When upsampling to power of 2, I was getting an occasional 1/2 second drop-out. I switched to a custom upsample where everything upsamples to 176.4 or 192 and no drop-outs so far.
In what context… USB or UPnP?
UPnP. I won’t be testing USB anymore.
1 Like
If you disable Convolution… can you play files up-sampled with Power of Two without interrupts?
I’m testing that, but convolution is the main benefit. My Dell is i7 with 32 Gb.
1 Like
I get it … Is there an output gain parameter in the convolution engine?
-6 db
Is that automatically applied or is this your output volume setting…? Do you get the interrupts if you lower your output volume by -10dB when convolution is enabled and upsampling with Power of Two?
The -6 is built in. I’ll add some more negative db after I test the no convolution option for a while.
2 Likes
Pardon me for causing confusion. It’s now been clarified by @chris (given a thumbs up by Antoine) that convolution (though not in real time, as I currently understand it?) does work with UPnP, so I’ve deleted my initial response that it does not.
1 Like
Hi Antoine, I appreciate that you put the sync issue in the list.
The realtime aspect is inconsequential with respect to convolution.
1 Like
My limited knowledge doesn’t permit me to eliminate either possibility at this point. Switching DACs is inconvenient for me, so I hope I or someone else comes up with a different method. 
1 Like
So… have you tried reducing the output gain? It appears that a default -6dB reduction in gain is applied automatically when convolution is engaged… What happens when you reduce the output level to -10dB or -12dB?
It may be inconsequential with respect to Audirvana’s native convolution black box… It happens to be a very nice feature of The Hangloose Convolver plugin, seeing and adjusting signal levels and switching FIR filter banks to do A/B/C/D/E/F comparisons of FIR filters all in Realtime…
Evidently, this requires a USB connection.
Are you saying that you cannot reduce the output volume to -10dB or -12dB with UPnP enabled?
It is highly recommended to reduce the gain (apply negative gain/attenuation) when using the r8brain upscaler—or any high-quality sample rate converter—especially if your source audio is normalized to or near 0 dBFS… I wonder if Audirvana manages this within the upscaler DSP itself?
If one is creating their own FIR filters attenuation can be applied within the FIR filter itself…
Okay… I see the note about applying Software Volume control when UPnP is enabled… the latency involved creates problems in the handshake…
Edited:
So to be clear, the distortion you perceive occurs when using convolution and up-sampling with Power of Two in the context of UPnP, but not USB… Yes? 
That’s an important consideration… There is a gain control when modulating to DSD, but none when up-sampling to higher PCM sample-rates.