Hey guys…over the past months, we’ve tried to reproduce this several times without success. Tried all three platforms (win, mac, linux, including with NAA), and several DACs including the HOLO Spring (which was mentioned by someone in one of these threads as having this issue). I’ve tried PCM upsampling, DSD upsampling, and replicating settings from the screenshots above.
In general, just from the symptom description, I am running short of explanations as to what Roon could be doing to cause this. Our job is to deliver the bits from the file, over TCP, to HQPlayer, which is not a very error-prone task. 95% of that code is shared when playing via RAAT–most of the complexity is in decoding the file and moving the bits to the right place for transmission to the device. In both cases, the bits are moving out over TCP…very similar.
The idea that Roon is actually somehow generating “pink noise” and transmitting it does not make any sense to me. We don’t have a noise generator, and the results of the kinds of programming errors that we could have in theory do not match the description of “pink noise” at all.
In any case, we were working on this because I was hoping to hear the static, in case the particular sound triggered any explanations or ideas. Sometimes the sound of an artifact or failure is a big clue as to what/where it is. Lots of people use words like “static” or “pink noise” but there are a lot of flavors .
I was also hoping to maybe make this reproducible to help give @jussi_laako something to try out. But we are not getting anywhere because it doesn’t happen in front of us.
If this is an issue with Roon, we need a technical bug report from @jussi_laako that describes what we’re doing wrong. I am very happy to make any changes needed to fix this, but we really don’t know what those would be yet.
I’m not necessarily sure this is his problem either, btw. It could be an issue with the DAC (I know some DACs that have failure modes that sound like white noise, and some drivers too…not the ones mentioned in these threads, though). But I think the investigation has to start from the DAC side and work backwards.