus.com/products/cs47l63/ -> Software & Tools -> CS47L63 Setup
The 2nd install packages seems to put the CS47L63 necessary files into the devices subdir. Without this package it won't work.
Now just connect the DK, power it and start the app. It should show you all the components and let's you interact. No idea how it works, but it seems the FTDI chip proxies the access live to the codec via SPI. Playing around with the tool showed me how some of the constants are composed and helped me better understand the overall process.

You CAN also program the DSP there, but you'll already need a FW file for this and I'm not sure how to create that one.

[OevreFlataeker]

Thanks for the other repo, which indeed will be helpful though it looks quite a lot more complicated right away on first side. I'll give it a try. i2s processing is new to me but I'll see if I can get it working and take it from there towards my overall use case. Maybe that project also avoids the problem I encountered doing it the naive way as documented here: https://devzone.nordicsemi.com/f/nordic-q-a/118551/output-of-my-audio-dk-is-noisy-clipping-when-coming-via-i2s

[ace-johnny]

Yes, as I previously mentioned, I've already installed WISCE and its CS47L63 package, but did not notice that the settings were live since I did not have any audio routed through it. I need to revisit and experiment with that further.

WISCE was also very illustrative for me when composing register configurations, and I too was hopeful that I could use it to program the DSP peripheral, but I've exhaustively searched for existing firmware binaries or creation techniques with no luck whatsoever, and Cirrus Logic has been less than responsive.

The nrf5340_i2s_echo repo I suggested is based on the nRF Connect SDK's behemoth BLE nrf5340_audio sample application, and is indeed a challenging study due to its overwhelming complexity, which was the entire impetus for creating my much simpler hello_codec example when getting started.

I'll respond to your DevZone ticket questions there, so that our conversation is more likely to be seen and participated in by other experienced users.

[OevreFlataeker]

Re: WISCW. Yes, I too was astonished that my changes actually had a live effect on the audio result. For verifying that it works, just load your bootstrap app and run the code until you enable the audio output. You should then be interactively able to turn on the test tone generator, noise, etc.

It seems the DK somehow passes the discovery requests directly to the codec. I don't understand yet how this works as I thought we would just have the usual Jlink to interface the PC.

I think i found SOME reference to programming the DSP the other day but this is totally of limits for me right now. First of all I need to get the basics right which for me means so far to get crisp audio from the various sources and more important understand where we have which channel. As you saw in my Devzone ticket even this is not yet clear for me. I agree the info on the Cirrus side apart from the datasheet is very sparse, and you don't find proper examples anywhere... Let me try to dig out again what I think I found regarding the DSP.