Delay between two audio outputs after several days

[claudius06]

Hello,
While testing PFOA for several consecutive days by running tests on PFOA, I noticed, after a few days, a delay (almost a second) between the sound coming from the Monitor output and the sound from the pre-listen output. However, this was not the case a few days prior. It seems like the Monitor output is gradually "slipping." How can I prevent this without having to restart PFOA?
I have PFOA - latest version running on Windows 11 Pro, with an I7 processor and 32 GB of RAM.
Best,
Claude.

[radio42]

I assume you are using two different audio devices, which have each their individual drivers and most important their own clock.
Ie. the sample rate of the two devices might be different and is probably not synchronized by a dedicated world clock.
As such, they might drift apart over time.

When the devices additionally (depending on the device) use the system clock instead of having their own internal clock (as most professional sound cards do have) this clock precision might also be dependent on the overall system load and pressure.

This could be a reasonable explanation for what you discovered.
However, if this is the case, then there is nothing I can do to adjust this.

Note, that instead of restarting the application you can also use the ‚Recreate and Reset‘ resp. the ‚Reset Full-Duplex‘ function in the context menu of the related mixer channel (make a right-click on the name of the mixer channel).

[claudius06]

Hello Bernd, thank you for your response.
I only use one device, so I don't think that's the cause of the issue. If you have another idea, I'm open to it.
I tested the "Recreate and Reset" feature, and it caused a deterioration in sound quality and an acceleration in media playback. Very strange. I had to restart PFOA.
The "Reset Full-Duplex" feature did not have any impact on the sound (tested before "Recreate and Reset"). Perhaps it will be interesting to use it if the delay issue happens again.

[radio42]

I am sorry, but what device is it and what driver model are you using, as there is. I other explanation.
The audio processing is actually driven by the driver‘s clock.
The Reset Full-Duplex will only work for certain use cases when forwarding the audio in certain SND2 cases.
But the Recreate and Reset function actually clears any possible the buffer, so if that doesn’t help or even make the sound worse, it can only be due to the clearing of the buffer.
Else than that I would have no idea what could cause the issue, except, that the device driver is using the system clock and not having its own clock.

So what device are you using?

[claudius06]

I am using a Sound Blaster Audigy Fx V2 card with Wasapi driver on a Windows 11 Pro PC with 32 GB of RAM.

[radio42]

Please explain your detailed setup of the related mixer channels incl. any DSP/VST, your routing and how you discover your audio signals not being in sync. Does the mixer channels use the same settings, also in the windows control panel, eg. the same sample rate even in the windows setting?
Please post screenshots of your settings.
However, I don’t personally own such a card, so I am not sure, if I can resolve your issue finally. I’m
ProppFrexx itself is 100% bit accurate, so the issue is for sure somewhere else.
And what version are you using.

[claudius06]

Hello Bernd,
Here is my detailed setup.
I'm using Stereo Tools as DSP on the PLAY channel.
Everything goes through the Sound Blaster card I mentioned.
Everything appears normal on Windows, with the same sampling frequency (48k 16b) matching what I see on PFOA.





I discovered that the signals were not synchronized by chance: when I increased the volume of the headphones and the speakers. This happened approximately after 3-4 days of continuous operation of PFOA. However, I need to specify that, since I am in a testing phase, I am performing a lot of manipulations.
May be the cause of the problem ?
Currently, I am continuing my tests (just playing a test playlist non-stop without any other manipulations except tagging my library on Tagger), and I will check at the end of this week if the issue occurs again.

[radio42]

No, volume changes can not be any reason.
I more assume, that StereoTools might be a reason, as this introduces an additional output buffer, which is then only local to that output channel.

So can you maybe test this without the VST to exclude this, as I have done a test myself and can not reproduce this after now 4 days…

[claudius06]

I understand your hypothesis about Stereo Tools. However, in my case, the PFL output comes from PLAY (PLAY sends the sound to the Monitor and PFL output), so I have the sound processed by Stereo Tools in the headphones as well (except, of course, when pre-listening to a sound). That's why I have some doubts about this idea. Before testing without the VST, I will keep it running in this configuration until the end of the week without working on anything else (except Tagger), and based on the result, I will test it without the VST and let you know. Thanks again for your help.

[radio42]

Please read my first post. This is the explanation.

When you route your playout all to the PLAY mixer channel and PLAY is using the NoSound device. And PLAY is then routed to OUT (also using NoSound). And OUT is using the None driver (which means it is using the internal system clock of the PC).
This means, your main playout path is NOT using your SoundBlaster device in the chain at all.

Ie. the PFL channel is using the SoundBlaster device.
BUT the PLAY channel is NOT using it. It is using the system clock instead. So you are using two different clocks to drive the sample rate. These are not synchronized and might drift apart over time.

So we are back to my initial assumption I made in my first post.
Which would again explain what you experience.