New subject: [SOLVED] Crackles in audio, drifting intermittent noise etc.

31 Oct 2016

A follow-up to a thread in August.
Recall: My Delta1010-LT card was not OK in my new PC so I tried it
 in another PC at the shop and it tested OK for 15 minutes,
 so I deemed that PC to be OK and my dealer would give me /that/ PC instead.
Imagine my surprise (anger) when I finally brought the other PC home
 and set it up, and the sound was fine...  For 15 minutes!
Then suddenly the same noises started again!
This time however, I noticed something: When using the card in an 'effect'
 situation, where audio input is passed through to an audio output, either
 directly or with some effect, horrible noise would appear after 15 minutes
 and then... slowly drift into working again for 15 minutes, and then slowly
 drift into noise again for 15 minutes, then working again for 15 minutes
 and so on.
This was a HUGE clue. This slow drifting in and out of noise is familiar.
It is instantly recognizable to me (as should you) as a synchronization
 problem, not a PCI voltage or bus problem.
Anyone who has worked with syncing multiple sound cards together,
 or with ADC + DAC mutual clock sync problems, or re-sampling up/down
 converters should recognize this slow drifting in and out of noise.
At last there was hope that this could be diagnosed and possibly solved.
So... on this replacement PC there is a setting in the BIOS for the number
 of CPU cores to activate. The cheaper original new PC had NO such setting.
* That setting has solved it! *
I had to set TWO CPU cores (or ONE) instead of FOUR CPU cores.
Working solidly for 3 days now!
*NO* other settings helped. Not even Speed Step et al.
They're all turned on now!
This is my first multi-core CPU. I'm late to the party.
Ironically I had told my dealer this was one of the reasons to get a new PC,
 so that I could test our app to see how it works with such newer CPUs.
I had even researched these CPUs and had read that it is possible that
 the CPU Clock signals can be independent for each core, thus causing
 some synchronization problems.
At this point I'm not sure if this a clock signal sync problem or a
 timer/counter problem. I had also read that 'local' timers/counters
 can be a problem with multi-core CPUs - that time is not quite the
 same in each core.
IIRC Someone told me that a quad core is two chips each with two cores ??
That might explain why sound works fine with TWO cores but not FOUR ??
My interrupt affinity list showed only core 3 was getting them,
 and forcing them to core 1 didn't help.
Can anyone shed some more light on these multi-core problems?
Comments welcome, please!
Thanks.
Tim.
On Thursday, August 25, 2016 3:20:40 PM EDT jonetsu(a)teksavvy.com wrote:
...
  > Hello,
 >
 > Suddenly the audio started crackling.  I was listening to a youtube
 > tutorial (firefox) and it started.  Bitwig and Renoise were running but
 > not playing anything.  I looked at  the log files (syslog, kern) nothing
 > relevant.  Card is 10101LT.
 >
 > jackd runs at about 10.3ms latency (last time I checked using
 > jack_iodelay), so it run as:
 >
 > /usr/bin/jackd --sync -T -P95 -ndefault -dalsa -dhw:M1010LT -r44100
 > -p128 -n2
 >
 > scaling_gouvernor all set to performance.  Interrupt prio looks OK:
 >   PID CLS RTPRIO  NI PRI %CPU STAT COMMAND
 >   436 FF      90   - 130  1.6 S    irq/18-snd_ice1
 >
 >    47 FF      50   -  90  0.0 S    irq/9-acpi
 >
 > limits audio.conf is:
 >
 > @audio   -  rtprio     95
 > @audio   -  memlock    unlimited
 >
 > Interrupts servicing for all 4 CPUs (i5) :
 >   0:  IR-IO-APIC-edge      i8042
 >   4:  IR-IO-APIC-edge
 >   8:  IR-IO-APIC-edge      rtc0
 >   9:  IR-IO-APIC-fasteoi   acpi
 >
 >  12:  IR-IO-APIC-edge      i8042
 >  18:  IR-IO-APIC-fasteoi   snd_ice1712
 >
 > jackd and pulse processes:
 >  /usr/bin/pulseaudio --start --log-target=syslog
 >
 >  [pulseaudio] <defunct>
 >
 >  /usr/bin/jackd -T -ndefault --sync -T -P95 -ndefault -dalsa
 >  -dhw:M1010LT -r44100 -p128 -n2
 >
 > The pa zombie was there way before the crackling started.  Not sure if
 > this is related.
