Agreed then, but with caveats…
I think I recall a post that referenced improvements to the Roon mp3 decoding algorithm, but can’t see it at the moment… if I do find it (or find it was something else entirely… ) I’ll update.
To me it seems that this is a bit of a pointless question. Encoding in mp3 changes the data - audibly or not (and depending on encoder and settings). Surely anyone anxious about minute differences shouldn’t be using mp3 in the first place.
Converting to 16bit would probably require dithering and possibly noise shaping. You don’t need those for 24bit.
Actually, I was expressing a tentative opinion - not a core belief. From the feedback here I think I can conclude (and please correct me if I’m wrong):
-
Provided I’m not using a lossy storage format, the actual format of the files makes no difference. Roon core has the horsepower and the clock cycles to do any conversions required, and the result will be top notch no matter what.
-
If I’m using DSP, the waters may be murkier. The DSP will require more computing horsepower to implement, but (again) shouldn’t affect the sound quality in any negative way? I’m not referring to the sound quality changes that occur as a result of the filters themselves (I’m responsible for those), but the Roon core and the CPU shouldn’t be significantly stressed by the DSP requirements?
Thanks again for the VERY helpful feedback - Glenn
- Basically true, although if you already have files using lossy compression, the “damage” has already been done, so it won’t matter all the same. Also, decoding those doesn’t require significant CPU either, especially if followed by DSP.
- If you use DSP, you will certainly stress the CPU more or less, but that won’t make any sound quality difference.
1 Like
WAV sounds better on Naim 1st generation streamers. FLAC sounds noticeably ‘sharper’ and a little harsh.
I ended up with an inadvertent blind test where I was noticing the above characteristics but couldn’t work out why and then checked the ND5 display and saw FLAC. 
Water
