Using Roon Muse parametric filters with MSB DACs

General
1

I am writing to get advice on the best ways to use Roon Muse parametric filters with MSB DACs. The use case is poor recordings that need help. My knowledge of DSP theory is moderate.

  1. I find up sampling in Roon makes the Roon Muse parametric filters sound better, particularly if the filter is operating in the upper frequencies. Perhaps pushing the Nyquist frequency further from the filter’s range. Is this supported by theory and experience?

  2. If up sampling is helpful, I am mindful of MSB’s advice that it is better to feed a native rate signal to MSB DACs to allow them to work their own oversampling magic. Does it help to feed an up sampled 192khz signal rather than a max upsampled 768khz signal so the DAC still has some room to add magic?

  3. Any other advice from DSP experts on the best Roon filter settings for this use case of Roon parametric filters used with MSB DACs.

I have tried various options. To my ears some of the differences are subtle. But for people with better ears than mine, I thought it would be useful to get the views who know DSP theory and the inner workings of MSB DACs well.

2

The reason Roon sounds better upsampled when you are doing DSP is that Roon built in DSP is rather poor and upsampling helps its own (Roons) algorithms a small amount. It is not increasing the quality of the conversion in a MSB DAC at all. You will get much better quality completely disabling all DSP in Roon, just turning on any DSP in Roon will cause a drop in quality, regardless of upsampling.

2 Likes
3

My anecdotal experience using Roon upsampling supports Dustin’s comment. So, I do not use Roon upsampling but, instead, pass the signal at its native frequency to the DAC, which in my case is a Premier + Digital Director.

Tangentially, I do use a DSP convolution file (provided by Audiolense) for my room but I instruct Audiolense to generate convolution files at all the standard frequencies so I avoid Roon upsampling. Of course, Roon does a bit depth conversion prior to applying convolution. I can’t speak to any aspects of that type of conversion, but whatever negatives it may or may not contribute is vastly outweighed by the sick, positive benefits of using convolution filters to “correct” for my room.

4

I understand that native signal to the MSB DAC is the preferred mode, and that is what I do for 80% plus of recordings. I am looking for ways to improve the poor recordings. Preferably without adding hardware such as an equaliser after the DAC, especially as that hardware would only be used for a minority of recordings. So the question is if I am trying to salvage a poor recording with DSP, how to best do it.