Featured Articles
Announcing Pi
Technical Specifications
Raspberry Pi FAQs
Win a Pi
Raspberry Pi Wishlist
Actions
Forum Thread Details
  • Replies 2 replies
  • Subscribers 665 subscribers
  • Views 606 views
  • Users 0 members are here
Related

Using Cirrus Logic Audio Card for PDM Signal?

Former Member

Hello,

 

Newbie disclaimer:

I am working on my first audio project, which is to manipulate some ultrasound signals coming from a MEMS microphone via the RPI and playing them back again in real time at audible frequencies, and I am learning a lot as I go along, so I may make obvious oversights in this post. With that said, here is my question:

 

Is there any way for the Cirrus Logic Audio Card to handle the transfer of PDM data to the Raspberry Pi 2? The microphone I'm working with (Knowles SPH0641LU4H-1) outputs bits at a rate of 3.072 to 4.8 MHz (the clock rate required by the device in ultrasonic mode). My goal is to get this signal to the Raspberry Pi, where I can use Csound to run it through a low-pass filter to turn it into an analog signal. Currently, my plan is to try to get ALSA, and by extension Csound, to use one of the GPIO pins available on the audio card's expansion header as a form of input for the PDM signal. Does this seem plausible? Is there a better way?

 

Thank you!

  • I believe this:

    My goal is to get this signal to the Raspberry Pi, where I can use Csound to run it through a low-pass filter to turn it into an analog signal.

    May be incorect because https://en.wikipedia.org/wiki/Low-pass_filter 

    would block the high waves. You really want to shift down the ultrasound into the audio range by altering the frequency with a divider circuit see https://en.wikipedia.org/wiki/Frequency_divider.

    C

  • in reply to clem57

    The signal from the microphone is a PDM signal, so according to the Wikipedia page on PDM (https://en.wikipedia.org/wiki/Pulse-density_modulation), a low-pass filter should serve the purpose of averaging the signal into a PCM format (and smoothing out the high frequency noise), which I guess is what Csound can manipulate. I plan to shift the frequency into the audible range once it is in a PCM format.

     

    I still haven't found a solution yet, but I'm reading through these posts with the hope that I'll find something useful: http://http://curiouser.cheshireeng.com/category/projects/pdm-microphone-toys/page/3/

     

    I'm also hoping to minimize my external hardware use by converting the PDM signal to PCM within the RPI if I can, which seems possible given that I2S pins are available on the Cirrus Logic Audio Card's expansion header, but I have to do some more reading to wrap my head around how that might be done.

     

     

     

    EDIT:

    "run it through a low-pass filter to turn it into an analog signal."

     

    What I just said in this post conflicts with this quote, so I'll try to clarify my thoughts about the low-pass filter.

     

    One solution I can think of is to implement a hardware low-pass filter to convert the PDM signal to an analog signal, then to sample that analog signal with the audio card. This, however, seems a little convoluted given that PDM is already in a digital format, and there are other factors in support of keeping the signal digital, but I suppose I can always fall back on this.

     

    If I keep the signal digital, I still have to run it through a low-pass filter to turn it into PCM data, so that's why I used the phrase "low-pass filter" to refer to both concepts.