1

DAC+ ADC pop / click sound

theblond

November 05, 2020 19:33

None

Follow

Hi,

I've searched the forum, but I didn't find exactly the same problem or any similar solution to this.

I am using RPi4 with DAC+ ADC running Hifiberry OS the latest build.

Independet of the source I noticed well audible pop / click sounds under the following circumstances:

iTunes - everytime I hit play or jump to another timestamp while playing I, here this sound:

 

Spotify - I think it's the same problem, but not as audible as with iTunes (maybe because iTunes has this significant delay before playing):

Analog Input - Here it's a different noise when start playing:

 

RPi start up

1) plugging in power supply (noise not really an issue)
2) red LED on DAC+ ADC lights up (noise not an issue)
3) but than comes this cleary audible sine wave

 

I here no issues, when I play continuously. I tried several audio files and input hardware.

• Facebook
• Twitter
• LinkedIn

28 comments

• 1
  
  Domenico Guerra November 06, 2020 08:03

  i have similar clicks in spotify and dnla when change manually the song, not tried others sources

  using last hifiberryos ans dac+ dsp

• 0
  
  HiFiBerry team November 06, 2020 08:11

  The DAC stage shuts down the output when the sound cards isn't playing something. Combined with a amplifier with an AC coupled input with an internal DC offset, this can result in these clicks. 
  Some applications offer a way to always output data without interrupting the data stream to the sound card. This can help. However, not all applications offer this option. If you don't have this option, there is nothing you can do about this.

• 1
  
  theblond November 06, 2020 11:05

  Well, i did some more research. This seems to be quiet old problem, independent of the OS.

  I found this 4 year old forum thread ending up without any solution.

  Texas Instrument stated in the datasheet of the PCM5122 DAC on the first page an "Intelligent Muting System; Soft Up or Down Ramp", hence this pop sound should not be there when powering up.

  It's annoying, because even most cheap sound cards don't produce such a loud pop sound.

  Is there something wrong with the software samplerate handling? Because the datasheet says "Deleted Internal Pop-Free Control For Sample-Rate Changes Or Clock Halts" in the revision history. The DAC cannot handle pop-free clock changes or halts.

  Maybe another idea:
  Does Hifiberry OS has an interface to set the control registers of the DAC, so I can turn of the Zero Data Detection Mode and prevent the DAC from shutting down?

• 0
  
  HiFiBerry team November 06, 2020 11:13

  When switching sample rates, the DAC has to be stopped. There's nothing you can do about this. 

  HiFiBerryOS does not have any interface to deal with internals of the chip. Feel free to experiment with i2cget/i2cset from command line.

• 0
  
  theblond November 06, 2020 12:44

  Thank you for the hint with i2cget and set! I'll try this.

  I understand that the DAC has to be stopped for sample rate changes. But is it possible that ALSA / Hifiberry OS or what ever part of the software is responsible for the sample rate stops the clock as soon as nothing plays? In that case there could be a pop sound too.

  In other words. Would it be a solution that the clock runs, although nothing plays?

  In my tests I encountered the pop sound without changing the software and music file, so there was no sample rate change I induced.

• 0
  
  HiFiBerry team November 06, 2020 13:04

  No, you can't simply change this as this would require a different driver. You might try audio isolation transformer that will make sure, there's always a relatively low-impedance connected to your amplifier:
  https://www.amazon.com/PAC-SNI-1-Noise-Isolator/dp/B000K50HJE
  
  
  

• 0
  
  Roderick van Domburg November 07, 2020 23:12

  Just to chime in here: same thing on a DAC2 Pro's headphone out. It's distracting and sounds cheap. Sure, HiFiBerry has an excellent price point and we love you for it. But pops like these distract from a "HiFi" experience. Something handy the PCM5412 offers:

  Intelligent muting system; soft up or down ramp and analog mute for 120 dB mute SNR with popless operation

  Over at Mouser, it's € 0,55 more expensive than a PCM5122 per piece.

