Piano World Home Page Forums Our Most Popular Forums Digital Pianos - Electronic Pianos - Synths & Keyboards Finding About Solving Issues with Dropouts / Crackling

I was suffering from crackling and dropouts when playing VSTs on my digital piano USB audio interface. I thought it was my CPU, so I upgraded the CPU and it was fixed (or so I thought)
I then upgraded something else on my PC and the crackling came back.

After trying a million things, I realised that every time I upgraded something, I also unplugged all USB devices and plugged them back in different USB slots.
So, I tried plugging my digital piano USB on different USB slots.

It ends up that there is 1 usb slot that all dropouts are gone.
Maybe thiscan help some of you, if are having any issues.

Do they operate at USB 3.0 speeds?

Is it a Dell? If affirmative, have you got Dell services running (supportassist, vault). If affirmative, disable these services. This may not suffice to solve latency on Dells, but it is a must have, in my experience.

Yes, 6 out of 8 of them. Haven't tried the USB 2 ports. The problem was solved on the last USB 3 port I tried.


No it's a desktop I built myself.

Somethimg interesting I just discovered. This is the first USB 3 slot on my motherboard.
The first 2 USB slots are named USB port 1 and USB port 2. The one I have plugged my digital piano is USB port 3.

Thanks for the heads up because I never cared to check for that: always assumed that all usb ports were handled the same way.

Two things to clarify: these are all usb ports on the back panel of the computer, correct? None of them is one of the additional usb ports cannected through optional cables, right?

How exactly are you connecting? Piano Midi usb to one port and the AI into another, or midi cable (or audio cable) to interface and only the interface to the computer?

These same thanks and questions are on my mind!

Cheers and happy interfacing,

HZ

Yes. that's all USB ports at the back panel. There's only one USB cable, both for midi and audio

Good find! We often see USB as "fungible" but they sometimes run on different buses and may actually have different drivers...

Ok. Because you said "digital piano audio interface" I thought you had a separate interface. Out of curiosity because there are not a lot of user reports of windows compared to macs: do you have a Yamaha or one of the newer Rolands with integrated interface? How good is latency?

No I have Korg LP-380U. The latency is really low. I have it at 64 buffer size in ASIO4ALL

Thanks. I was not aware that the 380U had usb audio, I thought in only added bluetooth.

And are you running your VSTs in a standalone program or in a DAW ? I am asking because I recently had crackling issues that were related to handling in my DAW (fixed meanwhile).

And just wondering how you are testing for how close you are for crackling/dropouts ? I typically look at my CPU monitor/VST activity monitor in FL Studio or in the CPU monitor in Pianoteq within the DAW. Then I slide my fingers across the whole keyboard to generate as many sounds/voices as I can while holding the sustain pedal down. If no crackling/dropouts during this process appears (and CPU activities stays reasonably low), then I feel good about it. What is your technique ? The reason why I am asking is to determine if the plugging into the one "good" USB port just removed you from the "edge/border" where crackling starts or whether it completely solves the problem.

I was checking the CPU activity in task manager and it was really low. But at the same time when I was playing pianoteq I was getting some CPU overload warnings.
I think it's CPU spikes causing the issue.
I don't think it completely removes the issue. I think it just moves the "edge/border" and makes it really hard to happen.

What would normally happen on the bad usb ports is that when I first load a VST for about 3-4 minutes I would get really bad dropouts to a point that I would have no sound for 2 seconds.
Then after 4 minutes it would become almost playable, but would get drop outs and crackles if I played a demanding piece with lots of sustain and chords, like Claire De Lune.

On the good USB I don't have to wait for 3-4 minutes. I can play immediately. I also don't get any dropouts even with something like Claire De Lune.

did you turn on high performance mode in power settings ? you can also try disabling throttling in bios.

Yes, I have it in high performance mode. Throttling is still enabled, but I was checking the temperature and it was low, about 45 C.

I had all the other suggested optimisations, like adjust for better performance of background services, untick "allow computer to turn off this device to save power", etc. from this link:
https://bandzoogle.com/blog/20-ways-to-optimize-your-windows-10-pc-for-music-production

One thing that would also cause the PC to have dropouts is when it was automatically downloading updates. I also disabled that.

I've been toying with Pianoteq again the last two weeks to get it to run reliably on Linux.

Sometimes it seems to run fine and I can do whatever the heck I want; sometimes it crackles even with single chords.

I am not yet completely certain (more testing is needed), but it looks like it's x11vnc vs. Pianoteq causing this issue.

I'm running Pianoteq on a NUC (J5040, so not the most powerful option; but 30% more powerful than the previous i3-6100U it ran on), and it has no monitor attached. It uses an HDMI monitor dummy so it boots as if it does have a display attached. x11vnc scrapes/shares this display.

It seems the output crackles when there's a lot to redraw on screen (Pianoteq updating its interface, KSysGuard drawing CPU graphs, etc). I'm suspecting that x11vnc is getting in Pianoteq's way, because they're probably both using interrupts to get their stuff handled as quickly as possible.

There's still more testing I need to do, but if I'm right, disconnecting the x11 session after I choose my preset should solve the problem.

This would also explain why the slower i3-6300U never had issues: it did have a display attached; a TV actually, mounted behind my MP7 at the time, so it never needed VNC or other screen sharing software. It also ran Windows, which is a bit easier in this regard; just install the interface's ASIO drivers and if there's no weird stuff going on, you'd normally be good to go. On Linux, it can be... more involved with more margin for error.

FWIW--

I found that turning off the Pianoteq keyboard display, or minimizing Pianoteq altogether, reduced dropouts on my old laptop.

With that change, and disabling the WiFi network adapter, dropouts disappeared.

@Charles Cohen : on Linux or Windows ?

Windows can give the audio thread an higher priority than the GUI, but on the other hand, a display driver access can insert a DPC (Deferred Procedure Call) which is considered to have a very high priority. We can measure the induced latency with latencymon which indicates the culprit driver if a DPC takes too much time.

On Linux, I know there are some priority handling, but I am less aware of how this works.