Need help for ASRC lock noise and GPIO output

Hello Power Pan,

I had worked up this project several years ago to monitor the ASRC lock bit and mute the audio when it is not locked.

The ASRC can be setup to mute when the sample rate is changed and that works fairly well. The issues are when clocks are lost or when they have been lost for a while then you relock the audio in the ASRC memory is old. When you power up it is random numbers so that is also a problem.

You will have to adapt this program to respond to the correct ASRC bit for the ASRC that is being used.

I attached two versions. A simpler version that only uses a MUX to switch that has no slew. The second one uses a slewed volume control and it also has a delay so the ASRC lock detect will be detected before the audio has a chance to make any noises.

You can also export this lock detect control signal over to a GPIO if you like. But it is best to just let the DSP handle it.

ADAU1452 ASRC Lock Detect to Mute Simple Example.dspproj

4213.ADAU1452 ASRC Lock Detect Example.zip

Dave T

Hello David

many thanks for your kind and prompt reply, will try and send the feed back here ASAP.

actually, I don't really need the lock bit to export when the ASRC auto mute could work properly.

Best regards,

Power Pan

Tested, and it works very well. but occasionally it will unlock unexpectedly, don't know why.

Hello PowerPan,

I suggest you look at the waveforms on a scope. The BCLK, LRCLK and SDATA and see if the waveform has signal integrity issues. This can cause the ASRC to think the sampling rate has changed and trigger the mute.

This post might help you in setting up the scope properly and what to expect.

(+) How to Take Meaningful Screenshots of I2S Audio Signals - Q&A - Audio - EngineerZone (analog.com)

So that is the first place to look.

One question: What is the sampling rate that is coming into the ASRC?

Dave T

when set DSP READ module as continuous read back, sometimes it goes wrong. if use manual click to read, it's fine.

Hello PowerPan,

You may be having some issues with the SPI signal when reading or with USB delays. I suggest you setup a counter in the DSP program to count the number of times you lose lock. If these events are happening quickly then you can miss it easily with a manual read or even a continuous read. That only happens every 10ms which is a long time to the DSP.

I will attach a program that shows a few different methods for counting pulses. You can do something like this and see what you find.

6013.ADAU1452 GPIO Pulse length and pulse toggle counter.dspproj

Also, you could output the result of the ASRC lock bit over to a GPIO and watch it on a scope. These are all real time methods to monitor what is going on in the DSP rather than a readback. A readback is not very good for possibly sample length events.

Dave T

Hi Dave

thanks for your suggestion. I'm using I2C, not SPI. after I stop continuous read, everything goes fine. that's why I reply so late cause I keep testing this for at lease 24 hours.

I might not get into something that's hard to catch. cause the problem is solved. so many works line up to finish.

Best regards,

Power Pan