  Edited by Roderick van Domburg November 07, 2020 23:13
• 0
  
  Domenico Guerra November 09, 2020 10:18

  i have "solved" the problem using squeezebox...

• 0
  
  theblond November 09, 2020 12:11

  Roderick van Domburg the datasheet of the PCM5122 says the same. Both DACs have the intelligent muting system.

  Domenico Guerra thank you for the input. I am still trying to change some settings through the I2C interface, but no look. The DAC is always busy.

• 0
  
  HiFiBerry team November 09, 2020 12:17

  You need to use the "-f" to access I2C addresses that are connected to kernel drivers. Note that this will change the state of the device without the Linux kernel knowing about it. This can result in all kinds of unexpected behavior - even to kernel crashes.

• 0
  
  Roderick van Domburg November 09, 2020 13:59

  theblond you're right, thanks for that pointer. Following this to see how your software quest goes.

• 0
  
  Mark RC November 17, 2020 04:36

  So, after updating my pi (still stretch, but wanted to get all the exploits and bugs patched), my DAC+ADC Pro is experiencing what sounds like a single sample a full-scale pop (running into the Hifiberry AMP shield, but it's just doing what it's told by the DAC), but it happening at random during program material, and randomly on left or right (never both, oddly).  

  What's the deal with this?

• 0
  
  HiFiBerry team November 17, 2020 06:33

  Haven't heard about this. What's your player software?

• 0
  
  Mark RC November 17, 2020 07:20

  Max2play, running through squeezelite.  The period on average is about 90 seconds, but anywhere from 1 second to a couple minutes.  Like I say, sometimes on the left, sometimes on the right.  They always seem to be full-scale and since the aamp60 hat is directly connected to the dac with it's own header, it's basically a full power hit..

  I'm going to try to find another amp to plug into the analog outputs to make sure it's not the hat, but both just arrived a couple days ago.

  Frustrating, because I just got this turntable connected through the ADC, and now the output has this loud snap, I might have to switch back to AMP2.  The pop happens regardless of source, could be a local file, the ADC, or an internet radio station...

  Could it have anything to do with updating stretch?

  Edited by Mark RC November 17, 2020 07:22
• 0
  
  HiFiBerry team November 17, 2020 07:22

  Sounds like some kind of buffer underrun. You might check the system logs for underruns. 
  Akternatively you can try to use HiFiBerryOS (https://www.hifiberry.com/os) as a Squeezelite player to check if this works better.

• 0
  
  Mark RC November 17, 2020 07:24

  Which log should I look at?

• 0
  
  HiFiBerry team November 17, 2020 07:25

  Unfortunately I can't say what logs exist on Max2Play. You might check with the developers if they can help debugging this.

• 0
  
  Mark RC November 17, 2020 18:34

  Another day, more debugging...  Nothing I can find in logs in the usual places (debug, kernal, etc...) and the slim logs are pretty generic to file operations, nothing at the driver level.

  I'm going to cross-post this to M2P as well, because it was a hack to get DAC+ADC Pro working to this point.  M2P did not have an interface instantiation for DAC+ADC Pro, so your only choices are to select DAC+ Pro or DAC+ADC...which sets those specific dtoverlays in /boot/config.txt.  

  When set to DAC+ Pro, the DAC works fine, no popping.  Of course, there is no ADC function with that set of drivers installed.

  When set to DAC+ADC, both the DAC and ADC work without popping, however the A/D defaults are super-wonky: PGA is set to max,  input selection fixed, mic bias set to on and other weirdness.  With the gain so high, you can only connect the turntable as direct MM (no RIAA equaliztion!) and the bias does weird things to the cartridge (shifts the BH curve so there's dynamic range issues).  Ever hear a cartridge without RIAA?  It sounds real special;)  Running the phono preamp resulted in solid noise as the input was (shall we say) 'saturated' since the PGA was set to max.