 >
 > The pulseaudio jackd sink is active and can be seen in qjackctl.
 >
 > So it seems OK.  But the audio is full of crackles.  In firefox, as
 > well as Bitwig and Renoise when now something is played.  Of course,
 > when the machine was started some audio was played in both Bitwig and
 > Renoise and it was fine.  Then I watched a youtube tutorial and bam,
 > after maybe 15 minutes, all audio output is full of crackling.
 >
 > Since the log files shows nothing.  How is troubleshooting information
 > gathered ?  There should be some SW system component that can be
 > probed, somwthing that could be observed.  - or is it that the 1010LT
 > is going awry ?  Can this happen at all ?
 >
 > Thanks for suggestions and comments !
 >
 > Cheers. 
> Hi, this is very, very ironic for me.
> Is this a new or replacement PC?
>
> I hope the following helps provide at least one answer for others having
>
>  these problems:
> My new i5 Acer M3910 PC is doing the same thing with my 1010LT.
> Right from day one. Digital noise, crackles, and pops, on the analog output.
>
> This is in BOTH Linux and Windows 7. The same noises in each.
>
> To be sure, I reinstalled the card on the older PC from which it came,
>
>  and it is fine. I also installed a SBLive! PCI card on the new PC and
>  it is fine. Seems this 1010LT doesn't like this new PC.
>
> I have noticed many people complaining about such noises,
>
>  especially the clicks and pops. But as you know this can be muddy
>  territory, with several different causes that are hard to pin down.
>
> So, being heavily involved with these ice1712 cards
>
>  (I helped with mudita24 mixer), I have some test results
>
>  that should interest owners of this card:
> Go to this site:
> http://onlinetonegenerator.com
>
>  and input a high frequency of say 15000Hz, so that you don't
>  hear the tone so much as any noise that may accompany it.
>
> Or, if so desired, use Jack and a tone generator plugin.
>
> On this new PC I get really horrible digital noise, clicks and pops.
> The digital noise, oddly, is worse with higher frequency test tones.
> From listening carefully to the noise, it seems something to do
>
>  with the PCI bus timing or voltage, or poor digital shielding such that
>  noise is getting into the D/A converters before conversion to analog.
>
> This 1010LT card is revision 'C', the last and latest revision.
>
> I tried everything. Turned off SpeedStep, adjusted Jack buffer size etc.
>
> This PC's BIOS is not very friendly, it is a so-called 'locked BIOS'
>
>  where you don't have much in the way of critical timing parameters.
>
> ----
>
> So, today I brought the card to my dealer and we tried it on
>
>  another i5 PC, having a much better Asus MB.
>
> Result: Perfect behaviour. Even at the lowest power saving settings,
>
>  on that test-tone website the card only made a couple of pops but
>  no digital noise. I would expect a couple of pops here and there
>  at these settings, so I concluded the card is fine there.
>
> Conclusion: My dealer is replacing this Acer MB with the better one.
>
> Believe me if there was something I could do to make it work
>
>  I would do it and report back. I was contemplating replacing the
>  electrolytic capacitors on the card just in case. You'd be surprised just
>  how weak caps can be, allowing digital ripple on supply lines etc.
>
> It's possible that may help.
> But given that the card works OK elsewhere, and I don't want to waste
>
>  any more time on this, I am just going for the replacement PC instead.
>
> Tim.