TR Forums

just brew it! wrote:

ludi wrote:

The Swamp wrote:

So far, so good. There is a big difference. There's a slight pop every now and again, but it's much better. So, it seems the problem was the PCI sound card after all. Oddly, I had no issues with it in my old system, which also ran Win7-64. I'm not sure what is causing issues in the current system.

Wait, "also ran Win7-64"? So you're using Win7-64 on this system? That might be your problem, the driver scene for Win7-64 was never spectacular.

Err... what? Are you sure you aren't thinking of XP-64? Win7-64 had reasonably good driver support IME.

Gah. That's what I get for staying up late. Yes, Win7 x64 should be fine.

The Swamp wrote:

The only two cards I have in use are the sound card (disabled in Device Manager) and the Radeon RX-550. The Radeon driver is recent. The popping seems more pronounced during heavy CPU usage. Maybe a bus issue?

I read this and thought of changing the IRQ to see if that improved performance, but no, we don’t do that any more, do we?

I remember reading somewhere that modern motherboards/chipsets don't have a true PCI bus; instead, it's bridged from PCIe. Which makes some sound cards (Creative X-Fi was mentioned) nigh unusable on some newer motherboards.

I have been struggling with a similar issue and it is starting to drive me crazy. There was a lot of noise in the audio from my X-Fi Fatal1ty (PCI device). It was not analog noise but clearly some buffer problems, so it pointed towards driver or data-path problems. Initially, I thought the aging sound card had developed a fault, but after switching to the onboard Realtek audio, the problem was still there. It becomes worse when the CPU is loaded, and it is also much worse in e.g. 24 bit/96Hz than 16 bit/48kHz. I have removed the X-Fi from the system completely and also tried a USB Audio DAC, but that also shows the issue so I am basically not capable of producing clean sound from my machine at the moment. While I have not found the smoking gun yet, it seems to affect the issue when I change the MB BIOS (with the newest one being the worst). Since both the Realtek and my USB Audio DAC is using MS standard drivers I doubt that is the place to look, but I am almost out of ideas now. There is no PCI Latency timer setting in my BIOS.

I don't remember doing anything special when I started having this issue. I have been using my X-Fi for a couple of years without issues in this system and did no perform any hardware changes at the time when it started (except maybe plugging some USB-devices in/out which happens daily). So I am contemplating whether some Windows update could have started this.

I think my next step would be to try and install a fresh Win10 to see if the issue is still there or if it perhaps could be some other software or driver causing the problems.

My system: i7-6700k, ASUS Z170-A.

setaG_lliB wrote:

I remember reading somewhere that modern motherboards/chipsets don't have a true PCI bus; instead, it's bridged from PCIe. Which makes some sound cards (Creative X-Fi was mentioned) nigh unusable on some newer motherboards.

Yes, my Z170-A uses a bridge chip to provide its PCI slot, however I have been using an X-Fi PCI sound card for a long time without problems until above issues started to appear.

I haven't tried the on-board sound, but I suspect I'm going to find the same problem popping up. It definitely seems to be tied to the system waking up from sleep mode. Back before I used sleep mode all the time, I didn't notice a problem with the sound. So that may be where the problem is coming from. When I want to watch streaming video on the system, I may have to do a reboot each time.

I hate to be the one to suggest it, but maybe live boot a Linux distro just for giggles?

I'm leaning towards something on the mobo itself, affecting the signal path, especially since you said it got worse with a BIOS update. DPC Latency, traces too close, USB 3 or Wi-Fi, bad ground - you could chase all of that and it could still end up being your OS. Or mobo. Or even something else in your audio chain after the signal leaves the PC.

Oh yeah, sleep mode. Can you just not use it for a while, or will the extra 30W of idling computer destroy the electric bill?

@MOSFET - Trace routing, bad ground, other audio chain issues would not be affected by sleep. It is almost certainly a driver issue of some sort.

I rarely use sleep mode for the same reason. There are almost always issues with stuff not working properly after waking up from sleep mode. Or at least that is how it has been for a long time. I know they have improved a lot of this stuff in Win10 lately to avoid those kinds of errors, but I still just mainly use hibernation instead of sleep mode.

JustAnEngineer wrote:

Aranarth wrote:

The popping and clicking sound is the powerful, two-cycle, briggs and stratton engine

The Otto cycle is four-stroke: suck, squeeze, bang then blow..

That's what she said!

just brew it! wrote:

@MOSFET - Trace routing, bad ground, other audio chain issues would not be affected by sleep. It is almost certainly a driver issue of some sort.

I'd agree with that.
Try a complete wipe of the driver and any old drivers in the system and then install fresh.
Does safe-mode still work for this?