  To get DAC+ADC Pro to work, I went into the u/i and instantiated a copy of DAC+ADC, calling it DAC+ADC PRO, while substituting the array entry for modifying config.text specifiying hifiberry-dacplusadcpro.  I verified that config.txt was set properly, added amixer to it configuring mic bias off, input settings, and a nominal input gain of 12db to (hopefully) match the output of the phono preamp.  This configuration I get both ADC and DAC working, but the popping is ridiculous.

  Checking /boot/overlays, I see hifiberry-dacplusadcpro.dtbo, but don't know how to determine it's version.  All file timestamps are as-of when I did the updates yesterday, and I can only assume they're the latest from the APT repos.  According to git, the most recent update was adding a parameter to shut off the status LEDs, which seems innocuous enough.

  Thinking about the popping a bit more, the full-scale behavior is troubling...if it were frame slips, there would be a noticeable hiccup in the program along with a pop of an amplitude connecting the last good sample with the next valid sample after the gap (or if time is slewing the other way, splicing a section out.  The hiccup would be in the form of a time squeeze or stretch with the 'splice', and the amplitude of the splice dependent on the program itself.  I heard a lot of this with a USB ADC, where the Pi's ethernet would cause underruns fairly often...you could hear it in the audio (not a POP!) and in the tempo of the music.  Full-scale excursions tend to imply something else is going on, don't know what at this point.

  Thanks,

  -Mark

• 0
  
  HiFiBerry team November 17, 2020 20:57

  It might be caused by internals of Max2Play. Both ADC and DAC run on exactly the same clock. If any process is using a different clock, samples will be either dropped or there won't be a sample. The system itself doesn't interpolate. Regular pops are usually a sign of two independent clocks that aren't synchronized. 
  
  The difference between the DAC+ ADC and DAC+ ADC Pro is in fact the clocking. For the standard version, the Pi provides all clocks, for the Pro version, the application has to synchronize on the clocks of the DAC. Looks like the process that's doing this on M2P doesn't do it correctly. It might be an issue that might require bigger code changes in the software. 

  Edited by HiFiBerry team November 17, 2020 21:00
• 0
  
  Mark RC November 17, 2020 21:04

  Okay, I'll pass that info along to M2P, but does the DAC+Pro (which works without popping) also use an internal clock or does it use the rpi clock? 

  I see no difference in the M2P configuration array entries between a DAC+ADC, and a DAC+PRO, other than the dtoverlay parameter.

• 0
  
  HiFiBerry team November 18, 2020 07:13

  The DAC+ Pro has only an output, not an input and output. That's a completely different thing. The dtoverlay doesn't have to do anything with this. The dtoverlay is just a description of the hardware, it has nothing to do with this.

• 0
  
  Mark RC November 18, 2020 16:52

  DAC+ADC does not pop, while DAC+ADC PRO has the popping problem.

  To confirm: DAC+ADC uses a sample clock generated from the RPi, and DAC+ADC PRO uses an internal clock, is that correct?

  If that is the case, we'll dig into that deeper...

• 0
  
  HiFiBerry team November 18, 2020 16:53

  Yes, the DAC+ ADC uses the Pi clock while the DAC+ ADC Pro uses it's own clock

• 0
  
  corentin December 13, 2020 23:42

  Good evening, I have the same concerns. Did you find a solution to the problem?

• 1
  
  theblond December 14, 2020 17:13

  No, not yet. The things I tried weren't successful. I have not had the time for further investigations.

• 1
  
  corentin December 14, 2020 17:18

  ok, i just received my card. I think I'll send her back. It's very disturbing

• 0
  
  mmo March 29, 2021 13:16

  Hi theblond,

  did you by any chance make some progress in the meantime. I have the same problem with a DAC2 Pro.

  Thanks and best regards,

  Martin

• 0
  
  theblond March 30, 2021 08:40

  Hi Martin,

  im sorry, I have not worked any further on this problem. But it is on my todo list. I did not forget it. As soon as I make any progress or have a conclusion, I will post it here.