Is There A Problem With Mirroring?

1

I don’t know what to make of this or if it is a problem or not, but here it is. When I turn mirroring on and loop back into my QA403 with the mirrored return going through an external DAC into the left channel and the feed directly from the QA403 going into the right channel I get the below, with the skirt of the left (mirrored) channel being wider, and the right channel OK.

Mirroring Through QA
Mirroring Through QA1511×1002 181 KB

However if I then turn mirroring off and feed that exact same DAC from another program on my computer and set the frequency for exactly that of the QA403, the skirt is fine as below.

Mirroring Through MI
Mirroring Through MI1920×1068 242 KB

My question - could it be somehow that the mirrored output from the QA403 is at a slightly different frequency as that directly from the QA403? I think I’ve eliminated the possibility that the DAC could have an off-frequency clock or anything untoward with the computer since the only difference between the two tests is that the mirrored output is feeding the same DAC in one test and another program is generating the exact same frequency indicated on the QA403 and going through the same DAC in the other test.

2

Hi @Rammis, yes, the USB DAC has its own timing reference. So, the difference in frequency will be related to the oscillator error in the QA403 and the USB DAC. Generally these are small (+/- 50 ppm which would manifest as 0.05 Hz error). But I think the problem you are seeing here is that transcoding is happening in Windows. This is a wicked problem, because so many are spending lots of money on a USB DAC, and the performance is getting destroyed by Windows.

Make sure your USB DAC under Windows is set to 48K, 32-bit. Stay with 48K on the QA403 and your DAC until you can get a measurement you trust.

image

Next, make sure the USB DAC you want to test is the primary device. That is, if you go to youtube, the video will play audio from that DAC.

If you USB DAC has a front-panel sampling rate display, make sure it’s at 48K.

Then, in the QA403 software, there are 3 ways to open the audio drive: Default, shared and exclusive. Try each, and look at the spectrum. As the Topping E70 Velvet thread has shown, there can be dramatic differences depending on the audio interface selected.

image

Because windows allows several apps to share the USB DAC, Windows has some extra work to do. For example, let’s say you are listening to lossless 192Ksps on your audio player. Theoretically, that app would want to open the USB DAC at 192. And then, you switch to song you ripped long ago at 44.1k. For that, you’d want the USB DAC at 44.1.

And while listening to the 44.1 track, your mail has to play a “new mail” notification at 48k. Now Windows has a problem, becuase the DAC is opened at 192 or 44.1, but you have a WAV with a 48K sample rate to play. So, windows will downsample the 48K to 44.1k (or upsample to 192), and then sum the two audio streams. But the process of upsampling/downsampling creates distortion artifacts. And that’s what your waveform looks like.

So, start at 48K for everything. And then try different audio interfaces. As the E70 Velvet thread shows, when everything is set up just right you should be able to see TDN+N of -124 dB if your DAC can do it.

3

Thanks Matt,

I had no idea the Windows audio was such a mess. Ugh. I’ll give your suggestions a try.