If it was me, and the driver allows for it, remove the card, all drivers, then install the latest one only, then reinstall the card.
Trouble is some drivers check to see if the device is installed before allowing the driver install to be completed.
When that happens sometimes you run into issues where M$ decides to install a generic driver for the device or install an older driver for you.

With it getting worse across a sleep it sounds like some driver is not restoring something properly after the low power state. N.B. This could be a completely different driver in the system that ends up hogging the bus so the CPU doesn't get to talk to the sound card before the sound card's buffers underflow. The change with BIOS versions could be that the default after a sleep changed in the BIOS so it is now worse.

Try removing the overclock, or at least any PCI bus overclock (set the PCI bus to back to "standard" speed). If that doesn't clear it up, and it's not some weird interrupt sharing/conflict problem, then it's likely a driver problem.

Aranarth wrote:

Try a complete wipe of the driver and any old drivers in the system and then install fresh.

IIRC he did a fresh install on this platform, so that's probably a dry hole...

Aranarth wrote:

When that happens sometimes you run into issues where M$ decides to install a generic driver for the device or install an older driver for you.

...investigating whether this may have happened may be worthwhile though.

I just tried installing a blank Win10 1803 on an empty SSD to see if my audio glitch issue was caused by some software or drivers. Unfortunately the problem was still there.. Again, I tried with both the onboard RTK audio and an external USB Audio DAC. It is very bad in high load situations, e.g. in a game, but it is still noticable when just playing a music file in Windows with a glitch maybe every couple of seconds.

Now I am thinking about installing an older Win10 version since I didn't always had this problem and my HW config is exactly the same.

Maybe this already came up, but are you running through any sort of power conditioner for power coming into your computer? I'm almost thinking you've got some dirty power coming in. I've see houses with old knob-and-tube wiring that adding a conditioner was almost required for anything audio related.

Usacomp2k3 wrote:

Maybe this already came up, but are you running through any sort of power conditioner for power coming into your computer? I'm almost thinking you've got some dirty power coming in. I've see houses with old knob-and-tube wiring that adding a conditioner was almost required for anything audio related.

Was also starting to suspect-POWER.
Try swapping out PSU if you have a spare,or can borrow one.

Before that thro,I'd do a visual of any electro caps on board.
Also disconnect EVERYTHING,except bare minimum to see if it's still there (including fans and pumps)
good luck.

You guys are ignoring the fact that the problem only seems to show up after the system has been through a sleep/wake cycle. If it was a power conditioning issue it would not behave like that.

just brew it! wrote:

You guys are ignoring the fact that the problem only seems to show up after the system has been through a sleep/wake cycle. If it was a power conditioning issue it would not behave like that.

Oh if life was so simple.
The way I've been reading it-It's there all the time,but worse after a sleep................

Few other diagnostics-Source
Is it still there ? Audio in browser-Games-Audio from file(mp3 etc)
Have you set BIOS to default-save -exit ?

HERETIC wrote:

The way I've been reading it-It's there all the time,but worse after a sleep................

Whether it occurs only after a sleep, or gets worse after a sleep... either way, that's evidence that it is not due to an external cause like dirty power.

Download a DPC Latency checker, run it, and see if that gives you any clues. It could lead you to the offending driver/firmware. It appears the audio buffers being starved by overaggressive/badly behaving drivers. When the buffers starve, you get an audio interruption that ends up as a popping sound or some other kind of momentary distortion.

roncat wrote:

Download a DPC Latency checker, run it, and see if that gives you any clues. It could lead you to the offending driver/firmware. It appears the audio buffers being starved by overaggressive/badly behaving drivers. When the buffers starve, you get an audio interruption that ends up as a popping sound or some other kind of momentary distortion.

That's a very good idea. Seconded.

As mentioned, I had similar issues with audio on my PC and yesterday I was able to strip it down to the bare minimum in effort to troubleshoot it.

Although I didn't really suspect the PSU, I also tried another brand new PSU just to be able to rule that one out. Even with just CPU+MB+RAM and an empty SSD where I installed a clean Win10, the issue remained. I had also ditched the Gfx card and was running with the integrated GPU.
Since this is unlikely to be a RAM issue (I ran a memtest on it and this is not the kind of error you would see with bad memory since it is not crashing), I am down to this being an issue with either the MB, the CPU or some incompatibility with the latest Win10 releases. I did not try installing an ancient Win10 version to check that and instead decided to buy a new MB+CPU. First I thought I could do it stepwise and buy a Z370 MB and reuse/test my old i7-6700k CPU in that and perhaps later replace the CPU with a never one, however I quickly found out that Z370 only supports Gen 8 CPUs.. So if I want a newer chipset, I am pretty much also forced to replace the CPU. Since I didn't want to sit with another Z170 motherboard in case the issue is somehow related to my CPU, I decided to just upgrade both. This issue has annoyed me long enough and I'd rather just replace the dodgy parts than spend more time debugging it.

So now I had a choice between X470+2700X or Z370+i7-8700k which were in a similar price range. I like the new Zen+, however I ended up going for the Intel solution since the IPC and single-threaded performance seems to be a little better. Hopefully this upgrade will solve all my problems with the weird audio problems (it also seemed like the system had other weird lag issues at some times which might have been related).

Based on what I am reading in this thread, it seems like the MB had a serious hardware design issue with the interrupt system. Something was stalling the bus. If I were you I would have RMA-ed the motherboard if able.

I thought these sort of problems were dead 10 years ago with much better chip sets. Guess not!

I feel almost like giving up now.. Despite replacing MB and CPU, I still get the same popping sounds in my audio. It doesn't seem as bad as before but it is definitely still there.
Now I have tried replacing motherboard, CPU, power supply and boot device (SSD). It is still present on the onboard Realtek audio on the new MB and also through a USB Audio DAC.

The only things left are the RAM, keyboard, mouse and monitor.. Well, and the case and 230V from the socket. The new SSD got a fresh Win10 installation so that was replaced as well, but now I am thinking about trying out some Linux dist. to check if the same issue is present there as well. It might just be a problem with recent Windows versions, although I would have expected to see a lot of other people complaining about the same then.

demolition wrote:

Well, and the case and 230V from the socket.

I'm sticking with my guess

HERETIC wrote:

Usacomp2k3 wrote:

Maybe this already came up, but are you running through any sort of power conditioner for power coming into your computer? I'm almost thinking you've got some dirty power coming in. I've see houses with old knob-and-tube wiring that adding a conditioner was almost required for anything audio related.

Was also starting to suspect-POWER.
Try swapping out PSU if you have a spare,or can borrow one.

Before that thro,I'd do a visual of any electro caps on board.
Also disconnect EVERYTHING,except bare minimum to see if it's still there (including fans and pumps)
good luck.

Usacomp2k3 wrote:

demolition wrote:

Well, and the case and 230V from the socket.

I'm sticking with my guess
Was also starting to suspect-POWER.
Try swapping out PSU if you have a spare,or can borrow one.

Although I don't have any proof, I have a hard time seeing that it could be caused by noisy mains in my case. In that case, I would expect it to sound like analog noise while it clearly is stuttering / missing samples which to me points to some DMA/buffering issue. I work with audio professionally, so I can easily tell the difference between these kinds of noise.

I guess if the analog noise on the supply is big enough to actually cause data corruption on a bus or interfere with a clock then all sorts of things could happen, but then I would expect the PC to also become generally unstable, corrupt the HDD etc. and it doesn't seem to do that.

Do you have a UPS? Does the issue happen if running direct from the wall power without the UPS? Does it happen if you unplug the UPS from the wall and run from the UPS battery?

I'm still suspecting something else stealing cycles like a graphics card or even a bad USB device where the processor has to keep retrying.

demolition wrote:

Although I don't have any proof, I have a hard time seeing that it could be caused by noisy mains in my case. In that case, I would expect it to sound like analog noise while it clearly is stuttering / missing samples which to me points to some DMA/buffering issue. I work with audio professionally, so I can easily tell the difference between these kinds of noise.

I guess if the analog noise on the supply is big enough to actually cause data corruption on a bus or interfere with a clock then all sorts of things could happen, but then I would expect the PC to also become generally unstable, corrupt the HDD etc. and it doesn't seem to do that.

Agreed. This doesn't sound like a power issue at all.

Reminds me of the bad old days of Darth VIA.

notfred wrote:

Do you have a UPS? Does the issue happen if running direct from the wall power without the UPS? Does it happen if you unplug the UPS from the wall and run from the UPS battery?
I'm still suspecting something else stealing cycles like a graphics card or even a bad USB device where the processor has to keep retrying.

No, I do not have a UPS.

I would also consider the graphics card or a USB device if it hasn't been for the fact that I had already tried stripping the system down to the bare minimum without it resolving the issue. That means removing the gfx card and just using the onboard (Intel Gfx), only having a single SATA device connected with a fresh Win10 install, and no USB devices except keyboard and mouse. And also replacing the PSU with a brand new Seasonic one that I was able to borrow for the test.

I found posts in several forums about people having similar issues with the audio (mostly with the onboard Realtek audio) and some of them were able fix it by disabling some CPU power saving features. One method which seemed to help many was to create a new power scheme in Windows and set the minimum processor state to 100%. I tried this with no improvement. I also went through the BIOS to try and find and disable any power saving / throttling features there as well, but could not find much.